인공지능은 의료를 어떻게 혁신하는가 (2017년 11월) 최윤섭
8 likes4,797 views
The document discusses the potential of artificial intelligence (AI) to revolutionize the medical field, highlighting the replacement of some medical professionals by AI technologies like IBM Watson. It outlines advancements in AI applications for medical data analysis, treatment recommendations, and overall healthcare efficiency, including substantial cost savings and improved patient outcomes. The text also addresses the ongoing debate regarding the reliability of AI in medicine and the need for validation through clinical trials and guidelines.
![Can machine-learning improve cardiovascular
risk prediction using routine clinical data?
Stephen F.Weng et al PLoS One 2017
correctly predicted compared to the baseline ACC/AHA model ranged from 191 non-cases for
the random forest algorithm to 355 non-cases for the neural networks. Full details on classifi-
cation analysis can be found in S2 Table.
Discussion
Compared to an established AHA/ACC risk prediction algorithm, we found all machine-
learning algorithms tested were better at identifying individuals who will develop CVD and
those that will not. Unlike established approaches to risk prediction, the machine-learning
methods used were not limited to a small set of risk factors, and incorporated more pre-exist-
Table 4. Performance of the machine-learning (ML) algorithms predicting 10-year cardiovascular disease (CVD) risk derived from applying train-
ing algorithms on the validation cohort of 82,989 patients. Higher c-statistics results in better algorithm discrimination. The baseline (BL) ACC/AHA
10-year risk prediction algorithm is provided for comparative purposes.
Algorithms AUC c-statistic Standard Error* 95% Confidence
Interval
Absolute Change from Baseline
LCL UCL
BL: ACC/AHA 0.728 0.002 0.723 0.735 —
ML: Random Forest 0.745 0.003 0.739 0.750 +1.7%
ML: Logistic Regression 0.760 0.003 0.755 0.766 +3.2%
ML: Gradient Boosting Machines 0.761 0.002 0.755 0.766 +3.3%
ML: Neural Networks 0.764 0.002 0.759 0.769 +3.6%
*Standard error estimated by jack-knife procedure [30]
https://doi.org/10.1371/journal.pone.0174944.t004
Can machine-learning improve cardiovascular risk prediction using routine clinical data?
•네 가지 기계학습 모델 모두 기존의 ACC/AHA 가이드라인 대비 더 정확했다.
•Neural Networks 이 AUC=0.764 로 가장 정확했다.
•“이 모델을 활용했더라면 355 명의 추가적인 cardiovascular event 를 예방했을 것”
•Deep Learning 을 활용하면 정확도는 더 높아질 수 있을 것
•Genetic information 등의 추가적인 risk factor 를 활용해볼 수 있다.](https://image.slidesharecdn.com/howaiwouldinnovatethemedicineoffuture171105-171105143858/85/2017-11-91-320.jpg)
![DeepFace: Closing the Gap to Human-Level
Performance in FaceVerification
Taigman,Y. et al. (2014). DeepFace: Closing the Gap to Human-Level Performance in FaceVerification, CVPR’14.
Figure 2. Outline of the DeepFace architecture. A front-end of a single convolution-pooling-convolution filtering on the rectified input, followed by three
locally-connected layers and two fully-connected layers. Colors illustrate feature maps produced at each layer. The net includes more than 120 million
parameters, where more than 95% come from the local and fully connected layers.
very few parameters. These layers merely expand the input
into a set of simple local features.
The subsequent layers (L4, L5 and L6) are instead lo-
cally connected [13, 16], like a convolutional layer they ap-
ply a filter bank, but every location in the feature map learns
a different set of filters. Since different regions of an aligned
image have different local statistics, the spatial stationarity
The goal of training is to maximize the probability of
the correct class (face id). We achieve this by minimiz-
ing the cross-entropy loss for each training sample. If k
is the index of the true label for a given input, the loss is:
L = log pk. The loss is minimized over the parameters
by computing the gradient of L w.r.t. the parameters and
Human: 95% vs. DeepFace in Facebook: 97.35%
Recognition Accuracy for Labeled Faces in the Wild (LFW) dataset (13,233 images, 5,749 people)](https://image.slidesharecdn.com/howaiwouldinnovatethemedicineoffuture171105-171105143858/85/2017-11-104-320.jpg)
![FaceNet:A Unified Embedding for Face
Recognition and Clustering
Schroff, F. et al. (2015). FaceNet:A Unified Embedding for Face Recognition and Clustering
Human: 95% vs. FaceNet of Google: 99.63%
Recognition Accuracy for Labeled Faces in the Wild (LFW) dataset (13,233 images, 5,749 people)
False accept
False reject
s. This shows all pairs of images that were
on LFW. Only eight of the 13 errors shown
he other four are mislabeled in LFW.
on Youtube Faces DB
ge similarity of all pairs of the first one
our face detector detects in each video.
False accept
False reject
Figure 6. LFW errors. This shows all pairs of images that were
incorrectly classified on LFW. Only eight of the 13 errors shown
here are actual errors the other four are mislabeled in LFW.
5.7. Performance on Youtube Faces DB
We use the average similarity of all pairs of the first one
hundred frames that our face detector detects in each video.
This gives us a classification accuracy of 95.12%±0.39.
Using the first one thousand frames results in 95.18%.
Compared to [17] 91.4% who also evaluate one hundred
frames per video we reduce the error rate by almost half.
DeepId2+ [15] achieved 93.2% and our method reduces this
error by 30%, comparable to our improvement on LFW.
5.8. Face Clustering
Our compact embedding lends itself to be used in order
to cluster a users personal photos into groups of people with
the same identity. The constraints in assignment imposed
by clustering faces, compared to the pure verification task,
lead to truly amazing results. Figure 7 shows one cluster in
a users personal photo collection, generated using agglom-
erative clustering. It is a clear showcase of the incredible
invariance to occlusion, lighting, pose and even age.
Figure 7. Face Clustering. Shown is an exemplar cluster for one
user. All these images in the users personal photo collection were
clustered together.
6. Summary
We provide a method to directly learn an embedding into
an Euclidean space for face verification. This sets it apart
from other methods [15, 17] who use the CNN bottleneck
layer, or require additional post-processing such as concate-
nation of multiple models and PCA, as well as SVM clas-
sification. Our end-to-end training both simplifies the setup
and shows that directly optimizing a loss relevant to the task
at hand improves performance.
Another strength of our model is that it only requires
False accept
False reject
Figure 6. LFW errors. This shows all pairs of images that were
incorrectly classified on LFW. Only eight of the 13 errors shown
here are actual errors the other four are mislabeled in LFW.
5.7. Performance on Youtube Faces DB
We use the average similarity of all pairs of the first one
hundred frames that our face detector detects in each video.
This gives us a classification accuracy of 95.12%±0.39.
Using the first one thousand frames results in 95.18%.
Compared to [17] 91.4% who also evaluate one hundred
frames per video we reduce the error rate by almost half.
DeepId2+ [15] achieved 93.2% and our method reduces this
error by 30%, comparable to our improvement on LFW.
5.8. Face Clustering
Our compact embedding lends itself to be used in order
to cluster a users personal photos into groups of people with
the same identity. The constraints in assignment imposed
by clustering faces, compared to the pure verification task,
Figure 7. Face Clustering. Shown is an exemplar cluster for one
user. All these images in the users personal photo collection were
clustered together.
6. Summary
We provide a method to directly learn an embedding into
an Euclidean space for face verification. This sets it apart
from other methods [15, 17] who use the CNN bottleneck
layer, or require additional post-processing such as concate-
nation of multiple models and PCA, as well as SVM clas-](https://image.slidesharecdn.com/howaiwouldinnovatethemedicineoffuture171105-171105143858/85/2017-11-105-320.jpg)
![Copyright 2016 American Medical Association. All rights reserved.
Development and Validation of a Deep Learning Algorithm
for Detection of Diabetic Retinopathy
in Retinal Fundus Photographs
Varun Gulshan, PhD; Lily Peng, MD, PhD; Marc Coram, PhD; Martin C. Stumpe, PhD; Derek Wu, BS; Arunachalam Narayanaswamy, PhD;
Subhashini Venugopalan, MS; Kasumi Widner, MS; Tom Madams, MEng; Jorge Cuadros, OD, PhD; Ramasamy Kim, OD, DNB;
Rajiv Raman, MS, DNB; Philip C. Nelson, BS; Jessica L. Mega, MD, MPH; Dale R. Webster, PhD
IMPORTANCE Deep learning is a family of computational methods that allow an algorithm to
program itself by learning from a large set of examples that demonstrate the desired
behavior, removing the need to specify rules explicitly. Application of these methods to
medical imaging requires further assessment and validation.
OBJECTIVE To apply deep learning to create an algorithm for automated detection of diabetic
retinopathy and diabetic macular edema in retinal fundus photographs.
DESIGN AND SETTING A specific type of neural network optimized for image classification
called a deep convolutional neural network was trained using a retrospective development
data set of 128 175 retinal images, which were graded 3 to 7 times for diabetic retinopathy,
diabetic macular edema, and image gradability by a panel of 54 US licensed ophthalmologists
and ophthalmology senior residents between May and December 2015. The resultant
algorithm was validated in January and February 2016 using 2 separate data sets, both
graded by at least 7 US board-certified ophthalmologists with high intragrader consistency.
EXPOSURE Deep learning–trained algorithm.
MAIN OUTCOMES AND MEASURES The sensitivity and specificity of the algorithm for detecting
referable diabetic retinopathy (RDR), defined as moderate and worse diabetic retinopathy,
referable diabetic macular edema, or both, were generated based on the reference standard
of the majority decision of the ophthalmologist panel. The algorithm was evaluated at 2
operating points selected from the development set, one selected for high specificity and
another for high sensitivity.
RESULTS TheEyePACS-1datasetconsistedof9963imagesfrom4997patients(meanage,54.4
years;62.2%women;prevalenceofRDR,683/8878fullygradableimages[7.8%]);the
Messidor-2datasethad1748imagesfrom874patients(meanage,57.6years;42.6%women;
prevalenceofRDR,254/1745fullygradableimages[14.6%]).FordetectingRDR,thealgorithm
hadanareaunderthereceiveroperatingcurveof0.991(95%CI,0.988-0.993)forEyePACS-1and
0.990(95%CI,0.986-0.995)forMessidor-2.Usingthefirstoperatingcutpointwithhigh
specificity,forEyePACS-1,thesensitivitywas90.3%(95%CI,87.5%-92.7%)andthespecificity
was98.1%(95%CI,97.8%-98.5%).ForMessidor-2,thesensitivitywas87.0%(95%CI,81.1%-
91.0%)andthespecificitywas98.5%(95%CI,97.7%-99.1%).Usingasecondoperatingpoint
withhighsensitivityinthedevelopmentset,forEyePACS-1thesensitivitywas97.5%and
specificitywas93.4%andforMessidor-2thesensitivitywas96.1%andspecificitywas93.9%.
CONCLUSIONS AND RELEVANCE In this evaluation of retinal fundus photographs from adults
with diabetes, an algorithm based on deep machine learning had high sensitivity and
specificity for detecting referable diabetic retinopathy. Further research is necessary to
determine the feasibility of applying this algorithm in the clinical setting and to determine
whether use of the algorithm could lead to improved care and outcomes compared with
current ophthalmologic assessment.
JAMA. doi:10.1001/jama.2016.17216
Published online November 29, 2016.
Editorial
Supplemental content
Author Affiliations: Google Inc,
Mountain View, California (Gulshan,
Peng, Coram, Stumpe, Wu,
Narayanaswamy, Venugopalan,
Widner, Madams, Nelson, Webster);
Department of Computer Science,
University of Texas, Austin
(Venugopalan); EyePACS LLC,
San Jose, California (Cuadros); School
of Optometry, Vision Science
Graduate Group, University of
California, Berkeley (Cuadros);
Aravind Medical Research
Foundation, Aravind Eye Care
System, Madurai, India (Kim); Shri
Bhagwan Mahavir Vitreoretinal
Services, Sankara Nethralaya,
Chennai, Tamil Nadu, India (Raman);
Verily Life Sciences, Mountain View,
California (Mega); Cardiovascular
Division, Department of Medicine,
Brigham and Women’s Hospital and
Harvard Medical School, Boston,
Massachusetts (Mega).
Corresponding Author: Lily Peng,
MD, PhD, Google Research, 1600
Amphitheatre Way, Mountain View,
CA 94043 (lhpeng@google.com).
Research
JAMA | Original Investigation | INNOVATIONS IN HEALTH CARE DELIVERY
(Reprinted) E1
Copyright 2016 American Medical Association. All rights reserved.](https://image.slidesharecdn.com/howaiwouldinnovatethemedicineoffuture171105-171105143858/85/2017-11-123-320.jpg)
![Assisting Pathologists in Detecting
Cancer with Deep Learning
6
Input & Validation Test
model size FROC @8FP AUC FROC @8FP AUC
40X 98.1 100 99.0 87.3 (83.2, 91.1) 91.1 (87.2, 94.5) 96.7 (92.6, 99.6)
40X-pretrained 99.3 100 100 85.5 (81.0, 89.5) 91.1 (86.8, 94.6) 97.5 (93.8, 99.8)
40X-small 99.3 100 100 86.4 (82.2, 90.4) 92.4 (88.8, 95.7) 97.1 (93.2, 99.8)
ensemble-of-3 - - - 88.5 (84.3, 92.2) 92.4 (88.7, 95.6) 97.7 (93.0, 100)
20X-small 94.7 100 99.6 85.5 (81.0, 89.7) 91.1 (86.9, 94.8) 98.6 (96.7, 100)
10X-small 88.7 97.2 97.7 79.3 (74.2, 84.1) 84.9 (80.0, 89.4) 96.5 (91.9, 99.7)
40X+20X-small 94.9 98.6 99.0 85.9 (81.6, 89.9) 92.9 (89.3, 96.1) 97.0 (93.1, 99.9)
40X+10X-small 93.8 98.6 100 82.2 (77.0, 86.7) 87.6 (83.2, 91.7) 98.6 (96.2, 99.9)
Pathologist [1] - - - 73.3* 73.3* 96.6
Camelyon16 winner [1, 23] - - - 80.7 82.7 99.4
Table 1. Results on Camelyon16 dataset (95% confidence intervals, CI). Bold indicates
results within the CI of the best model. “Small” models contain 300K parameters per
Inception tower instead of 20M. -: not reported. *A pathologist achieved this sensitivity
(with no FP) using 30 hours.
to 10 20% variance), and can confound evaluation of model improvements
by grouping multiple nearby tumors as one. By contrast, our non-maxima sup-
pression approach is relatively insensitive to r between 4 and 6, although less
accurate models benefited from tuning r using the validation set (e.g., 8). Fi-
The FROC evaluates tumor detection and localization
The FROC is defined as the sensitivity at 0.25,0.5,1,2,4,8 average FPs per tumor-negative slide.
Yun Liu et al. Detecting Cancer Metastases on Gigapixel Pathology Images (2017)
Sensitivity at 8 false positives per image](https://image.slidesharecdn.com/howaiwouldinnovatethemedicineoffuture171105-171105143858/85/2017-11-151-320.jpg)
![Fig 1. What can consumer wearables do? Heart rate can be measured with an oximeter built into a ring [3], muscle activity with an electromyographi
sensor embedded into clothing [4], stress with an electodermal sensor incorporated into a wristband [5], and physical activity or sleep patterns via an
accelerometer in a watch [6,7]. In addition, a female’s most fertile period can be identified with detailed body temperature tracking [8], while levels of me
attention can be monitored with a small number of non-gelled electroencephalogram (EEG) electrodes [9]. Levels of social interaction (also known to a
PLOS Medicine 2016](https://image.slidesharecdn.com/howaiwouldinnovatethemedicineoffuture171105-171105143858/85/2017-11-157-320.jpg)
![S E P S I S
A targeted real-time early warning score (TREWScore)
for septic shock
Katharine E. Henry,1
David N. Hager,2
Peter J. Pronovost,3,4,5
Suchi Saria1,3,5,6
*
Sepsis is a leading cause of death in the United States, with mortality highest among patients who develop septic
shock. Early aggressive treatment decreases morbidity and mortality. Although automated screening tools can detect
patients currently experiencing severe sepsis and septic shock, none predict those at greatest risk of developing
shock. We analyzed routinely available physiological and laboratory data from intensive care unit patients and devel-
oped “TREWScore,” a targeted real-time early warning score that predicts which patients will develop septic shock.
TREWScore identified patients before the onset of septic shock with an area under the ROC (receiver operating
characteristic) curve (AUC) of 0.83 [95% confidence interval (CI), 0.81 to 0.85]. At a specificity of 0.67, TREWScore
achieved a sensitivity of 0.85 and identified patients a median of 28.2 [interquartile range (IQR), 10.6 to 94.2] hours
before onset. Of those identified, two-thirds were identified before any sepsis-related organ dysfunction. In compar-
ison, the Modified Early Warning Score, which has been used clinically for septic shock prediction, achieved a lower
AUC of 0.73 (95% CI, 0.71 to 0.76). A routine screening protocol based on the presence of two of the systemic inflam-
matory response syndrome criteria, suspicion of infection, and either hypotension or hyperlactatemia achieved a low-
er sensitivity of 0.74 at a comparable specificity of 0.64. Continuous sampling of data from the electronic health
records and calculation of TREWScore may allow clinicians to identify patients at risk for septic shock and provide
earlier interventions that would prevent or mitigate the associated morbidity and mortality.
INTRODUCTION
Seven hundred fifty thousand patients develop severe sepsis and septic
shock in the United States each year. More than half of them are
admitted to an intensive care unit (ICU), accounting for 10% of all
ICU admissions, 20 to 30% of hospital deaths, and $15.4 billion in an-
nual health care costs (1–3). Several studies have demonstrated that
morbidity, mortality, and length of stay are decreased when severe sep-
sis and septic shock are identified and treated early (4–8). In particular,
one study showed that mortality from septic shock increased by 7.6%
with every hour that treatment was delayed after the onset of hypo-
tension (9).
More recent studies comparing protocolized care, usual care, and
early goal-directed therapy (EGDT) for patients with septic shock sug-
gest that usual care is as effective as EGDT (10–12). Some have inter-
preted this to mean that usual care has improved over time and reflects
important aspects of EGDT, such as early antibiotics and early ag-
gressive fluid resuscitation (13). It is likely that continued early identi-
fication and treatment will further improve outcomes. However, the
Acute Physiology Score (SAPS II), SequentialOrgan Failure Assessment
(SOFA) scores, Modified Early Warning Score (MEWS), and Simple
Clinical Score (SCS) have been validated to assess illness severity and
risk of death among septic patients (14–17). Although these scores
are useful for predicting general deterioration or mortality, they typical-
ly cannot distinguish with high sensitivity and specificity which patients
are at highest risk of developing a specific acute condition.
The increased use of electronic health records (EHRs), which can be
queried in real time, has generated interest in automating tools that
identify patients at risk for septic shock (18–20). A number of “early
warning systems,” “track and trigger” initiatives, “listening applica-
tions,” and “sniffers” have been implemented to improve detection
andtimelinessof therapy forpatients with severe sepsis andseptic shock
(18, 20–23). Although these tools have been successful at detecting pa-
tients currently experiencing severe sepsis or septic shock, none predict
which patients are at highest risk of developing septic shock.
The adoption of the Affordable Care Act has added to the growing
excitement around predictive models derived from electronic health
R E S E A R C H A R T I C L E
onNovember3,2016http://stm.sciencemag.org/Downloadedfrom](https://image.slidesharecdn.com/howaiwouldinnovatethemedicineoffuture171105-171105143858/85/2017-11-163-320.jpg)
![puted as new data became avail
when his or her score crossed t
dation set, the AUC obtained f
0.81 to 0.85) (Fig. 2). At a spec
of 0.33], TREWScore achieved a s
a median of 28.2 hours (IQR, 10
Identification of patients b
A critical event in the developme
related organ dysfunction (seve
been shown to increase after th
more than two-thirds (68.8%) o
were identified before any sepsi
tients were identified a median
(Fig. 3B).
Comparison of TREWScore
Weevaluatedtheperformanceof
methods for the purpose of provid
use of TREWScore. We first com
to MEWS, a general metric used
of catastrophic deterioration (17)
oped for tracking sepsis, MEWS
tion of patients at risk for severe
Fig. 2. ROC for detection of septic shock before onset in the validation
set. The ROC curve for TREWScore is shown in blue, with the ROC curve for
MEWS in red. The sensitivity and specificity performance of the routine
screening criteria is indicated by the purple dot. Normal 95% CIs are shown
for TREWScore and MEWS. TPR, true-positive rate; FPR, false-positive rate.
R E S E A R C H A R T I C L E
A targeted real-time early warning score (TREWScore)
for septic shock
AUC=0.83
At a specificity of 0.67,TREWScore achieved a sensitivity of 0.85
and identified patients a median of 28.2 hours before onset.](https://image.slidesharecdn.com/howaiwouldinnovatethemedicineoffuture171105-171105143858/85/2017-11-164-320.jpg)
![Cardiologist-Level Arrhythmia Detection with Convolutional Neural Networks
Pranav Rajpurkar⇤
PRANAVSR@CS.STANFORD.EDU
Awni Y. Hannun⇤
AWNI@CS.STANFORD.EDU
Masoumeh Haghpanahi MHAGHPANAHI@IRHYTHMTECH.COM
Codie Bourn CBOURN@IRHYTHMTECH.COM
Andrew Y. Ng ANG@CS.STANFORD.EDU
Abstract
We develop an algorithm which exceeds the per-
formance of board certified cardiologists in de-
tecting a wide range of heart arrhythmias from
electrocardiograms recorded with a single-lead
wearable monitor. We build a dataset with more
than 500 times the number of unique patients
than previously studied corpora. On this dataset,
we train a 34-layer convolutional neural network
which maps a sequence of ECG samples to a se-
quence of rhythm classes. Committees of board-
certified cardiologists annotate a gold standard
test set on which we compare the performance of
our model to that of 6 other individual cardiolo-
gists. We exceed the average cardiologist perfor-
mance in both recall (sensitivity) and precision
(positive predictive value).
1. Introduction
We develop a model which can diagnose irregular heart
Figure 1. Our trained convolutional neural network correctly de-
tecting the sinus rhythm (SINUS) and Atrial Fibrillation (AFIB)
from this ECG recorded with a single-lead wearable heart moni-
tor.
Arrhythmia detection from ECG recordings is usually per-
formed by expert technicians and cardiologists given the
1707.01836v1[cs.CV]6Jul2017](https://image.slidesharecdn.com/howaiwouldinnovatethemedicineoffuture171105-171105143858/85/2017-11-183-320.jpg)
![NCCN Guidelines Version 4.2014
Non-Small Cell Lung Cancer
NCCN Guidelines Index
NSCLC Table of Contents
Discussion
Version 4.2014, 06/05/14 © National Comprehensive Cancer Network, Inc. 2014, All rights reserved. The NCCN Guidelines®
and this illustration may not be reproduced in any form without the express written permission of NCCN®
.
Note: All recommendations are category 2A unless otherwise indicated.
Clinical Trials: NCCN believes that the best management of any cancer patient is in a clinical trial. Participation in clinical trials is especially encouraged.
NSCL-2
dT3, N0 related to size or satellite nodules.
fTesting is not listed in order of priority and is dependent upon clinical
circumstances, institutional processes, and judicious use of resources.
gMethods for evaluation include mediastinoscopy, mediastinotomy, EBUS, EUS,
and CT-guided biopsy.
hPositive PET/CT scan findings for distant disease need pathologic or other
radiologic confirmation. If PET/CT scan is positive in the mediastinum, lymph
node status needs pathologic confirmation.
iSee Principles of Surgical Therapy (NSCL-B).
jSee Principles of Radiation Therapy (NSCL-C).
kSee Chemotherapy Regimens for Neoadjuvant and Adjuvant Therapy (NSCL-D).
lExamples of high-risk factors may include poorly differentiated tumors (including
lung neuroendocrine tumors [excluding well-differentiated neuroendocrine tumors]),
vascular invasion, wedge resection, tumors >4 cm, visceral pleural involvement,
and incomplete lymph node sampling (Nx). These factors independently may not
be an indication and may be considered when determining treatment with adjuvant
chemotherapy.
mSee Chemotherapy Regimens Used with Radiation Therapy (NSCL-E).
CLINICAL ASSESSMENT PRETREATMENT EVALUATIONf INITIAL TREATMENT
Stage IA
(peripheral T1ab, N0)
Stage IB
(peripheral T2a, N0)
Stage I
(central T1ab–T2a, N0)
Stage II
(T1ab–2ab, N1; T2b, N0)
Stage IIB
(T3, N0)d
• PFTs (if not previously
done)
• Bronchoscopy
(intraoperative
preferred)
• Pathologic mediastinal
lymph node evaluationg
(category 2B)
• PET/CT scanh (if not
previously done)
• PFTs (if not previously
done)
• Bronchoscopy
• Pathologic mediastinal
lymph node evaluationg
• PET/CT scanh (if not
previously done)
• Brain MRI (Stage II,
Stage IB [category 2B])
Negative
mediastinal
nodes
Positive
mediastinal
nodes
Operable
Medically
inoperable
Negative
mediastinal
nodes
Positive
mediastinal
nodes
Operable
Medically
inoperable
Surgical exploration and
resectioni + mediastinal lymph
node dissection or systematic
lymph node sampling
Definitive RT including stereotactic
ablative radiotherapyj (SABR)
See Stage IIIA (NSCL-8) or Stage IIIB (NSCL-11)
Surgical exploration and
resectioni + mediastinal lymph
node dissection or systematic
lymph node sampling
N0
N1
See Stage IIIA (NSCL-8) or Stage IIIB (NSCL-11)
Definitive RT
including SABRj
Definitive chemoradiationj,m
See Adjuvant
Treatment (NSCL-3)
See Adjuvant
Treatment (NSCL-3)
Consider adjuvant
chemotherapyk
(category 2B) for
high-risk stages IB-IIl
Printed by yoon sup choi on 6/19/2014 8:23:15 PM. For personal use only. Not approved for distribution. Copyright © 2014 National Comprehensive Cancer Network, Inc., All Rights Reserved.](https://image.slidesharecdn.com/howaiwouldinnovatethemedicineoffuture171105-171105143858/85/2017-11-205-320.jpg)
![Azad Shademan et al. Sci Transl Med 2016
•Ex vivo 와 in vivo end-to-end 의 경우 모두 STAR 는 기존의 수술 방법들에 비해 대부분 더 나은 성과를 보임
•외과의사가 손으로 문합하는 것 (OPEN)
•복강경 수술 (laparoscopy, LAP)
•다빈치를 이용한 RAS
•비교 기준
•비교 기준은 봉합한 공간의 일관성 (spacing)
•얼마나 압력을 가했을 때 복강경이 새어 나오는가 (leak pressure)
•조직에서 바늘을 제거하는 것이 필요했던 실수의 횟수 (number of mistakes)
•총 수술 시간 (completion time)
•내강의 수축 여부 (lumen reduction)
smallest for STAR, indicating a greater consistency for semiautonomous
suture placement compared to other techniques (table S1). Moreover,
leak pressure reflects the functional quality of suturing. The linear
closure from STAR was able to withstand a higher average leak pressure
than all other techniques (Fig. 2B).
coordinates, intersuture spacing, bite depths, pull tension forces, and
suturing tool maneuvers before piercing. Using the NIRF markers as
reference points, the plan interpolated intermediate suture placements
on the bowel and adjusted placement of each suture, knot, and corner
slidetoaccommodatedeformationsandinducedscenerotations(Fig.1F).
Fig. 2. Ex vivo linear suturing under deformations. The experiment con-
sisted of closing a longitudinal cut along pig intestine, whereas the tissue
was deformed by pulling on stay sutures. Five samples were tested per tech-
nique (OPEN, LAP, RAS, and STAR). (A) Suture spacing. Central mark is the
median; box edges are the 25th and 75th percentiles, error bars are the range
excluding outliers, and red dots are outliers. The whiskers represent the range
not including outliers. There is a different N number for each boxplot because
eachsurgeonuseda different number of sutures [OPEN (n =174), LAP(n= 128),
RAS (n = 176), and STAR (n = 206)]. These data are presented numerically in
table S2, including the SDs. P values determined by ANOVA with post hoc
Games-Howell. (B and C) Leak pressures and number of mistakes (reposi-
tioned stitches or robot reboot). Data are from individual tissue samples
(n = 5) with averages marked by a horizontal line. P values determined by
independent samples t test. (D) Completion times separated into knot-tying
and suturing, and other time was spent restaging or changing sutures. Data
are averages (n = 5). P values determined by independent samples t test.
www.ScienceTranslationalMedicine.org 4 May 2016 Vol 8 Issue 337 337ra64 3
onMay7,2016rg/](https://image.slidesharecdn.com/howaiwouldinnovatethemedicineoffuture171105-171105143858/85/2017-11-222-320.jpg)
![Azad Shademan et al. Sci Transl Med 2016
LAP (Fig. 3E). Ex vivo studies simulating luminal strictures from anasto- whereas only 16 of 59 (27.1%) sutures required adjustmen
maining circumference (Fig. 3
ing that different levels of auto
be used effectively for differ
Overall, 57.8% of the procedu
fully autonomously with no
Alternatively, in the current
autonomous mode without an
teraction would require suture
in 42.2% of sutures placed, m
corners. The completion tim
also included supervisory ac
surgeon, which accounted f
the total time (7% for suture
justment, 3.3% for confirmati
location, and 2.6% for mistake
In vivo end-to-end anast
Finally, we performed in viv
autonomous surgery in pig i
cessed through a laparotomy
(n = 4) and compared these a
an OPEN control (n = 1) (fig
used the same suture algorith
ex vivo trials (Fig. 1, G and
OPEN control, the surgeon u
surgical hand tools to open th
exposed the intestine, and sutu
a transverse incision. The av
STAR procedure time was 50.0
where 77.4% was anastomos
22.6% was restaging time be
and front walls, which inclu
2.16 min for marking the tis
B and E, and Table 1). Al
OPEN timewasonly 8min,th
was comparable to the averag
laparoscopic anastomoses tha
30 min for vesicourethral (25
for aortic (26), to 90 min for
constructions (27).
Fig. 3. End-to-end anastomosis ex vivo. The experiment
consisted of closing a transverse cut in pig intestine. Five
samples were tested per technique (OPEN, LAP, RAS, and STAR).
(A) Suture spacing. Central mark is the median; box edges are
the 25th and 75th percentiles; and red dots are outliers. The
whiskers represent the range not including outliers. There is a
different N number for each boxplot because each surgeon
used a different number of sutures [OPEN (n = 138), LAP (n =
98), RAS (n = 132), and STAR (n = 180)]. The average spacing
betweenconsecutive sutures was calculated and compared be-
tween STAR and other modalities. The variance of suture
spacing is presented numerically in table S2, including the SD.
P values determined by ANOVA with post hoc Games-Howell. (B) Exvivo end-to-end anastomosis leak pressures.
Dataareindividualtissuesamples,withmeansdisplayedashorizontallines(n=4to5).Onesample was sutured
closed and thus could not be tested for leak pressure. P values determined by independent samples t test.
(C) The leak pressure as a function of maximum suture spacing. Data are individual tissue samples that were fit
to a rational function (y = 0.854/x) (n = 4 to 5). (D) Number of mistakes (repositioned stitches or robot reboot).
Data are individual tissue samples with means displayed as horizontal lines (n = 5). P values determined by
independent samples t test. (E) Ex vivo end-to-end anastomosis completion times. Average times for n = 5
•Ex vivo 와 in vivo end-to-end 의 경우 모두 STAR 는 기존의 수술 방법들에 비해 대부분 더 나은 성과를 보임
•외과의사가 손으로 문합하는 것 (OPEN)
•복강경 수술 (laparoscopy, LAP)
•다빈치를 이용한 RAS
•비교 기준
•비교 기준은 봉합한 공간의 일관성 (spacing)
•얼마나 압력을 가했을 때 복강경이 새어 나오는가 (leak pressure)
•조직에서 바늘을 제거하는 것이 필요했던 실수의 횟수 (number of mistakes)
•총 수술 시간 (completion time)
•내강의 수축 여부 (lumen reduction)](https://image.slidesharecdn.com/howaiwouldinnovatethemedicineoffuture171105-171105143858/85/2017-11-223-320.jpg)
인공지능은 의료를 어떻게 혁신하는가 (2017년 11월) 최윤섭
- 1. How Artificial Intelligence would Innovate the Medicine in the Future. Professor, SAHIST, Sungkyunkwan University Director, Digital Healthcare Institute Yoon Sup Choi, Ph.D. 인공지능은 의료를 어떻게 혁신할 것인가
- 2. “It's in Apple's DNA that technology alone is not enough. It's technology married with liberal arts.”
- 3. The Convergence of IT, BT and Medicine
- 6. Inevitable Tsunami of Change
- 8. Vinod Khosla Founder, 1st CEO of Sun Microsystems Partner of KPCB, CEO of KhoslaVentures LegendaryVenture Capitalist in SiliconValley
- 9. “Technology will replace 80% of doctors”
- 10. https://www.youtube.com/watch?time_continue=70&v=2HMPRXstSvQ “영상의학과 전문의를 양성하는 것을 당장 그만둬야 한다. 5년 안에 딥러닝이 영상의학과 전문의를 능가할 것은 자명하다.” Hinton on Radiology
- 13. Luddites in the 1810’s
- 14. and/or
- 16. •AP 통신: 로봇이 인간 대신 기사를 작성 •초당 2,000 개의 기사 작성 가능 •기존에 300개 기업의 실적 ➞ 3,000 개 기업을 커버
- 17. • 1978 • As part of the obscure task of “discovery” — providing documents relevant to a lawsuit — the studios examined six million documents at a cost of more than $2.2 million, much of it to pay for a platoon of lawyers and paralegals who worked for months at high hourly rates. • 2011 • Now, thanks to advances in artificial intelligence, “e-discovery” software can analyze documents in a fraction of the time for a fraction of the cost. • In January, for example, Blackstone Discovery of Palo Alto, Calif., helped analyze 1.5 million documents for less than $100,000.
- 18. “At its height back in 2000, the U.S. cash equities trading desk at Goldman Sachs’s New York headquarters employed 600 traders, buying and selling stock on the orders of the investment bank’s large clients. Today there are just two equity traders left”
- 19. •일본의 Fukoku 생명보험에서는 보험금 지급 여부를 심사하 는 사람을 30명 이상 해고하고, IBM Watson Explorer 에게 맡기기로 결정 •의료 기록을 바탕으로 Watson이 보험금 지급 여부를 판단 •인공지능으로 교체하여 생산성을 30% 향상 •2년 안에 ROI 가 나올 것이라고 예상 •1년차: 140m yen •2년차: 200m yen
- 22. No choice but to bring AI into the medicine
- 23. Martin Duggan,“IBM Watson Health - Integrated Care & the Evolution to Cognitive Computing”
- 24. •약한 인공 지능 (Artificial Narrow Intelligence) • 특정 방면에서 잘하는 인공지능 • 체스, 퀴즈, 메일 필터링, 상품 추천, 자율 운전 •강한 인공 지능 (Artificial General Intelligence) • 모든 방면에서 인간 급의 인공 지능 • 사고, 계획, 문제해결, 추상화, 복잡한 개념 학습 •초 인공 지능 (Artificial Super Intelligence) • 과학기술, 사회적 능력 등 모든 영역에서 인간보다 뛰어난 인공 지능 • “충분히 발달한 과학은 마법과 구분할 수 없다” - 아서 C. 클라크
- 26. 2010 2020 2030 2040 2050 2060 2070 2080 2090 2100 90% 50% 10% PT-AI AGI EETNTOP100 Combined 언제쯤 기계가 인간 수준의 지능을 획득할 것인가? Philosophy and Theory of AI (2011) Artificial General Intelligence (2012) Greek Association for Artificial Intelligence Survey of most frequently cited 100 authors (2013) Combined 응답자 누적 비율 Superintelligence, Nick Bostrom (2014)
- 27. Superintelligence: Science of fiction? Panelists: Elon Musk (Tesla, SpaceX), Bart Selman (Cornell), Ray Kurzweil (Google), David Chalmers (NYU), Nick Bostrom(FHI), Demis Hassabis (Deep Mind), Stuart Russell (Berkeley), Sam Harris, and Jaan Tallinn (CSER/FLI) January 6-8, 2017, Asilomar, CA https://brunch.co.kr/@kakao-it/49 https://www.youtube.com/watch?v=h0962biiZa4
- 28. Superintelligence: Science of fiction? Panelists: Elon Musk (Tesla, SpaceX), Bart Selman (Cornell), Ray Kurzweil (Google), David Chalmers (NYU), Nick Bostrom(FHI), Demis Hassabis (Deep Mind), Stuart Russell (Berkeley), Sam Harris, and Jaan Tallinn (CSER/FLI) January 6-8, 2017, Asilomar, CA Q: 초인공지능이란 영역은 도달 가능한 것인가? Q: 초지능을 가진 개체의 출현이 가능할 것이라고 생각하는가? Table 1 Elon Musk Start Russell Bart Selman Ray Kurzweil David Chalmers Nick Bostrom DemisHassabis Sam Harris Jaan Tallinn YES YES YES YES YES YES YES YES YES Table 1-1 Elon Musk Start Russell Bart Selman Ray Kurzweil David Chalmers Nick Bostrom DemisHassabis Sam Harris Jaan Tallinn YES YES YES YES YES YES YES YES YES Q: 초지능의 실현이 일어나기를 희망하는가? Table 1-1-1 Elon Musk Start Russell Bart Selman Ray Kurzweil David Chalmers Nick Bostrom DemisHassabis Sam Harris Jaan Tallinn Complicated Complicated Complicated YES Complicated YES YES Complicated Complicated https://brunch.co.kr/@kakao-it/49 https://www.youtube.com/watch?v=h0962biiZa4
- 31. Superintelligence, Nick Bostrom (2014) 일단 인간 수준(human baseline)의 강한 인공지능이 구현되면, 이후 초지능(superintelligence)로 도약(take off)하기까지는 극히 짧은 시간이 걸릴 수 있다. How far to superintelligence
- 32. •약한 인공 지능 (Artificial Narrow Intelligence) • 특정 방면에서 잘하는 인공지능 • 체스, 퀴즈, 메일 필터링, 상품 추천, 자율 운전 •강한 인공 지능 (Artificial General Intelligence) • 모든 방면에서 인간 급의 인공 지능 • 사고, 계획, 문제해결, 추상화, 복잡한 개념 학습 •초 인공 지능 (Artificial Super Intelligence) • 과학기술, 사회적 능력 등 모든 영역에서 인간보다 뛰어난 인공 지능 • “충분히 발달한 과학은 마법과 구분할 수 없다” - 아서 C. 클라크
- 40. •복잡한 의료 데이터의 분석 및 insight 도출 •영상 의료/병리 데이터의 분석/판독 •연속 데이터의 모니터링 및 예방/예측 인공지능의 의료 활용
- 41. •복잡한 의료 데이터의 분석 및 insight 도출 •영상 의료/병리 데이터의 분석/판독 •연속 데이터의 모니터링 및 예방/예측 인공지능의 의료 활용
- 42. Jeopardy! 2011년 인간 챔피언 두 명 과 퀴즈 대결을 벌여서 압도적인 우승을 차지
- 43. IBM Watson on Jeopardy!
- 44. 600,000 pieces of medical evidence 2 million pages of text from 42 medical journals and clinical trials 69 guidelines, 61,540 clinical trials IBM Watson on Medicine Watson learned... + 1,500 lung cancer cases physician notes, lab results and clinical research + 14,700 hours of hands-on training
- 50. Annals of Oncology (2016) 27 (suppl_9): ix179-ix180. 10.1093/annonc/mdw601 Validation study to assess performance of IBM cognitive computing system Watson for oncology with Manipal multidisciplinary tumour board for 1000 consecutive cases: An Indian experience • MMDT(Manipal multidisciplinary tumour board) treatment recommendation and data of 1000 cases of 4 different cancers breast (638), colon (126), rectum (124) and lung (112) which were treated in last 3 years was collected. • Of the treatment recommendations given by MMDT, WFO provided 50% in REC, 28% in FC, 17% in NREC • Nearly 80% of the recommendations were in WFO REC and FC group • 5% of the treatment provided by MMDT was not available with WFO • The degree of concordance varied depending on the type of cancer • WFO-REC was high in Rectum (85%) and least in Lung (17.8%) • high with TNBC (67.9%); HER2 negative (35%) • WFO took a median of 40 sec to capture, analyze and give the treatment. (vs MMDT took the median time of 15 min)
- 51. San Antonio Breast Cancer Symposium—December 6-10, 2016 Concordance WFO (@T2) and MMDT (@T1* v. T2**) (N= 638 Breast Cancer Cases) Time Point /Concordance REC REC + FC n % n % T1* 296 46 463 73 T2** 381 60 574 90 This presentation is the intellectual property of the author/presenter.Contact [email protected] for permission to reprint and/or distribute.26 * T1 Time of original treatment decision by MMDT in the past (last 1-3 years) ** T2 Time (2016) of WFO’s treatment advice and of MMDT’s treatment decision upon blinded re-review of non-concordant cases
- 52. Sung Won Park,APFCP, 2017 Assessing the performance of Watson for Oncology using colon cancer cases treated with surgery and adjuvant chemotherapy at Gachon University Gil Medical Center • Stage II with high risk and stage III colon cancer patients (N=162) • Retrospective study: From September 1, 2014 to August 31, 2016 • Gachon University Gil Medical Center (GMC) • Generally accepted by GMC-recommendation in 83.3% • Concordant with • WFO-Rec: 53.1% • WFO-FC: 30.2% • WFO-NREC: 13.0% • Not included: 3.7%
- 53. WFO in ASCO 2017 • Concordance assessment of a cognitive computing system in Thailand (범룽랏) • 2015-2016 patients 211명 (92명 retrospective; 119 prospective) • Concordance • Overall: 83% • Colorectal 89%, lung 91%, breast 76%, Gastric 78%
- 54. WFO in ASCO 2017 • Early experience with IBM WFO cognitive computing system for lung and colorectal cancer treatment (마니팔 병원) • 지난 3년간: lung cancer(112), colon cancer(126), rectum cancer(124) • lung cancer: localized 88.9%, meta 97.9% • colon cancer: localized 85.5%, meta 76.6% • rectum cancer: localized 96.8%, meta 80.6%
- 55. WFO in ASCO 2017 • Early experience with IBM WFO cognitive computing system for lung and colorectal cancer treatment (마니팔 병원) • 지난 3년간: lung cancer(112), colon cancer(126), rectum cancer(124) • lung cancer: localized 88.9%, meta 97.9% • colon cancer: localized 85.5%, meta 76.6% • rectum cancer: localized 96.8%, meta 80.6% Performance of WFO in India 2017 ASCO annual Meeting, J Clin Oncol 35, 2017 (suppl; abstr 8527)
- 56. WFO in ASCO 2017 •Use of a cognitive computing system for treatment of colon and gastric cancer in South Korea (가천대길병원) • 2012-2016 • 대장암 환자(stage II-IV) 340명 • 진행성 위암 환자 185명 (Retrospective) • Concordance • 대장암 환자 전체 (340명): 73% • 보조 (adjuvant) 항암치료를 받은 250명: 85% • 전이성 환자 90명: 40% • 위암 환자 전체: 49% • Trastzumab/FOLFOX 가 국민 건강 보험 수가를 받지 못함 • S-1(tegafur, gimeracil and oteracil)+cisplatin): • 국내는 매우 루틴; 미국에서는 X
- 57. 잠정적 결론 •왓슨 포 온콜로지와 의사의 일치율: •암종별로 다르다. •같은 암종에서도 병기별로 다르다. •같은 암종에 대해서도 병원별/국가별로 다르다. •시간이 흐름에 따라 달라질 가능성이 있다.
- 58. WHY? •국가별 가이드라인의 차이 • WFO는 기본적으로 MSKCC 기준 • 인종적 차이, 인허가 약물의 차이, 보험 제도의 차이 •NCCN 가이드라인의 업데이트 •암종별 치료 가능한 옵션의 다양성 차이 • 폐암: 다양함 vs 직장암: 다양하지 않음 • TNBC: 다양하지 않음 vs HER2 (-): 다양함
- 59. •WFO 의 정확성을 어떻게 증명해야 할까?
- 60. •WFO 의 정확성을 어떻게 증명해야 할까? 임상시험! vs. Watson for Oncology종양내과 전문의(들) 암환자 10,000명 암환자 10,000명 이러한 임상 시험은 가능할까? • Prospective, single blind randomized trial • Primary endpoint: overall survival (OS) • Secondary endpoint: progression-free survival (PFS)
- 61. •WFO 의 정확성을 어떻게 증명해야 할까? 임상시험! vs. 종양내과 전문의(들) Watson for Oncology •WFO 만으로 환자를 진료하는 것이 비윤리적이다. (신약의 경우 전임상에서 검증) •의사들의 실력이 heterogeneous 하므로 임상시험의 결과가 일반화되기 어렵다. •왓슨은 계속 진화하므로, 과거의 임상시험 결과가 현재에 적용되기 어렵다. 암환자 10,000명 암환자 10,000명 • Prospective, single blind randomized trial • Primary endpoint: overall survival (OS) • Secondary endpoint: progression-free survival (PFS)
- 62. •WFO 의 정확성을 어떻게 증명해야 할까? 임상시험! vs. NCCN 가이드라인 NCCN 가이드라인 + Watson for Oncology 이러한 임상 연구는 가능할 것이나, 결과 예측은 어렵다. •WFO 만으로 환자를 진료하는 것이 비윤리적이다. (신약의 경우 전임상에서 검증) •의사들의 실력이 heterogeneous 하므로 임상시험의 결과가 일반화되기 어렵다. •왓슨은 계속 진화하므로, 과거의 임상시험 결과가 현재에 적용되기 어렵다. •IBM에서 이 임상연구를 진행하기를 과연 원할 것인가? 암환자 10,000명 암환자 10,000명 • Prospective, single blind randomized trial • Primary endpoint: overall survival (OS) • Secondary endpoint: progression-free survival (PFS)
- 63. Meeting with Dr. Kyu Rhee, Chief Health Officer of Watson Health (2017. 7. 4.)
- 64. Four factors to validate clinical benefits of WFO •Patients Outcome •human doctor vs human + WFO •mortality and morbidity •Cost (의료비용 절감) •낮은 재발율 및 재입원율을 통한 비용 절감 효과가 있는가 •Doctor’s Satisfaction (의료진의 만족도) •WFO을 이용함으로써 의사들의 진료 프로세스가 개선되는가? •WFO를 사용하는 의료진의 사용자 경험은 어떠한가? •Patients Satisfaction (환자의 만족도) •환자들이 WFO을 사용하기를 원하는가?
- 65. 원칙이 필요하다 •어떤 환자의 경우, 왓슨에게 의견을 물을 것인가? •왓슨을 (암종별로) 얼마나 신뢰할 것인가? •왓슨의 의견을 환자에게 공개할 것인가? •왓슨과 의료진의 판단이 다른 경우 어떻게 할 것인가? •왓슨에게 보험 급여를 매길 수 있는가? 이러한 기준에 따라 의료의 질/치료효과가 달라질 수 있으나, 현재 개별 병원이 개별적인 기준으로 활용하게 됨
- 66. Watson for Oncology 는 수가를 받을 수 있을까?
- 67. Empowering the Oncology Community for Cancer Care Genomics Oncology Clinical Trial Matching Watson Health’s oncology clients span more than 35 hospital systems “Empowering the Oncology Community for Cancer Care” Andrew Norden, KOTRA Conference, March 2017, “The Future of Health is Cognitive”
- 68. IBM Watson Health Watson for Clinical Trial Matching (CTM) 18 1. According to the National Comprehensive Cancer Network (NCCN) 2. http://csdd.tufts.edu/files/uploads/02_-_jan_15,_2013_-_recruitment-retention.pdf© 2015 International Business Machines Corporation Searching across eligibility criteria of clinical trials is time consuming and labor intensive Current Challenges Fewer than 5% of adult cancer patients participate in clinical trials1 37% of sites fail to meet minimum enrollment targets. 11% of sites fail to enroll a single patient 2 The Watson solution • Uses structured and unstructured patient data to quickly check eligibility across relevant clinical trials • Provides eligible trial considerations ranked by relevance • Increases speed to qualify patients Clinical Investigators (Opportunity) • Trials to Patient: Perform feasibility analysis for a trial • Identify sites with most potential for patient enrollment • Optimize inclusion/exclusion criteria in protocols Faster, more efficient recruitment strategies, better designed protocols Point of Care (Offering) • Patient to Trials: Quickly find the right trial that a patient might be eligible for amongst 100s of open trials available Improve patient care quality, consistency, increased efficiencyIBM Confidential
- 69. •총 16주간 HOG( Highlands Oncology Group)의 폐암과 유방암 환자 2,620명을 대상 •90명의 환자를 3개의 노바티스 유방암 임상 프로토콜에 따라 선별 •임상 시험 코디네이터: 1시간 50분 •Watson CTM: 24분 (78% 시간 단축) •Watson CTM은 임상 시험 기준에 해당되지 않는 환자 94%를 자동으로 스크리닝
- 70. Watson Genomics Overview 20 Watson Genomics Content • 20+ Content Sources Including: • Medical Articles (23Million) • Drug Information • Clinical Trial Information • Genomic Information Case Sequenced VCF / MAF, Log2, Dge Encryption Molecular Profile Analysis Pathway Analysis Drug Analysis Service Analysis, Reports, & Visualizations
- 71. 2017 HIMSS, courtesy of Hyejin Kam (Asan Medical Center)
- 72. 2017 HIMSS, courtesy of Hyejin Kam (Asan Medical Center)
- 74. 2015.10.4.Transforming Medicine, San Diego 의료 데이터 의료 기기
- 75. IBM Watson Health Organizations Leveraging Watson Watson for Oncology Best Doctors (second opinion) Bumrungrad International Hospital Confidential client (Bangladesh and Nepal) Gachon University Gil Medical Center (Korea) Hangzhou Cognitive Care – 50+ Chinese hospitals Jupiter Medical Center Manipal Hospitals – 16 Indian Hospitals MD Anderson (**Oncology Expert Advisor) Memorial Sloan Kettering Cancer Center MRDM - Zorg (Netherlands) Pusan National University Hospital Clinical Trial Matching Best Doctors (second opinion) Confidential – Major Academic Center Highlands Oncology Group Froedtert & Medical College of Wisconsin Mayo Clinic Multiple Life Sciences pilots 24 Watson Genomic Analytics Ann & Robert H Lurie Children’s Hospital of Chicago BC Cancer Agency City of Hope Cleveland Clinic Columbia University, Irwing Cancer Center Duke Cancer Institute Fred & Pamela Buffett Cancer Center Fleury (Brazil) Illumina 170 Gene Panel NIH Japan McDonnell Institute at Washington University in St. Louis New York Genome Center Pusan National University Hospital Quest Diagnostics Stanford Health University of Kansas Cancer Center University of North Carolina Lineberger Cancer Center University of Southern California University of Washington Medical Center University of Tokyo Yale Cancer Center Andrew Norden, KOTRA Conference, March 2017, “The Future of Health is Cognitive”
- 76. 한국에서도 Watson을 볼 수 있을까? 2015.7.9. 서울대학병원
- 78. (Source: Yoon Sup Choi)
- 79. 길병원 인공지능 암센터 다학제진료실 (Source: Stat News)
- 80. 계명대학교 동산병원 (Source: Yoon Sup Choi)
- 81. 대구가톨릭대학병원 (Source: Yoon Sup Choi)
- 82. • 인공지능으로 인한 인간 의사의 권위 약화 • 환자의 자기 결정권 및 권익 증대 • 의사의 진료 방식 및 교육 방식의 변화 필요 http://news.donga.com/3/all/20170320/83400087/1
- 83. • 의사와 Watson의 판단이 다른 경우? • NCCN 가이드라인과 다른 판단을 주기는 것으로 보임 • 100 여명 중에 5 case. • 환자의 판단이 합리적이라고 볼 수 있는가? • Watson의 정확도는 검증되지 않았음 • ‘제 4차 산업혁명’ 등의 buzz word의 영향으로 보임 • 임상 시험이 필요하지 않은가? • 환자들의 선호는 인공지능의 adoption rate 에 영향 • 병원 도입에 영향을 미치는 요인들 • analytical validity • clinical validity/utility • 의사들의 인식/심리적 요인 • 환자들의 인식/심리적 요인 • 규제 환경 (인허가, 수가 등등) • 결국 환자가 원하면 (그것이 의학적으로 타당한지를 떠나서) 병원 도입은 더욱 늘어날 수 밖에 없음
- 84. • Watson에 대한 환자 반응이 생각보다 매우 좋음 • 도입 2개월만에 85명 암 환자 진료 • 기존의 길병원 예측보다는 더 빠른 수치일 듯 • Big5 에서도 길병원으로 전원 문의 증가 한다는 후문 • 교수들이 더 열심히 상의하고 환자 본다고 함
- 85. • 부산대학병원 (2017년 1월) • Watson의 솔루션 두 가지를 도입 • Watson for Oncology • Watson for Genomics
- 86. • 건양대학병원 Watson for Oncology 도입 • 2017년 3월 • “최원준 건양대병원장은 "지역 환자들은 수도권의 여러 병원을 찾아다닐 필요가 없어질 것"이라며 "병원의 우수 한 협진 팀과 인공지능 의료 시스템의 시너지를 바탕으 로 암 환자에게 최상의 의료 서비스를 제공하겠다"고 약 속했다."
- 89. •“향후 10년 동안 첫번째 cardiovascular event 가 올 것인가” 예측 •전향적 코호트 스터디: 영국 환자 378,256 명 •일상적 의료 데이터를 바탕으로 기계학습으로 질병을 예측하는 첫번째 대규모 스터디 •기존의 ACC/AHA 가이드라인과 4가지 기계학습 알고리즘의 정확도를 비교 •Random forest; Logistic regression; Gradient bossting; Neural network
- 90. Can machine-learning improve cardiovascular risk prediction using routine clinical data? Stephen F.Weng et al PLoS One 2017 in a sensitivity of 62.7% and PPV of 17.1%. The random forest algorithm resulted in a net increase of 191 CVD cases from the baseline model, increasing the sensitivity to 65.3% and PPV to 17.8% while logistic regression resulted in a net increase of 324 CVD cases (sensitivity 67.1%; PPV 18.3%). Gradient boosting machines and neural networks performed best, result- ing in a net increase of 354 (sensitivity 67.5%; PPV 18.4%) and 355 CVD (sensitivity 67.5%; PPV 18.4%) cases correctly predicted, respectively. The ACC/AHA baseline model correctly predicted 53,106 non-cases from 75,585 total non- cases, resulting in a specificity of 70.3% and NPV of 95.1%. The net increase in non-cases Table 3. Top 10 risk factor variables for CVD algorithms listed in descending order of coefficient effect size (ACC/AHA; logistic regression), weighting (neural networks), or selection frequency (random forest, gradient boosting machines). Algorithms were derived from training cohort of 295,267 patients. ACC/AHA Algorithm Machine-learning Algorithms Men Women ML: Logistic Regression ML: Random Forest ML: Gradient Boosting Machines ML: Neural Networks Age Age Ethnicity Age Age Atrial Fibrillation Total Cholesterol HDL Cholesterol Age Gender Gender Ethnicity HDL Cholesterol Total Cholesterol SES: Townsend Deprivation Index Ethnicity Ethnicity Oral Corticosteroid Prescribed Smoking Smoking Gender Smoking Smoking Age Age x Total Cholesterol Age x HDL Cholesterol Smoking HDL cholesterol HDL cholesterol Severe Mental Illness Treated Systolic Blood Pressure Age x Total Cholesterol Atrial Fibrillation HbA1c Triglycerides SES: Townsend Deprivation Index Age x Smoking Treated Systolic Blood Pressure Chronic Kidney Disease Triglycerides Total Cholesterol Chronic Kidney Disease Age x HDL Cholesterol Untreated Systolic Blood Pressure Rheumatoid Arthritis SES: Townsend Deprivation Index HbA1c BMI missing Untreated Systolic Blood Pressure Age x Smoking Family history of premature CHD BMI Systolic Blood Pressure Smoking Diabetes Diabetes COPD Total Cholesterol SES: Townsend Deprivation Index Gender Italics: Protective Factors https://doi.org/10.1371/journal.pone.0174944.t003 PLOS ONE | https://doi.org/10.1371/journal.pone.0174944 April 4, 2017 8 / 14 •기존 ACC/AHA 가이드라인의 위험 요소의 일부분만 기계학습 알고리즘에도 포함 •하지만, Diabetes는 네 모델 모두에서 포함되지 않았다. •기존의 위험 예측 툴에는 포함되지 않던, 아래와 같은 새로운 요소들이 포함되었다. •COPD, severe mental illness, prescribing of oral corticosteroids •triglyceride level 등의 바이오 마커
- 91. Can machine-learning improve cardiovascular risk prediction using routine clinical data? Stephen F.Weng et al PLoS One 2017 correctly predicted compared to the baseline ACC/AHA model ranged from 191 non-cases for the random forest algorithm to 355 non-cases for the neural networks. Full details on classifi- cation analysis can be found in S2 Table. Discussion Compared to an established AHA/ACC risk prediction algorithm, we found all machine- learning algorithms tested were better at identifying individuals who will develop CVD and those that will not. Unlike established approaches to risk prediction, the machine-learning methods used were not limited to a small set of risk factors, and incorporated more pre-exist- Table 4. Performance of the machine-learning (ML) algorithms predicting 10-year cardiovascular disease (CVD) risk derived from applying train- ing algorithms on the validation cohort of 82,989 patients. Higher c-statistics results in better algorithm discrimination. The baseline (BL) ACC/AHA 10-year risk prediction algorithm is provided for comparative purposes. Algorithms AUC c-statistic Standard Error* 95% Confidence Interval Absolute Change from Baseline LCL UCL BL: ACC/AHA 0.728 0.002 0.723 0.735 — ML: Random Forest 0.745 0.003 0.739 0.750 +1.7% ML: Logistic Regression 0.760 0.003 0.755 0.766 +3.2% ML: Gradient Boosting Machines 0.761 0.002 0.755 0.766 +3.3% ML: Neural Networks 0.764 0.002 0.759 0.769 +3.6% *Standard error estimated by jack-knife procedure [30] https://doi.org/10.1371/journal.pone.0174944.t004 Can machine-learning improve cardiovascular risk prediction using routine clinical data? •네 가지 기계학습 모델 모두 기존의 ACC/AHA 가이드라인 대비 더 정확했다. •Neural Networks 이 AUC=0.764 로 가장 정확했다. •“이 모델을 활용했더라면 355 명의 추가적인 cardiovascular event 를 예방했을 것” •Deep Learning 을 활용하면 정확도는 더 높아질 수 있을 것 •Genetic information 등의 추가적인 risk factor 를 활용해볼 수 있다.
- 92. 3 output max-pooling convolution -- motif detection embedding sequencing medical record visits/admissions time gaps/transferphrase/admission prediction 1 2 3 4 5 time gap record vector word vector ? prediction point Figure 1. Overview of Deepr for predicting future risk from medical record. Top-left box depicts an example of medical record with multiple visits, each of which has multiple coded objects (diagnosis & procedure). The future risk is unknown (question mark (?)). Steps from-left-to-right: (1) Medical record is sequenced into phrases separated by coded time-gaps/transfers; then from-bottom-to-top: (2) Words are embedded into continuous vectors, (3) local word vectors are convoluted to detect local motifs, (4) max-pooling to derive record-level vector, (5) classifier is applied to predict an output, which is a future event. Best viewed in color. B. Sequencing EMR This task refers to transforming an EMR into a sentence, which is essentially a sequence of words. We present here how the words are defined and arranged in the sentence. procedures are in digits: 1910 Z83 911 1008 D12 K31 1-3m R94 RAREWORD H53 Y83 M62 Y92 E87 T81 RAREWORD RAREWORD 1893 D12 Deepr:A Convolutional Net for Medical Records •“퇴원한 환자가 6개월 이내에 다시 입원할 것인가?” 예측 •EMR 데이터를 바탕으로 예측 •호주의 환자 30만명을 대상으로 검증
- 93. •복잡한 의료 데이터의 분석 및 insight 도출 •영상 의료/병리 데이터의 분석/판독 •연속 데이터의 모니터링 및 예방/예측 인공지능의 의료 활용
- 96. 12 Olga Russakovsky* et al. Fig. 4 Random selection of images in ILSVRC detection validation set. The images in the top 4 rows were taken from ILSVRC2012 single-object localization validation set, and the images in the bottom 4 rows were collected from Flickr using scene-level queries. tage of all the positive examples available. The second is images collected from Flickr specifically for the de- http://arxiv.org/pdf/1409.0575.pdf
- 97. • Main competition • 객체 분류 (Classification): 그림 속의 객체를 분류 • 객체 위치 (localization): 그림 속 ‘하나’의 객체를 분류하고 위치를 파악 • 객체 인식 (object detection): 그림 속 ‘모든’ 객체를 분류하고 위치 파악 16 Olga Russakovsky* et al. Fig. 7 Tasks in ILSVRC. The first column shows the ground truth labeling on an example image, and the next three show three sample outputs with the corresponding evaluation score. http://arxiv.org/pdf/1409.0575.pdf
- 98. Performance of winning entries in the ILSVRC2010-2015 competitions in each of the three tasks http://image-net.org/challenges/LSVRC/2015/results#loc Single-object localization Localizationerror 0 10 20 30 40 50 2011 2012 2013 2014 2015 Object detection Averageprecision 0.0 17.5 35.0 52.5 70.0 2013 2014 2015 Image classification Classificationerror 0 10 20 30 2010 2011 2012 2013 2014 2015
- 100. Kaiming He, Xiangyu Zhang, Shaoqing Ren, Jian Sun, “Deep Residual Learning for Image Recognition”, 2015 How deep is deep?
- 102. Classification http://image-net.org/challenges/LSVRC/2015/results
- 104. DeepFace: Closing the Gap to Human-Level Performance in FaceVerification Taigman,Y. et al. (2014). DeepFace: Closing the Gap to Human-Level Performance in FaceVerification, CVPR’14. Figure 2. Outline of the DeepFace architecture. A front-end of a single convolution-pooling-convolution filtering on the rectified input, followed by three locally-connected layers and two fully-connected layers. Colors illustrate feature maps produced at each layer. The net includes more than 120 million parameters, where more than 95% come from the local and fully connected layers. very few parameters. These layers merely expand the input into a set of simple local features. The subsequent layers (L4, L5 and L6) are instead lo- cally connected [13, 16], like a convolutional layer they ap- ply a filter bank, but every location in the feature map learns a different set of filters. Since different regions of an aligned image have different local statistics, the spatial stationarity The goal of training is to maximize the probability of the correct class (face id). We achieve this by minimiz- ing the cross-entropy loss for each training sample. If k is the index of the true label for a given input, the loss is: L = log pk. The loss is minimized over the parameters by computing the gradient of L w.r.t. the parameters and Human: 95% vs. DeepFace in Facebook: 97.35% Recognition Accuracy for Labeled Faces in the Wild (LFW) dataset (13,233 images, 5,749 people)
- 105. FaceNet:A Unified Embedding for Face Recognition and Clustering Schroff, F. et al. (2015). FaceNet:A Unified Embedding for Face Recognition and Clustering Human: 95% vs. FaceNet of Google: 99.63% Recognition Accuracy for Labeled Faces in the Wild (LFW) dataset (13,233 images, 5,749 people) False accept False reject s. This shows all pairs of images that were on LFW. Only eight of the 13 errors shown he other four are mislabeled in LFW. on Youtube Faces DB ge similarity of all pairs of the first one our face detector detects in each video. False accept False reject Figure 6. LFW errors. This shows all pairs of images that were incorrectly classified on LFW. Only eight of the 13 errors shown here are actual errors the other four are mislabeled in LFW. 5.7. Performance on Youtube Faces DB We use the average similarity of all pairs of the first one hundred frames that our face detector detects in each video. This gives us a classification accuracy of 95.12%±0.39. Using the first one thousand frames results in 95.18%. Compared to [17] 91.4% who also evaluate one hundred frames per video we reduce the error rate by almost half. DeepId2+ [15] achieved 93.2% and our method reduces this error by 30%, comparable to our improvement on LFW. 5.8. Face Clustering Our compact embedding lends itself to be used in order to cluster a users personal photos into groups of people with the same identity. The constraints in assignment imposed by clustering faces, compared to the pure verification task, lead to truly amazing results. Figure 7 shows one cluster in a users personal photo collection, generated using agglom- erative clustering. It is a clear showcase of the incredible invariance to occlusion, lighting, pose and even age. Figure 7. Face Clustering. Shown is an exemplar cluster for one user. All these images in the users personal photo collection were clustered together. 6. Summary We provide a method to directly learn an embedding into an Euclidean space for face verification. This sets it apart from other methods [15, 17] who use the CNN bottleneck layer, or require additional post-processing such as concate- nation of multiple models and PCA, as well as SVM clas- sification. Our end-to-end training both simplifies the setup and shows that directly optimizing a loss relevant to the task at hand improves performance. Another strength of our model is that it only requires False accept False reject Figure 6. LFW errors. This shows all pairs of images that were incorrectly classified on LFW. Only eight of the 13 errors shown here are actual errors the other four are mislabeled in LFW. 5.7. Performance on Youtube Faces DB We use the average similarity of all pairs of the first one hundred frames that our face detector detects in each video. This gives us a classification accuracy of 95.12%±0.39. Using the first one thousand frames results in 95.18%. Compared to [17] 91.4% who also evaluate one hundred frames per video we reduce the error rate by almost half. DeepId2+ [15] achieved 93.2% and our method reduces this error by 30%, comparable to our improvement on LFW. 5.8. Face Clustering Our compact embedding lends itself to be used in order to cluster a users personal photos into groups of people with the same identity. The constraints in assignment imposed by clustering faces, compared to the pure verification task, Figure 7. Face Clustering. Shown is an exemplar cluster for one user. All these images in the users personal photo collection were clustered together. 6. Summary We provide a method to directly learn an embedding into an Euclidean space for face verification. This sets it apart from other methods [15, 17] who use the CNN bottleneck layer, or require additional post-processing such as concate- nation of multiple models and PCA, as well as SVM clas-
- 106. Show and Tell: A Neural Image Caption Generator Vinyals, O. et al. (2015). Show and Tell:A Neural Image Caption Generator, arXiv:1411.4555 v om Samy Bengio Google [email protected] Dumitru Erhan Google [email protected] s a cts his re- m- ed he de- nts A group of people shopping at an outdoor market. ! There are many vegetables at the fruit stand. Vision! Deep CNN Language ! Generating! RNN Figure 1. NIC, our model, is based end-to-end on a neural net- work consisting of a vision CNN followed by a language gener-
- 107. Show and Tell: A Neural Image Caption Generator Vinyals, O. et al. (2015). Show and Tell:A Neural Image Caption Generator, arXiv:1411.4555 Figure 5. A selection of evaluation results, grouped by human rating.
- 108. Radiologist
- 109. Medical Imaging AI Startups by Applications
- 110. Bone Age Assessment • M: 28 Classes • F: 20 Classes • Method: G.P. • Top3-95.28% (F) • Top3-81.55% (M)
- 112. Business Area Medical Image Analysis VUNOnet and our machine learning technology will help doctors and hospitals manage medical scans and images intelligently to make diagnosis faster and more accurately. Original Image Automatic Segmentation EmphysemaNormal ReticularOpacity Our system finds DILDs at the highest accuracy * DILDs: Diffuse Interstitial Lung Disease Digital Radiologist Collaboration with Prof. Joon Beom Seo (Asan Medical Center) Analysed 1200 patients for 3 months
- 113. Digital Radiologist Collaboration with Prof. Joon Beom Seo (Asan Medical Center) Analysed 1200 patients for 3 months
- 114. Digital Radiologist Med Phys. 2013 May;40(5):051912. doi: 10.1118/1.4802214. Collaboration with Prof. Joon Beom Seo (Asan Medical Center) Analysed 1200 patients for 3 months
- 115. Digital Radiologist Med Phys. 2013 May;40(5):051912. doi: 10.1118/1.4802214. Collaboration with Prof. Joon Beom Seo (Asan Medical Center) Analysed 1200 patients for 3 months
- 116. Digital Radiologist Med Phys. 2013 May;40(5):051912. doi: 10.1118/1.4802214. Collaboration with Prof. Joon Beom Seo (Asan Medical Center) Analysed 1200 patients for 3 months Feature Engineering vs Feature Learning alization of Hand-crafted Feature vs Learned Feature in 2D Feature Engineering vs Feature Learning • Visualization of Hand-crafted Feature vs Learned Feature in 2D Visualization of Hand-crafted Feature vs Learned Feature in 2D
- 117. Bench to Bedside : Practical Applications • Contents-based Case Retrieval –Finding similar cases with the clinically matching context - Search engine for medical images. –Clinicians can refer the diagnosis, prognosis of past similar patients to make better clinical decision. –Accepted to present at RSNA 2017 Digital Radiologist
- 118. https://www.nih.gov/news-events/news-releases/nih-clinical-center-provides-one-largest-publicly-available-chest-x-ray-datasets-scientific-community •112,120 장의 질병군이 포함된 chest frontal 영상 •그 중에 984장에 대한 bounding box 정보 포함 •14개의 질환 카테고리 •ROI, CT confirmed result, Pathological result 많지 않음
- 119. •Zebra Medical Vision에서 $1 에 영상의학데이터를 판독해주는 서비스를 런칭 (2017년 10월) •항목은 확정되지는 않았으나, Pulmonary Hypertension, Lung Nodule, Fatty Liver, Emphysema, Coronary Calcium Scoring, Bone Mineral Density, Aortic Aneurysm 등으로 예상 https://www.zebra-med.com/aione/
- 120. Zebra Medical Vision’s AI1: AI at Your Fingertips https://www.youtube.com/watch?v=0PGgCpXa-Fs
- 121. Detection of Diabetic Retinopathy
- 122. 당뇨성 망막병증 • 당뇨병의 대표적 합병증: 당뇨병력이 30년 이상 환자 90% 발병 • 안과 전문의들이 안저(안구의 안쪽)를 사진으로 찍어서 판독 • 망막 내 미세혈관 생성, 출혈, 삼출물 정도를 파악하여 진단
- 123. Copyright 2016 American Medical Association. All rights reserved. Development and Validation of a Deep Learning Algorithm for Detection of Diabetic Retinopathy in Retinal Fundus Photographs Varun Gulshan, PhD; Lily Peng, MD, PhD; Marc Coram, PhD; Martin C. Stumpe, PhD; Derek Wu, BS; Arunachalam Narayanaswamy, PhD; Subhashini Venugopalan, MS; Kasumi Widner, MS; Tom Madams, MEng; Jorge Cuadros, OD, PhD; Ramasamy Kim, OD, DNB; Rajiv Raman, MS, DNB; Philip C. Nelson, BS; Jessica L. Mega, MD, MPH; Dale R. Webster, PhD IMPORTANCE Deep learning is a family of computational methods that allow an algorithm to program itself by learning from a large set of examples that demonstrate the desired behavior, removing the need to specify rules explicitly. Application of these methods to medical imaging requires further assessment and validation. OBJECTIVE To apply deep learning to create an algorithm for automated detection of diabetic retinopathy and diabetic macular edema in retinal fundus photographs. DESIGN AND SETTING A specific type of neural network optimized for image classification called a deep convolutional neural network was trained using a retrospective development data set of 128 175 retinal images, which were graded 3 to 7 times for diabetic retinopathy, diabetic macular edema, and image gradability by a panel of 54 US licensed ophthalmologists and ophthalmology senior residents between May and December 2015. The resultant algorithm was validated in January and February 2016 using 2 separate data sets, both graded by at least 7 US board-certified ophthalmologists with high intragrader consistency. EXPOSURE Deep learning–trained algorithm. MAIN OUTCOMES AND MEASURES The sensitivity and specificity of the algorithm for detecting referable diabetic retinopathy (RDR), defined as moderate and worse diabetic retinopathy, referable diabetic macular edema, or both, were generated based on the reference standard of the majority decision of the ophthalmologist panel. The algorithm was evaluated at 2 operating points selected from the development set, one selected for high specificity and another for high sensitivity. RESULTS TheEyePACS-1datasetconsistedof9963imagesfrom4997patients(meanage,54.4 years;62.2%women;prevalenceofRDR,683/8878fullygradableimages[7.8%]);the Messidor-2datasethad1748imagesfrom874patients(meanage,57.6years;42.6%women; prevalenceofRDR,254/1745fullygradableimages[14.6%]).FordetectingRDR,thealgorithm hadanareaunderthereceiveroperatingcurveof0.991(95%CI,0.988-0.993)forEyePACS-1and 0.990(95%CI,0.986-0.995)forMessidor-2.Usingthefirstoperatingcutpointwithhigh specificity,forEyePACS-1,thesensitivitywas90.3%(95%CI,87.5%-92.7%)andthespecificity was98.1%(95%CI,97.8%-98.5%).ForMessidor-2,thesensitivitywas87.0%(95%CI,81.1%- 91.0%)andthespecificitywas98.5%(95%CI,97.7%-99.1%).Usingasecondoperatingpoint withhighsensitivityinthedevelopmentset,forEyePACS-1thesensitivitywas97.5%and specificitywas93.4%andforMessidor-2thesensitivitywas96.1%andspecificitywas93.9%. CONCLUSIONS AND RELEVANCE In this evaluation of retinal fundus photographs from adults with diabetes, an algorithm based on deep machine learning had high sensitivity and specificity for detecting referable diabetic retinopathy. Further research is necessary to determine the feasibility of applying this algorithm in the clinical setting and to determine whether use of the algorithm could lead to improved care and outcomes compared with current ophthalmologic assessment. JAMA. doi:10.1001/jama.2016.17216 Published online November 29, 2016. Editorial Supplemental content Author Affiliations: Google Inc, Mountain View, California (Gulshan, Peng, Coram, Stumpe, Wu, Narayanaswamy, Venugopalan, Widner, Madams, Nelson, Webster); Department of Computer Science, University of Texas, Austin (Venugopalan); EyePACS LLC, San Jose, California (Cuadros); School of Optometry, Vision Science Graduate Group, University of California, Berkeley (Cuadros); Aravind Medical Research Foundation, Aravind Eye Care System, Madurai, India (Kim); Shri Bhagwan Mahavir Vitreoretinal Services, Sankara Nethralaya, Chennai, Tamil Nadu, India (Raman); Verily Life Sciences, Mountain View, California (Mega); Cardiovascular Division, Department of Medicine, Brigham and Women’s Hospital and Harvard Medical School, Boston, Massachusetts (Mega). Corresponding Author: Lily Peng, MD, PhD, Google Research, 1600 Amphitheatre Way, Mountain View, CA 94043 ([email protected]). Research JAMA | Original Investigation | INNOVATIONS IN HEALTH CARE DELIVERY (Reprinted) E1 Copyright 2016 American Medical Association. All rights reserved.
- 124. Case Study: TensorFlow in Medicine - Retinal Imaging (TensorFlow Dev Summit 2017)
- 126. Training Set / Test Set • CNN으로 후향적으로 128,175개의 안저 이미지 학습 • 미국의 안과전문의 54명이 3-7회 판독한 데이터 • 우수한 안과전문의들 7-8명의 판독 결과와 인공지능의 판독 결과 비교 • EyePACS-1 (9,963 개), Messidor-2 (1,748 개)a) Fullscreen mode b) Hit reset to reload this image. This will reset all of the grading. c) Comment box for other pathologies you see eFigure 2. Screenshot of the Second Screen of the Grading Tool, Which Asks Graders to Assess the Image for DR, DME and Other Notable Conditions or Findings
- 127. • EyePACS-1 과 Messidor-2 의 AUC = 0.991, 0.990 • 7-8명의 안과 전문의와 sensitivity, specificity 가 동일한 수준 • F-score: 0.95 (vs. 인간 의사는 0.91) Additional sensitivity analyses were conducted for sev- eralsubcategories:(1)detectingmoderateorworsediabeticreti- effects of data set size on algorithm performance were exam- ined and shown to plateau at around 60 000 images (or ap- Figure 2. Validation Set Performance for Referable Diabetic Retinopathy 100 80 60 40 20 0 0 70 80 85 95 90 75 0 5 10 15 20 25 30 100806040 Sensitivity,% 1 – Specificity, % 20 EyePACS-1: AUC, 99.1%; 95% CI, 98.8%-99.3%A 100 High-sensitivity operating point High-specificity operating point 100 80 60 40 20 0 0 70 80 85 95 90 75 0 5 10 15 20 25 30 100806040 Sensitivity,% 1 – Specificity, % 20 Messidor-2: AUC, 99.0%; 95% CI, 98.6%-99.5%B 100 High-specificity operating point High-sensitivity operating point Performance of the algorithm (black curve) and ophthalmologists (colored circles) for the presence of referable diabetic retinopathy (moderate or worse diabetic retinopathy or referable diabetic macular edema) on A, EyePACS-1 (8788 fully gradable images) and B, Messidor-2 (1745 fully gradable images). The black diamonds on the graph correspond to the sensitivity and specificity of the algorithm at the high-sensitivity and high-specificity operating points. In A, for the high-sensitivity operating point, specificity was 93.4% (95% CI, 92.8%-94.0%) and sensitivity was 97.5% (95% CI, 95.8%-98.7%); for the high-specificity operating point, specificity was 98.1% (95% CI, 97.8%-98.5%) and sensitivity was 90.3% (95% CI, 87.5%-92.7%). In B, for the high-sensitivity operating point, specificity was 93.9% (95% CI, 92.4%-95.3%) and sensitivity was 96.1% (95% CI, 92.4%-98.3%); for the high-specificity operating point, specificity was 98.5% (95% CI, 97.7%-99.1%) and sensitivity was 87.0% (95% CI, 81.1%-91.0%). There were 8 ophthalmologists who graded EyePACS-1 and 7 ophthalmologists who graded Messidor-2. AUC indicates area under the receiver operating characteristic curve. Research Original Investigation Accuracy of a Deep Learning Algorithm for Detection of Diabetic Retinopathy Results
- 128. Skin Cancer
- 129. ABCDE checklist
- 130. 0 0 M O N T H 2 0 1 7 | V O L 0 0 0 | N A T U R E | 1 LETTER doi:10.1038/nature21056 Dermatologist-level classification of skin cancer with deep neural networks Andre Esteva1 *, Brett Kuprel1 *, Roberto A. Novoa2,3 , Justin Ko2 , Susan M. Swetter2,4 , Helen M. Blau5 & Sebastian Thrun6 Skin cancer, the most common human malignancy1–3 , is primarily diagnosed visually, beginning with an initial clinical screening and followed potentially by dermoscopic analysis, a biopsy and histopathological examination. Automated classification of skin lesions using images is a challenging task owing to the fine-grained variability in the appearance of skin lesions. Deep convolutional neural networks (CNNs)4,5 show potential for general and highly variable tasks across many fine-grained object categories6–11 . Here we demonstrate classification of skin lesions using a single CNN, trained end-to-end from images directly, using only pixels and disease labels as inputs. We train a CNN using a dataset of 129,450 clinical images—two orders of magnitude larger than previous datasets12 —consisting of 2,032 different diseases. We test its performance against 21 board-certified dermatologists on biopsy-proven clinical images with two critical binary classification use cases: keratinocyte carcinomas versus benign seborrheic keratoses; and malignant melanomas versus benign nevi. The first case represents the identification of the most common cancers, the second represents the identification of the deadliest skin cancer. The CNN achieves performance on par with all tested experts across both tasks, demonstrating an artificial intelligence capable of classifying skin cancer with a level of competence comparable to dermatologists. Outfitted with deep neural networks, mobile devices can potentially extend the reach of dermatologists outside of the clinic. It is projected that 6.3 billion smartphone subscriptions will exist by the year 2021 (ref. 13) and can therefore potentially provide low-cost universal access to vital diagnostic care. There are 5.4 million new cases of skin cancer in the United States2 every year. One in five Americans will be diagnosed with a cutaneous malignancy in their lifetime. Although melanomas represent fewer than 5% of all skin cancers in the United States, they account for approxi- mately 75% of all skin-cancer-related deaths, and are responsible for over 10,000 deaths annually in the United States alone. Early detection is critical, as the estimated 5-year survival rate for melanoma drops from over 99% if detected in its earliest stages to about 14% if detected in its latest stages. We developed a computational method which may allow medical practitioners and patients to proactively track skin lesions and detect cancer earlier. By creating a novel disease taxonomy, and a disease-partitioning algorithm that maps individual diseases into training classes, we are able to build a deep learning system for auto- mated dermatology. Previous work in dermatological computer-aided classification12,14,15 has lacked the generalization capability of medical practitioners owing to insufficient data and a focus on standardized tasks such as dermoscopy16–18 and histological image classification19–22 . Dermoscopy images are acquired via a specialized instrument and histological images are acquired via invasive biopsy and microscopy; whereby both modalities yield highly standardized images. Photographic images (for example, smartphone images) exhibit variability in factors such as zoom, angle and lighting, making classification substantially more challenging23,24 . We overcome this challenge by using a data- driven approach—1.41 million pre-training and training images make classification robust to photographic variability. Many previous techniques require extensive preprocessing, lesion segmentation and extraction of domain-specific visual features before classification. By contrast, our system requires no hand-crafted features; it is trained end-to-end directly from image labels and raw pixels, with a single network for both photographic and dermoscopic images. The existing body of work uses small datasets of typically less than a thousand images of skin lesions16,18,19 , which, as a result, do not generalize well to new images. We demonstrate generalizable classification with a new dermatologist-labelled dataset of 129,450 clinical images, including 3,374 dermoscopy images. Deep learning algorithms, powered by advances in computation and very large datasets25 , have recently been shown to exceed human performance in visual tasks such as playing Atari games26 , strategic board games like Go27 and object recognition6 . In this paper we outline the development of a CNN that matches the performance of dermatologists at three key diagnostic tasks: melanoma classification, melanoma classification using dermoscopy and carcinoma classification. We restrict the comparisons to image-based classification. We utilize a GoogleNet Inception v3 CNN architecture9 that was pre- trained on approximately 1.28 million images (1,000 object categories) from the 2014 ImageNet Large Scale Visual Recognition Challenge6 , and train it on our dataset using transfer learning28 . Figure 1 shows the working system. The CNN is trained using 757 disease classes. Our dataset is composed of dermatologist-labelled images organized in a tree-structured taxonomy of 2,032 diseases, in which the individual diseases form the leaf nodes. The images come from 18 different clinician-curated, open-access online repositories, as well as from clinical data from Stanford University Medical Center. Figure 2a shows a subset of the full taxonomy, which has been organized clinically and visually by medical experts. We split our dataset into 127,463 training and validation images and 1,942 biopsy-labelled test images. To take advantage of fine-grained information contained within the taxonomy structure, we develop an algorithm (Extended Data Table 1) to partition diseases into fine-grained training classes (for example, amelanotic melanoma and acrolentiginous melanoma). During inference, the CNN outputs a probability distribution over these fine classes. To recover the probabilities for coarser-level classes of interest (for example, melanoma) we sum the probabilities of their descendants (see Methods and Extended Data Fig. 1 for more details). We validate the effectiveness of the algorithm in two ways, using nine-fold cross-validation. First, we validate the algorithm using a three-class disease partition—the first-level nodes of the taxonomy, which represent benign lesions, malignant lesions and non-neoplastic 1 Department of Electrical Engineering, Stanford University, Stanford, California, USA. 2 Department of Dermatology, Stanford University, Stanford, California, USA. 3 Department of Pathology, Stanford University, Stanford, California, USA. 4 Dermatology Service, Veterans Affairs Palo Alto Health Care System, Palo Alto, California, USA. 5 Baxter Laboratory for Stem Cell Biology, Department of Microbiology and Immunology, Institute for Stem Cell Biology and Regenerative Medicine, Stanford University, Stanford, California, USA. 6 Department of Computer Science, Stanford University, Stanford, California, USA. *These authors contributed equally to this work. © 2017 Macmillan Publishers Limited, part of Springer Nature. All rights reserved.
- 131. LETTERH his task, the CNN achieves 72.1±0.9% (mean±s.d.) overall he average of individual inference class accuracies) and two gists attain 65.56% and 66.0% accuracy on a subset of the set. Second, we validate the algorithm using a nine-class rtition—the second-level nodes—so that the diseases of have similar medical treatment plans. The CNN achieves two trials, one using standard images and the other using images, which reflect the two steps that a dermatologist m to obtain a clinical impression. The same CNN is used for a Figure 2b shows a few example images, demonstrating th distinguishing between malignant and benign lesions, whic visual features. Our comparison metrics are sensitivity an Acral-lentiginous melanoma Amelanotic melanoma Lentigo melanoma … Blue nevus Halo nevus Mongolian spot … Training classes (757)Deep convolutional neural network (Inception v3) Inference classes (varies by task) 92% malignant melanocytic lesion 8% benign melanocytic lesion Skin lesion image Convolution AvgPool MaxPool Concat Dropout Fully connected Softmax Deep CNN layout. Our classification technique is a Data flow is from left to right: an image of a skin lesion e, melanoma) is sequentially warped into a probability over clinical classes of skin disease using Google Inception hitecture pretrained on the ImageNet dataset (1.28 million 1,000 generic object classes) and fine-tuned on our own 29,450 skin lesions comprising 2,032 different diseases. ning classes are defined using a novel taxonomy of skin disease oning algorithm that maps diseases into training classes (for example, acrolentiginous melanoma, amelanotic melano melanoma). Inference classes are more general and are comp or more training classes (for example, malignant melanocytic class of melanomas). The probability of an inference class is c summing the probabilities of the training classes according to structure (see Methods). Inception v3 CNN architecture repr from https://research.googleblog.com/2016/03/train-your-ow classifier-with.html GoogleNet Inception v3 • 129,450개의 피부과 병변 이미지 데이터를 자체 제작 • 미국의 피부과 전문의 18명이 데이터 curation • CNN (Inception v3)으로 이미지를 학습 • 피부과 전문의들 21명과 인공지능의 판독 결과 비교 • 표피세포 암 (keratinocyte carcinoma)과 지루각화증(benign seborrheic keratosis)의 구분 • 악성 흑색종과 양성 병변 구분 (표준 이미지 데이터 기반) • 악성 흑색종과 양성 병변 구분 (더마토스코프로 찍은 이미지 기반)
- 132. Skin cancer classification performance of the CNN and dermatologists. LETT a b 0 1 Sensitivity 0 1 Specificity Melanoma: 130 images 0 1 Sensitivity 0 1 Specificity Melanoma: 225 images Algorithm: AUC = 0.96 0 1 Sensitivity 0 1 Specificity Melanoma: 111 dermoscopy images 0 1 Sensitivity 0 1 Specificity Carcinoma: 707 images Algorithm: AUC = 0.96 0 1 Sensitivity 0 1 Specificity Melanoma: 1,010 dermoscopy images Algorithm: AUC = 0.94 0 1 Sensitivity 0 1 Specificity Carcinoma: 135 images Algorithm: AUC = 0.96 Dermatologists (25) Average dermatologist Algorithm: AUC = 0.94 Dermatologists (22) Average dermatologist Algorithm: AUC = 0.91 Dermatologists (21) Average dermatologist cancer classification performance of the CNN and 21명 중에 인공지능보다 정확성이 떨어지는 피부과 전문의들이 상당수 있었음 피부과 전문의들의 평균 성적도 인공지능보다 좋지 않았음
- 133. Skin cancer classification performance of the CNN and dermatologists. LETT a b 0 1 Sensitivity 0 1 Specificity Melanoma: 130 images 0 1 Sensitivity 0 1 Specificity Melanoma: 225 images Algorithm: AUC = 0.96 0 1 Sensitivity 0 1 Specificity Melanoma: 111 dermoscopy images 0 1 Sensitivity 0 1 Specificity Carcinoma: 707 images Algorithm: AUC = 0.96 0 1 Sensitivity 0 1 Specificity Melanoma: 1,010 dermoscopy images Algorithm: AUC = 0.94 0 1 Sensitivity 0 1 Specificity Carcinoma: 135 images Algorithm: AUC = 0.96 Dermatologists (25) Average dermatologist Algorithm: AUC = 0.94 Dermatologists (22) Average dermatologist Algorithm: AUC = 0.91 Dermatologists (21) Average dermatologist cancer classification performance of the CNN and
- 134. Skin Cancer Image Classification (TensorFlow Dev Summit 2017) Skin cancer classification performance of the CNN and dermatologists. https://www.youtube.com/watch?v=toK1OSLep3s&t=419s
- 135. Digital Pathologist
- 136. Diagnostic Concordance Among Pathologists Interpreting Breast Biopsy Specimens A B DC Benign without atypia / Atypic / DCIS (ductal carcinoma in situ) / Invasive Carcinoma Interpretation? Elmore etl al. JAMA 2015
- 137. Figure 4. Participating Pathologists’ Interpretations of Each of the 240 Breast Biopsy Test Cases 0 25 50 75 100 Interpretations, % 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 34 36 38 40 42 44 46 48 50 52 54 56 58 60 62 64 66 68 70 72 Case Benign without atypia 72 Cases 2070 Total interpretations A 0 25 50 75 100 Interpretations, % 218 220 222 224 226 228 230 232 234 236 238 240 Case Invasive carcinoma 23 Cases 663 Total interpretations D 0 25 50 75 100 Interpretations, % 147 145 149 151 153 155 157 159 161 163 165 167 169 171 173 175 177 179 181 183 185 187 189 191 193 195 197 199 201 203 205 207 209 211 213 215 217 Case DCIS 73 Cases 2097 Total interpretations C 0 25 50 75 100 Interpretations, % 74 76 78 80 82 84 86 88 90 92 94 96 98 100 102 104 106 108 110 112 114 116 118 120 122 124 126 128 130 132 134 136 138 140 142 144 Case Atypia 72 Cases 2070 Total interpretations B Benign without atypia Atypia DCIS Invasive carcinoma Pathologist interpretation DCIS indicates ductal carcinoma in situ. Diagnostic Concordance in Interpreting Breast Biopsies Original Investigation Research Elmore etl al. JAMA 2015 Diagnostic Concordance Among Pathologists Interpreting Breast Biopsy Specimens
- 138. Figure 4. Participating Pathologists’ Interpretations of Each of the 240 Breast Biopsy Test Cases 0 25 50 75 100 Interpretations, % 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 34 36 38 40 42 44 46 48 50 52 54 56 58 60 62 64 66 68 70 72 Case Benign without atypia 72 Cases 2070 Total interpretations A 0 25 50 75 100 Interpretations, % 218 220 222 224 226 228 230 232 234 236 238 240 Case Invasive carcinoma 23 Cases 663 Total interpretations D 0 25 50 75 100 Interpretations, % 147 145 149 151 153 155 157 159 161 163 165 167 169 171 173 175 177 179 181 183 185 187 189 191 193 195 197 199 201 203 205 207 209 211 213 215 217 Case DCIS 73 Cases 2097 Total interpretations C 0 25 50 75 100 Interpretations, % 74 76 78 80 82 84 86 88 90 92 94 96 98 100 102 104 106 108 110 112 114 116 118 120 122 124 126 128 130 132 134 136 138 140 142 144 Case Atypia 72 Cases 2070 Total interpretations B Benign without atypia Atypia DCIS Invasive carcinoma Pathologist interpretation DCIS indicates ductal carcinoma in situ. Diagnostic Concordance in Interpreting Breast Biopsies Original Investigation Research Elmore etl al. JAMA 2015 Diagnostic Concordance Among Pathologists Interpreting Breast Biopsy Specimens
- 139. Elmore etl al. JAMA 2015 Diagnostic Concordance Among Pathologists Interpreting Breast Biopsy Specimens • Concordance noted in 5194 of 6900 case interpretations or 75.3%. • Reference diagnosis was obtained from consensus of 3 experienced breast pathologists. spentonthisactivitywas16(95%CI,15-17);43participantswere awarded the maximum 20 hours. Pathologists’ Diagnoses Compared With Consensus-Derived Reference Diagnoses The 115 participants each interpreted 60 cases, providing 6900 total individual interpretations for comparison with the con- sensus-derived reference diagnoses (Figure 3). Participants agreed with the consensus-derived reference diagnosis for 75.3% of the interpretations (95% CI, 73.4%-77.0%). Partici- pants (n = 94) who completed the CME activity reported that Patient and Pathologist Characteristics Associated With Overinterpretation and Underinterpretation The association of breast density with overall pathologists’ concordance (as well as both overinterpretation and under- interpretation rates) was statistically significant, as shown in Table 3 when comparing mammographic density grouped into 2 categories (low density vs high density). The overall concordance estimates also decreased consistently with increasing breast density across all 4 Breast Imaging- Reporting and Data System (BI-RADS) density categories: BI-RADS A, 81% (95% CI, 75%-86%); BI-RADS B, 77% (95% Figure 3. Comparison of 115 Participating Pathologists’ Interpretations vs the Consensus-Derived Reference Diagnosis for 6900 Total Case Interpretationsa Participating Pathologists’ Interpretation ConsensusReference Diagnosisb Benign without atypia Atypia DCIS Invasive carcinoma Total Benign without atypia 1803 200 46 21 2070 Atypia 719 990 353 8 2070 DCIS 133 146 1764 54 2097 Invasive carcinoma 3 0 23 637 663 Total 2658 1336 2186 720 6900 DCIS indicates ductal carcinoma in situ. a Concordance noted in 5194 of 6900 case interpretations or 75.3%. b Reference diagnosis was obtained from consensus of 3 experienced breast pathologists. Diagnostic Concordance in Interpreting Breast Biopsies Original Investigation Research Comparison of 115 Participating Pathologists’ Interpretations vs the Consensus-Derived Reference Diagnosis for 6900 Total Case Interpretations
- 140. Constructing higher-level contextual/relational features: Relationships between epithelial nuclear neighbors Relationships between morphologically regular and irregular nuclei Relationships between epithelial and stromal objects Relationships between epithelial nuclei and cytoplasm Characteristics of stromal nuclei and stromal matrix Characteristics of epithelial nuclei and epithelial cytoplasm Building an epithelial/stromal classifier: Epithelial vs.stroma classifier Epithelial vs.stroma classifier B Basic image processing and feature construction: H&E image Image broken into superpixels Nuclei identified within each superpixel A Relationships of contiguous epithelial regions with underlying nuclear objects Learning an image-based model to predict survival Processed images from patients Processed images from patients C D onNovember17,2011stm.sciencemag.orgwnloadedfrom TMAs contain 0.6-mm-diameter cores (median of two cores per case) that represent only a small sample of the full tumor. We acquired data from two separate and independent cohorts: Nether- lands Cancer Institute (NKI; 248 patients) and Vancouver General Hospital (VGH; 328 patients). Unlike previous work in cancer morphom- etry (18–21), our image analysis pipeline was not limited to a predefined set of morphometric features selected by pathologists. Rather, C-Path measures an extensive, quantitative feature set from the breast cancer epithelium and the stro- ma (Fig. 1). Our image processing system first performed an automated, hierarchical scene seg- mentation that generated thousands of measure- ments, including both standard morphometric descriptors of image objects and higher-level contextual, relational, and global image features. The pipeline consisted of three stages (Fig. 1, A to C, and tables S8 and S9). First, we used a set of processing steps to separate the tissue from the background, partition the image into small regions of coherent appearance known as superpixels, find nuclei within the superpixels, and construct Constructing higher-level contextual/relational features: Relationships between epithelial nuclear neighbors Relationships between morphologically regular and irregular nuclei Relationships between epithelial and stromal objects Relationships between epithelial nuclei and cytoplasm Characteristics of stromal nuclei and stromal matrix Characteristics of epithelial nuclei and epithelial cytoplasm Epithelial vs.stroma classifier Epithelial vs.stroma classifier Relationships of contiguous epithelial regions with underlying nuclear objects Learning an image-based model to predict survival Processed images from patients alive at 5 years Processed images from patients deceased at 5 years L1-regularized logisticregression modelbuilding 5YS predictive model Unlabeled images Time P(survival) C D Identification of novel prognostically important morphologic features basic cellular morphologic properties (epithelial reg- ular nuclei = red; epithelial atypical nuclei = pale blue; epithelial cytoplasm = purple; stromal matrix = green; stromal round nuclei = dark green; stromal spindled nuclei = teal blue; unclassified regions = dark gray; spindled nuclei in unclassified regions = yellow; round nuclei in unclassified regions = gray; background = white). (Left panel) After the classification of each image object, a rich feature set is constructed. (D) Learning an image-based model to predict survival. Processed images from patients alive at 5 years after surgery and from patients deceased at 5 years after surgery were used to construct an image-based prog- nostic model. After construction of the model, it was applied to a test set of breast cancer images (not used in model building) to classify patients as high or low risk of death by 5 years. www.ScienceTranslationalMedicine.org 9 November 2011 Vol 3 Issue 108 108ra113 2 onNovember17,2011stm.sciencemag.orgDownloadedfrom Digital Pathologist Sci Transl Med. 2011 Nov 9;3(108):108ra113
- 141. Digital Pathologist Sci Transl Med. 2011 Nov 9;3(108):108ra113 Top stromal features associated with survival. primarily characterizing epithelial nuclear characteristics, such as size, color, and texture (21, 36). In contrast, after initial filtering of im- ages to ensure high-quality TMA images and training of the C-Path models using expert-derived image annotations (epithelium and stroma labels to build the epithelial-stromal classifier and survival time and survival status to build the prognostic model), our image analysis system is automated with no manual steps, which greatly in- creases its scalability. Additionally, in contrast to previous approaches, our system measures thousands of morphologic descriptors of diverse identification of prognostic features whose significance was not pre- viously recognized. Using our system, we built an image-based prognostic model on the NKI data set and showed that in this patient cohort the model was a strong predictor of survival and provided significant additional prognostic information to clinical, molecular, and pathological prog- nostic factors in a multivariate model. We also demonstrated that the image-based prognostic model, built using the NKI data set, is a strong prognostic factor on another, independent data set with very different SD of the ratio of the pixel intensity SD to the mean intensity for pixels within a ring of the center of epithelial nuclei A The sum of the number of unclassified objects SD of the maximum blue pixel value for atypical epithelial nuclei Maximum distance between atypical epithelial nuclei B C D Maximum value of the minimum green pixel intensity value in epithelial contiguous regions Minimum elliptic fit of epithelial contiguous regions SD of distance between epithelial cytoplasmic and nuclear objects Average border between epithelial cytoplasmic objects E F G H Fig. 5. Top epithelial features. The eight panels in the figure (A to H) each shows one of the top-ranking epithelial features from the bootstrap anal- ysis. Left panels, improved prognosis; right panels, worse prognosis. (A) SD of the (SD of intensity/mean intensity) for pixels within a ring of the center of epithelial nuclei. Left, relatively consistent nuclear intensity pattern (low score); right, great nuclear intensity diversity (high score). (B) Sum of the number of unclassified objects. Red, epithelial regions; green, stromal re- gions; no overlaid color, unclassified region. Left, few unclassified objects (low score); right, higher number of unclassified objects (high score). (C) SD of the maximum blue pixel value for atypical epithelial nuclei. Left, high score; right, low score. (D) Maximum distance between atypical epithe- lial nuclei. Left, high score; right, low score. (Insets) Red, atypical epithelial nuclei; black, typical epithelial nuclei. (E) Minimum elliptic fit of epithelial contiguous regions. Left, high score; right, low score. (F) SD of distance between epithelial cytoplasmic and nuclear objects. Left, high score; right, low score. (G) Average border between epithelial cytoplasmic objects. Left, high score; right, low score. (H) Maximum value of the minimum green pixel intensity value in epithelial contiguous regions. Left, low score indi- cating black pixels within epithelial region; right, higher score indicating presence of epithelial regions lacking black pixels. onNovember17,2011stm.sciencemag.orgDownloadedfrom and stromal matrix throughout the image, with thin cords of epithe- lial cells infiltrating through stroma across the image, so that each stromal matrix region borders a relatively constant proportion of ep- ithelial and stromal regions. The stromal feature with the second largest coefficient (Fig. 4B) was the sum of the minimum green in- tensity value of stromal-contiguous regions. This feature received a value of zero when stromal regions contained dark pixels (such as inflammatory nuclei). The feature received a positive value when stromal objects were devoid of dark pixels. This feature provided in- formation about the relationship between stromal cellular composi- tion and prognosis and suggested that the presence of inflammatory cells in the stroma is associated with poor prognosis, a finding con- sistent with previous observations (32). The third most significant stromal feature (Fig. 4C) was a measure of the relative border between spindled stromal nuclei to round stromal nuclei, with an increased rel- ative border of spindled stromal nuclei to round stromal nuclei asso- ciated with worse overall survival. Although the biological underpinning of this morphologic feature is currently not known, this analysis sug- gested that spatial relationships between different populations of stro- mal cell types are associated with breast cancer progression. Reproducibility of C-Path 5YS model predictions on samples with multiple TMA cores For the C-Path 5YS model (which was trained on the full NKI data set), we assessed the intrapatient agreement of model predictions when predictions were made separately on each image contributed by pa- tients in the VGH data set. For the 190 VGH patients who contributed two images with complete image data, the binary predictions (high or low risk) on the individual images agreed with each other for 69% (131 of 190) of the cases and agreed with the prediction on the aver- aged data for 84% (319 of 380) of the images. Using the continuous prediction score (which ranged from 0 to 100), the median of the ab- solute difference in prediction score among the patients with replicate images was 5%, and the Spearman correlation among replicates was 0.27 (P = 0.0002) (fig. S3). This degree of intrapatient agreement is only moderate, and these findings suggest significant intrapatient tumor heterogeneity, which is a cardinal feature of breast carcinomas (33–35). Qualitative visual inspection of images receiving discordant scores suggested that intrapatient variability in both the epithelial and the stromal components is likely to contribute to discordant scores for the individual images. These differences appeared to relate both to the proportions of the epithelium and stroma and to the appearance of the epithelium and stroma. Last, we sought to analyze whether sur- vival predictions were more accurate on the VGH cases that contributed multiple cores compared to the cases that contributed only a single core. This analysis showed that the C-Path 5YS model showed signif- icantly improved prognostic prediction accuracy on the VGH cases for which we had multiple images compared to the cases that con- tributed only a single image (Fig. 7). Together, these findings show a significant degree of intrapatient variability and indicate that increased tumor sampling is associated with improved model performance. DISCUSSION Heat map of stromal matrix objects mean abs.diff to neighbors H&E image separated into epithelial and stromal objects A B C Worse prognosis Improved prognosis Improved prognosis Improved prognosis Worse prognosis Worse prognosis Fig. 4. Top stromal features associated with survival. (A) Variability in ab- solute difference in intensity between stromal matrix regions and neigh- bors. Top panel, high score (24.1); bottom panel, low score (10.5). (Insets) Top panel, high score; bottom panel; low score. Right panels, stromal matrix objects colored blue (low), green (medium), or white (high) according to each object’s absolute difference in intensity to neighbors. (B) Presence R E S E A R C H A R T I C L E onNovember17,2011stm.sciencemag.orgDownloadedfrom Top epithelial features.The eight panels in the figure (A to H) each shows one of the top-ranking epithelial features from the bootstrap anal- ysis. Left panels, improved prognosis; right panels, worse prognosis.
- 142. ISBI Grand Challenge on Cancer Metastases Detection in Lymph Node
- 145. International Symposium on Biomedical Imaging 2016 H&E Image Processing Framework Train whole slide image sample sample training data normaltumor Test whole slide image overlapping image patches tumor prob. map 1.0 0.0 0.5 Convolutional Neural Network P(tumor)
- 147. Clinical study on ISBI dataset Error Rate Pathologist in competition setting 3.5% Pathologists in clinical practice (n = 12) 13% - 26% Pathologists on micro-metastasis(small tumors) 23% - 42% Beck Lab Deep Learning Model 0.65% Beck Lab’s deep learning model now outperforms pathologist Andrew Beck, Machine Learning for Healthcare, MIT 2017
- 148. Andrew Beck, Advancing medicine with intelligent pathology, AACR 2017
- 149. Assisting Pathologists in Detecting Cancer with Deep Learning • The localization score(FROC) for the algorithm reached 89%, which significantly exceeded the score of 73% for a pathologist with no time constraint.
- 150. Assisting Pathologists in Detecting Cancer with Deep Learning • Algorithms need to be incorporated in a way that complements the pathologist’s workflow. • Algorithms could improve the efficiency and consistency of pathologists. • For example, pathologists could reduce their false negative rates (percentage of undetected tumors) by reviewing the top ranked predicted tumor regions including up to 8 false positive regions per slide.
- 151. Assisting Pathologists in Detecting Cancer with Deep Learning 6 Input & Validation Test model size FROC @8FP AUC FROC @8FP AUC 40X 98.1 100 99.0 87.3 (83.2, 91.1) 91.1 (87.2, 94.5) 96.7 (92.6, 99.6) 40X-pretrained 99.3 100 100 85.5 (81.0, 89.5) 91.1 (86.8, 94.6) 97.5 (93.8, 99.8) 40X-small 99.3 100 100 86.4 (82.2, 90.4) 92.4 (88.8, 95.7) 97.1 (93.2, 99.8) ensemble-of-3 - - - 88.5 (84.3, 92.2) 92.4 (88.7, 95.6) 97.7 (93.0, 100) 20X-small 94.7 100 99.6 85.5 (81.0, 89.7) 91.1 (86.9, 94.8) 98.6 (96.7, 100) 10X-small 88.7 97.2 97.7 79.3 (74.2, 84.1) 84.9 (80.0, 89.4) 96.5 (91.9, 99.7) 40X+20X-small 94.9 98.6 99.0 85.9 (81.6, 89.9) 92.9 (89.3, 96.1) 97.0 (93.1, 99.9) 40X+10X-small 93.8 98.6 100 82.2 (77.0, 86.7) 87.6 (83.2, 91.7) 98.6 (96.2, 99.9) Pathologist [1] - - - 73.3* 73.3* 96.6 Camelyon16 winner [1, 23] - - - 80.7 82.7 99.4 Table 1. Results on Camelyon16 dataset (95% confidence intervals, CI). Bold indicates results within the CI of the best model. “Small” models contain 300K parameters per Inception tower instead of 20M. -: not reported. *A pathologist achieved this sensitivity (with no FP) using 30 hours. to 10 20% variance), and can confound evaluation of model improvements by grouping multiple nearby tumors as one. By contrast, our non-maxima sup- pression approach is relatively insensitive to r between 4 and 6, although less accurate models benefited from tuning r using the validation set (e.g., 8). Fi- The FROC evaluates tumor detection and localization The FROC is defined as the sensitivity at 0.25,0.5,1,2,4,8 average FPs per tumor-negative slide. Yun Liu et al. Detecting Cancer Metastases on Gigapixel Pathology Images (2017) Sensitivity at 8 false positives per image
- 152. Assisting Pathologists in Detecting Cancer with Deep Learning Yun Liu et al. Detecting Cancer Metastases on Gigapixel Pathology Images (2017) • 구글의 인공지능은 @8FP 및 FROC에서 큰 개선 (92.9%, 88.5%) •@8FP: FP를 8개까지 봐주면서, 달성할 수 있는 sensitivity •FROC: FP를 슬라이드당 1/4, 1/2, 1, 2, 4, 8개를 허용한 민감도의 평균 •즉, FP를 조금 봐준다면, 인공지능은 매우 높은 민감도를 달성 •인간 병리학자는 민감도 73%에 반해 특이도는 거의 100% 달성 •인간 병리학자와 인공지능 병리학자는 서로 잘하는 것이 다름 •양쪽이 협력하면 판독 효율성, 일관성, 민감도 등에서 개선 기대 가능
- 156. http://www.rolls-royce.com/about/our-technology/enabling-technologies/engine-health-management.aspx#sense 250 sensors to monitor the “health” of the GE turbines
- 157. Fig 1. What can consumer wearables do? Heart rate can be measured with an oximeter built into a ring [3], muscle activity with an electromyographi sensor embedded into clothing [4], stress with an electodermal sensor incorporated into a wristband [5], and physical activity or sleep patterns via an accelerometer in a watch [6,7]. In addition, a female’s most fertile period can be identified with detailed body temperature tracking [8], while levels of me attention can be monitored with a small number of non-gelled electroencephalogram (EEG) electrodes [9]. Levels of social interaction (also known to a PLOS Medicine 2016
- 158. •복잡한 의료 데이터의 분석 및 insight 도출 •영상 의료/병리 데이터의 분석/판독 •연속 데이터의 모니터링 및 예방/예측 인공지능의 의료 활용
- 159. Project Artemis at UIOT
- 163. S E P S I S A targeted real-time early warning score (TREWScore) for septic shock Katharine E. Henry,1 David N. Hager,2 Peter J. Pronovost,3,4,5 Suchi Saria1,3,5,6 * Sepsis is a leading cause of death in the United States, with mortality highest among patients who develop septic shock. Early aggressive treatment decreases morbidity and mortality. Although automated screening tools can detect patients currently experiencing severe sepsis and septic shock, none predict those at greatest risk of developing shock. We analyzed routinely available physiological and laboratory data from intensive care unit patients and devel- oped “TREWScore,” a targeted real-time early warning score that predicts which patients will develop septic shock. TREWScore identified patients before the onset of septic shock with an area under the ROC (receiver operating characteristic) curve (AUC) of 0.83 [95% confidence interval (CI), 0.81 to 0.85]. At a specificity of 0.67, TREWScore achieved a sensitivity of 0.85 and identified patients a median of 28.2 [interquartile range (IQR), 10.6 to 94.2] hours before onset. Of those identified, two-thirds were identified before any sepsis-related organ dysfunction. In compar- ison, the Modified Early Warning Score, which has been used clinically for septic shock prediction, achieved a lower AUC of 0.73 (95% CI, 0.71 to 0.76). A routine screening protocol based on the presence of two of the systemic inflam- matory response syndrome criteria, suspicion of infection, and either hypotension or hyperlactatemia achieved a low- er sensitivity of 0.74 at a comparable specificity of 0.64. Continuous sampling of data from the electronic health records and calculation of TREWScore may allow clinicians to identify patients at risk for septic shock and provide earlier interventions that would prevent or mitigate the associated morbidity and mortality. INTRODUCTION Seven hundred fifty thousand patients develop severe sepsis and septic shock in the United States each year. More than half of them are admitted to an intensive care unit (ICU), accounting for 10% of all ICU admissions, 20 to 30% of hospital deaths, and $15.4 billion in an- nual health care costs (1–3). Several studies have demonstrated that morbidity, mortality, and length of stay are decreased when severe sep- sis and septic shock are identified and treated early (4–8). In particular, one study showed that mortality from septic shock increased by 7.6% with every hour that treatment was delayed after the onset of hypo- tension (9). More recent studies comparing protocolized care, usual care, and early goal-directed therapy (EGDT) for patients with septic shock sug- gest that usual care is as effective as EGDT (10–12). Some have inter- preted this to mean that usual care has improved over time and reflects important aspects of EGDT, such as early antibiotics and early ag- gressive fluid resuscitation (13). It is likely that continued early identi- fication and treatment will further improve outcomes. However, the Acute Physiology Score (SAPS II), SequentialOrgan Failure Assessment (SOFA) scores, Modified Early Warning Score (MEWS), and Simple Clinical Score (SCS) have been validated to assess illness severity and risk of death among septic patients (14–17). Although these scores are useful for predicting general deterioration or mortality, they typical- ly cannot distinguish with high sensitivity and specificity which patients are at highest risk of developing a specific acute condition. The increased use of electronic health records (EHRs), which can be queried in real time, has generated interest in automating tools that identify patients at risk for septic shock (18–20). A number of “early warning systems,” “track and trigger” initiatives, “listening applica- tions,” and “sniffers” have been implemented to improve detection andtimelinessof therapy forpatients with severe sepsis andseptic shock (18, 20–23). Although these tools have been successful at detecting pa- tients currently experiencing severe sepsis or septic shock, none predict which patients are at highest risk of developing septic shock. The adoption of the Affordable Care Act has added to the growing excitement around predictive models derived from electronic health R E S E A R C H A R T I C L E onNovember3,2016http://stm.sciencemag.org/Downloadedfrom
- 164. puted as new data became avail when his or her score crossed t dation set, the AUC obtained f 0.81 to 0.85) (Fig. 2). At a spec of 0.33], TREWScore achieved a s a median of 28.2 hours (IQR, 10 Identification of patients b A critical event in the developme related organ dysfunction (seve been shown to increase after th more than two-thirds (68.8%) o were identified before any sepsi tients were identified a median (Fig. 3B). Comparison of TREWScore Weevaluatedtheperformanceof methods for the purpose of provid use of TREWScore. We first com to MEWS, a general metric used of catastrophic deterioration (17) oped for tracking sepsis, MEWS tion of patients at risk for severe Fig. 2. ROC for detection of septic shock before onset in the validation set. The ROC curve for TREWScore is shown in blue, with the ROC curve for MEWS in red. The sensitivity and specificity performance of the routine screening criteria is indicated by the purple dot. Normal 95% CIs are shown for TREWScore and MEWS. TPR, true-positive rate; FPR, false-positive rate. R E S E A R C H A R T I C L E A targeted real-time early warning score (TREWScore) for septic shock AUC=0.83 At a specificity of 0.67,TREWScore achieved a sensitivity of 0.85 and identified patients a median of 28.2 hours before onset.
- 166. Jan 7, 2016
- 167. In an early research project involving 600 patient cases, the team was able to predict near-term hypoglycemic events up to 3 hours in advance of the symptoms. IBM Watson-Medtronic Jan 7, 2016
- 168. Sugar.IQ 사용자의 음식 섭취와 그에 따른 혈당 변 화, 인슐린 주입 등의 과거 기록 기반 식후 사용자의 혈당이 어떻게 변화할지 Watson 이 예측
- 169. ADA 2017, San Diego, Courtesy of Taeho Kim (Seoul Medical Center)
- 170. ADA 2017, San Diego, Courtesy of Taeho Kim (Seoul Medical Center)
- 171. ADA 2017, San Diego, Courtesy of Taeho Kim (Seoul Medical Center)
- 172. ADA 2017, San Diego, Courtesy of Taeho Kim (Seoul Medical Center)
- 174. Prediction ofVentricular Arrhythmia Collaboration with Prof. Segyeong Joo (Asan Medical Center) Analysed “Physionet Spontaneous Ventricular Tachyarrhythmia Database” for 2.5 months (on going project) Joo S, Choi KJ, Huh SJ, 2012, Expert Systems with Applications (Vol 39, Issue 3) ▪ Recurrent Neural Network with Only Frequency Domain Transform • Input : Spectrogram with 129 features obtained after ectopic beats removal • Stack of LSTM Networks • Binary cross-entropy loss • Trained with RMSprop • Prediction Accuracy : 76.6% ➞ 89.6% Dropout Dropout
- 175. Prediction ofVentricular TachycardiaOne Hour before Occurrence UsingArtificial Neural Networks Hyojeong Lee1,* , Soo-Yong Shin2,* , Myeongsook Seo3 ,Gi-Byoung Nam3 & Segyeong Joo1,4 Ventricular tachycardia (VT) is a potentially fatal tachyarrhythmia, which causes a rapid heartbeat as a result of improper electrical activity of the heart.This is a potentially life-threatening arrhythmia because it can cause low blood pressure and may lead to ventricular fibrillation, asystole, and sudden cardiac death.To preventVT, we developed an early prediction model that can predict this event one hour before its onset using an artificial neural network (ANN) generated using 14 parameters obtained from heart rate variability (HRV) and respiratory rate variability (RRV) analysis. De-identified raw data from the monitors of patients admitted to the cardiovascular intensive care unit atAsan Medical Center between September 2013 andApril 2015 were collected.The dataset consisted of 52 recordings obtained one hour prior toVT events and 52 control recordings.Two-thirds of the extracted parameters were used to train theANN, and the remaining third was used to evaluate performance of the learned ANN.The developedVT prediction model proved its performance by achieving a sensitivity of 0.88, specificity of 0.82, andAUC of 0.93. Sudden cardiac death (SCD) causes more than 300,000 deaths annually in the United States1 . Coronary artery disease, cardiomyopathy, structural heart problems, Brugada syndrome, and long QT syndrome are well known causes of SCD1–4 . In addition, spontaneous ventricular tachyarrhythmia (VTA) is a main cause of SCD, contrib- uting to about 80% of SCDs5 . Ventricular tachycardia (VT) and ventricular fibrillation (VF) comprise VTA. VT is defined as a very rapid heartbeat (more than 100 times per minute), which does not allow enough time for the ventricles to fill with blood between beats. VT may terminate spontaneously after a few seconds; however, in some cases, VT can progress to more dangerous or fatal arrhythmia, VF. Accordingly, early prediction of VT will help in reducing mortality from SCD by allowing for preventive care of VTA. Several studies have reported attempts at predicting VTAs by assessing the occurrence of syncope, left ventricu- lar systolic dysfunction, QRS (Q, R, and S wave in electrocardiogram) duration, QT (Q and T wave) dispersion, Holter monitoring, signal averaged electrocardiograms (ECGs), heart rate variability (HRV), T wave alternans, electrophysiologic testing, B-type natriuretic peptides, and other parameters or method6–10 . Among these studies, prediction of VTAs based on HRV analysis has recently emerged and shown potential for predicting VTA11–13 . Previous studies have focused on the prediction of VT using HRV analysis. In addition, most studies assessed the statistical value of each parameter calculated on or prior to the VT event and parameters of control data, which were collected from Holter recordings and implantable cardioverter defibrillators (ICDs)12,14,15 . However, the results were not satisfactory in predicting fatal events like VT. To make a better prediction model of VT, it is essential to utilize multiple parameters from various methods of HRV analysis and to generate a classifier that can deal with complex patterns composed of such parameters7 . Artificial neural network (ANN) is a valuable tool for classification of a database with multiple parameters. ANN is a kind of machine learning algorithms, which can be trained using data with multiple parameters16 . After training, the ANN calculates an output value according to the input parameters, and this output value can be used 1 Department of Biomedical Engineering, University of Ulsan College of Medicine, Seoul, Republic of Korea. 2 Department of Biomedical Informatics, Asan Medical Center, Seoul, Republic of Korea. 3 Department of Internal Re e e : 26 pr 2016 A ep e : 03 s 2016 P s e : 26 s 2016 OPEN Lee H. et al, Scientific Report, 2016
- 176. Prediction of Ventricular Tachycardia One Hour before Occurrence Using Artificial Neural Networks ww.nature.com/scientificreports/ in pattern recognition or classification. ANN has not been widely used in medical analysis since the algorithm is not intuitive for physicians. However, utilization of ANN in medical research has recently emerged17–19 . Our Parameters Control dataset (n=110) VTs dataset (n=110) Mean±SD Mean±SD p-Value Mean NN (ms) 0.709±0.149 0.718±0.158 0.304 SDNN (ms) 0.061±0.042 0.073±0.045 0.013 RMSSD (ms) 0.068±0.053 0.081±0.057 0.031 pNN50 (%) 0.209±0.224 0.239±0.205 0.067 VLF (ms2 ) 4.1E-05±6.54E-05 6.23E-05±9.81E-05 0.057 LF (ms2 ) 7.61E-04±1.16E-03 1.04E-03±1.15E-03 0.084 HF (ms2 ) 1.53E-03±2.02E-03 1.96E-03±2.16E-03 0.088 LF/HF 0.498±0.372 0.533±0.435 0.315 SD1 (ms) 0.039±0.029 0.047±0.032 0.031 SD2 (ms) 0.081±0.057 0.098±0.06 0.012 SD1/SD2 0.466±0.169 0.469±0.164 0.426 RPdM (ms) 2.73±0.817 2.95±0.871 0.038 RPdSD (ms) 0.721±0.578 0.915±0.868 0.075 RPdV 28.4±5.31 25.4±3.56 <0.002 Table 1. Comparison of HRV and RRV parameters between the control and VT dataset. ANN with Input Sensitivity (%) Specificity (%) Accuracy (%) PPV (%) NPV (%) AUC HRV parameters 11 70.6(12/17) 76.5(13/17) 73.5(25/34) 75.0(12/16) 72.2(13/18) 0.75 RRV parameters 3 82.4(14/17) 82.4(14/17) 82.4(28/34) 82.4(14/17) 82.4(14/17) 0.83 HRV+RRV parameters 14 88.2(15/17) 82.4(14/17) 85.3(29/34) 83.3(15/18) 87.5(14/16) 0.93 Table 2. Performance of three ANNs in predicting a VT event 1hour before onset for the test dataset. Lee H. et al, Scientific Report, 2016 This ANN with 13 hidden neurons in one hidden layer showed the best performance.
- 177. www.nature.com/scientificreports/ Discussion Figure 1. ROC curve of three ANNs (dashed line, with only HRV parameters; dashdot line, with parameters; solid line, with HRV and RRV parameters; dotted line, reference) used in the predict VT event one hour before onset. ROC curve of three ANNs (dashed line, with only HRV parameters; dashdot line, with only RRV parameters; solid line, with HRV and RRV parameters; dotted line, reference) used in the prediction of aVT event one hour before onset. Prediction of Ventricular Tachycardia One Hour before Occurrence Using Artificial Neural Networks Lee H. et al, Scientific Report, 2016
- 178. •아주대병원 외상센터, 응급실, 내과계 중환자실 등 3곳의 80개 병상 •산소포화도, 혈압, 맥박, 뇌파, 체온 등 8가지 환자 생체 데이터를 하나로 통합 저장 •생체 정보를 인공지능으로 실시간 모니터링+분석하여 1-3시간 전에 예측 •부정맥, 패혈증, 급성호흡곤란증후군(ARDS), 계획되지 않은 기도삽관 등의 질병
- 180. 서울아산병원 딥러닝 워크샵 2017. 9.
- 181. (source: VUNO) APPH(Alarms Per Patients Per Hour)
- 182. 서울아산병원 딥러닝 워크샵 2017. 9.
- 183. Cardiologist-Level Arrhythmia Detection with Convolutional Neural Networks Pranav Rajpurkar⇤ [email protected] Awni Y. Hannun⇤ [email protected] Masoumeh Haghpanahi [email protected] Codie Bourn [email protected] Andrew Y. Ng [email protected] Abstract We develop an algorithm which exceeds the per- formance of board certified cardiologists in de- tecting a wide range of heart arrhythmias from electrocardiograms recorded with a single-lead wearable monitor. We build a dataset with more than 500 times the number of unique patients than previously studied corpora. On this dataset, we train a 34-layer convolutional neural network which maps a sequence of ECG samples to a se- quence of rhythm classes. Committees of board- certified cardiologists annotate a gold standard test set on which we compare the performance of our model to that of 6 other individual cardiolo- gists. We exceed the average cardiologist perfor- mance in both recall (sensitivity) and precision (positive predictive value). 1. Introduction We develop a model which can diagnose irregular heart Figure 1. Our trained convolutional neural network correctly de- tecting the sinus rhythm (SINUS) and Atrial Fibrillation (AFIB) from this ECG recorded with a single-lead wearable heart moni- tor. Arrhythmia detection from ECG recordings is usually per- formed by expert technicians and cardiologists given the 1707.01836v1[cs.CV]6Jul2017
- 184. •ZIO Patch •2009년에 FDA에서 인허가 받은 의료기기 •최대 2주까지 붙이고 다니면서 지속적으로 심전도를 측정
- 185. Cardiologist-Level Arrhythmia Detection with Convolutional Neural Networks • Training Set • 약 3만 명의 환자에게서 얻은 64,000여 건의 심전도 데이터 • 34층 깊이의 CNN으로 학습 • Test Set • 336명에게서 얻은 Zio Patch의 심전도 데이터 • 세 명의 심장내과 전문의들이 상의하여 정답 • 총 12가지 종류의 부정맥으로 분류 • 총 6명의 심장내과 전문의들 vs 인공지능 • 부정맥의 발생 여부 • 부정맥 종류
- 186. Cardiologist-Level Arrhythmia Detection with Convolutional Neural Networks Cardiologist-Level Arrhythmia Detection with Convoluti Figure 3. Evaluated on the test set, the model outperforms the Patch which h et al., 2013). seconds long Each record is pert highlights responding to The 30 second ECG annotatio tations were d Technicians w rhythmia detec ination by Car technicians we could annotate labeled from t
- 187. Cardiologist-Level Arrhythmia Detection with Convolutional Neural Networks Cardiologist-Level Arrhythmia Detection with Convolutional Neural Networks Seq Set Model Cardiol. Model Cardiol. Class-level F1 Score AFIB 0.604 0.515 0.667 0.544 AFL 0.687 0.635 0.679 0.646 AVB TYPE2 0.689 0.535 0.656 0.529 BIGEMINY 0.897 0.837 0.870 0.849 CHB 0.843 0.701 0.852 0.685 EAR 0.519 0.476 0.571 0.529 IVR 0.761 0.632 0.774 0.720 JUNCTIONAL 0.670 0.684 0.783 0.674 NOISE 0.823 0.768 0.704 0.689 SINUS 0.879 0.847 0.939 0.907 SVT 0.477 0.449 0.658 0.556 TRIGEMINY 0.908 0.843 0.870 0.816 VT 0.506 0.566 0.694 0.769 WENCKEBACH 0.709 0.593 0.806 0.736 Aggregate Results Precision (PPV) 0.800 0.723 0.809 0.763 Recall (Sensitivity) 0.784 0.724 0.827 0.744 F1 0.776 0.719 0.809 0.751 Table 1. The top part of the table gives a class-level comparison of the expert to the model F1 score for both the Sequence and the Set In both the Sequence and the Set ca F1 score for each class separately. W overall F1 (and precision and recall) a weighted mean. Model vs. Cardiologist Performance We assess the cardiologist performanc call that each of the records in the t truth label from a committee of three as individual labels from a disjoint set gists. To assess cardiologist performan take the average of all the individual c using the group label as the ground tru Table 1 shows the breakdown of bo model scores across the different rh model outperforms the average cardi on most rhythms, noticeably outperfo gists in the AV Block set of arrhythm Mobitz I (Wenckebach), Mobitz II (AV plete heart block (CHB). This is esp the severity of Mobitz II and complete importance of distinguishing these tw
- 188. Cardiologist-Level Arrhythmia Detection with Convolutional Neural Networks Cardiologist-Level Arrhythmia Detection with Convolutional Neur Figure 4. A confusion matrix for the model predictions on the test beat detection (Coast et works have also been u (Melo et al., 2000). W high-accuracy for some cient for high-accuracy segmentation. For exa neural network to distin and Sinus Rhythm on network can distinguish high-accuracy, it does n recordings or classify am available in MIT-BIH. training data, and becau information in the featur The most common datas algorithms is the MIT-B & Mark, 2001) which ECG data. Other com
- 189. 인공지능의 의료 활용 •복잡한 의료 데이터의 분석 및 insight 도출 •영상 의료/병리 데이터의 분석/판독 •연속 데이터의 모니터링 및 예방/예측
- 190. •인공지능은 의사를 대체하는가 •결과에 대한 책임은 누가 지는가 •인공지능의 의학적 효용을 어떻게 증명할 것인가 •의료 인공지능을 어떻게 규제할 것인가 Issues
- 191. •인공지능은 의사를 대체하는가 •결과에 대한 책임은 누가 지는가 •인공지능의 의학적 효용을 어떻게 증명할 것인가 •의료 인공지능을 어떻게 규제할 것인가 Issues
- 192. •인공지능이 의사를 대체할 수 있을까? •인공지능이 의사를 모두 대체할 수 있을까? Yes or No?
- 193. •인공지능이 의사를 대체할 수 있을까? 있다. •인공지능이 의사를 모두 대체할 수 있을까? 없다. Yes or No?
- 194. •인공지능이 의사를 대체할 수 있을까? 있다. •인공지능이 의사를 모두 대체할 수 있을까? 없다. Yes or No?
- 195. •인공지능이 의사를 대체할 수 있을까? 있다. •인공지능이 의사를 모두 대체할 수 있을까? 없다. Yes or No?
- 196. Yes or No? •인공지능이 의사를 대체할 수 있을까? 있다. •인공지능이 의사를 모두 대체할 수 있을까? 없다. ‘대체’ 라는 급진적 표현을 사용하는 질문이 우리 고민의 본질을 가리지는 않는가?
- 197. 대한영상의학회 춘계학술대회 2017.6 장기적으로 의사가 선택 가능한 대처 방안 중에 ‘변화를 막는다’는 선택지는 없다.
- 199. •J&J이 수면 유도 마취로봇인 ‘세더시스(Sedasys)' 를 2014년 출시 •결장경, 내시경 검사 때 프로포폴을 주사해 환자 수면을 유도하는 마취용 의료로봇 •혈중 산소 함량, 심장박동 수 등 환자 신체 징후에 따라 투약량을 조절 •2013년 FDA가 승인하면서 미국, 호주, 캐나다 등 병원에 2014년부터 보급 •수면내시경 의료비를 1/10 로 낮춤 (2000달러 vs 150-200달러) •마취전문의협회 등은 대대적인 반대 캠페인을 벌이고 정치권에 규제 로비를 전개 •월스트리트 저널: “J&J가 수입원이 줄어들 위기에 처한 마취전문의들과 싸움에서 패한 것"
- 201. 기계적인 일을 모두 기계가 대신한다면, 과연 인간의 역할은 무엇일까? 현재 의사의 역할에는 어떤 것들이 있을까?
- 202. 기계적인 일을 모두 기계가 대신한다면, 과연 인간의 역할은 무엇일까? 현재 의사의 역할에는 어떤 것들이 있을까? •사라질 역할 •유지될 역할 •새로운 역할
- 203. •사라질 역할 •기계적인 역할: 기계가 더 쉽고 정확하게 할 수 있는 일 •근거 및 논리에 의한 판단 •순서도로 도식화할 수 있는 것 •‘왜 그런 결정을 내렸는지 논리적으로 설명할 수 있는가?’ •‘다른 의사들에게 가도 비슷한 결정을 내릴 것인가?’ •‘내가 한 달 뒤에 보더라도 같은 결정을 내릴까?’ •의료 데이터 모니터링 및 해석, 판독
- 205. NCCN Guidelines Version 4.2014 Non-Small Cell Lung Cancer NCCN Guidelines Index NSCLC Table of Contents Discussion Version 4.2014, 06/05/14 © National Comprehensive Cancer Network, Inc. 2014, All rights reserved. The NCCN Guidelines® and this illustration may not be reproduced in any form without the express written permission of NCCN® . Note: All recommendations are category 2A unless otherwise indicated. Clinical Trials: NCCN believes that the best management of any cancer patient is in a clinical trial. Participation in clinical trials is especially encouraged. NSCL-2 dT3, N0 related to size or satellite nodules. fTesting is not listed in order of priority and is dependent upon clinical circumstances, institutional processes, and judicious use of resources. gMethods for evaluation include mediastinoscopy, mediastinotomy, EBUS, EUS, and CT-guided biopsy. hPositive PET/CT scan findings for distant disease need pathologic or other radiologic confirmation. If PET/CT scan is positive in the mediastinum, lymph node status needs pathologic confirmation. iSee Principles of Surgical Therapy (NSCL-B). jSee Principles of Radiation Therapy (NSCL-C). kSee Chemotherapy Regimens for Neoadjuvant and Adjuvant Therapy (NSCL-D). lExamples of high-risk factors may include poorly differentiated tumors (including lung neuroendocrine tumors [excluding well-differentiated neuroendocrine tumors]), vascular invasion, wedge resection, tumors >4 cm, visceral pleural involvement, and incomplete lymph node sampling (Nx). These factors independently may not be an indication and may be considered when determining treatment with adjuvant chemotherapy. mSee Chemotherapy Regimens Used with Radiation Therapy (NSCL-E). CLINICAL ASSESSMENT PRETREATMENT EVALUATIONf INITIAL TREATMENT Stage IA (peripheral T1ab, N0) Stage IB (peripheral T2a, N0) Stage I (central T1ab–T2a, N0) Stage II (T1ab–2ab, N1; T2b, N0) Stage IIB (T3, N0)d • PFTs (if not previously done) • Bronchoscopy (intraoperative preferred) • Pathologic mediastinal lymph node evaluationg (category 2B) • PET/CT scanh (if not previously done) • PFTs (if not previously done) • Bronchoscopy • Pathologic mediastinal lymph node evaluationg • PET/CT scanh (if not previously done) • Brain MRI (Stage II, Stage IB [category 2B]) Negative mediastinal nodes Positive mediastinal nodes Operable Medically inoperable Negative mediastinal nodes Positive mediastinal nodes Operable Medically inoperable Surgical exploration and resectioni + mediastinal lymph node dissection or systematic lymph node sampling Definitive RT including stereotactic ablative radiotherapyj (SABR) See Stage IIIA (NSCL-8) or Stage IIIB (NSCL-11) Surgical exploration and resectioni + mediastinal lymph node dissection or systematic lymph node sampling N0 N1 See Stage IIIA (NSCL-8) or Stage IIIB (NSCL-11) Definitive RT including SABRj Definitive chemoradiationj,m See Adjuvant Treatment (NSCL-3) See Adjuvant Treatment (NSCL-3) Consider adjuvant chemotherapyk (category 2B) for high-risk stages IB-IIl Printed by yoon sup choi on 6/19/2014 8:23:15 PM. For personal use only. Not approved for distribution. Copyright © 2014 National Comprehensive Cancer Network, Inc., All Rights Reserved.
- 206. NCCN Guidelines Version 4.2014 Non-Small Cell Lung Cancer NCCN Guidelines Index NSCLC Table of Contents Discussion Version 4.2014, 06/05/14 © National Comprehensive Cancer Network, Inc. 2014, All rights reserved. The NCCN Guidelines® and this illustration may not be reproduced in any form without the express written permission of NCCN® . Note: All recommendations are category 2A unless otherwise indicated. Clinical Trials: NCCN believes that the best management of any cancer patient is in a clinical trial. Participation in clinical trials is especially encouraged. NSCL-8 hPositive PET/CT scan findings for distant disease need pathologic or other radiologic confirmation. If PET/CT scan is positive in the mediastinum, lymph node status needs pathologic confirmation. iSee Principles of Surgical Therapy (NSCL-B). jSee Principles of Radiation Therapy (NSCL-C). kSee Chemotherapy Regimens for Neoadjuvant and Adjuvant Therapy (NSCL-D). mSee Chemotherapy Regimens Used with Radiation Therapy (NSCL-E). nR0 = no residual tumor, R1 = microscopic residual tumor, R2 = macroscopic residual tumor. sPatients likely to receive adjuvant chemotherapy may be treated with induction chemotherapy as an alternative. MEDIASTINAL BIOPSY FINDINGS INITIAL TREATMENT ADJUVANT TREATMENT T1-3, N0-1 (including T3 with multiple nodules in same lobe) Surgeryi,s Resectable Medically inoperable Surgical resectioni + mediastinal lymph node dissection or systematic lymph node sampling See Treatment according to clinical stage (NSCL-2) N0–1 N2 See NSCL-3 Margins negative (R0)n Sequential chemotherapyk (category 1) + RTj Margins positiven Surveillance (NSCL-14) R1n R2n Chemoradiationj (sequentialk or concurrentm) Surveillance (NSCL-14) Concurrent chemoradiationj,m Surveillance (NSCL-14) T1-2, T3 (≥7 cm), N2 nodes positivei • Brain MRI • PET/CT scan,h if not previously done Negative for M1 disease Positive Definitive concurrent chemoradiationj,m (category 1) or Induction chemotherapyk ± RTj See Treatment for Metastasis solitary site (NSCL-13) or distant disease (NSCL-15) No apparent progression Progression Surgeryi ± chemotherapyk (category 2B) ± RTj (if not given) RTj (if not given) ± chemotherapykLocal Systemic See Treatment for Metastasis solitary site (NSCL-13) or distant disease (NSCL-15) T3 (invasion), N2 nodes positive • Brain MRI • PET/CT scan,h if not previously done Negative for M1 disease Positive Definitive concurrent chemoradiationj,m See Treatment for Metastasis solitary site (NSCL-13) or distant disease (NSCL-15) Printed by yoon sup choi on 6/19/2014 8:23:15 PM. For personal use only. Not approved for distribution. Copyright © 2014 National Comprehensive Cancer Network, Inc., All Rights Reserved.
- 207. •유지/강조될 역할 •마지막 의료적 의사 결정 •인간만이 할 수 있는 인간적인 일 •Human touch •커뮤니케이션, 공감, care … •환자를 진료/치료하는 이외의 역할 •기초 연구 •새로운 데이터와 기준을 만들어내는 일 ➞ 기계에 반영
- 208. Over the course of a career, an oncologist may impart bad news an average of 20,000 times, but most practicing oncologists have never received any formal training to help them prepare for such conversations.
- 209. High levels of empathy in primary care physicians correlate with better clinical outcomes for their patients with diabetes
- 210. 의사는 진료시에 얼마나 공감능력을 발휘하나?
- 211. We identified 384 empathic opportunities and found that physicians had responded empathically to 39 (10%) of them
- 212. In 398 conversations, the total empathic opportunities was 292. When they occurred, oncologists responded with continuers 22% of the time.
- 213. G.M. Lucas et al. / Computers in Human Behavior 37 (2014) 94–100 인간 의사와 인공지능 의사 중 누가 라뽀 형성을 더 잘 할까?
- 215. It’s only a computer: Virtual humans increase willingness to disclose G.M. Lucas et al. / Computers in Human Behavior 37 (2014) 94–100 인공지능이 상담한다고 믿음 (computer frame) 사람이 원격으로 상담한다고 믿음 (human frame) 실제로 인공지능이 상담 (AI) 실제로 사람이 상담 (Tele-operated) Method Frame
- 216. It’s only a computer: Virtual humans increase willingness to disclose G.M. Lucas et al. / Computers in Human Behavior 37 (2014) 94–100 ‘‘How close are you to your family?’’ ‘‘Tell me about a situation that you wish you had handled differently.’’ ‘‘Tell me about an event, or something that you wish you could erase from your memory.’’ ‘‘Tell me about the hardest decision you’ve ever had to make.’’ ‘‘Tell me about the last time you felt really happy.’’ ‘‘What are you most proud of in your life?’’ ‘‘What’s something you feel guilty about?’’ ‘‘When was the last time you argued with someone and what was it about?’’
- 217. It’s only a computer: Virtual humans increase willingness to disclose G.M. Lucas et al. / Computers in Human Behavior 37 (2014) 94–100 0 5 10 15 20 Computer frame Human frame 0 15.25 30.5 45.75 61 Computer frame Human frame 0 0.033 0.065 0.098 0.13 Computer frame Human frame 0 0.3 0.6 0.9 1.2 Computer frame Human frame Fear of Self-disclosure Impression Management Sadness Displays Willingness to Disclosure 환자는 인간 상담사보다 인공지능을 더 신뢰하고, 감정과 자신의 치부를 더 솔직하게 드러냈다.
- 218. It’s only a computer: Virtual humans increase willingness to disclose G.M. Lucas et al. / Computers in Human Behavior 37 (2014) 94–100 ‘‘This is way better than talking to a person. I don’t really feel comfortable talking about personal stuff to other people.’’ ‘‘A human being would be judgmental. I shared a lot of personal things, and it was because of that.’’
- 219. 로봇이 사람처럼 수술할 수 있을까?
- 220. Supervised autonomous robotic soft tissue surgery Children’s National Health System 소아외과에서 개발한 Smart Tissue Autonomous Robot (STAR) Azad Shademan et al. Sci Transl Med 2016
- 221. Supervised autonomous robotic soft tissue surgery Azad Shademan et al. Sci Transl Med 2016
- 222. Azad Shademan et al. Sci Transl Med 2016 •Ex vivo 와 in vivo end-to-end 의 경우 모두 STAR 는 기존의 수술 방법들에 비해 대부분 더 나은 성과를 보임 •외과의사가 손으로 문합하는 것 (OPEN) •복강경 수술 (laparoscopy, LAP) •다빈치를 이용한 RAS •비교 기준 •비교 기준은 봉합한 공간의 일관성 (spacing) •얼마나 압력을 가했을 때 복강경이 새어 나오는가 (leak pressure) •조직에서 바늘을 제거하는 것이 필요했던 실수의 횟수 (number of mistakes) •총 수술 시간 (completion time) •내강의 수축 여부 (lumen reduction) smallest for STAR, indicating a greater consistency for semiautonomous suture placement compared to other techniques (table S1). Moreover, leak pressure reflects the functional quality of suturing. The linear closure from STAR was able to withstand a higher average leak pressure than all other techniques (Fig. 2B). coordinates, intersuture spacing, bite depths, pull tension forces, and suturing tool maneuvers before piercing. Using the NIRF markers as reference points, the plan interpolated intermediate suture placements on the bowel and adjusted placement of each suture, knot, and corner slidetoaccommodatedeformationsandinducedscenerotations(Fig.1F). Fig. 2. Ex vivo linear suturing under deformations. The experiment con- sisted of closing a longitudinal cut along pig intestine, whereas the tissue was deformed by pulling on stay sutures. Five samples were tested per tech- nique (OPEN, LAP, RAS, and STAR). (A) Suture spacing. Central mark is the median; box edges are the 25th and 75th percentiles, error bars are the range excluding outliers, and red dots are outliers. The whiskers represent the range not including outliers. There is a different N number for each boxplot because eachsurgeonuseda different number of sutures [OPEN (n =174), LAP(n= 128), RAS (n = 176), and STAR (n = 206)]. These data are presented numerically in table S2, including the SDs. P values determined by ANOVA with post hoc Games-Howell. (B and C) Leak pressures and number of mistakes (reposi- tioned stitches or robot reboot). Data are from individual tissue samples (n = 5) with averages marked by a horizontal line. P values determined by independent samples t test. (D) Completion times separated into knot-tying and suturing, and other time was spent restaging or changing sutures. Data are averages (n = 5). P values determined by independent samples t test. www.ScienceTranslationalMedicine.org 4 May 2016 Vol 8 Issue 337 337ra64 3 onMay7,2016rg/
- 223. Azad Shademan et al. Sci Transl Med 2016 LAP (Fig. 3E). Ex vivo studies simulating luminal strictures from anasto- whereas only 16 of 59 (27.1%) sutures required adjustmen maining circumference (Fig. 3 ing that different levels of auto be used effectively for differ Overall, 57.8% of the procedu fully autonomously with no Alternatively, in the current autonomous mode without an teraction would require suture in 42.2% of sutures placed, m corners. The completion tim also included supervisory ac surgeon, which accounted f the total time (7% for suture justment, 3.3% for confirmati location, and 2.6% for mistake In vivo end-to-end anast Finally, we performed in viv autonomous surgery in pig i cessed through a laparotomy (n = 4) and compared these a an OPEN control (n = 1) (fig used the same suture algorith ex vivo trials (Fig. 1, G and OPEN control, the surgeon u surgical hand tools to open th exposed the intestine, and sutu a transverse incision. The av STAR procedure time was 50.0 where 77.4% was anastomos 22.6% was restaging time be and front walls, which inclu 2.16 min for marking the tis B and E, and Table 1). Al OPEN timewasonly 8min,th was comparable to the averag laparoscopic anastomoses tha 30 min for vesicourethral (25 for aortic (26), to 90 min for constructions (27). Fig. 3. End-to-end anastomosis ex vivo. The experiment consisted of closing a transverse cut in pig intestine. Five samples were tested per technique (OPEN, LAP, RAS, and STAR). (A) Suture spacing. Central mark is the median; box edges are the 25th and 75th percentiles; and red dots are outliers. The whiskers represent the range not including outliers. There is a different N number for each boxplot because each surgeon used a different number of sutures [OPEN (n = 138), LAP (n = 98), RAS (n = 132), and STAR (n = 180)]. The average spacing betweenconsecutive sutures was calculated and compared be- tween STAR and other modalities. The variance of suture spacing is presented numerically in table S2, including the SD. P values determined by ANOVA with post hoc Games-Howell. (B) Exvivo end-to-end anastomosis leak pressures. Dataareindividualtissuesamples,withmeansdisplayedashorizontallines(n=4to5).Onesample was sutured closed and thus could not be tested for leak pressure. P values determined by independent samples t test. (C) The leak pressure as a function of maximum suture spacing. Data are individual tissue samples that were fit to a rational function (y = 0.854/x) (n = 4 to 5). (D) Number of mistakes (repositioned stitches or robot reboot). Data are individual tissue samples with means displayed as horizontal lines (n = 5). P values determined by independent samples t test. (E) Ex vivo end-to-end anastomosis completion times. Average times for n = 5 •Ex vivo 와 in vivo end-to-end 의 경우 모두 STAR 는 기존의 수술 방법들에 비해 대부분 더 나은 성과를 보임 •외과의사가 손으로 문합하는 것 (OPEN) •복강경 수술 (laparoscopy, LAP) •다빈치를 이용한 RAS •비교 기준 •비교 기준은 봉합한 공간의 일관성 (spacing) •얼마나 압력을 가했을 때 복강경이 새어 나오는가 (leak pressure) •조직에서 바늘을 제거하는 것이 필요했던 실수의 횟수 (number of mistakes) •총 수술 시간 (completion time) •내강의 수축 여부 (lumen reduction)
- 224. In Flesh-Cutting Task, Autonomous Robot Surgeon Beats Human Surgeons •의사보다 조직 절개를 더 잘 하는 로봇 •피부, 지방, 근육의 세 가지 돼지 조직을 이용해 STAR의 자동 절개 검증 •돼지 피부의 절개 실력을 외과 의사와 비교 •일직선으로 5cm 절개 비교 •절개 라인으로부터 얼마나 벗어났는가 •수술 부위에 상처를 남겼는가
- 225. In Flesh-Cutting Task, Autonomous Robot Surgeon Beats Human Surgeons •의사보다 조직 절개를 더 잘 하는 로봇 •피부, 지방, 근육의 세 가지 돼지 조직을 이용해 STAR의 자동 절개 검증 •돼지 피부의 절개 실력을 외과 의사와 비교 •일직선으로 5cm 절개 비교 •절개 라인으로부터 얼마나 벗어났는가 •수술 부위에 상처를 남겼는가
- 226. In Flesh-Cutting Task, Autonomous Robot Surgeon Beats Human Surgeons •Pseudo Tumor Resection •돼지 지방 조직에 점토로 만든 모형 종양을 설치 •STAR가 이 모형 종양을 잘 제거하는지를 테스트 •4mm 마진을 남기고 떼어내기 •향후 단계적으로 발전하면서 자동 수술의 기반이 될 것
- 227. In Flesh-Cutting Task, Autonomous Robot Surgeon Beats Human Surgeons •Pseudo Tumor Resection •돼지 지방 조직에 점토로 만든 모형 종양을 설치 •STAR가 이 모형 종양을 잘 제거하는지를 테스트 •4mm 마진을 남기고 떼어내기 •향후 단계적으로 발전하면서 자동 수술의 기반이 될 것
- 228. •새로운 역할 •임상에 인공지능을 활용하는 방법에 대한 트레이닝 •구체적으로 어떻게 활용할지에 대한 연구 및 가이드라인 필요 •clinical outcome •quality of care •cost effectiveness •이러한 역할에 맞게 의학 교육도 바뀌어야 할 것
- 229. •인공지능은 의사를 대체하는가 •결과에 대한 책임은 누가 지는가 •인공지능의 의학적 효용을 어떻게 증명할 것인가 •의료 인공지능을 어떻게 규제할 것인가 Issues
- 230. 결과에 대한 책임은 누가 지는가 •가장 민감한 부분이며, 실제 적용에 가장 큰 걸림돌 •여러 변수가 있기 때문에 간단한 문제가 아니다. •Bottom Line: 최종 의사결정은 인간 의사가 내린다. •일단 현재는 책임 소재는 누구에게? •진단 및 의학적 결정의 책임은 누가 지는가 •현실적으로 의학적 결정은 의사만 내리는 것인가
- 231. •가장 민감한 부분이며, 실제 적용에 가장 큰 걸림돌 •여러 변수가 있기 때문에 간단한 문제가 아니다. •Bottom Line: 최종 의사결정은 인간 의사가 내린다. •일단 현재는 책임 소재는 누구에게? •진단 및 의학적 결정의 책임은 누가 지는가 •현실적으로 의학적 결정은 의사만 내리는 것인가 결과에 대한 책임은 누가 지는가
- 232. •인공지능의 형식과 활용 방법에 따라서 달라질 수 있다. •결과 양식: 등수 / 점수 / 신호등 (상/중/하) •근거/과정의 투명성: 근거의 유무 / 판단 과정 투명 or 블랙박스 •인간 의사의 개입 시점 •pre-screening: AI, then human doctor •double reading: AI + human doctor •double check (second opinion): human doctor, then AI
- 235. Assisting Pathologists in Detecting Cancer with Deep Learning 윤곽선, 색상, heat map 등으로 표시 질병의 유무, 질병의 중증도 등을 제시
- 236. •인공지능의 형식과 활용 방법에 따라서 달라질 수 있다. •결과 양식: 등수 / 점수 / 신호등 (상/중/하) •근거/과정의 투명성: 근거의 유무 / 판단 과정 투명 or 블랙박스 •인간 의사의 개입 시점 •pre-screening: AI, then human doctor •double reading: AI + human doctor •double check (second opinion): human doctor, then AI
- 238. 떡수인가 묘수인가?
- 239. THEBLACKBOX 2 0 | N A T U R E | V O L 5 3 8 | 6 O C T O B E R 2 0 1 6 THEBLACKBOX OFAI
- 240. Animal Intelligence: Clever Hans
- 241. Animal Intelligence: Clever Hans • Clever Hans was an horse that was claimed to have been able to perform arithmetic. • After a formal investigation in 1907, psychologist Oskar Pfungst demonstrated that the horse was not actually performing these mental tasks, but was watching the reactions of his human observers. • The trainer was entirely unaware that he was providing such cues.
- 242. https://namkugkim.wordpress.com/2017/05/15/ “실제로 서울아산병원 흉부 X-ray를 학습시켰을때, 다른 질 환과 다른게 Cardiomegaly (심장 비대)의 경우 학습결과는 좋았으나 전혀 다른 것을 학습한다는 것을 이것을 통해 알게 되었다. Cardiomegaly는 심장이 커지는 것을 X-ray로 진단하는 것 인데, 딥러닝이 실제로 심장의 크기를 보는 것이 아니라, 이런 질환을 가진 환자의 X-ray에 있는 특징인 수술자국을 보고 있 다는 것을 CAM으로 확인하였다." - 서울아산병원 김남국 교수님 블로그
- 243. •인공지능의 형식과 활용 방법에 따라서 달라질 수 있다. •결과 양식: 등수 / 점수 / 신호등 (상/중/하) •근거/과정의 투명성: 근거의 유무 / 판단 과정 투명 or 블랙박스 •인간 의사의 개입 시점 •pre-screening: AI ➞ then human doctor •double reading: AI + human doctor •double check: human doctor ➞ AI
- 244. •인공지능은 의사를 대체하는가 •결과에 대한 책임은 누가 지는가 •인공지능의 의학적 효용을 어떻게 증명할 것인가 •의료 인공지능을 어떻게 규제할 것인가 Issues
- 245. 아직은 근거가 부족하다 • Analytical validity • Clinical validity • Clinical utility + • Cost-effectiveness • Efficiency of clinical practice 결과 형식 근거 유무/과정 활용 방식
- 246. 인공지능의 대명사 Watson의 경우에도 아직 충분한 근거를 보여준 바 없다. 정확성 / 의학적 효용 / 진료 효율성 / 비용 절감
- 247. Q: Watson이 MSKCC에 들어간지 이제 5년째지만, 아직 Watson 의 정확성이나 효과에 대해서는 보여준 데이터나 근거가 별로 없다. 왜 그런가? A: 아직까지 효과성을 검증하기 위한 기간이 충분하지 않았던 것 같다. Q: 그 근거를 혹시 언제쯤 볼 수 있는지 아는가? A: 확실하지 않다. 우리도 그러한 근거가 나오기를 기다리고 있다.
- 248. •Watson Oncology 의 임상 시험 디자인을 한다면, • Primary / secondary outcome 을 무엇으로 해야할까 • Cost-effectiveness 를 어떻게 증명할까 • 개별 병원에 특화된 시스템: 연구의 범용성 이슈
- 249. Annals of Oncology (2016) 27 (suppl_9): ix179-ix180. 10.1093/annonc/mdw601 Validation study to assess performance of IBM cognitive computing system Watson for oncology with Manipal multidisciplinary tumour board for 1000 consecutive cases: An Indian experience • MMDT(Manipal multidisciplinary tumour board) treatment recommendation and data of 1000 cases of 4 different cancers breast (638), colon (126), rectum (124) and lung (112) which were treated in last 3 years was collected. • Of the treatment recommendations given by MMDT, WFO provided 50% in REC, 28% in FC, 17% in NREC • Nearly 80% of the recommendations were in WFO REC and FC group • 5% of the treatment provided by MMDT was not available with WFO • The degree of concordance varied depending on the type of cancer • WFO-REC was high in Rectum (85%) and least in Lung (17.8%) • high with TNBC (67.9%); HER2 negative (35%) 후향적으로 의료진 판단과의 ‘일치도’를 보는 것 이상으로, 전향적 임상 연구로 정확성/효과성/안전성을 증명할 수 있을까?
- 250. IBM Watson Health Data and Evidence Strategy • Concordance • Decision impact • Pre-/post- assessment with focus on outcomes such as: Ø Guidelines adherence Ø Cost Ø Time savings Ø Toxicity, hospitalizations, emergency visits Ø Physician and patient satisfaction Ø Tumor response Ø Survival
- 251. The new engl and jour nal of medicine original article Single Reading with Computer-Aided Detection for Screening Mammography Fiona J. Gilbert, F.R.C.R., Susan M. Astley, Ph.D., Maureen G.C. Gillan, Ph.D., Olorunsola F. Agbaje, Ph.D., Matthew G. Wallis, F.R.C.R., Jonathan James, F.R.C.R., Caroline R.M. Boggis, F.R.C.R., and Stephen W. Duffy, M.Sc., for the CADET II Group* From the Aberdeen Biomedical Imaging Centre, University of Aberdeen, Aberdeen (F.J.G., M.G.C.G.); the Department of Im- aging Science and Biomedical Engineer- ing,UniversityofManchester,Manchester (S.M.A.); the Department of Epidemiolo- gy, Mathematics, and Statistics, Wolfson Institute of Preventive Medicine, London (O.F.A., S.W.D.); the Cambridge Breast Unit, Addenbrookes Hospital, Cambridge (M.G.W.); the Nottingham Breast Insti- tute, Nottingham City Hospital, Notting- ham (J.J.); and the Nightingale Breast Screening Unit, Wythenshawe Hospital, Manchester (C.R.M.B.) — all in the Unit- ed Kingdom. Address reprint requests to Dr. Gilbert at the Aberdeen Biomedical Imaging Centre, University of Aberdeen, Lilian Sutton Bldg., Foresterhill, Aberdeen AB25 2ZD, Scotland, United Kingdom, or at [email protected]. *The members of the Computer-Aided Detection Evaluation Trial II (CADET II) group are listed in the Appendix. This article (10.1056/NEJMoa0803545) was published at www.nejm.org on Oc- tober 1, 2008. N Engl J Med 2008;359:1675-84. Copyright © 2008 Massachusetts Medical Society. ABSTR ACT Background The sensitivity of screening mammography for the detection of small breast can- cers is higher when the mammogram is read by two readers rather than by a single reader. We conducted a trial to determine whether the performance of a single reader using a computer-aided detection system would match the performance achieved by two readers. Methods The trial was designed as an equivalence trial, with matched-pair comparisons be- tween the cancer-detection rates achieved by single reading with computer-aided de- tection and those achieved by double reading. We randomly assigned 31,057 women undergoing routine screening by film mammography at three centers in England to double reading, single reading with computer-aided detection, or both double read- ing and single reading with computer-aided detection, at a ratio of 1:1:28. The pri- mary outcome measures were the proportion of cancers detected according to regi- men and the recall rates within the group receiving both reading regimens. Results The proportion of cancers detected was 199 of 227 (87.7%) for double reading and 198 of 227 (87.2%) for single reading with computer-aided detection (P=0.89). The overall recall rates were 3.4% for double reading and 3.9% for single reading with computer-aided detection; the difference between the rates was small but significant (P<0.001). The estimated sensitivity, specificity, and positive predictive value for single reading with computer-aided detection were 87.2%, 96.9%, and 18.0%, respectively. The corresponding values for double reading were 87.7%, 97.4%, and 21.1%. There were no significant differences between the pathological attributes of tumors de- tected by single reading with computer-aided detection alone and those of tumors detected by double reading alone. Conclusions Single reading with computer-aided detection could be an alternative to double read- ing and could improve the rate of detection of cancer from screening mammograms read by a single reader. (ClinicalTrials.gov number, NCT00450359.) Mammography • single reading+CAD vs. double reading • Outcome: Cancer detection rate / Recall rate
- 252. The new engl and jour nal of medicine original article Single Reading with Computer-Aided Detection for Screening Mammography Fiona J. Gilbert, F.R.C.R., Susan M. Astley, Ph.D., Maureen G.C. Gillan, Ph.D., Olorunsola F. Agbaje, Ph.D., Matthew G. Wallis, F.R.C.R., Jonathan James, F.R.C.R., Caroline R.M. Boggis, F.R.C.R., and Stephen W. Duffy, M.Sc., for the CADET II Group* From the Aberdeen Biomedical Imaging Centre, University of Aberdeen, Aberdeen (F.J.G., M.G.C.G.); the Department of Im- aging Science and Biomedical Engineer- ing,UniversityofManchester,Manchester (S.M.A.); the Department of Epidemiolo- gy, Mathematics, and Statistics, Wolfson Institute of Preventive Medicine, London (O.F.A., S.W.D.); the Cambridge Breast Unit, Addenbrookes Hospital, Cambridge (M.G.W.); the Nottingham Breast Insti- tute, Nottingham City Hospital, Notting- ham (J.J.); and the Nightingale Breast Screening Unit, Wythenshawe Hospital, Manchester (C.R.M.B.) — all in the Unit- ed Kingdom. Address reprint requests to Dr. Gilbert at the Aberdeen Biomedical Imaging Centre, University of Aberdeen, Lilian Sutton Bldg., Foresterhill, Aberdeen AB25 2ZD, Scotland, United Kingdom, or at [email protected]. *The members of the Computer-Aided Detection Evaluation Trial II (CADET II) group are listed in the Appendix. This article (10.1056/NEJMoa0803545) was published at www.nejm.org on Oc- tober 1, 2008. N Engl J Med 2008;359:1675-84. Copyright © 2008 Massachusetts Medical Society. ABSTR ACT Background The sensitivity of screening mammography for the detection of small breast can- cers is higher when the mammogram is read by two readers rather than by a single reader. We conducted a trial to determine whether the performance of a single reader using a computer-aided detection system would match the performance achieved by two readers. Methods The trial was designed as an equivalence trial, with matched-pair comparisons be- tween the cancer-detection rates achieved by single reading with computer-aided de- tection and those achieved by double reading. We randomly assigned 31,057 women undergoing routine screening by film mammography at three centers in England to double reading, single reading with computer-aided detection, or both double read- ing and single reading with computer-aided detection, at a ratio of 1:1:28. The pri- mary outcome measures were the proportion of cancers detected according to regi- men and the recall rates within the group receiving both reading regimens. Results The proportion of cancers detected was 199 of 227 (87.7%) for double reading and 198 of 227 (87.2%) for single reading with computer-aided detection (P=0.89). The overall recall rates were 3.4% for double reading and 3.9% for single reading with computer-aided detection; the difference between the rates was small but significant (P<0.001). The estimated sensitivity, specificity, and positive predictive value for single reading with computer-aided detection were 87.2%, 96.9%, and 18.0%, respectively. The corresponding values for double reading were 87.7%, 97.4%, and 21.1%. There were no significant differences between the pathological attributes of tumors de- tected by single reading with computer-aided detection alone and those of tumors detected by double reading alone. Conclusions Single reading with computer-aided detection could be an alternative to double read- ing and could improve the rate of detection of cancer from screening mammograms read by a single reader. (ClinicalTrials.gov number, NCT00450359.) Table 1 double reading single reading & CAD proportion of cancers detected 87.7% 87.2% overall recall rates 3.4% 3.9% sensitivity 87.2% 87.8% specificity 96.9% 97.7% positive predicted value 18.0% 21.1% Conclusion: Single reading with computer-aided detection could be an alternative to double reading and could improve the rate of detection of cancer from screening mammograms read by a single reader.
- 253. Copyright 2015 American Medical Association. All rights reserved. Diagnostic Accuracy of Digital Screening Mammography With and Without Computer-Aided Detection Constance D. Lehman, MD, PhD; Robert D. Wellman, MS; Diana S. M. Buist, PhD; Karla Kerlikowske, MD; Anna N. A. Tosteson, ScD; Diana L. Miglioretti, PhD; for the Breast Cancer Surveillance Consortium IMPORTANCE After the US Food and Drug Administration (FDA) approved computer-aided detection (CAD) for mammography in 1998, and the Centers for Medicare and Medicaid Services (CMS) provided increased payment in 2002, CAD technology disseminated rapidly. Despite sparse evidence that CAD improves accuracy of mammographic interpretations and costs over $400 million a year, CAD is currently used for most screening mammograms in the United States. OBJECTIVE To measure performance of digital screening mammography with and without CAD in US community practice. DESIGN, SETTING, AND PARTICIPANTS We compared the accuracy of digital screening mammography interpreted with (n = 495 818) vs without (n = 129 807) CAD from 2003 through 2009 in 323 973 women. Mammograms were interpreted by 271 radiologists from 66 facilities in the Breast Cancer Surveillance Consortium. Linkage with tumor registries identified 3159 breast cancers in 323 973 women within 1 year of the screening. MAIN OUTCOMES AND MEASURES Mammography performance (sensitivity, specificity, and screen-detected and interval cancers per 1000 women) was modeled using logistic regression with radiologist-specific random effects to account for correlation among examinations interpreted by the same radiologist, adjusting for patient age, race/ethnicity, time since prior mammogram, examination year, and registry. Conditional logistic regression was used to compare performance among 107 radiologists who interpreted mammograms both with and without CAD. RESULTS Screening performance was not improved with CAD on any metric assessed. Mammography sensitivity was 85.3% (95% CI, 83.6%-86.9%) with and 87.3% (95% CI, 84.5%-89.7%) without CAD. Specificity was 91.6% (95% CI, 91.0%-92.2%) with and 91.4% (95% CI, 90.6%-92.0%) without CAD. There was no difference in cancer detection rate (4.1 in 1000 women screened with and without CAD). Computer-aided detection did not improve intraradiologist performance. Sensitivity was significantly decreased for mammograms interpreted with vs without CAD in the subset of radiologists who interpreted both with and without CAD (odds ratio, 0.53; 95% CI, 0.29-0.97). CONCLUSIONS AND RELEVANCE Computer-aided detection does not improve diagnostic accuracy of mammography. These results suggest that insurers pay more for CAD with no established benefit to women. JAMA Intern Med. 2015;175(11):1828-1837. doi:10.1001/jamainternmed.2015.5231 Published online September 28, 2015. Invited Commentary page 1837 Author Affiliations: Department of Radiology, Massachusetts General Hospital, Boston (Lehman); Group Health Research Institute, Seattle, Washington (Wellman, Buist, Miglioretti); Departments of Medicine and Epidemiology and Biostatistics, University of California, San Francisco, San Francisco (Kerlikowske); Norris Cotton Cancer Center, Geisel School of Medicine at Dartmouth, Dartmouth College, Lebanon, New Hampshire (Tosteson); Department of Public Health Sciences, School of Medicine, University of California, Davis (Miglioretti). Corresponding Author: Constance D. Lehman, MD, PhD, Department of Radiology, Massachusetts General Hospital, Avon Comprehensive Breast Evaluation Center, 55 Fruit St, WAC 240, Boston, MA 02114 (clehman @mgh.harvard.edu). Research Original Investigation | LESS IS MORE 1828 (Reprinted) jamainternalmedicine.com Copyright 2015 American Medical Association. All rights reserved. CAD for Mammography in US • 1998년 미국 FDA에서 승인 • 2002년 Centers for Medicare and Medicaid Services (CMS) 혜택 시작 • 연간 $400m의 비용 발생
- 254. • 2002년 Centers for Medicare and Medicaid Services (CMS) 혜택 시작 • 2012년 전체 mammogram 중 83%가 CAD를 사용cer diagnosis within the follow-up period. True-positive examination results were defined as those with a positive examination assessment and breast cancer diagnosis. False- positive examination results were examinations with a posi- Mammography performan using logistic regression, includ diologist-specific random effect tion among examinations read b dom effects were allowed to vary the reading. Performance measu dian of the random effects distrib specific relative performance wa (OR) with 95% CIs comparing C for patient age at diagnosis, time year of examination, and the BC Receiver operating characte mated from 135 radiologists wh mogram associated with a cance cal logistic regression model tha accuracy parameters to depend o ing examination interpretation. racy among radiologists for exa the same condition (with or wi threshold for recall to vary acro mally distributed, radiologist-spe ied by whether the radiologist us We estimated the normalized mary ROC curves across the obs rates from this model.26 We plo the false-positive rate and supe curves. Two separate main sensitiv in subsets of total examinations Figure 1. Screening Mammography Patterns From 2000 to 2012 in US Community Practices in the Breast Cancer Surveillance Consortium (BCSC) 100 80 60 40 20 0 TypeofMammography,% Year 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 Film Digital with CAD Type Digital without CAD Data are provided from the larger BCSC population including all screening mammograms (5.2 million mammograms) for the indicated time period. Research Original Investigation Digital Screening Mammog CMS 보험 혜택 5% 83% 74%
- 255. Diagnostic accuracy was not improved with CAD on any performance metric assessed w/ CAD w/o CAD sensitivity 85.3% 87.3% sensitivity for invasive cancer 82.1% 85.0% sensitivity for DCIS 93.2% 94.3% specificity 91.6% 91.4% Detection Rate (Overall) 4.1 per 1000 4.1 per 1000 Detection Rate in DCIS 1.2 per 1000 0.9 per 1000 < < < > • CAD를 사용할 때가 사용하지 않을 때보다 민감도가 더 좋지 않았다! • 시사점 • 의학적 효용은 연구 결과 우리의 기대와 다를 수 있다. • 의학적 효용을 증명하기 위해서는 오랜 시간이 걸린다. • 딥러닝 기반의 CAD는 결과가 달라질 것으로 예상
- 256. From the ROC analysis, the accuracy of mammographic interpretations with CAD was significantly lower than for those without CAD (P = .002). The normalized partial area under the summary ROC curve was 0.84 for interpretations with CAD and 0.88 for interpretations without CAD (Figure 2). In this subset of 135 radi at least 1 mammogram associated sensitivity of mammography was 86.9%) with and 89.3%% (95% CI CAD. Specificity of mammograp 90.4%-91.8%) with and 91.3% (95% out CAD. Differences by Age, Breast Density, Men and Time Since Last Mammogram We found no differences in diagnos graphic interpretations with and w subgroups assessed, including pat menopausal status, and time si (Table 3). Intraradiologist Performance Measures f With and Without CAD Among 107 radiologists who interpr with and without CAD, intraradiolog improved with CAD, and CAD was a sensitivity. Sensitivity of mammogr 81.0%-85.6%) with and 89.6% (95% out CAD. Specificity of mammogra 89.8%-91.7%) with and 89.6% (95% out CAD. The OR for specificity b interpreted with CAD and those inte the same radiologist was 1.02 (95% C was significantly decreased for ma Figure 2. Receiver Operating Characteristic Curves for Digital Screening Mammography With and Without the Use of CAD, Estimated From 135 Radiologists Who Interpreted at Least 1 Examination Associated With Cancer 100 80 60 40 20 0 0 403020 True-PositiveRate,% False-Positive Rate, % 10 No CAD use (PAUC, 0.88) CAD use (PAUC, 0.84) Each circle represents the true-positive or false-positive rate for a single radiologist, for examinations interpreted with (orange) or without (blue) computer-aided detection (CAD). Circle size is proportional to the number of mammograms associated with cancer interpreted by that radiologist with or without CAD. PAUC indicates partial area under the curve. DCIS, ductal carcinoma in situ; exam, examination. a Odds ratio for CAD vs No CAD adjusted for site, age group, race/ethnicity, time since prior mammogram, and calendar year of the examination using with CAD use. b The 95% CIs for sensitivity and specificity are The accuracy of mammographic interpretations with CAD was significantly lower than for those without CAD (P = .002)
- 257. •인공지능은 의사를 대체하는가 •결과에 대한 책임은 누가 지는가 •인공지능의 의학적 효용을 어떻게 증명할 것인가 •의료 인공지능을 어떻게 규제할 것인가 Issues
- 258. • 의료기기가 아니라고 하기는 어려운 부분이 있다. • 치료 방법 선별 및 prioritisation는 의료 행위로 봐야 하는 것 아닌가? • 하지만, 의료기기라고 하더라도 현실적으로 관리가 쉽지 않다. • Evidence DB와 알고리즘의 업데이트: NCCN, 논문 + feedback loop • 업데이트 주기는 어떻게 되는가? • 피드백과 근거 자료의 퀄리티는 누가 어떤 기준으로 판단하는가? • 업데이트를 할 때마다 의료기기 변경 인허가를 새롭게 받아야 하는가? Watson for Oncology 규제의 어려움
- 259. Medical AI based on Deep learning • 임상시험: 전향적 or 후향적 • 얼마나 정확해야 정확한 것인가 • 학습 및 검증 조건 • 몇명의 중복 판독 • multi-center 임상을 요구? • 보험 수가를 매길 수 있는가 • cost-effectiveness 의 증명 • ‘행위별 수가제’ 의 이슈
- 263. 21st Century Cure Act •아래와 같은 소프트웨어는 의료기기에서 제외 •환자의 의료 정보 또는 임상 연구 결과나 가이드라인 등의 의료 정보를 표시, 분석, 출력하는 소프트웨어 •질병 예방, 진단, 치료에 관해 의료 전문가에게 권고 사항을 지원하거나 제공하는 소프트웨어로 •의료전문가가 그러한 권고의 근거를 독립적으로 검토할 수 있도록 제작된 소프트웨어 •다만, 의료 영상이나 체외 진단 기기로부터 나온 신호, 신호 획득 시스템으로부터 나온 신호나 패턴 등을 획득, 처리, 분석하기 위한 경우에는 의료기기로 분류 •위에 해당하는 비의료용 SW라도 다음에 관한 사실이 발견될 경우 FDA 관할 하에 둘 수 있음 •소프트웨어가 의도한대로 작동하지 않아 환자에게 위해를 끼칠 가능성이 있는 경우 •소프트웨어가 의료 전문가의 임상적 판단을 보장하지 않는 경우 •의료 전문가가 정보 또는 치료 권고의 근거를 검토할 수 있는 합리적 기회가 부여되지 않는 경우 •본래 제조자가 의도한 사용자와 사용 환경이 아닌 경우 SEC. 3060. “CLARIFYING MEDICAL SOFTWARE REGULATION.”
- 264. •기존에 Class III 로 관리되었던 CADx가 “암의 의심 병변을 찾아내는 목적”에 한해서 Class II 의료기기로 하향조정 (2017. 7. 19) •"Radiological computer-assisted diagnostic (CADx) software for lesions suspicious for cancer” •by Center for Devices and Radiological Health (CDRH) •Quantitative Insights의 QuantX라는 CADx 기기의 classification의 de novo request 에 대한 결정 •여타 목적의 CADx 에 대해서도 향후 영향을 미칠 수도 있을 것으로 예상 https://www.accessdata.fda.gov/cdrh_docs/pdf17/DEN170022.pdf 인공지능 의료기기에 대한 FDA의 최근 규제 완화
- 265. • 의사는 정말 인공지능의 오류 가능성에 대한 최후의 보루인가? • 의사를 보조 및 강화(augment)하는 인공지능은 검증이 필요 없는가? • 인공지능의 사용에 따라 의사의 숙련도에 이상이 생길 가능성은 없나? One more thing,
- 269. 1940년대 1950년대 1960년대 1980년대 •조종사1 •조종사2 •항공기관사 •항공사 •무선통신사 •조종사1 •조종사2 •항공기관사 •항공사 •조종사1 •조종사2 •항공기관사 •조종사1 •조종사2
- 270. “조종사가 없는 비행기의 시대가 열릴 것이다. 그건 단지 시간 문제일 뿐이다.” James Albaugh, Boeing, 2011
- 272. 3분
- 273. “항공 자동화의 정밀도가 높아지면서 조종사의 역할은 기계의 감시자나 감독관으로 전락했다” Hehmant Bhana, Advisor of Flight Safety Foundation (FSF)
- 274. 2명133명 2002-2011년1962-1971년 100만명 당 100만명 당 비행 자동화
- 275. 조종사들의 탈숙련화 (diskilling of the Crew) 자동화에 대한 지나친 의존이 조종사의 전문지식과 반사신경이 감퇴, 수동 비행 기술이 퇴화 •66명의 베테랑 조종사로 실험 •엔진이 폭발한 보잉737기를 조종 •수동 조종으로 착륙시키는 시뮬레이션 대부분 간신히 통과
- 276. 조종사들의 탈숙련화 (diskilling of the Crew) 자동화에 대한 지나친 의존이 조종사의 전문지식과 반사신경이 감퇴, 수동 비행 기술이 퇴화 •66명의 베테랑 조종사로 실험 •엔진이 폭발한 보잉737기를 조종 •수동 조종으로 착륙시키는 시뮬레이션 대부분 간신히 통과 실험 직전 두 달동안의 수동 비행 시간의 양과 조종능력이 상관관계
- 277. 의사들의 탈숙련화를 야기할까?
- 278. •인공지능의 의료 활용 •복잡한 데이터의 분석 및 권고안 도출 •영상 의료/병리 데이터의 분석/판독 •연속 데이터의 모니터링 및 예측 •새로운 이슈 •의사의 대체 가능 여부 •결과의 책임 소재 •근거 창출의 필요성과 어려움 •규제의 어려움과 원칙의 필요성 인공지능은 미래의 의료를 어떻게 혁신할 것인가
- 280. Feedback/Questions • Email: [email protected] • Blog: http://www.yoonsupchoi.com • Facebook: Yoon Sup Choi
- 281. Feedback/Questions • Email: [email protected] • Blog: http://www.yoonsupchoi.com • Facebook: Yoon Sup Choi