A COMPENSIVE STUDY ON AI FOR CANCER TREATMENTS

Soundarya B C,Bharatkumar KS, Hemankrishna S, PavanPani, Gururaghavendra, GaneshRaj

Artificial Intelligence and Machine Learning Department, Alva’s Institute of Engineering and Technology,
Mangalore, 574225, India

therapies. Finally, we will examine the effects

ABSTRACT: of various therapies on patient outcomes and
This study's objective is to demonstrate new the possibility for other medical uses. A type
deep learning applications in cancer imaging. of computer technology known as artificial
An algorithm may learn to program itself intelligence (AI) enables robots to carry out
using a vast body of samples that show the activities that would typically need human
desired behavior thanks to a family of intellect. Predictive analytics, medical
computing techniques called deeplearning. diagnosis, and robotics are just a few of the
Commonly used pharmacological effect uses for it. AI is an effective tool for illness
screening. Thodsdilute natural materials detection and medical research because it can
without separating out the active elements. swiftly and effectively examine vast volumes
Over the previous two decades, Strong active of data. By examining medical imaging and
isomeric molecules have been found and genetic data to find indicators of cancer, AI
isolated in modern medicine. Strong logical may be utilized to support diagnosis and
reasoning skills and autonomous learning therapy in the field of cancer treatment.
capabilities of artificial intelligence AI allow it Chemotherapy and radiation are the two
to mimic the way the human brain thinks. The most often used AI-assisted cancer therapies.
current method of developing anticancer AI is used in radiotherapy to determine the
drugs may be greatly optimized by using AI best radiation dosages and angular positions
technologies like machine learning. However, to target cancer cells. In AI-assisted
AI currently has certain relative limitations. chemotherapy, computer algorithms are used
This paper reviews the advancement of to examine In addition to these currently
artificial intelligence technologies in the study available AI-assisted cancer therapies, current
of anticancer drugs, including Deep Learning research is looking into methods to use AI
(DL) and Machine Learning (ML). We technology even more effectively. The
anticipate the development of Artificial creation of algorithms that can more precisely
Intelligence (AI) in the future at the same and early identify cancer is one potential field
time. of study. Additional research focuses include
KEYWORDS: AI, Cancer, drug discovery, developing tailored cancer medicines and
computational method, therapeutic target medication therapies utilising AI as well as
medications, AI Anticancer therapy Drug determining the best techniques for observing
development Precision therapy, Artificial and controlling adverse [Link]-assisted
intelligence Anticancer therapy Drug cancer therapies have a wide range of
development Precision therapy possible uses. With more precise diagnoses,
earlier detection, and more individualised
INTRODUCTION: treatment options, AI has the potential to
For decades, research into the application of
enhance patient outcomes. Also, it can lower
artificial intelligence (AI) in the medical arena
the likelihood of a false positive and lower the
has been expanding quickly. We may now
number of pointless procedures or
examine the potential of AI technology to help
[Link] technology may also aid in the
with cancer therapies thanks to recent
discovery of novel therapies and offer
advancements in the field. The present state
knowledge about the progression of cancer
of AI-assisted cancer therapies is examined, as
and the most effective ways to treat it. Finally,
well as their potential to enhance patient
AI has the ability to completely change how
outcomes. We will first go over some
cancer is treated. AI-assisted cancer therapies
fundamental AI principles and how they relate
might become more effective with more
to the treatment of cancer. The current
study and development.
cancer therapies that use AI will next be
discussed, along with their benefits and LITERATURE:
drawbacks. After that, we'll talk about the
latest advancements in AI-assisted cancer
These malignant cells might spread to those identified and treated. AI frequently uses
other bodily areas through the lymphatic Multilayer Neural Network Architecture
system or circulation. 2020 [Vesteghem C] The (MNNA) with logical reasoning and learning
rise of big data in biomedical research has capabilities, which may closely mimic human
changed cancer research. For scientists, thought processes.[5]AI can forecast the
handling complicated biological issues and efficacy of anticancer drugs or help in their
gathering information from numerous sources development. High-throughput

comes naturally. [1] Cancerous cells can screening data typically show a correlation
develop through a number of stages of between the genetic diversity of cancer cells
processes, including inherited features and and treatment efficacy. Different cancers and
environmental variables. The most prevalent the same drugs may have various response
malignancies to be discovered by 2020 are patterns. Lind et al.[6]The model predicts that
those of the breast, lung, colon, rectal, patients who have had paclitaxel for
prostate, skin, and stomach . [WHO 2020] endometrial cancer, 5-FU for gastric cancer, or
Every year, cancer claims the lives of one in six tamoxifen for ovarian cancer will be resistant.
people. About one-third of all cancer related The prognosis for each of these individuals
deaths are caused by high BMI, alcohol use, was found to be dismal. This study
poor diets high in processed foods, and demonstrates the huge potential of artificial
insufficient exercise.[2]By 2020, there will be intelligence in forecasting the sensitivity of
10 million cancer deaths, making it the top anticancer treatments.[7 ]The model predicts
cause of death among newly diagnosed that patients who have had paclitaxel for
cancers. The most common cancers are, lung endometrial cancer, 5-FU for gastric cancer, or
cancer (2.2 million cases),Breast cancer (2.2 tamoxifen for ovarian cancer will be resistant.
million cases), colon and rectum cancer (1.9 It was determined that each of these people
million cases), stomach cancer(1.0 million had a poor outlook. This work highlights the
cases), non-melanoma skin cancer (1.2 million enormous potential of artificial intelligence
cases), and prostate cancer(1.4 million cases). for predicting the sensitivity of anticancer
These are the most frequent reasons for therapy.[8]The explosive increase of imaging
cancer deaths [3] .Drugs that target cancerous data has significant effects on cancer
cells while preserving healthy cells are used in research, especially when [Link]
targeted treatment. Gene abnormalities that more metadata, such as genomes and
distinguish cancer cells from healthy ones are outcomes. Large sample numbers provide
common. A cell's DNA contains genes that significant power for the detection and
instruct it how to carry out numerous jobs. measurement of histologic patterns
Ones with gene mutations behave differently associated to clinically and physiologically
from healthy cells.[4] This has major relevant aspects, as demonstrated by recent
implications for early detection of cancers research. These approaches have the
using image data. Using AI, cancers may be potential to improve clinical and diagnostic
procedures[9].Even if a tool is scalable and already been discussed in a number of review
uniform, it won't speed up research until papers[20, 21]. However, we will concentrate
users incorporate it into their daily routines. on Artificial Intelligence applications for NGS
In order to make tools usable by those with and cancer genomes testing in our evaluation
little or no programming experience, software [22]. AI techniques for visualising cancer
should be created from the ground up with (cancer imaging)Recent advancements in
the target audience in mind. Application cancer imaging technology and analysis aid in
development centred on well defined APIs the visualisation of tumour spread, compound
can facilitate this [10]. effect screening, staging, and cancer therapy
PathML: A Computational Pathology Toolkit: decision-making. Since it deals with a variety
To address various needs, a number of tools of applications, such as breast cancer risk
are currently available in computational prognosis, radiomics and AI is an excellent
pathology. Some tools are intended to be fully platform for breast image analysis . A drug's
customizable, general-purpose libraries, while discovery and development can take over ten
others are just intended to execute certain years and cost an average of US$2.8 billion.
flows or work components. There are also Even then, nine out of 10 medicinal
commercial tools for digital pathology, some compounds fall short of passing regulatory
of which support machine learning analysis. approval and Phase II clinical
However, these proprietary tools are not trials[31,32].Increasing screening for high-risk
always the best choice for researchers groups is the current emphasis of breast
because they are more expensive and less cancer prevention and control measures, and
adaptable to develop than open source tools, early detection is a crucial part of such
which allow complete disclosure into the strategies[33].While no acknowledged
underpinning source code [11].TECHNIQUES: industry-accepted gold standard for image
These are some techniques which are used in segmentation methods has been found,
fields of cancer treatment such as Lung automatic image segmentation methods
Cancer, Brain tumour ,Breast cancer and often comprise thresholding-based, region-
prostate cancer AI and precision medicine based, edge-based, clustering-based, or other
Precision medicine, also known as segmentation approaches. Convolutional,
personalised medicine, treats illnesses by nonlinear, pooling, and fully connected layers
customising a patient's care based on their are all present in the computational route of
genomic, lifestyle, and environmental CNN, one of the most advanced deep learning
features [17]. the development of NGS and (DL) [Link] used in breast cancer
precision medicine, genetic profiles of screening can lessen or perhaps eliminate
patients are being employed more frequently certain apparent tumours being missed or
for disease diagnosis, risk assessment, and the misconstrued [34]
creation of focused treatments [17]. The ].Using Radiomics to diagnose
genomic profiles of patients include gene
expression in significant amounts, and it's Lung Cancer
interesting to note that ML classification Low-dose computed tomography (LDCT)
techniques applied to gene expression data images are utilised to determine radiomic
are not new [17,18]. In the past, microarrays characteristics that are often employed in
were used for thorough gene expression lung cancer screening and diagnosis.
analysis, but RNA-seq has since replaced them Numerous studies have demonstrated that, in
[19]. When outcome data or diagnosis contrast to molecular or blood testing, this
information is available, expression data are method helps the early diagnosis of lung
evaluated to identify the relevant genes in the cancer (25). With an accuracy of 79.06%, a
upregulated or downregulated pathways and sensitivity of 78.00%, and a specificity of
are also trained to predict the cancer 78.01%, Kumar et al. showed that radiomics is
subtypes and prognosis. Different ML a useful technique for discriminating between
applications on gene expression data have malignant and benign tumours(26).The
radiomic characteristics of lung cancer

3
patients' tumour, node, and metastasis (TNM) with PyQt5 and MariaDB, both of which are
stages have been established by Chaddad et intended for use in medical environments like
al(27). CNNs are used in DL radiomics, where hospitals and clinical pictures[37].
the model learns in a cascade manner without Breast cancer screening using
any previous feature definition and
necessitates a substantial quantity of data .By ultrasound and MRI
removing redundancies and preventing The incidence and mortality from breast
overfitting, the cascade approach analyses cancer have been steadily rising in recent
data to yield usable information[28,29].In years[38].Breast ultrasonography is another
order to give information that is invisible to widely used imaging diagnostic tool for breast
the human eye and enable more precise cancer because of its non-radioactive nature
grading, AI algorithms are able to assess and straightforward operation. Due to the
prognostication, monitoring, and treatment recent development of ultrasound
response evaluation photos at the pixel elastography, the accuracy of ultrasonography
level[30]. in the diagnosis of breast cancer has
increased, and the ability to distinguish
Brain tumour detection and between benign and malignant lesions with
identification using MRI greater specificity is now possible due to the
imaging semi-quantitative assessment of lesion
stiffness [39].In contrast to the traditional
The median filter is used to reprocess the
practices of sending pictures from the
MRI brain images for the analysis of the brain
ultrasound machine to a workstation for
tissue utilising morphological segmentation
image preprocessing, a DL technique (S-
and tumour extraction methods [35]. The
detect) may directly identify and highlight
study of tumours is now being done using 3D
breast tumors on the ultrasound system[40].
scanning thanks to recent technological
MRI, which unlike other modalities like
advancements. Brain tumour identification
magnetic resonance imaging, has no ionising
and classification using 3D image processing
radiation or other health dangers, is the
has been presented [36]. Many Deep learning
greatest approach to identify breast cancer in
frameworks were employed, including
women with dense breasts
MobileNetV2, MobileNetV3 small,
(MMs)[41,42]
MobileNetV3 big, VGG16,VGG19, and unique
CNN models. The most accurate was .The Use of AI in Prostate
CNN(Convolutional Neural Network). It Cancer Diagnostic Imaging
provides a solution combining a CNN created
Depending on how the disease presents,
with Keras and Tensor flow with a fully
prostate cancer can be low-risk (indolent) or
functional cross-platform application created
high-risk (aggressive), with a wide range of of the method. MRI-TRUS fusion-guided
risk. Although most locally contained diseases biopsies[15].
are expected to have excellent cancer-specific
survival rates, early identification and Conclusion
localisation of prostate cancer at a stage that
Artificial intelligence has made significant
may be treated is crucial. Transrectal
contributions to the creation and use of
ultrasonography (TRUS), the most popular
anticancer medications. The ongoing
procedure for determining whether or not a
advancement of AI and its incredible tools
man has prostate cancer, is used to sample
aims to decrease the difficulties faced by
twelve different areas of the prostate .It is
pharmaceutical companies, impacting both
typically advised following an elevated PSA
the process of developing new medications
blood test or an abnormal digital rectal
and the lifespan of the final product, which
examination[12].The use of MRI to
may explain the growth in the number of
distinguishing between clinically relevant and
start-ups in this sector. Machine learning is
non-significant PCs has been demonstrated in
becoming a crucial tool for cancer researchers
a number of noteworthy studies. Was
as the size and complexity of biomedical
developed as an alternative to transrectal
imaging datasets skyrocket (i.e., the amount
ultrasonography-guided biopsy to direct the
of data contained in each sample). Software
pathologist to the precise place for removing
solutions must be created with an emphasis
the tissue samples[13,14,15].MRI-TRUS
on scalability, standardization, and
fusion-guided needle biopsies to efficiently
accessibility in order to support these
predict prostate cancer This method is
activities. Pathologists and healthcare experts
arduous and time-consuming, so AI has been
may profit from deep learning technology by
added to automate the procedure to lighten
working more productively, boosting up
the load on clinicians. Three institutions using
standards for quality, and offering decent
an Aixplorer (Supersonic Imagine, Aix-en-
prognosis. AI has considerably aided in the
Provence, France), an iU22 (Philips
design and development of novel cancer
Healthcare, Bothell, WA, USA),and a Pro Focus
drugs by successfully using imaging, genomic,
2202a (BK Medical) ultrasound scanner,
and cellular cancer data. The proper
respectively, provided a dataset with TRUS
combination of radiomics and deep learning
images. 436 photos of 181 guys were found in
will surely result in the production of a strong,
the databases that were used in the study.
sensitive, and accurate computer-aided
This deep learning technique was created to
diagnosis tool for radiologists in the not too
automatically segment TRUS pictures for MRI.
distant future. Growing data streams and
The model was assessed using the Hausdorff
advancements in algorithms enable AI to
distance (3.0 mm), the Jaccard index, and its
improve medical oncology through properly
median accuracy (98%). (0.93).
scrutinised, specific applications engaging at
The periphery and transition zones' pixel-wise
specific points along the cancer therapy
segmentation accuracy rates were reported to
continuum.
be 97 and 98%, respectively. Targeting
malignant lesions has become faster because
Reference
[1] Amara R. 1984. New directions for futures research: setting the [Link] 36 (1–2), 43–47.
[2] Ambrosini R, Wang P, Kolar B, O’Dell W. 2008. Computer-aideddetection of metastatic brain
tumors using automated 3-D templatematching. Proc Intl Soc Mag Reson Med., 31(1):85-93. doi:
10.1002/jmri.22009.
[3] Aresta G., Jacobs C., Araújo T., Cunha A., Ramos I., van Ginneken B.,Campilho A. 2019. iW-Net: an
automatic and minimalistic interactivelung nodule segmentation deep network. Scientific Reports,
vol. 9, n. 1,n. article 11591, DOI: 10.1038/s41598-019-48004-8 Bi W. L., Hosny A.,Schabath M. B.,
Giger M. L., Birkbak N. J., Mehrtash A., Allison T.,Arnaout O., Abbosh C., Dunn.

5
[4] Bray F., Ferlay J., Soerjomataram I., Siegel R.L., Torre L.A., Jemal A. [Link] cancer statistics
2018: GLOBOCAN estimates of incidence andmortality worldwide for 36 cancers in 185 countries. CA
Cancer J [Link];68(6):394-424. doi: 10.3322/caac.21492. Epub 2018 Sep 12.
[5] Chagpar A. B., Coccia M. 2019. Factors associated with breast cancer mortality-per-incident case
in low-tomiddle income countries (LMICs).Journal of Clinical Oncology, vol. 37, no. 15, suppl. pp.
15661566,DOI: 10.1200/JCO.2019.37.15
[6] C.M. Lo, U. Iqbal, Y.J. Li, Cancer quantification from data mining to artificial intelligence, Comput.
Methods Programs Biomed. 145 (2017) A1.
[7] J. Abbasi, Artificial intelligence tools for Sepsis and Cancer, JAMA 320 (22) (2018) 2303.
[8] C.M. Lo, U. Iqbal, Y.J. Li, Cancer quantification from data mining to
artificial intelligence, Comput. Methods Programs Biomed. 145 (2017) A1.
[9] J. Abbasi, Artificial intelligence tools for Sepsis and Cancer, JAMA 320(22) (2018) 2303.

[10] W.L. Bi, A. Hosny, M.B. Schabath, M.L. Giger, N.J. Birkbak, A. Mehrtash,T. Allison, O. Arnaout,
[Link], I.F. Dunn, R.H. Mak, R.M. Tamimi, C.M. Tempany, C. Swanton, [Link], L.H. Schwartz,
R.J. Gillies, R.Y. Huang, H. Aerts, Artificial intelligence in cancer imaging: clinical challenges and
applications, CA Cancer J. Clin. 69 (2) (2019) 127–157.
[11] L. Lin, Q. Dou, Y.M. Jin, G.Q. Zhou, Y.Q. Tang, W.L. Chen, B.A. Su, F. Liu, C.J. Tao, N. Jiang, J.Y. Li,
L.L. Tang, C.M. Xie, S.M. Huang, J. Ma, P.A. Heng, J.T.S. Wee, M.L.K. Chua, H. Chen, Y. Sun, Deep
learning for automated contouring of primary tumor volumes by MRI for nasopharyngeal carcinoma,
Radiology 291 (3) (2019) 677–686.
[12] Tegmark M. Life 3.0. UK: Penguin Random House; 2017.
[13] Hinton GE, Salakhutdinov RR. Reducing the dimensionality of data with neural networks. Science
2006;313(5786):504e7.[Link] science.1127647.
[14] Heathfield H, Kirkham N. A cooperative approach to decision support in the differential
diagnosis of breast disease. Med Inf 1992;17(1):21e33.
[15] Heathfield HA, Winstanley G, Kirkham N. Computer-assisted breast cancer grading. J Biomed
Eng 1988;10(5):379e86. [9] Leaning MS, Ng KE, Cramp DG. Decision support for patient management
in oncology. Med Info 1992;17(1):35e46.
[16] Houssami N, Kirkpatrick-Jones G, Noguchi N, Lee CI. Artificial Intelligence (AI) for the early
detection of breast cancer: a scoping review to assess AI’s potential in breast screening practice.
Expert Rev Med Devices 2019;16(5): 351e62. [Link]
[17] Lin E, Kuo PH, Liu YL, Yu YW, Yang AC, Tsai SJ (2018) A deep learning approach for predicting
antidepressant response in major depression using clinical and genetic biomarkers. Front
Psychiantry 9:290.[Link]
[18] Liu Y, Liang Y, Wishart D (2015) PolySearch2: a significantly improved text-mining system for
discovering associations between human diseases, genes, drugs, metabolites, toxins and more.
Nucleic Acids Res 43:W535–W542. [Link] org/10.1093/nar/gkv383

[19] Low SK, Zembutsu H, Nakamura Y (2018) Breast cancer: The translation of big genomic data to
cancer precision medicine. Cancer Sci 109:497–506. [Link]
[20] Lyu B, Haque A (2018) Deep learning based tumor type classification using gene expression
data. [Link]
[21] Ma C, Zhang Y, Zhang M (2015) Tree Kernel-based Protein-Protein interaction extraction
considering both modal verb phrases and appositive dependency features. In: Paper presented at
the proceedings of the 24th international conference on world wide web—WWW ’15 companion
MacArthur DG et al (2014) Guidelines for investigating causality of sequence variants in human
disease. Nature 508:469–476. https ://[Link]/10.1038/nature13127
[22] Mahmood AS, Wu TJ, Mazumder R, Vijay-Shanker K (2016) DiMeX: a text mining system for
mutation-disease association extraction. PLoS One
11:e0152725. [Link]
[23] Mallory EK, Zhang C, Re C, Altman RB (2016) Large-scale extraction of gene interactions from
full-text literature using DeepDive. Bioinformatics 32:106–113. [Link]
tics/btv476
[24] Meng C, Zeleznik OA, Thallinger GG, Kuster B, Gholami AM, Culhane AC (2016) Dimension
reduction techniques for the integrative analysis of multi-omics data. Brief Bioinform 17:628–641.
[Link]
25. National Lung Screening Trial Research Team . Reduced lung-cancer mortality with low-dose
computed tomographic screening. 10.1056/NEJMoa1102873 [PMC free article] [PubMed] [CrossRef]
[Google Scholar]
26. Kumar D, Chung AG, Shaifee MJ, et al. Discovery radiomics for pathologically-proven computed
tomography lung cancer prediction. In International Conference Image Analysis and Recognition
2017 (pp. 54-62). Springer, Cham. [Google Scholar]
27. Chaddad A, Desrosiers C, Toews M, et al. Predicting survival time of lung cancer patients using
radiomic analysis. Oncotarget 2017;8:104393. 10.18632/oncotarget.22251 [PMC free article]
[PubMed] [CrossRef] [Google Scholar]

[31]. Álvarez-Machancoses Ó, Fernández-Martínez J.L. Using artificial intelligence methods to speed

up drug discovery. Expert Opin. Drug Discovery. 2019;14:769–777. [PubMed] [Google Scholar]
[32]. Fleming N. How artificial intelligence is changing drug discovery. Nature.
2018;557 S55–S55. [PubMed] [Google Scholar]
[33]Cedolini, C.; Bertozzi, S.; Londero, A.P.; Bernardi, S.; Seriau, L.; Concina, S.; Cattin, F.; Risaliti, A.
Type of breast cancer diagnosis, screening, and [Link]. Breast Cancer 2014, 14, 235–240.
[Google Scholar] [CrossRef]
[34].Rodriguez-Ruiz, A.; Lang, K.; Gubern-Merida, A.; Broeders, M.; Gennaro, G.; Clauser, P.; Helbich,
T.H.; Chevalier, M.; Tan, T.; Mertelmeier, T.; et al. Stand-Alone Artificial Intelligence for Breast
Cancer Detection in Mammography: Comparison with 101 Radiologists. J. Natl. Cancer Inst. 2019,
111, 916–922. [Google Scholar] [CrossRef]
[35] D. V. Gore and V. Deshpande, “Comparative study of various techniques using deep learning for
brain tumor detection,” in 2020 International Conference for Emerging Technology (INCET), Jun.
2020, pp. 1–4, doi: 10.1109/INCET49848.2020.9154030.
[36] Tuhin [Link]., “Detection and 3d visualization of brain tumor usingdeep learning and
polynomial interpolation,” IEEE Asia-Pacific Conferenceon Computer Science and Data Engineering,
pp. 1-6, 2020, doi:[Link]
[37] S. Pokhrel. [Link]., "Brain Tumor Detection Application Based onConvolutional Neural
Network," International Conference on IntelligentTechnologies (CONIT), pp.1-5, 2021, doi:
[Link] 10.1109/CONIT55038.2022.9848177.
[28]Lecun, Y.; Bengio, Y.; Hinton, G. Deep learning. Nature 2015, 521, 436–444.[Google Scholar]
[CrossRef]
[29]Vaz, J.M.; Balaji, S. Convolutional neural networks (CNNs): Concepts and applications in
pharmacogenomics. Mol. Divers. 2021, 25, 1569–1584. [Google Scholar] [CrossRef]
[30]Aneja, S.; Chang, E.; Omuro, A. Applications of artificial intelligence in neuro-oncology. Curr.
Opin. Neurol. 2019, 32, 850–856. [Google Scholar] [CrossRef] [PubMed]

[39].Giuseppetti, G.M.; Martegani, A.; Di Cioccio, B.; Baldassarre, [Link] in the diagnosis
of the nodular breast lesions: Preliminary report. Radiol. Med. 2005, 110, 69–76. [Google Scholar]
[38].Ferlay J, Soerjomataram I, Dikshit R, Eser S, Mathers C, Rebelo M, Parkin DM, Forman D, Bray F.
Cancer incidence and mortality worldwide: sources,methods and major patterns in GLOBOCAN
2012. Int J Cancer.2015;136:E359–E386. [PubMed] [Google Scholar]
[40].Di Segni et al[32] also evaluated the diagnostic performance of S-detect in the assessment of
focal breast lesions. S-detect showed a sensitivity > 90% and a 70.8% specificity, with inter-rater

7
agreement ranging from moderate to good.S-detect may be a feasible tool for the characterization
of breast lesionsand assist physicians in making clinical decisions.
[41].Teh, W.; Wilson, A.R.M. The role of ultrasound in breast cancer screening. A consensus
statement by the European Group for breast cancer screening. Eur. J. Cancer 1998,34, 449–450.
[Google Scholar] [CrossRef]
[42].Sardanelli, F.; Giuseppetti, G.M.; Panizza, P.; Bazzocchi, M.; Fausto, A.;Simonetti, G.; Lattanzio,
V.; Del Maschio, A. Sensitivity of MRI Versus Mammography for Detecting Foci of Multifocal,
Multicentric Breast Cancer in Fatty and Dense Breasts Using the Whole-Breast Pathologic
Examination as a Gold Standard. Am. J. Roentgenol. 2004, 183, 1149–1157. [Google
Scholar][CrossRef].