Microwave Imaging and other Imaging Modalities for Breast Cancer

Detection

1. INTRODUCTION

Breast cancer, the second leading cause of death around the world, is uncontrolled growth of
abnormal breast cells that have the ability to spread to other sites in the body if not treated
[1]. A total of 2.3 million cases of breast cancer and 685000 related deaths were reported in
2022 [2]. From 1990s to 2000s the breast cancer mortality rate has been decreasing slowly
[3], by the improvement in early diagnosis and treatment through many conventional to
digital technologies like mammography, computed tomography, ultrasonography, magnetic
resonance, nuclear medicine and positron emission tomography and microwave imaging
(MWI) [4].

Contrast enhanced mammography is a standard method for early detection of breast cancer
providing significant morphologic enhancement combined with low dose [5], low mortality
rate with early screening [6], and diagnosing the type and the malignancy of lesion [7]. An
analysis occurred on 29 radiological methods concluded that although ultrasound is a good
modality for breast assessment, however for clinical implications more advancement in
diagnostic methods is required [8]. Considering computed tomography for breast screening
and staging, fluorodeoxyglucose positron emission tomography coupled with CT (FDG-
PET/CT) is effective for the staging of ductal and lobular carcinomas [9], and dual-energy
CT for accurate tumor size assessment [10]. With the advancement in technology CT images
detection using deep learning, helps to detect breast cancer on CT [11], and machine
learning, differentiate between breast lymphoma & carcinoma [12], technologies have also
showed significant improvement.

Besides the above modalities and their effectiveness in breast cancer screening another
modality named microwave imaging (MWI) is currently being used as an advanced
technology. MWI is a non-ionizing, non-invasive modality that uses low power microwaves
to diagnose the tumor cells with electromagnetic waves through antenna [13], differentiating
between the dielectric properties of normal and cancerous tissues [14]. The First-In-Human
clinical investigation of the wavelia system, which is the first generation breast imaging
device, using a low-power electromagnetic wave for microwave breast imaging (MBI)
demonstrated that rather than conventional mammography this MBI prototype provides a
great potential for the detection of breast pathologies [15].

Radar based techniques to detect tumors with breast’s interior [16] and scattering based
methods of MWI to reconstruct the imaging tissues [17] have been experimented and proved
to be effective as a non-invasive diagnostic modality. An advanced technology of MI called
SAFE (Scan and Find Early) is also a useful modality to investigate the abnormal breast
mass because of non ionizing radiation and no compression of breast is required [18].
Further, we are going to evaluate different types of reconstructing algorithms used in MI of
breast and a comparative analysis of microwave breast imaging to other modalities is being
discussed.

2 Imaging modalities:

2.1 MAMMOGRAPHY:

A randomized control study showed mammography screening reduces BC mortality risk in all
ages women but significantly in women aged 5o-69 [19]. Breast varies in density making it
harder to detect and categorize masses unlike calcification, a systematic study between
traditional ML and DL methods for BC diagnosis shows that traditional ML is limited to specific
density types or datasets while DL methods offer promise in breast cancer diagnosis [20]. An
experiment was conducted on a hybrid transfer learning model in which different classification
algorithms including ALEXNET, ResNet50, Mobile Net, VGG16, VGG19, etc comparison
implemented on datasets of 2D and 3D images are studied using data argumentation and pre-
training techniques. The result showed that the MVGG network, pre-trained on ImageNet has a
high accuracy rate and AUC value and an algorithm in which the clinical classification threshold,
set below the mathematical threshold, decreases false negative instances in mammograms,
leading to enhanced 5-year survival rates [21].
An experiment was conducted for classifying mass lesions in mammography images using data
in a supervised manner instead of designing specific descriptors and conventional neural
networks (CNNs). The result proved that both learned and hand-crafted features combination
proves to be more effective for classifying mass lesions in mammography images [22].
Consecutive mammography screening examinations before the breast cancer diagnosis provide
the highest protection against BC mortality as compared to non-consecutive mammography
screening examinations [23].
A cross-sectional analytical study was conducted to assess the efficacy of digital mammography
in the diagnosis of breast cancer. The result revealed digital mammography is highly effective
with a sensitivity of 97% and specificity of 64.5% having a positive predicted value of 89% and
a negative predicted value of 90.9%. The overall diagnostic accuracy was 89.3% and also the
radiation dosage to the patient was very low compared to screen-film mammography. Based on
these results the adoption of digital mammography over screen-film mammography is
recommended to detect and diagnose breast cancer. [24]

2.2 CONTRAST-ENHANCED MAMMOGRAPHY(CEM):

Contrast-enhanced mammography has higher sensitivity (p<0.001), specificity (p = 0.016), and

overall accuracy (p < 0.001) than conventional mammography. The adoption of CESM could
potentially reduce false negatives, particularly among women with dense breast tissue [25].
CESM could be regarded as a valuable tool for the preliminary evaluation of breast lesions [26].
In the case of detecting cancerous lesions, contrast-enhanced mammography, molecular breast
imaging, and MRI possesses the same abilities. However, MRI tends to identify more additional
suspicious benign lesions unrelated to the primary tumor as compared to contrast-enhanced
mammography or molecular breast imaging resulting in lower accuracy in predicting true
positive cases causing unnecessary biopsies. Additionally, all three modalities tend to provide
larger measurements for the size of the primary tumor comparable to its actual size especially
MRI[27].

A study was conducted to check the sensitivity of Contrast-enhanced mammography to MR

imaging both are similar effective for assessing the spread of newly diagnosed breast cancer but
contrast-enhanced mammography has high PPV and less false positive findings as compared to
MR imaging. It has the ability to illustrate lesion shape and blood flow characteristics at a lower
cost[28]. Another analytical study comparing contrast-enhanced mammography and MRI for
breast cancer detection revealed that both techniques have high sensitivity but their specificity is
relatively modest resulting in higher false positive findings and unnecessary procedures like
biopsies. The Alternative diagnostic method confirmed that AUC values for CESM and MRI
were both near 1.0, showing excellent diagnostic accuracy. CESM appeared to outperform MRI
in breast cancer detection based on AUC analysis. Contrast-enhanced spectral mammography in
patients referred from the breast cancer screening program[29].

A pilot study comparing CEM and conventional digital mammography revealed that CEM
appeared to enhance confidence levels, particularly in dense breast tissues regardless of the level
of enhancement seen in the background [30]. A study was conducted using major databases
comparing CEM and CE-MRI in breast cancer diagnosis. The quality of the studies was assessed
based on the QUADAS-2 criteria and combined sensitivities and specificities of CEM and CE-
MRI were determined through the application of a random-effects model using the STATA
software's "metaprop" function. The result revealed the effectiveness of CEM as a screening
method extends to intermediate-risk women, including those with dense breasts and a history of
breast cancer. Both CE-MRI and CEM demonstrate strong sensitivity in detecting breast
abnormalities. However, CE-MRI shows a comparatively lower specificity in identifying breast
lesions yielding more false positive results [31].

However, CEM might not be as effective as MRI at finding hidden areas of cancer in a pre-
surgery phase that could influence the surgical approach and patient outcomes [32].

2.3 ULTRASONOGRAPGY(USG):

A supplemental USG can detect the hidden breast cancer that mammography might overlook.
Primary USG and primary mammography have proven to be same effective in women with
dense breasts regarding sensitivity, specificity, cancer detection rates, and biopsy rates. However,
P-US has elevated recall rates and a higher frequency of invasive breast cancer [33]. In dense
breasts, mammography is not efficient enough to detect breast cancer because both dense tissues
and tumors appear white on images. A study was conducted on the base of a Multi-modality
Independent Screening Trial (MIST), among Chinese women to check the efficacy of additional
USG after mammography. The result shows that the additional USG after mammography
benefits women with BIRADS 0 to 2 scores. It is advised not only in women with dense breasts
but also in women with non-cancerous breast conditions [34].

In the case of breast tumor detection, a study using diverse datasets conducted compared the
effectiveness of three imaging methods mammography, MRI, and ultrasonography. It found that
MRI was the most accurate than mammography and USG. However, USG proved to be little
more effective than mammography [35]. Contrast-enhanced ultrasonography is effective in
distinguishing among breast cancer subtypes. Each subtype demonstrates distinct enhancement
characteristics and perfusion parameters aiding in precise diagnosis [36]. The retrospective study
was performed comparing the size measurements of breast masses by 2D US and 3D UG. The
result found big differences in size measurements especially in cancerous tumors. The CEUS
seemed to be more accurate in tumor size evaluation optimizing clinical results and patient
treatment decisions in breast cancer care[37].

A study was conducted using a computer-aided diagnosis (CAD) model integrating a wavelet
neural network (WNN) and the grey wolf optimization (GWO) algorithm for improved breast
cancer diagnosis using USG. The result found that the integration of grey wolf optimization
(GWO) with wavelet neural network (WNN) in the proposed CAD model markedly enhances
the efficiency and accuracy of breast cancer diagnosis using ultrasound imaging [38]. A
retrospective analysis proved that the USG has greater sensitivity in detecting suspicious breast
lesions. This suggests that incorporating USG with mammography can enhance the detection and
diagnostic accuracy of breast cancer [39].

A study conducted to assess the diagnostic precision of conventional ultrasound (US),

contrast-enhanced US (CEUS), and dynamic contrast-enhanced magnetic resonance
imaging (DCE-MRI) in evaluating the breast tumor dimensions proved that the CEUS
and DCE-MRI are similarly accurate in evaluating breast cancer size, with CEUS being
more convenient and faster. Conventional US often underdiagnose lesion size, while
 CEUS and DCE-MRI are more likely to overdiagnose[40]. CEUS proved very effective
in distinguishing small breast neoplasms[41]. CEUS with SonoVue in more effective in
detecting penetrating ductal carcinoma than gre- scale ultrasound before surgery thus
optimizing surgical planning and treatment outcomes[42].

2.4 COMPUTED TOMOGRAPHY (CT)

CT scan is a best modality for imaging the soft tissues in a 3D manner. CT chest, for breast
diagnosis, investigates the soft tissues deeply by removing the overlapping tissues. CT chest
is a helpful modality for patients having dense breast parenchyma as it shows more false-
negative results (cancer and CT interpreted as negative) as compared to mammography
having more false-positive results ( no cancer and mammogram interpreted as positive) [43],
suggesting biopsy for 60% of cases.

An advanced molecular imaging technique conducted experiment with 18F-fluorodeoxy

glucose (18F-FDG) on positron emission tomography/computed tomography (PET/CT) for
breast Ca diagnosis. The effect of 18F-FDG PET/CT is evaluated based on its metabolic
parameters including the maximum standardized uptake value (SUVmax), metabolic tumor
volume (MTV) and total lesion glycolysis (TLG) indicating that the higher the levels of
metabolic parameters, the more the chances of recurrence of Ca will be. For this analysis 2
groups were made which showed that there were high levels of metabolic parameters in the
event group (the group with chances of recurrence) with p-value (p < 0.05) as compared to
the event-free group [44]. Moreover, 18F-FDG PET/CT radiomics based experiments, which
use mathematical analysis to identify the results that are not perceive by human eye, are
effective to differentiate between the breast carcinoma and lymphoma because of the
difference in the uptake of FDG by both abnormalities as analyzed through radiomics [12].
Similarly considering a comparison based study conducted between the [18F]-FDG
PET/MRI, CT and MRI clearly identified that PET/MRI shows more sensitivity and is more
effective in detecting nodal metastasis as compared to MRI or CT which are less sensitive
[45].
Another technique called Dual-energy CT (DECT) has been used for staging of invasive
carcinomas including invasive ductal carcinoma (IDC) and invasive lobular carcinoma (ILC)
as compared to other lesions. Based on the iodine concentration both IDC and other lesions
can be differentiated. Threshold value of 6.13 mg/mL is optimal for distinguishing both types
of lesions under receiver operating characteristic curve (ROC) [10].

2.5 MAGNETIC RESONANCE IMAGING (MRI)

A systemic comparison conducted to check the effectiveness of contrast-enhanced

mammography (CEM) and MRI of breast showed that the sensitivity was high 96% for both
modalities with 77% specificity however the diagnostic odds ratio (DOR) indicated better
diagnosis of MRI as compared to CEM [46]. A clinical study providing images at dynamic
contrast material-enhanced (DCE) breast MRI evaluated the images through conventional
computer aided software and also through artificial intelligence (AI) based software. Using
AI the area under the operating characteristic (ROC) curve (AUC) as from 0.71 to 0.76 and
also the sensitivity improved, while specificity remained constant showed that different
breast tumors can be differentiated with the use of advanced AI models [47].

The Dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI) based on deep

learning (DL) evaluated most of the cancers as benign and no need to perform biopsies for
such cancers thus benefiting the patient from further complications [48].

To detect breast cancer association with breast parenchymal enhancement (BPE) a study
analyzing different parameters of BPE in premenopausal and postmenopausal women
evaluated that with moderate BPE in premenopausal women there is risk of Ca [odds ratio
(OR) 1.49, 95% CI 1.05–2.11; p = 0.02] as compared to minimal to moderate BPE in
postmenopausal women with no statistically significant results (OR 1.45, 95% CI 0.92–2.27;
p = 0.1) [49]. To check the disease-free survival (DFS) after treatment in patients with triple-
negative breast cancer (TNBC), a radiomics study based on MRI features was developed in
order to check the DFS in patients. Radiomics study, based on tumor related factors from
MRI, showed the chances of recurrence of TNBC after treatment however when combined
with the clinical, patient related factors, called clinicopathologic-radiomic (CCR) model
showed effective results [50]. (add statistical values after evaluation)

2.6 DIGITAL BREAST TOMOSYNTHESIS

Breast tomosynthesis is a new type of digital x-ray mammogram that uses low dose x-rays
and computer reconstructions to create three-dimensional (3-D) images of the breasts, also
called 3-D mammography and digital breast tomosynthesis (DBT) reference

Synthesized mammograms (SM) constructed from DBT image acquisition also called
synthesized 2D (s2D) mammography has proved to be more effective than digital
mammography (DM) through various studies investigating that SM/DBT and DM/DBT has
lower recall rate [51, 52] but similar CDR as compared to DM alone [51, 52] and similar
CDR in comparison of SM/DBT with Full-Field Digital Mammography (FFDM)/DBT [53],
lower biopsy rate of SM/DBT as compared with DM/DBT and DM alone [52] and almost
similar interval cancer rate (ICR), that is the cancer diagnosed in the time interval between
the normal regular mammogram and the next screening mammogram [52]. Moreover, DBT
alone is more sensitive than DM alone however there is comparable specificity between
combined DBT and DM, and DM alone [54]. Investigating the positive predictive value 1
(PPV1), both methods combined DBT and DM, and combined DBT and S2D have similar
accuracy in predicting cancer [51] and similar accuracy in comparison between SM/DBT and
FFDM/DBT [53] whereas no difference is detected for DBT alone and DM alone [51].
SM/DBT images, as compared to FFDM, is efficient in detecting breast lesions using less
radiation dose [55].

2.6 USE OF AI

3. MICROWAVE BREAST IMAGING (MBI)

An emerging imaging technology named microwave imaging (MWI) aid in breast Ca

diagnosis using the microwaves frequency range from 0.5 to 30 GHz [56]. 300 MHz-300
GHz [57]Microwave breast imaging (MBI) as demonstrated from the First-In-Human
clinical investigation, is a non-invasive diagnostic modality to detect most of the breast
cancers [58]. Currently MRI derived breast models with benign and malignant tumors have
been developed in order to test for the microwave imaging research [59].

3.1 DIELECTRIC PROPERTIES OF BREAST

 Increases in real permittivity and conductivity in the breast over the areas
where benign and malignant anomalies were located [60] ( consider it
below)

Dielectric properties, which are the properties related to how the tissue responds to
electromagnetic fields reference missing, may help distinguish between the normal tissue and
the metastatic lesions of breast by evaluating the change in dielectric properties of different
tissues [61]. High dielectric differences between the metastasized and surrounding tissues
help determine the tumor location [62]. The permittivity εr and conductivity σ (S/m) values
of axillary lymph nodes (ALN) calculated with T1-w Dixon-W images at 6 GHz and
compared with different tissues have shown that the values for lymph nodes are below the
average values for fibro-glandular tissue as in table 1 and these values are further used to
check the efficiency of MWI system in detecting metastasized lymph node lesions [56].

Tissues

Fibro-glandular tissues Lymph nodes

εr 42.47 32.54

σ (S/m) 5.42 3.61

Table 1
Further through another methodology, numerical models of the axillary region developed
from MR images were evaluated which showed 29% contrast in dielectric properties between
healthy and metastasized lymph nodes concluding that the development of ALN-MWI can be
helpful to distinguish between the healthy and metastasized lesions [63]. The estimated
relative permittivity of those ALNs ranged from 16.6 to 49.3 at 5 GHz.

 Also read for frozen lymph node measurement

Ex vivo dielectric properties of induced benign tumors and healthy mammary tissues of rat,
studied for the first time indicated that the behavior of benign tissue differs from itself as
well as from the healthy tissues mainly affected by the adenosis tissue and similar behavior is
expected in human tissues also [64].

MI system using wide-band microstrip patch antenna detected the presence of breast tumors
based on the dielectric properties when experimented on a breast model, demonstrating that
the presence of tumor shows low value of reflection coefficient but high peak when
processed through the confocal microwave imaging (CMI) algorithm comparatively in the
absence of tumor. Moreover considering the thermal properties, cancerous tumor absorbs
more energy thus increasing the specific absorption rate (SAR) and resulting in high
temperature of breast model [65]. Using the mono-static radar imaging system, the position
of the malignant and benign tumor is determined through the difference in dielectric
properties. Due to the irregular morphology of the malignant tumor, there is more damping
factor and more noise accumulates around it thus distinguishing it from the benign tumor.
Further clearer images are obtained using time domain signal because of lower noise levels
rather than the frequency domain. Moreover images obtained from lower frequency provide
clear location as compared to at high frequency because of the frequency related effects of
antenna [62].

3.2 TYPES
IS FACTORIZATION METHOD A METHOD OF MICROWAVE IMAGING?

3.2.1 Radar Based Techniques

Radar based MWI device widely detect pathologically proven breast cancers in a safe
manner [66]. Radar based MBI conducted using ultrawide band (UWB) antipodal Vivaldi
antenna on 50 mm radius breast models detected the tumors at the exact location inside the
breast and proved effective in detecting cancer at early stage [67]. Portable microwave
imaging performed by using a field-programmable gate array (FPGA) proposed hardware,
speed up the processing of MIST algorithm [68] thus FPGA can be helpful in wearable MBI
because of its low-cost, low power consumption and fast processing of images. A microstrip
patch antenna of the size (30 mm × 30 mm × 1.6 mm) having the substrate material FR4 (fire
retardant 4) has developed and can be used for tumor detection of spherical breast as of its
minimum return loss, high gain and flexibility to distribute around the breast [69].
Introducing the inst-fed rectangular microstrip patch antenna utilizing RT substrate for the
development of realistic models exaggerate early cancer detection rate with the use of Delay
and Sum (DAS) algorithm reconstruction [70].

3.3 USE OF AI
Unlike the conventional MWI approaches to detect breast cancer, deep learning algorithm
has greatly improved the diagnosis particularly of small sized tumors by using breast masses
of different sizes [71]. 3D reconstruction of the images, obtained by contrast-source
inversion (CSI) method, using convolutional neural network (CNN) show that along with the
3D reconstruction, this deep learning algorithm also demonstrate the location of tumor [72].
1. Chaudhury, S., et al., Breast cancer calcifications: identification using a novel segmentation
approach. 2021. 2021: p. 1-13.
2. Arnold, M., et al., Current and future burden of breast cancer: Global statistics for 2020 and
2040. 2022. 66: p. 15-23.
3. Giaquinto, A.N., et al., Breast cancer statistics, 2022. 2022. 72(6): p. 524-541.
4. Younis, Y.S., et al., Early diagnosis of breast cancer using image processing techniques. 2022.
2022: p. 1-6.
5. Jochelson, M.S. and M.B.J.R. Lobbes, Contrast-enhanced mammography: state of the art. 2021.
299(1): p. 36-48.
6. Duffy, S.W., et al., Mammography screening reduces rates of advanced and fatal breast cancers:
Results in 549,091 women. 2020. 126(13): p. 2971-2979.
7. Nicosia, L., et al. Contrast-enhanced spectral mammography in the evaluation of breast
microcalcifications: controversies and diagnostic management. in Healthcare. 2023. MDPI.
8. Manhoobi, I.P., et al., Diagnostic accuracy of radiography, digital breast tomosynthesis, micro-
CT and ultrasound for margin assessment during breast surgery: A systematic review and meta-
analysis. 2022. 29(10): p. 1560-1572.
9. Groheux, D., E.J.C. Hindie, and T. Imaging, Breast cancer: initial workup and staging with FDG
PET/CT. 2021. 9(3): p. 221-231.
10. Volterrani, L., et al., Dual-energy CT for locoregional staging of breast cancer: preliminary
results. 2020. 214(3): p. 707-714.
11. Yasaka, K., et al., Impact of deep learning on radiologists and radiology residents in detecting
breast cancer on CT: a cross-vendor test study. 2024. 79(1): p. e41-e47.
12. Ou, X., et al., Radiomics based on 18F‐FDG PET/CT could differentiate breast carcinoma from
breast lymphoma using machine‐learning approach: A preliminary study. 2020. 9(2): p. 496-506.
13. Bicer, M.B.J.D., Radar-Based Microwave Breast Imaging Using Neurocomputational Models.
2023. 13(5): p. 930.
14. Sami, H., et al., Development of prototype microwave radar-based imaging system for breast
cancer detection. 2021. 39(2): p. 170-179.
15. Moloney, B.M., et al., Microwave Imaging in Breast Cancer - Results from the First-In-Human
Clinical Investigation of the Wavelia System. Acad Radiol, 2022. 29 Suppl 1: p. S211-s222.
16. O’Loughlin, D., et al., Microwave breast imaging: Clinical advances and remaining challenges.
2018. 65(11): p. 2580-2590.
17. Bolomey, J.-C.J.T.W.o.A.E.I.A.o.M.F.I., Crossed viewpoints on microwave-based imaging for
medical diagnosis: From genesis to earliest clinical outcomes. 2018: p. 369-414.
18. Janjic, A., et al., SAFE: A novel microwave imaging system design for breast cancer screening and
early detection—Clinical evaluation. 2021. 11(3): p. 533.
19. Canelo-Aybar, C., et al., Benefits and harms of breast cancer mammography screening for
women at average risk of breast cancer: a systematic review for the European Commission
Initiative on Breast Cancer. 2021. 28(4): p. 389-404.
20. Gardezi, S.J.S., et al., Breast cancer detection and diagnosis using mammographic data:
Systematic review. 2019. 21(7): p. e14464.
21. Khamparia, A., et al., Diagnosis of breast cancer based on modern mammography using hybrid
transfer learning. 2021. 32: p. 747-765.
22. Arevalo, J., et al., Representation learning for mammography mass lesion classification with
convolutional neural networks. 2016. 127: p. 248-257.
23. Duffy, S.W., et al., Beneficial effect of consecutive screening mammography examinations on
mortality from breast cancer: a prospective study. 2021. 299(3): p. 541-547.
24. Zeeshan, M., et al., Diagnostic accuracy of digital mammography in the detection of breast
cancer. 2018. 10(4).
25. Mori, M., et al., Diagnostic accuracy of contrast-enhanced spectral mammography in
comparison to conventional full-field digital mammography in a population of women with
dense breasts. 2017. 24: p. 104-110.
26. Zhu, X., et al., Diagnostic value of contrast-enhanced spectral mammography for screening
breast cancer: systematic review and meta-analysis. 2018. 18(5): p. e985-e995.
27. Sumkin, J.H., et al., Diagnostic performance of MRI, molecular breast imaging, and contrast-
enhanced mammography in women with newly diagnosed breast cancer. 2019. 293(3): p. 531-
540.
28. Lee-Felker, S.A., et al., Newly diagnosed breast cancer: comparison of contrast-enhanced
spectral mammography and breast MR imaging in the evaluation of extent of disease. 2017.
285(2): p. 389-400.
29. Xiang, W., H. Rao, and L.J.T.c. Zhou, A meta‐analysis of contrast‐enhanced spectral
mammography versus MRI in the diagnosis of breast cancer. 2020. 11(6): p. 1423-1432.
30. Kim, G., et al., Comparison of contrast-enhanced mammography with conventional digital
mammography in breast cancer screening: a pilot study. 2019. 16(10): p. 1456-1463.
31. Gelardi, F., et al., Contrast-enhanced mammography versus breast magnetic resonance imaging:
a systematic review and meta-analysis. 2022. 12(8): p. 1890.
32. Taylor, D.B., et al., Contrast-enhanced mammography (CEM) versus MRI for breast cancer
staging: detection of additional malignant lesions not seen on conventional imaging. 2023. 7(1):
p. 8.
33. Yang, L., et al., Performance of ultrasonography screening for breast cancer: a systematic review
and meta-analysis. 2020. 20: p. 1-15.
34. Dong, H., et al., Improved performance of adjunctive ultrasonography after mammography
screening for breast cancer among Chinese females. 2018. 18(3): p. e353-e361.
35. Rahaman, H.A. and S.K.J.M.R.C. Chinnikatti, Clinical and USG findings of patients presented with
breast pain. 2021. 8(3): p. 195-199.
36. Liang, X., et al., Application of contrast-enhanced ultrasound in the differential diagnosis of
different molecular subtypes of breast cancer. 2020. 42(6): p. 261-270.
37. Tian, Y., et al., The Size Differences of Breast Cancer and Benign Tumors Measured by Two‐
Dimensional Ultrasound and Contrast‐Enhanced Ultrasound. 2024.
38. Bourouis, S., et al., Meta-heuristic algorithm-tuned neural network for breast cancer diagnosis
using ultrasound images. 2022. 12: p. 834028.
39. Akhund, J., et al., ULTRASOUND VS. MAMMOGRAPHY IN EVALUATING SUSPICIOUS BREAST
LESIONS. 2023. 2023(1): p. 288-288.
40. Shi, X.-Q., et al., Accuracy of conventional ultrasound, contrast-enhanced ultrasound and
dynamic contrast-enhanced magnetic resonance imaging in assessing the size of breast cancer.
2022. 82(2): p. 157-168.
41. Du, Y.-R., et al., Application of contrast-enhanced ultrasound in the diagnosis of small breast
lesions. 2018. 70(3): p. 291-300.
42. Van Esser, S., et al., Accuracy of contrast-enhanced breast ultrasound for pre-operative tumor
size assessment in patients diagnosed with invasive ductal carcinoma of the breast. 2007. 7(1):
p. 63.
43. Desperito, E., et al., Chest CT for breast cancer diagnosis. 2022. 12(11): p. 1699.
44. Qu, Y.-H., et al., The correlation of 18 F-FDG PET/CT metabolic parameters, clinicopathological
factors, and prognosis in breast cancer. 2021. 23: p. 620-627.
45. Morawitz, J., et al., Comparison of nodal staging between CT, MRI, and [18 F]-FDG PET/MRI in
patients with newly diagnosed breast cancer. 2022: p. 1-10.
46. Neeter, L.M., et al., Comparing the diagnostic performance of contrast-enhanced
mammography and breast MRI: a systematic review and meta-analysis. 2023. 14(1): p. 174.
47. Jiang, Y., A.V. Edwards, and G.M.J.R. Newstead, Artificial intelligence applied to breast MRI for
improved diagnosis. 2021. 298(1): p. 38-46.
48. Witowski, J., et al., Improving breast cancer diagnostics with deep learning for MRI. 2022.
14(664): p. eabo4802.
49. Watt, G.P., et al., Association of breast cancer with MRI background parenchymal enhancement:
the IMAGINE case-control study. 2020. 22: p. 1-12.
50. Kim, S., et al., MRI radiomic features: association with disease-free survival in patients with
triple-negative breast cancer. 2020. 10(1): p. 3750.
51. Alabousi, M., et al., Performance of digital breast tomosynthesis, synthetic mammography, and
digital mammography in breast cancer screening: a systematic review and meta-analysis. 2021.
113(6): p. 680-690.
52. Zeng, B., et al., Breast cancer screening using synthesized two-dimensional mammography: a
systematic review and meta-analysis. 2021. 59: p. 270-278.
53. Huang, M.L., et al., Prospective comparison of synthesized mammography with DBT and full-field
digital mammography with DBT uncovers recall disagreements that may impact cancer
detection. 2022. 29(7): p. 1039-1045.
54. Alabousi, M., et al., Digital breast tomosynthesis for breast cancer detection: a diagnostic test
accuracy systematic review and meta-analysis. 2020. 30: p. 2058-2071.
55. You, C., et al., Comparison of the diagnostic performance of synthesized two-dimensional
mammography and full-field digital mammography alone or in combination with digital breast
tomosynthesis. 2020: p. 1-7.
56. Godinho, D.M., et al. Extracting dielectric properties for MRI-based phantoms for axillary
microwave imaging device. in 2020 14th European Conference on Antennas and Propagation
(EuCAP). 2020. IEEE.
57. Shao, W. and T.J.I.m.m. McCollough, Advances in microwave near-field imaging: Prototypes,
systems, and applications. 2020. 21(5): p. 94-119.
58. Moloney, B.M., et al., Microwave imaging in breast cancer–results from the first-in-human
clinical investigation of the wavelia system. 2022. 29: p. S211-S222.
59. Pelicano, A.C., et al., Repository of MRI-derived models of the breast with single and multiple
benign and malignant tumors for microwave imaging research. 2024. 19(5): p. e0302974.
60. Fernandez-Aranzamendi, E., et al. Breast tumor classification by age and size based on analysis
of dielectric properties performed on in vivo and ex vivo measurements. in 2022 International
Workshop on Antenna Technology (iWAT). 2022. IEEE.
61. Janjic, A., et al., Microwave breast lesion classification–results from clinical investigation of the
safe microwave breast cancer system. 2023. 30: p. S1-S8.
62. Zhang, H.J.I.a., Microwave imaging for breast cancer detection: The discrimination of breast
lesion morphology. 2020. 8: p. 107103-107111.
63. Godinho, D.M., et al., Development of MRI‐based axillary numerical models and estimation of
axillary lymph node dielectric properties for microwave imaging. 2021. 48(10): p. 5974-5990.
64. Yilmaz, T. and F.J.S. Ates Alkan, In vivo dielectric properties of healthy and benign rat mammary
tissues from 500 MHz to 18 GHz. 2020. 20(8): p. 2214.
65. Bhargava, D. and P.J.C.S.i.T.E. Rattanadecho, Microwave imaging of breast cancer: Simulation
analysis of SAR and temperature in tumors for different age and type. 2022. 31: p. 101843.
66. Adachi, M., et al., Feasibility of portable microwave imaging device for breast cancer detection.
2021. 12(1): p. 27.
67. Özmen, H., M.B.J.T.J.o.E.E. KURT, and C. Sciences, Radar-based microwave breast cancer
detection system with a high-performanceultrawide band antipodal Vivaldi antenna. 2021.
29(5): p. 2326-2345.
68. Saied, I., et al. Hardware accelerator for wearable and portable radar-based microwave breast
imaging systems. in 2021 IEEE International Symposium on Circuits and Systems (ISCAS). 2021.
IEEE.
69. Chowdhury, N.A., Microwave Imaging Technique for Breast Cancer Detection. 2024.
70. Yashaswini, P., U. Singh, and J. Mukherjee. Enhancing Early-Stage Breast Cancer Detection:
Antenna Design, Realistic Phantom Modeling, and Multi-Static Imaging with DAS Algorithm. in
2024 11th International Conference on Signal Processing and Integrated Networks (SPIN). 2024.
IEEE.
71. Franceschini, S., et al., A deep learning approach for diagnosis support in breast cancer
microwave tomography. 2023. 13(10): p. 1693.
72. Khoshdel, V., et al., Full 3D microwave breast imaging using a deep-learning technique. 2020.
6(8): p. 80.