Deep Learning for Segmentation and Classification in Mammograms for Breast Cancer Detection: A Systematic Literature Review

doi:10.37015/AUDT.2024.230051

Abstract

Abstract:

Integrating machine learning into medical diagnostics has revolutionized the field, particularly enhancing Computer-aided Diagnosis (CAD) systems. These systems assist healthcare professionals by leveraging medical data and machine learning algorithms for more accurate diagnosis and treatment plans. Mammography, an X-ray-based imaging technique, is pivotal in early breast cancer detection, enabling the differentiation between benign and malignant lesions. Recent studies have focused on developing deep learning-enabled mammography CAD systems, which have shown promising results in detecting, segmenting, and classifying anomalies in mammogram images. This comprehensive review presents an innovative system architecture for breast cancer detection, segmentation, and classification using deep learning within mammography CAD systems. It also explores publicly available mammogram datasets and the critical parameters for assessing deep learning system performance. The literature review is meticulously conducted using the PRISMA methodology to evaluate and synthesise novel research findings in this domain. This survey highlights the technological advancements and underlines the potential of deep learning in transforming mammographic analysis for breast cancer detection.

Key words: Computer-aided diagnosis; Deep learning; Mammography; Breast cancer; Detection; Classification

Raymond Sutjiadi, MS, Siti Sendari, PhD, Heru Wahyu Herwanto, PhD, Yosi Kristian, PhD. Deep Learning for Segmentation and Classification in Mammograms for Breast Cancer Detection: A Systematic Literature Review. Advanced Ultrasound in Diagnosis and Therapy, 2024, 8(3): 94-105.

Figures/Tables 13

Figure 1

Figure 2

Figure 3

Table 1

Figure 4

Figure 5

Table 2

Figure 6

Table 3

Figure 7

Table 4

Figure 8

Table 5

Comparison of deep learning architectures for breast cancer detection and performance metrics"

No.	Paper	Deep learning architecture	Performance metric
1	Pengcheng, X. et al., 2018 [48]	19-layer VGGNet CNN + 152-layer ResNet (class activation mapping for localising abnormalities)	92.53% Accuracy (CBIS-DDSM Dataset)
2	Jiao, Z. et al., 2018 [49]	10-layer CNN + Metric Learning Layers	97.4% Accuracy (DDSM Dataset) 96.7% Accuracy (MIAS Dataset)
3	Al-masni, M.A. et al., 2018 [57]	26-layer YOLO	99.7% Accuracy of Location Detection and 97% Accuracy of Mass Classification (DDSM Dataset)
4	Al-antari, M.A. et al., 2018 [32]	26-layer YOLO (mass detection) + 16-layer FrCN (mass segmentation) + 7-layer CNN (mass recognition and classification)	Mass Detection 98.96% Accuracy, 99.24% F1-score, 97.62% MCC Mass Segmentation 92.97% Accuracy, 92.69%F1-score, 85.93% MCC Mass Classification 95.64% Accuracy, 96.84% F1-score, 94.78% AUC, 89.91%(INbreast Dataset)
5	Ribli, D. et al., 2018 [53]	Faster R-CNN using 16-layer VGG16 model	95% AUC (INbreast Dataset)
6	Ragab, D.A. et al., 2019 [50]	DCNN using 12-layer AlexNet model	87.2% Accuracy, 94% AUC (CBIS-DDSM Dataset)
7	Shen, L. et al., 2019 [34]	CNN with end-to-end training using combination of 16-layer VGG-16 and 50-layer ResNet50	CBIS-DDSM Dataset 88% AUC (best single model), 91% AUC (four-model averaging) INbreast Dataset 95% AUC (best single model), 98% AUC (four-model averaging)
8	Li, H. et al., 2019 [35]	DenseNet-II Neural Network Model	94.55% Average Accuracy (Private Dataset)
9	Khan, H.N. et al., 2019 [36]	Multi-View Feature Fusion (MVFF): 4-layer Small VGGNet-like using Multi-View ROI as the input	93.2% AUC for mass and calcification classification 84% AUC for malignant and benign classification 93% AUC for normal and abnormal classification (CBIS-DDSM dan Mini-MIAS)
10	Yala, A. et al., 2019 [37]	18-layer ResNet18 and Risk Factor Logistic Regression (RF-LR) model	79% AUC for premenopausal patients & 70% AUC for postmenopausal patients (Private Dataset)
11	Xu, C. et al., 2021 [38]	Multi-Scale Attention Module (MSAM): constructed by stacking multiple MSA bottlenecks.	94.2% AUC (DDSM Dataset), 92.85% AUC (DDSM+INBreast Dataset)
12	Oyelade, O.N. et al., 2021 [39]	12-layer CNN with data augmentation	93.75% Accuracy (DDSM + CBIS, INbreast, and MIAS Dataset), 87.29 % Accuracy (CBIS-DDSM Dataset)
13	AlGhamdi, M. et al. [51]	Dual View-DCNN (DV-DCNN): a 4-layer dense block + neighbourhood patch matching layers with dual view image input	97.5% Accuracy, 95% Sensitivity, 96% Specificity, 98% AUC (CBIS-DDSM) 96% Accuracy, 94% Sensitivity, 95% Specificity, 97% AUC (INbreast)
14	Chouhan, N. et al. [52]	Diverse Features based Breast Cancer Detection (DFeBCD): DCNN (6 highway blocks + 3 fully connected layers) + Support Vector Machine (SVM) / Emotional Learning inspired Ensemble Classifier (ELiEC)	86.1% ROC-AUC (SVM) & 86.5% ROC-AUC (ELiEC) 93.2% PR-AUC (SVM) & 93.4% PR-AUC (ELiEC) 80.5% Accuracy (SVM) & 80.3% Accuracy (ELiEC) (IRMA - DDSM Dataset)
15	El Houby, E.M.F. et. al., 2021 [40]	10-layer CNN with image pre-processing	96.55% Sensitivity, 96.49% Specificity, 96.52% Accuracy, 98% AUC (INbreast Dataset) 98% Sensitivity, 92.6% Specificity, 95.3% Accuracy, 97.4% AUC (MIAS Dataset)
16	Salama, W.M. et al., 2021 [41]	Pre-trained modified U-Net model for segmentation + different deep learning models (InceptionV3, DenseNet121, ResNet50, VGG16, Mobile-NetV2)	98.87% Accuracy, 98.88% AUC, 98.98% Sensitivity, 98.79% Precision, 97.99% F1-Score (MLO DDSM datasets) 99.43% Accuracy, 99.22% AUC, 99.12% Sensitivity, 98.99% Precision, 98.98% F1-Score (MLO and CC DDSM dataset)
17	Oyelade, O.N. et al., 2022 [42]	Wavelet-CNN-Wavelet with augmented dataset using Generative Adversarial Network (GAN)	99% Accuracy, 99% Recall, 99% Precision, 100% Specificity, 99% F1-Score (MIAS Dataset)
18	Escorcia-Gutierrez, J. et al., 2022 [43]	Automated Deep Learning Based Breast Cancer Diagnosis (ADL-BCD): 34-layer ResNet34	96.07% Accuracy, 95.90% Specificity, 92.15% Recall, 93.54% Precision (MIAS Dataset)
19	Maqsood, S. et al., 2022 [54]	Transferable Texture Convolutional Neural Network (TTCNN) based on deep features of convolutional neural network models (InceptionResNet-V2, Inception-V3, VGG-16, VGG-19, GoogLeNet, ResNet-18, ResNet-50, and ResNet-101)	99.08% Accuracy, 98.96% Specificity, 99.19% Sensitivity (DDSM Dataset) 96.82% Accuracy, 97.68% Specificity, 95.99% Sensitivity (INbreast Dataset) 96.57% Accuracy, 97.03% Specificity, 96.11% Sensitivity (MIAS Dataset)
20	Adedigba, A.P. et al., 2022 [45]	Discriminative Fine-tuning Method using DenseNet & AlexNet	99.8% Accuracy (DenseNet) & 98.8% Accuracy (AlexNet) (INbreast Dataset)
21	Chakravarthy S.R., S. et al., 2022 [44]	18-layer ResNet-18 + Improved Crow-Search Optimized Extreme Learning Machine (ICS-ELM)	97.193% Accuracy (DDSM Dataset), 98.137% Accuracy (MIAS Dataset), 98.266% Accuracy (INbreast Dataset)
22	Rehman, K. et al., 2022 [55]	Depth-wise 2D V-net 64 Convolutional Neural Network	95% Accuracy (PINUM Private Dataset), 97% Accuracy (CBIS-DDSM Dataset), 98% Accuracy (DDSM Dataset)
23	Kavitha, T. et al., 2022 [56]	Optimal Multi-Level Thresholding-based Segmentation with DL-enabled Capsule Network (OMLTS-DLCN): OKMT-SGO (for segmentation) + CapsNet (feature extraction) + BPNN (classification)	98.50% Accuracy (Mini-MIAS Dataset) and 97.55% Accuracy (DDSM Dataset)
24	Elkorany, A.S. et al., 2023 [46]	CNNs (Inception-V3, ResNet50, and AlexNet) + Term Variance (feature selection) + Multiclass SVM (classifier)	97.81% Accuracy (70% training), 98% Accuracy (80% training), 100% Accuracy (90% training) (MIAS Dataset)
25	Bouzar-Benlabiod, L. et al., 2023 [47]	SE-ResNet101 (RoI extraction) + Case-Based Reasoning System/CBR (classification)	86.71% Accuracy, 91.34% Recall (CBIS-DDSM Dataset)

Table 5

References 57

[1]	Sung H, Ferlay J, Siegel RL, Laversanne M, Soerjomataram I, Jemal A, et al. Global Cancer Statistics 2020: GLOBOCAN Estimates of Incidence and Mortality Worldwide for 36 Cancers in 185 Countries. CA Cancer J Clin 2021;71:209-249.
[2]	Giaquinto AN, Sung H, Miller KD, Kramer JL, Newman LA, Minihan A, et al. Breast Cancer Statistics, 2022. CA Cancer J Clin 2022;72:524-541.
[3]	Guo F, Kuo YF, Shih YCT, Giordano SH, Berenson AB. Trends in breast cancer mortality by stage at diagnosis among young women in the United States. Cancer 2018;124:3500-3509.
[4]	Jeun JH, Lee JH, Cho E, Kim SJ, Park EH, Do Byun K. Invasive breast cancer presenting as a mass replaced by calcification on mammography: A report of two cases. Journal of the Korean Society of Radiology 2019;80:591-597.
[5]	Ekpo EU, Alakhras M, Brennan P. Errors in Mammography Cannot be Solved Through Technology Alone. Asian Pac J Cancer Prev 2018;19:291-301.
[6]	D. Anyfantis, A. Koutras, G. Apostolopoulos, I. Christoyianni. Breast density transformations using cycleGANs for revealing undetected findings in mammograms. Signals 2023;4:421-438.
[7]	Zhang J, Wu J, Zhou XS, Shi F, Shen D. Recent advancements in artificial intelligence for breast cancer: Image augmentation, segmentation, diagnosis, and prognosis approaches. Semin Cancer Biol 2023;96:11-25.
[8]	Chan HP, Hadjiiski LM, Samala RK. Computer-aided diagnosis in the era of deep learning. Med Phys 2020;47:e218-e227.
[9]	Huang Q, Zhang F, Li X. Machine Learning in ultrasound computer-aided diagnostic systems: A survey. Biomed Res In 2018;2018:5137904.
[10]	Visuña L, Yang D, Garcia-Blas J, Carretero J. Computer-aided diagnostic for classifying chest X-ray images using deep ensemble learning. BMC Med Imaging 2022;22:178.
[11]	Wahab Sait AR, Dutta AK. Developing a deep-learning-based coronary artery disease detection technique using computer tomography images. Diagnostics (Basel) 2023;13:1312.
[12]	He M, Cao Y, Chi C, Yang X, Ramin R, Wang S, et al. Research progress on deep learning in magnetic resonance imaging-based diagnosis and treatment of prostate cancer: a review on the current status and perspectives. Front Oncol 2023;13:1189370.
[13]	Kadhim YA, Khan MU, Mishra A. Deep learning-based computer-aided diagnosis (CAD): applications for medical image datasets. Sensors (Basel) 2022;22:8999.
[14]	Page MJ, McKenzie JE, Bossuyt PM, Boutron I, Hoffmann TC, Mulrow CD, et al. The PRISMA 2020 statement: an updated guideline for reporting systematic reviews. BMJ 2021;372:n71.
[15]	Pranolo A, Mao Y, Wibawa AP, Utama ABP, Dwiyanto FA. Optimized three deep learning models based-PSO hyperparameters for Beijing PM2.5 prediction. Knowledge Engineering and Data Science 2022.
[16]	Li H, Shen HW. Local latent representation based on geometric convolution for particle data feature exploration. IEEE Trans Vis Comput Graph 2023;29:3354-3367.
[17]	Taye MM. Understanding of machine learning with deep learning: architectures, workflow, applications and future directions. Computers 2023;12.
[18]	Sweeney RI, Lewis SJ, Hogg P, McEntee MF. A review of mammographic positioning image quality criteria for the craniocaudal projection. Br J Radiol 2018;91:20170611.
[19]	Mohamed AA, Luo Y, Peng H, Jankowitz RC, Wu S. Understanding clinical mammographic breast density assessment: a deep learning perspective. J Digit Imaging 2018;31:387-392.
[20]	Moreira IC, Amaral I, Domingues I, Cardoso A, Cardoso MJ, Cardoso JS. INbreast: toward a full-field digital mammographic database. Acad Radiol 2012;19:236-248.
[21]	Suckling J. Mammographic Image Analysis Society (MIAS) database v1.21. [Dataset]. Apollo-University of Cambridge Repository 2015.
[22]	Heath M, Bowyer K, Kopans D, Moore R, Kegelmeyer P. The digital database for screening mammography. 5th International Workshop on Digital Mammography Toronto 2001:212-218.
[23]	Lee RS, Gimenez F, Hoogi A, Miyake KK, Gorovoy M, Rubin DL. A curated mammography data set for use in computer-aided detection and diagnosis research. Sci Data 2017;4:170177.
[24]	Cui R, Wang L, Lin L, Li J, Lu R, Liu S, et al. Deep learning in barrett's esophagus diagnosis: current status and future directions. Bioengineering (Basel) 2023;10:1239.
[25]	Dewandra ARF, Wibawa AP, Pujianto U, Utama ABP, Nafalski A. Journal unique visitors forecasting based on multivariate attributes using CNN. International Journal of Artificial Intelligence Research 2022;6.
[26]	Orozco-Arias S, Piña JS, Tabares-Soto R, Castillo-Ossa LF, Guyot R, Isaza G. Measuring performance metrics of machine learning algorithms for detecting and classifying transposable elements. Processes 2020;8.
[27]	Duntsch I, Gediga G. Confusion matrices and rough set data analysis. J Phys Conf Ser 2019;1229.
[28]	Alzubaidi L, Zhang J, Humaidi AJ, Al-Dujaili A, Duan Y, Al-Shamma O, et al. Review of deep learning: concepts, CNN architectures, challenges, applications, future directions. J Big Data 2021;8:53.
[29]	Fränti P, Mariescu-Istodor R. Soft precision and recall. Pattern Recognit Lett 2023;167:115-121.
[30]	Steiner JM, Morse C, Lee RY, Curtis JR, Engelberg RA. Sensitivity and specificity of a machine learning algorithm to identify goals-of-care documentation for adults with congenital heart disease at the end of life. J Pain Symptom Manage 2020;60:e33-e36.
[31]	Hicks SA, Strümke I, Thambawita V, Hammou M, Riegler MA, Halvorsen P, et al. On evaluation metrics for medical applications of artificial intelligence. Sci Rep 2022;12:5979.
[32]	Al-Antari MA, Al-Masni MA, Choi MT, Han SM, Kim TS. A fully integrated computer-aided diagnosis system for digital X-ray mammograms via deep learning detection, segmentation, and classification. Int J Med Inform 2018;117:44-54.
[33]	Mrudula Devi K, Venkata Ramakrishna S, Rama Koteswara Rao G, Prasad C. Gradient-based optimization of the area under the minimum of false positive and false negative functions. 2nd International Conference on Smart Electronics and Communication (ICOSEC) 2021;779-785.
[34]	Shen L, Margolies LR, Rothstein JH, Fluder E, McBride R, Sieh W. Deep learning to improve breast cancer detection on screening mammography. Sci Rep 2019;9:12495.
[35]	Li H, Zhuang SS, Li DA, Zhao JM, Ma YY. Benign and malignant classification of mammogram images based on deep learning. Biomed Signal Process Control 2019;51:347-354.
[36]	Khan HN, Shahid AR, Raza B, Dar AH, Alquhayz H. Multi-view feature fusion based four views model for mammogram classification using convolutional neural network. IEEE Access 2019;7:165724-165733.
[37]	Yala A, Lehman C, Schuster T, Portnoi T, Barzilay R. A deep learning mammography-based model for improved breast cancer risk prediction. Radiology 2019;292:60-66.
[38]	Xu CB, Lou M, Qi YL, Wang YM, Pi JD, Ma YD. Multi-Scale Attention-Guided Network for mammograms classification. Biomed Signal Process Control 2021;68.
[39]	Oyelade ON, Ezugwu AE. A deep learning model using data augmentation for detection of architectural distortion in whole and patches of images. Biomed Signal Process Control 2021;65.
[40]	El Houby EMF, Yassin NIR. Malignant and nonmalignant classification of breast lesions in mammograms using convolutional neural networks. Biomed Signal Process Control 2021;70.
[41]	Salama WM, Aly MH. Deep learning in mammography images segmentation and classification: Automated CNN approach. Alexandria Engineering Journal 2021;60:4701-4709.
[42]	Oyelade ON, Ezugwu AE. A novel wavelet decomposition and transformation convolutional neural network with data augmentation for breast cancer detection using digital mammogram. Sci Rep 2022;12:5913.
[43]	Escorcia-Gutierrez J, Mansour RF, Beleño K, Jiménez-Cabas J, Pérez M, Madera N, et al. Automated deep learning empowered breast cancer diagnosis using biomedical mammogram images. Computers, Materials & Continua 2022;71:4221-4235.
[44]	Chakravarthy SRS, Rajaguru H. Automatic detection and classification of mammograms using improved extreme learning machine with deep learning. IRBM 2022;43:49-61.
[45]	Adedigba AP, Adeshina SA, Aibinu AM. Performance evaluation of deep learning models on mammogram classification using small dataset. Bioengineering 2022;9:161.
[46]	Elkorany AS, Elsharkawy ZF. Efficient breast cancer mammograms diagnosis using three deep neural networks and term variance. Sci Rep 2023;13:2663.
[47]	Bouzar-Benlabiod L, Harrar K, Yamoun L, Khodja MY, Akhloufi MA. A novel breast cancer detection architecture based on a CNN-CBR system for mammogram classification. Comput Biol Med 2023;163:107133.
[48]	Xi PC, Shu C, Goubran R. Abnormality Detection in Mammography using Deep Convolutional Neural Networks. 2018 IEEE International Symposium on Medical Measurements and Applications (MeMeA) 2018;1-6.
[49]	Jiao ZC, Gao XB, Wang Y, Li J. A parasitic metric learning net for breast mass classification based on mammography. Pattern Recognit 2018;75:292-301.
[50]	Ragab DA, Sharkas M, Marshall S, Ren J. Breast cancer detection using deep convolutional neural networks and support vector machines. PeerJ 2019;7:e6201.
[51]	AlGhamdi M, Abdel-Mottaleb M. DV-DCNN: Dual-view deep convolutional neural network for matching detected masses in mammograms. Comput Methods Programs Biomed 2021;207:106152.
[52]	Chouhan N, Khan A, Shah JZ, Hussnain M, Khan MW. Deep convolutional neural network and emotional learning based breast cancer detection using digital mammography. Comput Biol Med 2021;132:104318.
[53]	Ribli D, Horváth A, Unger Z, Pollner P, Csabai I. Detecting and classifying lesions in mammograms with Deep Learning. Sci Rep 2018;8:4165.
[54]	Maqsood S, Damasevicius R, Maskeliunas R. TTCNN: A breast cancer detection and classification towards computer-aided diagnosis using digital mammography in early stages. Applied Sciences 2022;12:3273.
[55]	Rehman KU, Li J, Pei Y, Yasin A, Ali S, Saeed Y. Architectural distortion-based digital mammograms classification using depth wise convolutional neural network. Biology (Basel) 2021;11:15.
[56]	Kavitha T, Mathai PP, Karthikeyan C, Ashok M, Kohar R, Avanija J, et al. Deep learning based capsule neural network model for breast cancer diagnosis using mammogram images. Interdiscip Sci 2022;14:113-129.
[57]	Al-Masni MA, Al-Antari MA, Park JM, Gi G, Kim TY, Rivera P, et al. Simultaneous detection and classification of breast masses in digital mammograms via a deep learning YOLO-based CAD system. Comput Methods Programs Biomed 2018;157:85-94.

Item	MIAS	DDSM	CBIS-DDSM	INbreast
Number of images	322	10,480	10,239	410
Resolution	50 Micron	42-50 Micron	42-50 Micron	70 Micron
Image format	PGM	LJPEG	DICOM	DICOM
Mammograms projection	MLO	MLO and CC	MLO and CC	MLO and CC
Overlay	Yes	Yes	Yes	Yes
Total size	3.16 GB	230.9 GB	163.6 GB	2 GB

No.	Method/Algorithm	Paper
1	Convolutional Neural Network (CNN)	[32], [34], [35], [36], [37], [38], [39], [40], [41], [42], [43], [44], [45], [46], [47]
2	Deep Convolutional Neural Network (DCNN)	[48], [49], [50], [51], [52]
3	Full-resolution Convolutional Networks (FrCN)	[32]
4	Faster R-CNN	[53]
5	Transferable Texture Convolutional Neural Network (TTCNN)	[54]
6	Depth-wise Convolutional Neural Network	[55]
7	Optimal Multi-Level Thresholding-based Segmentation with DL Enabled Capsule Network (OMLTS-DLCN)	[56]
8	You Only Look Once (YOLO)	[32], [57]