Next Article in Journal
Update on Corneal Confocal Microscopy Imaging
Next Article in Special Issue
Mathematical Assessment of Machine Learning Models Used for Brain Tumor Diagnosis
Previous Article in Journal
The Association between Diabetes Mellitus, High Monocyte/Lymphocyte Ratio, and Survival in Endometrial Cancer: A Retrospective Cohort Study
Font Type:
Arial Georgia Verdana
Font Size:
Aa Aa Aa
Line Spacing:
Column Width:

The Systematic Review of Artificial Intelligence Applications in Breast Cancer Diagnosis

Dilber Uzun Ozsahin
Declan Ikechukwu Emegano
Berna Uzun
2,3,4 and
Ilker Ozsahin
Medical Diagnostic Imaging Department, College of Health Sciences, University of Sharjah, Sharjah 27272, United Arab Emirates
Operational Research Center in Healthcare, Near East University, TRNC Mersin 10, 99138 Nicosia, Turkey
Department of Statistics, Carlos III University of Madrid, 28903 Madrid, Spain
Department of Mathematics, Near East University, TRNC Mersin 10, 99138 Nicosia, Turkey
Brain Health Imaging Institute, Department of Radiology, Weill Cornell Medicine, New York, NY 10065, USA
Author to whom correspondence should be addressed.
Diagnostics 2023, 13(1), 45;
Submission received: 7 December 2022 / Revised: 16 December 2022 / Accepted: 17 December 2022 / Published: 23 December 2022


Several studies have demonstrated the value of artificial intelligence (AI) applications in breast cancer diagnosis. The systematic review of AI applications in breast cancer diagnosis includes several studies that compare breast cancer diagnosis and AI. However, they lack systematization, and each study appears to be conducted uniquely. The purpose and contributions of this study are to offer elaborative knowledge on the applications of AI in the diagnosis of breast cancer through citation analysis in order to categorize the main area of specialization that attracts the attention of the academic community, as well as thematic issue analysis to identify the species being researched in each category. In this study, a total number of 17,900 studies addressing breast cancer and AI published between 2012 and 2022 were obtained from these databases: IEEE, Embase: Excerpta Medica Database Guide-Ovid, PubMed, Springer, Web of Science, and Google Scholar. We applied inclusion and exclusion criteria to the search; 36 studies were identified. The vast majority of AI applications used classification models for the prediction of breast cancer. Howbeit, accuracy (99%) has the highest number of performance metrics, followed by specificity (98%) and area under the curve (0.95). Additionally, the Convolutional Neural Network (CNN) was the best model of choice in several studies. This study shows that the quantity and caliber of studies that use AI applications in breast cancer diagnosis will continue to rise annually. As a result, AI-based applications are viewed as a supplement to doctors’ clinical reasoning, with the ultimate goal of providing quality healthcare that is both affordable and accessible to everyone worldwide.

1. Introduction

Breast cancer (BC) is one of the greatest threats to women in the 21st century. It has rendered many women mentally unstable [1] and many lives have been lost. Although early treatment reduces the mortality rate of this malignancy, a late diagnosis is potentially fatal. Breast cancer’s pathogenic effects include age [2], reproductive effects such as breastfeeding [3], testosterone levels and menopause [4], familial history, genetic disorders [5], and other environmental effects [6]. Anatomically, the breast is made up of healthy blood vessels, connective tissues, milk duct lobules, and lymph nodes. However, breast cancer occurs when abnormal cells (tumors) grow in the connective tissues, milk ducts, lymph nodes, and lobules of the breast [7]. Breast cancer can be benign or cancerous. Cancerous can be classified by invasive carcinoma and non-invasive carcinoma [8]. Invasive carcinoma is highly pathological with metastatic [9] adverse conditions while non-invasive breast cancer does not cause proliferation to the neighboring organs [10]. Meanwhile, according to a 2020 statistical evaluation [11], approximately 2.3 million women were diagnosed with breast cancer, resulting in 685,000 deaths, globally [12]. In the United States of America, approximately 13% of women will experience breast cancer in their lifetime. Annually, 287,850 invasive cases and 51,400 non-invasive cancers were recorded [11]. The incidence has increased tremendously, however, the mortality rate has recently been reduced [13]. Concurrently, the incorporation of artificial intelligence (AI) into the diagnosis of breast cancer is an emerging procedure that results in a better diagnosis of the disease. AI’s success rate has been attributed to its ability to reproduce, using high-resolution images from variable tissue specimens. AI is the application of computational models to simulate human intellectual abilities as well as resolving severe healthcare concerns, such as complicated biological anomalies such as cancers [14]. Numerous genomic and epigenomics variants contribute to cancer’s complexity and diversity; as a result, the use of AI algorithms to diagnose gene mutations or these abnormalities in protein complexes at a preliminary phase has enormous promise. This study’s objective is to investigate the recent developments in diagnostic applications of AI in breast cancer diagnosis by assiduously surveying the current literary studies according to preferred reporting items for systematic review and meta-analysis of diagnostic test accuracy (PRISMA-DTA).

AI in Breast Cancer

At present, AI has been introduced into practically every industry to maximize production, efficiency, and accuracy [15,16]. Advances in computation, data, and algorithm performance have made AI more powerful, user-friendly, and objectively directed than previously. It is also utilized for intrusion detection systems [16], pictorial synthesis [17], optical character recognition (OCR) [18], facial expression identification, etc. In healthcare, AI is applied in different domains, such as patient monitoring [19], drug dispensing, and hospital management [19,20]. AI has undeniably had a greater impact on complex image analysis, as well as providing data for quantitative assessment through automation and removing the radiation risk associated with breast radiological examination [21]. To mimic the rational decisions of humans with operational excellence, AI offers superintelligence; using AI techniques can be beneficial for the instant incorporation of feature learning, the ability to process and handle complex and multiple dimensional data, as well as its availability of diagnostic data from a variety of clinical experiments. Therefore, medical consultants, academicians, and oncologists have recognized the potential of developing and employing AI in many parts of the diagnosis of breast cancer disorders. This hope has been fueled by recent advancements in AI [21]. AI research often uses enhanced models with higher cancer incidence than screening programs. As a result, algorithms for imaging assessment and the diagnosis of breast cancer, in each category of AI algorithms, have demonstrated substantial improvements. Meanwhile, deep learning (DL) algorithms have also been discovered to be substantially more encouraging than conventional machine learning (ML) algorithms [22,23]. They have also demonstrated that they are strong candidates for pre-existing chronic imaging studies, and particularly for breast cancer image processing research [22,23]. Alternatively, Buckner et al. summarized ML, computer-assisted detection, and computer-assisted diagnosis as now being possible. It reduces the number of false positives (FP) that are produced by the outputs from the Chan-Vese segmentation technique after it has been initialized using the marker controller watershed algorithm (MCWS). The computer-assisted diagnosis technique makes use of blended learning, which consists of four Support Vector Machine (SVM)-based classifiers. Each of these base classifiers uses the features extracted from a certain tissue constituent [24]. There are possibilities, on the other hand, that mammography misses many malignancies because of several factors, such as breast density, tumor size, or subtle indications of cancer that are invisible in cases of metastasis. AI is accurate in detection using digital mammography. In addition, the sensitivity of AI, when evaluated in terms of its ability to pick up on the various breast cancer morphological characteristics, such as, asymmetry, and distortion, proves highly effective [25].
The AI algorithm is used in medical imaging. Although picture archiving and communication systems (PACSs) are now in use, and constitute a tremendous supply of information, the practical use of AI in medical imaging has been hampered by the absence of large public databases. Despite this, a significant number of software programs that are extremely helpful for diagnoses in general, and particularly for the identification of breast lesions, have been established. The AI detection of lesions is automatically applied in a variety of imaging techniques, and it is now the most prevalent AI application. It involves locating the regions of the image that have high and different lesions, based on the training of the models. Recent research has shown that convolutional neural networks (CNNs) are capable of matching the detection capability of an experienced radiologist [26]. CNN categorization has the benefit of removing any variabilities. The delineation of the boundaries of the lesions is a vital reason for the application CNN in lesions [27]. CNN, when simulated to complex tasks, performs with higher accuracy and speed than humans [28]. As a result, U-nets, which are utilized for the segmentation of the images, is a typical example of the kind of network that is employed for this objective; they can differentiate between tissues, such as glandular and adipose, in digital mammography after the volume of the lesion is calculated [29]. On the other hand, the diagnostic criteria for classifications in radiography and histopathology are widespread; nonetheless, the manual identification, classification, and grading employed by radiologists are time-consuming and more prone to inter and intra-observer variances. However, improved clinical outcomes rely solely on the early identification of breast cancer [30]. According to the National Breast Cancer Foundation, this early diagnosis involves breast lumps, mass, density, etc., being detected at an early stage, when it is still in the localized stage. If early detection is achieved the overall survival rate after five years is 99%. Therefore, performing monthly breast self-examinations, in addition to performing periodic clinical breast examinations such as mammography, is an important part of early detection [31] and it increases a patient’s chance of surviving the disease [32]. The study by Jan Witowski, Laura Heacock, et al. [33] created a DL methodology that increases the specificity of the dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI) of breast tissue. DCE-MRI is a technology that is occasionally utilized for women who are at a higher risk of developing breast cancer. These scientists verified the DL techniques on different cohorts, which demonstrated that this strategy has the potential to reduce the number of needless surgeries by minimizing the percentage of false positives [33]. The identification of breast cancer at an earlier stage is essential for improving treatment outcomes. On the other hand, AI is a tool that improves breast cancer screening. The application of screening strategies using AI on a population has proved effective in lowering the mortality rate caused by breast cancer. The use of AI technology in screening cancer achieves 89% specificity and 76% sensitivity, both of which are substantially higher compared with the figures for standard computer-aided diagnosis (CAD) systems, which are approximately 50% [34]. This was achieved using digital mammography, which serves as the screening method for the women invited to take part in the screening programs. The use of digital mammography allows for the ability to construct CAD systems [35] that have the potential to lessen the workload of radiologists. Studies have indicated that AI-driven technologies have superior diagnostic accuracy than conventional techniques, and this trend toward the widespread implementation of AI-based systems will likely continue [35]. When it comes to helping radiologists in the interpretation of digital mammograms in the Breast Cancer Early Detection Program (BCEDP), recent advancements in AI have opened up possibilities that extend farther than what is provided by standard CAD systems. The goals of AI systems are similar to CAD systems: enhancing the diagnosis of malignant tumors, minimizing the effects of interval malignancies, and, simultaneously, minimizing the amount of reading required. It is possible that, in the long run, new AI systems and methods will even increase the expense ratio of BCEDPs [36].
In another literary study, the distinction between benign and malignant tumors was made through the use of ultrasonography, which is a form of digital imaging. The identification of breast cancer by breast ultrasonography has been suggested to benefit from the application of several different AI techniques. At present, great classification performance in biomedical images can be achieved with the application of various learning approaches, particularly DL. The image classification model offers an accuracy of 97.18% and could classify breast cancer as normal, benign, or malignant [37]. In the same manner, traditionally, ultrasound has been the diagnostic method of choice for determining whether breast tumors are benign or cancerous. As it detects occult breast malignancies, it has become an emerging procedure in this modern time. In contrast to other techniques, such as mammography, DBT, and MRI, ultrasound has some advantages, for example, it is non-ionizing, affordable and has the capacity to provide detailed insights and surveillance [38]. According to the National Breast Cancer Foundation’s 2020 report, AI has been used successfully in the diagnosis of more than 276,000 breast cancer cases. In addition to this figure, 48,000 cases were diagnosed using the non-deep learning methodology, particularly in the diagnosis of aggressive cancerous types. Gene testing and histopathological imaging are two methods that can be used to identify breast cancer. As genetic analysis is not cost-effective, medical laboratory histological imaging is most frequently employed for breast cancer screening and diagnosis. The combination of action DL and ML enables the comprehensive analysis of the diagnosis and treatment of breast cancer using genetic sequencing or histopathological imaging. Breast cancer image analysis using AI can detect breast lumps (masses), mass segmentation, breast density, and the risk of breast cancer. In the majority of patients, breast lumps are the most frequent occurrence of breast cancer [39]; consequently, their detection is an essential step utilized in CAD [40]
The studies by Farahnaz Sadoughi et al., in which breast cancer was diagnosed using image acquisition, was the first step to be taken in the image process, then the image is processed and segmented. Different AI techniques, such as SVM, KNN, genetic and Naive Bayes, as well as DL [41,42,43], were employed for its categorization. According to the findings, SVM had the highest accuracy percentage across the board for all of the different image analysis tasks. In order to enhance the diagnostic efficacy, Shahid Munir Shah, et al. utilized a variety of imaging techniques to streamline the process of breast cancer identification. These imaging techniques include mammograms, ultrasound (US), magnetic resonance imaging (MRI), histopathological pictures, or any hybrid model of the aforementioned imaging techniques. The development of powerful AI algorithms, such as DL, and the accessibility of large datasets are two factors that have contributed to the recent uptick in scientific research [44]. In addition, Dileep G. [45] and S.M. Shah [46] used several imaging modalities in the diagnosis of breast cancer, such as mammography, X-ray, thermography, magnetic resonance imaging (MRI), Positron computed tomography (PET), computed tomography (CT), ultra-high-density ultrasound and histological examinations. The results from using these imaging modalities showed that most women have malignancy, whose etiologic factors are linked to heredity, lifestyle, and environmental. AI is employed in breast cancer diagnosis because it yields faster, more accurate diagnoses.
Sha et al. [46] summarized breast cancer in terms of etiology, diagnosis, and treatment, as well as preventive measures. These are called prediction classifiers and models that categorize a woman’s chances of acquiring breast cancer, as well as direct screening guidelines supporting the existence of established and quantified hormonal, ecological, behavioral, or familial risk variables. This familial factor contributes to the development of breast cancer in women. However, image modalities quantification such as magnetic resonance imaging (MRI) is a proven effective diagnostic model, particularly in the diagnosis of this malignancy. Mammography is the primary diagnostic choice [47] that many health personnel utilizes as it creates a significant level of specialization for instance molecular and genomic profiling that are essential in the management of breast cancer [46]. Furthermore, additional studies have shown the relationship between the image analysis of breast cancer and its application; the diagnosis of BC, segmentation techniques, camera calibration, and data processing are all parts of this process. Studies have also shown the successes and application span of supervised and unsupervised learning such as DL, CNN, and other related approaches in breast cancer evaluation. The combination of unsupervised and transfer learning (TL) in BC diagnosis is an emerging technique in AI. TL has the potential to ease, albeit only slightly, the problem of insufficient annotations of data. Utilizing a CNN that has already been trained on the approaches an organization takes throughout the procedure of TL is more effective than carrying out further supervisory training [48]. It is necessary to utilize numerous modes or types of images in the registration and fusion when performing diagnostic and therapeutic analysis of breast cancer. This helps clinicians to gain more knowledge, which helps them in diagnostic accuracy [49]. The registration of images in an AI breast cancer diagnosis locates reference points in a plurality of images; this is accomplished by performing rotation (spatial) on the images to place them to t in a coordinate. This must coincide exactly, one for one, for registration to be successful [50]. The utilization of computational methods including digital technologies (image analysis and AI), work in conjunction with X-rays to assist in the early detection of BC. The expansion of this, coupled with the development of high tech, has not only facilitated the disease’s earlier diagnosis but has also made it possible to treat a significant number of patients [51].
The diagnostic significance of images is improved by the process of fusion, which involves the extraction of meaningful information from many images, and the filtering of unnecessary information, and consequently the improvement of image quality. In general, signal level, data level, feature level, and decision level fusions make up the process of the fusion of images, in ascending order, from lowest to highest [52]. The use of meta-analysis in the diagnostic efficacy of DL has enabled the timely identification of breast cancer. The study used four categories: breast cancer; validation type; imaging modalities, such as, ultrasound; and DL algorithms versus healthcare professionals. The results showed that the pooled sensitivity was 88% (95% confidence interval: 85–90%), the specificity was 84% (79–87%), and the AUC was 0.92. (0.90–0.94). It was shown that all of the subgroups had a diagnostic accuracy that was satisfactory when using equivalent DL algorithms. Consequently, these techniques are beneficial for detecting breast cancer through the use of diagnostic imaging [52]. Again, Xue et al. [53], Freeman et al. [54], Mendes et al. [55] in all the three studies, AI was shown to detect 53, 45, and 50 percent of low-risk cancer, respectively. Additionally, AI identified 10%, 4%, and 0% of the breast cancer data from the data set that has already been screened and is utilized by radiologists. In the application of AI to BC images, AI presented a summary of the advantages and problems during the BC imaging survey, including prospective solutions, utilizing ML [56] and DL to forecast the risk of cancer, mammography appraisal, and data set labeling. In addition, the systematic and rapid segmentation of the region of interest (ROI) and breast density in magnetic resonance imaging (MRI), as reported by Pandey et al. [57], showed that the pictorial-based analysis displayed excellent segmentation in terms of accuracy, specificity, AUC, and sensitivity. The success was made evident by many technologically advanced devices which use the principle of DL for the diagnosis of BC. In particular, (AI)-CAD systems, such as iCAD’s PowerLook Tomo Detection and Screen Point Medical’s Transpara, are emerging into existence as the utilization of computed tomography becomes more widespread [57]. In summary, according to Shah et al. [44], many Al applications in BC diagnosis were reported by Nassif et al. [58], Dileep and Gianchandani Gyani [45], Huang et al. [32]., H-p et al. [59] and Shah et al. [59], whose evaluations offers a promising remark that the challenges of cancer prognosis and diagnosis are dealt with using advances in AI. These challenges are resolved through the use of AI, as AI has shown higher diagnostic accuracy than CAD in the detection of BC in mammograms. In addition, because the data set is readily available, AI may now be utilized to analyze mammograms using various image processing techniques. As a result of their lower cost and increasing prevalence, the images obtained from histological breast cancer tests are utilized in DL BC detection. AI has been integrated into several screening procedures to determine breast mass, density, and segmentation. The overall summative description of the various literary studies is given in Table 1.

2. Methodology

The methodology employed in this systematic review is devoid of any medical (either prospective or retrospective) data of patients; therefore, it was not necessary to obtain ethical approval to carry out the study. The data used in this study are articles from open-access publications based on details such as the dates of publication, authors’ names and methodologies used, type of AI models, dataset employed, and the overall results, such as the area under the curve (AUC), sensitivity, specificity, and accuracy. The literature searches were based on journals published in English between 2012 and 2022, all-inclusive. Breast cancer, breast cancer diagnosis, AI, and AI in breast cancer are the keywords used in the search. In the review, we searched these databases: IEEE, Embase, Excerpta Medica Database Guide Ovid, PubMed, Springer, Web of Science, and Google Scholar. A total of 17,900 results were found based on the aforementioned keywords. Some of the search results were out of the scope of breast cancer and/or AI. Those that were with the scope of this study had treatment options instead of diagnosis which led to further screening. Ultimately, 36 studies relevant to this review were used in this study. A detailed summary of the findings is shown in Table 1. A subcategory with a detailed explanation of BC diagnosis advantages, disadvantages/limitations, and AI in BC are detailed, as can be seen from the block diagram (Figure 1) of this review.

3. Results and Discussion

This study demonstrated the various articles that were systematically reviewed concerning the application of AI in breast cancer diagnosis, as published in various journals. Recently, there has been an increase in AI-based studies on the diagnosis of BC, which has demonstrated a significant value in this study. Rowland. W. [39] had a similar study on the concept and correlation of AI, ML, and DL, however, the application was geared towards other clinical predictions instead of BC specifically. Figure 1 shows the block diagram for this review. The selected studies were published between 2012 and 2022 and are all-inclusive. Figure 2 shows the annually published data on AI in BC diagnostics from the various aforementioned databases. As a result of AI’s promising use and modifications to suit breast tomosynthesis, there have been many studies on the applications of AI in BC diagnosis, from 2020 to date [40]. Meanwhile, the vast majority of the studies in Table 2 show prediction classification models in the evaluation of the parameters.
In addition, Table 3 shows the comparison of the latest related studies. However, accuracy (99%) has the highest number of performance metrics, followed by specificity (98%) and the area under the curve (0.95), as reported by Berker et al. [41]. These authors [42,43,44], use the CNN and the Digital Database for Screening Mammography (DDSM) in the comparative analysis of ImageNet and the classification of breast cancer [45,46]. Similarly, the author uses a Deep Neural Network (DNN) to classify breast cancer. As is evident from this review, the majority of the studies have focused on specificity or area under the curve without other parameters such as sensitivity, the accuracy of the diagnostic procedure, etc. being mentioned. This is a typical example of the limitations in medical studies. However, models such as CNN, Deep Convolutional Neural Networks (DCNN), Artificial Neural Networks (ANN), Digital Databases for Screening Mammography (DDSM), and others have been used in various studies. This is because these models are capable of automatic cancer detection as well as lesion interpretation. The studies provide independent advice for radiologists and oncologists to improve lesion identification and prognosis. Meanwhile, Figure 3 shows that the USA has the highest number of studies, followed by China and Japan. However, Figure 4 showed the summary of the task, date, and number of images or studies in the summary of Table 2. i.e., various algorithms of AI applications and their performances in breast cancer diagnosis. The figure shows that the classification algorithm was mostly used in the years between 2016 and 2019. In addition, in 2017 and 2019, 510,000 cases of fibro glandular breast density and 640,000 cases of breast imaging were studied using a classification algorithm in Switzerland and the United Kingdom, respectively.

3.1. Emerging Techniques of AI Applications in BC Diagnosis

Radiomics is an emerging technique that extracts variable quantitative features from images of medical origin [86]. Radiomics has been useful in the management and diagnosis of BC. The knowledge of radiomics helps in predictions, the staging of tumors, and the evaluation of therapeutic response [87]. Many academicians [88] have agreed that radiotherapy is very effective in the prognosis of BC. However, the only shortcoming of radiomics is that the data generated from locally advanced breast cancer (LABC) after non-adjuvant therapy and during post-surgical therapy cannot predict the survival rate of BC. Meanwhile, using AI’s application in BC diagnostics, reproducibility serves as a key point of experimental science, yet could be hampered by many factors such as the human or biological system, intrinsic variables, mislabeled samples, cross-contamination, and in some cases, over-passage in the cellular lines [89]. Aside from the generalizability issue, researchers use methods to remove these unnecessary (irrelevant) factors, which heavily rely on statistically applicable facts [88]. Generalizability issues are commendable trends, and the data set generated could be processed before mining [90,91,92]. This has demonstrated the importance of preprocessing medical data sets. However, this study has numerous unutilized social benefits. According to [93], AI applications in BC are becoming more common in developing countries. The reduced cost of the DeepMind automated system, for instance, offers relief to both developed and underdeveloped countries, as recognized by the World Health Organization (WHO) [93].

3.2. The Conception and Respective Correlations of AI, ML, and DL

AI includes techniques that enable computers to mimic the behavior of humans. The primary functions of AI in breast cancer screening are the segmentation and classification of benign or cancerous tumors [94]. ML is a learning algorithm whose characteristics and variables represent observable data [95]. DL is a typical example of ML, the methodology of which is dependent on deep neural networks that resemble but exaggerate human brain neurons [95] and are used in the classification and recognition of images [95]. DL uses a deep modular structure to promote hierarchies in learning and extracts information from simple to sophisticated models. However, there are several distinctions between ML and DL. Also, Figure 5 shows the relationship between AI, ML and DL. In terms of data dependencies, ML learns a mathematical model from training data [96]. The learned model forecasts the future by testing the data. Learning or training in ML denotes the collaborative technique of assessing the discrepancy between malignancy and benign using an assessment metric called the objective function [97]. This learning could be supervised learning [97], whereby the observed training data and the target are prerequisites for a training model. As a result, several research laboratories and corporations are attempting to build AI technologies for diverse healthcare domains. Supervised Machine Learning (SML) [98] may give healthcare professionals improved assistance in conducting differential diagnosis. The SML methodology also employs sophisticated ways to forecast health problems and alert the entire public towards impending danger [99]. In mammography, the cancer images are labels that allow the algorithm to learn the features of this malignancy. However, unsupervised learning [100] has no diagnostic features or abnormal labels. In semi-supervised learning, the information provided for the algorithm is not important for the training [101]. Simultaneously, the additional methodology could be employed to improve efficiency and cut down on the number of inaccurate predictions about breast cancer. The same can be said for breast pictures used to potentially detect breast cancer [35,102,103,104]. When compared to manual methods, AI-based automated image analysis helps eliminate laborious and time-consuming screening processes, while also efficiently capturing useful and relevant information from large amounts of image/picture data. This is accomplished in comparison to manual inspection.

3.3. BC Diagnosis Advantages

When AI is used in mammography, it can detect cancer up to two years earlier than a human oncologist. If diseases can be identified and treated sooner, more lives can be spared. Doctors can be more certain in their diagnoses with the aid of AI, thus increasing their efficiency [55]. AI also reduces the stress encountered by radiologists. Instead of spending hours reading mammogram images, AI is never tired. With AI, accuracy and earlier cancer detection are achieved. In addition, with the help of AI, medical care can be enhanced [56]. The use of AI-based diagnostic tools in the diagnosis of breast cancer has helped increase the efficiency of radiologists and produces results that are better than those obtained by radiologists working alone. While AI distinguishes between structures and image components using complex ML algorithms, its application in clinical practice is limited due to its low specificity [57]. In particular, with the use of AI, the accuracy of CAD has increased. AI uses CNN, a DL technology well-suited for image assessment and classification that has diagnostic accuracy that is approximate to or even better than that of radiologists [59] in cancer detection. AI could also help diagnose breast cancers that are not identified in the screening; as a result, there is a need for further screening, especially in dense breast tissues, which have low sensitivity. Some studies have shown that dense breasts have a higher rate of false-negative outcomes [58]. A negative mammography result may reassure women when they have cancer. The use of AI in the diagnosis of BC is saddled with problems. This makes AI limited, as can be seen in the subsequent section.

3.4. Limitation of AI in BC Diagnosis

This is the boundary between AI and human intellect. There is a need for AI to independently produce, replicate and accept data without human aid. AI machines could be considered a legitimate form of AI. This necessitates the existence of a universally accessible code, which can only be accomplished through the equitable distribution of data [105]. Databases that are simple to navigate and software that is intuitive to use need to be implemented into the information technology systems of hospitals across the globe. Additionally, there should be trust and confidence among clinicians to henceforth implement AI in all of their clinical decisions. Consequently, medical professionals should have adequate training on the use of AI innovations. The modern era and technology have provided us with apps that monitor diseases, such as heart rate and diabetes; however, this is yet to be seen in cases of BC. This will improve the quality of patient care and patient satisfaction if innovated. Ethically, the confidentiality of data, privacy violations, patient autonomy, consent, etc., are concerns that should be considered when utilizing AI in breast cancer diagnosis. Several precautionary measures are developed to prevent the disclosure of personal information, and legislation is in place to prohibit any misconduct. In addition, radiomics is still lacking widely in clinical practice today [106]. We, therefore, hope that decision-makers will implement these limitations so that AI will remain an effective diagnostic tool.

4. Conclusions

This study followed the standard method for the systematic review of papers. Stringent measures were considered in the inclusion criteria. In these criteria, geographically, the test scores were validated, i.e., the articles’ different centers in different countries. This excluded a large volume of studies, particularly in cases where the dataset was used for similar testing and validation. Internal validation overestimates the accuracy and has limited generalizability. This can also result in overfitting, and the loss of generalizability because the model’s performance is dependent on the data. Temporal validation offers a great capacity for the statistical models to forecast future circumstances for the entire population, based on which the model originally derived, after the observations that were used to generate the model. Nevertheless, the non-English studies were excluded because they have no relevance to the study. Meanwhile, the overall diagnostic performance can be improved through the use of AI in diagnostic approaches, for example, in detecting metastatic breast carcinoma in lymph node biopsies. Several models, such as CNN, and Digital Database for Screening Mammography (DDSM), were applied in order to achieve a timely and accurate result. In addition, AI achieves a significant result in images of breast cancer. This study showed that CNN was the most widely used algorithm. In addition, accuracy (98%) has the highest number of performance metrics, followed by specificity (99%) and area under the curve (0.95), and the findings from this study also showed that the majority of the studies were from the United States of America, followed by China and Japan.

Author Contributions

B.U. conceptualized the idea, and D.I.E. curated the data, analysis, and methodology, as well as writing the first draft. The validation, review, and editing of the manuscript were conducted by D.U.O. and I.O. The entire study was supervised by D.U.O. All authors have read and agreed to the published version of the manuscript.


This research received no external funding.

Informed Consent Statement

Not applicable.

Data Availability Statement

The data come from literary studies that were searched and reviewed between 2012 and 2022.

Conflicts of Interest

The authors declare no conflict of interest.


  1. Fortin, J.; Leblanc, M.; Elgbeili, G.; Cordova, M.J.; Marin, M.-F.; Brunet, A. The mental health impacts of receiving a breast cancer diagnosis: A meta-analysis. Br. J. Cancer 2021, 125, 1582–1592. [Google Scholar] [CrossRef] [PubMed]
  2. Siegel, R.; Naishadham, D.; Jemal, A. Cancer statistics, 2013. Cancer J. Clin. 2013, 63, 11–30. [Google Scholar] [CrossRef] [PubMed] [Green Version]
  3. Beral, V.; Bull, D.; Doll, R.; Peto, R.; Reeves, G. Breast cancer and breastfeeding: Collaborative reanalysis of individual data from 47 epidemiological studies in 30 countries, including 50302 women with breast cancer and 96973 women without the disease. Lancet 2002, 360, 187–195. [Google Scholar] [CrossRef]
  4. Sieri, S.; Krogh, V.; Bolelli, G.; Abagnato, C.A.; Grioni, S.; Pala, V.; Evangelista, A.; Allemani, C.; Micheli, A.; Tagliabue, G.; et al. Sex Hormone Levels, Breast Cancer Risk, and Cancer Receptor Status in Postmenopausal Women: The ORDET Cohort. Cancer Epidemiology Biomarkers Prev. 2009, 18, 169–176. [Google Scholar] [CrossRef] [PubMed] [Green Version]
  5. Colditz, G.A.; Kaphingst, K.A.; Hankinson, S.E.; Rosner, B. Family history and risk of breast cancer: Nurses’ health study. Breast Cancer Res. Treat. 2012, 133, 1097–1104. [Google Scholar] [CrossRef] [Green Version]
  6. Harbeck, N.; Penault-Llorca, F.; Cortes, J.; Gnant, M.; Houssami, N.; Poortmans, P.; Ruddy, K.; Tsang, J.; Cardoso, F. Breast cancer. Nat. Rev. Dis. Primers 2019, 5, 66. [Google Scholar] [CrossRef]
  7. Mahmood, T.; Li, J.; Pei, Y.; Akhtar, F.; Imran, A.; Rehman, K.U. A Brief Survey on Breast Cancer Diagnostic With Deep Learning Schemes Using Multi-Image Modalities. IEEE Access 2020, 8, 165779–165809. [Google Scholar] [CrossRef]
  8. Chiao, J.-Y.; Chen, K.-Y.; Liao, K.Y.-K.; Hsieh, P.-H.; Zhang, G.; Huang, T.-C. Detection and classification the breast tumors using mask R-CNN on sonograms. Medicine 2019, 98, e15200. [Google Scholar] [CrossRef]
  9. Cruz-Roa, A.; Gilmore, H.; Basavanhally, A.; Feldman, M.; Ganesan, S.; Shih, N.N.; Tomaszewski, J.; González, F.A.; Madabhushi, A. Accurate and reproducible invasive breast cancer detection in whole-slide images: A Deep Learning approach for quantifying tumor extent. Sci. Rep. 2017, 7, srep46450. [Google Scholar] [CrossRef] [Green Version]
  10. Wei, T.; Kanki, T.; Chikanari, M.; Uemura, T.; Sekitani, T.; Tanaka, H. Enhanced electronic-transport modulation in single-crystalline VO2 nanowire-based solid-state field-effect transistors. Sci. Rep. 2017, 7, 17215. [Google Scholar] [CrossRef]
  11. Breast Cancer Facts & Statistics 2022. Available online: (accessed on 4 November 2022).
  12. Breast Cancer. Available online: (accessed on 8 November 2022).
  13. Siegel, R.L.; Miller, K.D.; Fuchs, H.E.; Jemal, A. Cancer statistics, 2022. CA A Cancer J. Clin. 2022, 72, 7–33. [Google Scholar] [CrossRef]
  14. Iqbal, M.J.; Javed, Z.; Sadia, H.; Qureshi, I.A.; Irshad, A.; Ahmed, R.; Malik, K.; Raza, S.; Abbas, A.; Pezzani, R.; et al. Clinical applications of artificial intelligence and machine learning in cancer diagnosis: Looking into the future. Cancer Cell Int. 2021, 21, 1–11. [Google Scholar] [CrossRef] [PubMed]
  15. Prevedello, L.M.; Halabi, S.S.; Shih, G.; Wu, C.C.; Kohli, M.D.; Chokshi, F.H.; Erickson, B.J.; Kalpathy-Cramer, J.; Andriole, K.P.; Flanders, A.E. Challenges Related to Artificial Intelligence Research in Medical Imaging and the Importance of Image Analysis Competitions. Radiol. Artif. Intell. 2019, 1, e180031. [Google Scholar] [CrossRef] [PubMed]
  16. Nazir, A.; Khan, R.A. A novel combinatorial optimization based feature selection method for network intrusion detection. Comput. Secur. 2020, 102, 102164. [Google Scholar] [CrossRef]
  17. Crenn, A.; Meyer, A.; Konik, H.; Khan, R.A.; Bouakaz, S. Generic Body Expression Recognition Based on Synthesis of Realistic Neutral Motion. IEEE Access 2020, 8, 207758–207767. [Google Scholar] [CrossRef]
  18. Memon, J.; Sami, M.; Khan, R.A.; Uddin, M. Handwritten Optical Character Recognition (OCR): A Comprehensive Systematic Literature Review (SLR). IEEE Access 2020, 8, 142642–142668. [Google Scholar] [CrossRef]
  19. Marinovich, M.L.; Wylie, E.; Lotter, W.; Pearce, A.; Carter, S.M.; Lund, H.; Waddell, A.; Kim, J.G.; Pereira, G.F.; Lee, C.I.; et al. Artificial intelligence (AI) to enhance breast cancer screening: Protocol for population-based cohort study of cancer detection. BMJ Open 2022, 12, e054005. [Google Scholar] [CrossRef]
  20. Jaliaawala, M.S.; Khan, R.A. Can autism be catered with artificial intelligence-assisted intervention technology? A comprehensive survey. Artif. Intell. Rev. 2018, 53, 1039–1069. [Google Scholar] [CrossRef] [Green Version]
  21. Sun, P.; Feng, Y.; Chen, C.; Dekker, A.; Qian, L.; Wang, Z.; Guo, J. An AI model of sonographer’s evaluation+ S-Detect + elastography + clinical information improve the preoperative identification of benign and malignant breast masses. Front. Oncol. 2022, 12, 1022441. [Google Scholar] [CrossRef]
  22. Sharma, S.; Mehra, R. Conventional Machine Learning and Deep Learning Approach for Multi-Classification of Breast Cancer Histopathology Images—A Comparative Insight. J. Digit. Imaging 2020, 33, 632–654. [Google Scholar] [CrossRef]
  23. Burt, J.R.; Torosdagli, N.; Khosravan, N.; Raviprakash, H.; Mortazi, A.; Tissavirasingham, F.; Hussein, S.; Bagci, U. Deep learning beyond cats and dogs: Recent advances in diagnosing breast cancer with deep neural networks. Br. J. Radiol. 2018, 91, 20170545. [Google Scholar] [CrossRef] [PubMed]
  24. Buckner, C.A.; Lafrenie, R.; Dénommée, J.; Caswell, J.; Want, D.; Gan, G.; Leong, Y.; Bee, P.; Chin, E.; Teh, A.; et al. We are IntechOpen, the world’s leading publisher of Open Access books Built by scientists, for scientists TOP 1 %. Intech 2016, 11, 13, no. tourism. Available online: (accessed on 3 December 2022).
  25. Goyal, S.; Srivastava, B.R. Review of Artificial Intelligence Applicability of Various Diagnostic Modalities, their Advantages, Limitations, and Overcoming the Challenges in Breast Imaging. Int. J. Sci. Study 2021, 9, 16–21. [Google Scholar]
  26. Rodriguez-Ruiz, A.; Lång, 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. Gynecol. Oncol. 2019, 111, 916–922. [Google Scholar] [CrossRef]
  27. Liu, K.-L.; Wu, T.; Chen, P.-T.; Tsai, Y.M.; Roth, H.; Wu, M.-S.; Liao, W.-C.; Wang, W. Deep learning to distinguish pancreatic cancer tissue from non-cancerous pancreatic tissue: A retrospective study with cross-racial external validation. Lancet Digit. Health 2020, 2, e303–e313. [Google Scholar] [CrossRef]
  28. Balakrishnan, G.; Zhao, A.; Sabuncu, M.R.; Guttag, J.; Dalca, A.V. VoxelMorph: A Learning Framework for Deformable Medical Image Registration. IEEE Trans. Med. Imaging 2019, 38, 1788–1800. [Google Scholar] [CrossRef] [Green Version]
  29. Ronneberger, O.; Fischer, P.; Brox, T. U-net: Convolutional networks for biomedical image segmentation. In International Conference on Medical Image Computing and Computer-Assisted Intervention; Springer: Cham, Switzerland, 2015; Volume 9351, pp. 234–241. [Google Scholar] [CrossRef]
  30. Breast Cancer|Early Detection and Screening|CancerCare. Available online: (accessed on 3 December 2022).
  31. Breast Cancer Early Detection—National Breast Cancer Foundation. Available online: (accessed on 3 December 2022).
  32. Huang, S.; Yang, J.; Fong, S.; Zhao, Q. Artificial intelligence in cancer diagnosis and prognosis: Opportunities and challenges. Cancer Lett. 2019, 471, 61–71. [Google Scholar] [CrossRef]
  33. Witowski, J.; Heacock, L.; Reig, B.; Kang, S.K.; Lewin, A.; Pysarenko, K.; Patel, S.; Samreen, N.; Rudnicki, W.; Łuczyńska, E.; et al. Improving breast cancer diagnostics with DL for MRI. Sci. Transl. Med. 2 2022, 14, eabo4802. [Google Scholar] [CrossRef]
  34. Cole, E.B.; Zhang, Z.; Marques, H.S.; Nishikawa, R.M.; Hendrick, R.E.; Yaffe, M.J.; Padungchaichote, W.; Kuzmiak, C.; Chayakulkheeree, J.; Conant, E.F.; et al. Assessing the Stand-Alone Sensitivity of Computer-Aided Detection with Cancer Cases From the Digital Mammographic Imaging Screening Trial. AJR Am. J. Roentgenol. 2012, 199, W392. [Google Scholar] [CrossRef] [Green Version]
  35. Díaz, O.; Rodríguez-Ruiz, A.; Gubern-Mérida, A.; Martí, R.; Chevalier, M. Are artificial intelligence systems useful in breast cancer screening programs? Radiologia 2021, 63, 236–244. [Google Scholar] [CrossRef]
  36. Trister, A.D.; Buist, D.S.M.; Lee, C.I. Will Machine Learning Tip the Balance in Breast Cancer Screening? JAMA Oncol. 2017, 3, 1463. [Google Scholar] [CrossRef] [PubMed]
  37. Liu, H.; Cui, G.; Luo, Y.; Guo, Y.; Zhao, L.; Wang, Y.; Subasi, A.; Dogan, S.; Tuncer, T. Artificial Intelligence-Based Breast Cancer Diagnosis Using Ultrasound Images and Grid-Based Deep Feature Generator. Int. J. Gen. Med. 2022, 15, 2271–2282. [Google Scholar] [CrossRef] [PubMed]
  38. Goyal, S. Essentials of Abdomino-Pelvic Sonography: A Handbook for Practitioners; CRC Press: Boca Raton, FL, USA, 2020; Available online: (accessed on 3 December 2022).
  39. Lee, J.; Kang, B.J.; Kim, S.H.; Park, G.E. Evaluation of Computer-Aided Detection (CAD) in Screening Automated Breast Ultrasound Based on Characteristics of CAD Marks and False-Positive Marks. Diagnostics 2022, 12, 583. [Google Scholar] [CrossRef]
  40. Parvathavarthini, S.; Visalakshi, N.K.; Shanthi, S. Breast Cancer Detection using Crow Search Optimization based Intuitionistic Fuzzy Clustering with Neighborhood Attraction. Asian Pac. J. Cancer Prev. 2019, 20, 157–165. [Google Scholar] [CrossRef] [Green Version]
  41. Abdel-Zaher, A.M.; Eldeib, A.M. Breast cancer classification using deep belief networks. Expert Syst. Appl. 2016, 46, 139–144. [Google Scholar] [CrossRef]
  42. Karabatak, M. A new classifier for breast cancer detection based on Naïve Bayesian. Measurement 2015, 72, 32–36. [Google Scholar] [CrossRef]
  43. Zafiropoulos, E.; Maglogiannis, I.; Anagnostopoulos, I. A support vector machine approach to breast cancer diagnosis and prognosis. IFIP Int. Fed. Inf. Process. 2006, 204, 500–507. [Google Scholar] [CrossRef]
  44. Shah, S.M.; Khan, R.A.; Arif, S.; Sajid, U. Artificial intelligence for breast cancer analysis: Trends & directions. Comput. Biol. Med. 2022, 142, 105221. [Google Scholar] [CrossRef]
  45. Dileep, G.; Gyani, S.G.G. Artificial Intelligence in Breast Cancer Screening and Diagnosis. Cureus 2022, 14, e30318. [Google Scholar] [CrossRef]
  46. Shah, R.; Rosso, K.; Nathanson, S.D. Pathogenesis, prevention, diagnosis and treatment of breast cancer. World J. Clin. Oncol. 2014, 5, 283–298. [Google Scholar] [CrossRef] [Green Version]
  47. What Is Breast Cancer Screening?|CDC. Available online: (accessed on 3 December 2022).
  48. Zhang, Y.-N.; Xia, K.-R.; Li, C.-Y.; Wei, B.-L.; Zhang, B. Review of Breast Cancer Pathologigcal Image Processing. BioMed Res. Int. 2021, 2021, 1994764. [Google Scholar] [CrossRef] [PubMed]
  49. Maintz, J.; Viergever, M.A. A survey of medical image registration. Med Image Anal. 1998, 2, 1–36. [Google Scholar] [CrossRef] [PubMed]
  50. Hill, D.L.G.; Batchelor, P.G.; Holden, M.; Hawkes, D.J. Medical image registration. Phys. Med. Biol. 2001, 46, R1–R45. [Google Scholar] [CrossRef] [PubMed]
  51. Younis, Y.S.; Ali, A.H.; Alhafidhb, O.K.S.; Yahia, W.B.; Alazzam, M.B.; Hamad, A.A.; Meraf, Z. Early Diagnosis of Breast Cancer Using Image Processing Techniques. J. Nanomater 2022, 2022, 2641239. [Google Scholar] [CrossRef]
  52. Kooi, T.; Litjens, G.; van Ginneken, B.; Gubern-Mérida, A.; Sánchez, C.I.; Mann, R.; den Heeten, A.; Karssemeijer, N. Large scale deep learning for computer aided detection of mammographic lesions. Med. Image Anal. 2017, 35, 303–312. [Google Scholar] [CrossRef]
  53. Xue, P.; Wang, J.; Qin, D.; Yan, H.; Qu, Y.; Seery, S.; Jiang, Y.; Qiao, Y. Deep learning in image-based breast and cervical cancer detection: A systematic review and meta-analysis. NPJ Digit. Med. 2022, 5, 19. [Google Scholar] [CrossRef]
  54. Freeman, K.; Geppert, J.; Stinton, C.; Todkill, D.; Johnson, S.; Clarke, A.; Taylor-Phillips, S. Use of artificial intelligence for image analysis in breast cancer screening programmes: Systematic review of test accuracy. BMJ 2021, 374, n1872. [Google Scholar] [CrossRef]
  55. Mendes, J.; Domingues, J.; Aidos, H.; Garcia, N.; Matela, N. AI in Breast Cancer Imaging: A Survey of Different Applications. J. Imaging 2022, 8, 228. [Google Scholar] [CrossRef]
  56. Ozsahin, I.; Uzun Ozsahin, D. Applied Machine Learning and Multi-criteria Decision-making in Healthcare; Bentham Science Publishers: Sharjah, United Arab Emirates, 2021; ISBN 9781681088716. [Google Scholar]
  57. Pandey, D.; Yin, X.; Wang, H.; Su, M.-Y.; Chen, J.-H.; Wu, J.; Zhang, Y. Automatic and fast segmentation of breast region-of-interest (ROI) and density in MRIs. Heliyon 2018, 4, e01042. [Google Scholar] [CrossRef] [Green Version]
  58. The American Cancer Society. Available online: (accessed on 27 May 2020).
  59. Chan, H.P.; Samala, R.K.; Hadjiiski, L.M. CAD and Ai for breast cancer-recent development and challenges. Br. J. Radiol. 2020, 93, 20190580. [Google Scholar] [CrossRef]
  60. Sadoughi, F.; Kazemy, Z.; Hamedan, F.; Owji, L.; Rahmanikatigari, M.; Azadboni, T.T. Artificial intelligence methods for the diagnosis of breast cancer by image processing: A review. Breast Cancer Targets Ther. 2018, 10, 219–230. [Google Scholar] [CrossRef] [PubMed]
  61. Le, E.; Wang, Y.; Huang, Y.; Hickman, S.; Gilbert, F. Artificial intelligence in breast imaging. Clin. Radiol. 2019, 74, 357–366. [Google Scholar] [CrossRef] [PubMed]
  62. Shao, D.; Dai, Y.; Li, N.; Cao, X.; Zhao, W.; Cheng, L.; Rong, Z.; Huang, L.; Wang, Y.; Zhao, J. Artificial intelligence in clinical research of cancers. Briefings Bioinform. 2021, 23, bbab523. [Google Scholar] [CrossRef] [PubMed]
  63. Tan, M.; Zheng, B.; Ramalingam, P.; Gur, D. Prediction of Near-term Breast Cancer Risk Based on Bilateral Mammographic Feature Asymmetry. Acad. Radiol. 2013, 20, 1542–1550. [Google Scholar] [CrossRef] [PubMed] [Green Version]
  64. Qiu, Y.; Wang, Y.; Yan, S.; Tan, M.; Cheng, S.; Liu, H.; Zheng, B. An initial investigation on developing a new method to predict short-term breast cancer risk based on deep learning technology. Med. Imaging 2016 Comput. Aided Diagn. 2016, 9785, 978521. [Google Scholar] [CrossRef]
  65. Akselrod-Ballin, A.; Chorev, M.; Shoshan, Y.; Spiro, A.; Hazan, A.; Melamed, R.; Barkan, E.; Herzel, E.; Naor, S.; Karavani, E.; et al. Predicting breast cancer by applying DL to linked health records and mammograms. Radiology 2019, 292, 331–342. [Google Scholar] [CrossRef]
  66. Samala, R.K.; Chan, H.-P.; Hadjiiski, L.; Helvie, M.A.; Wei, J.; Cha, K. Mass detection in digital breast tomosynthesis: Deep convolutional neural network with transfer learning from mammography. Med. Phys. 2016, 43, 6654–6666. [Google Scholar] [CrossRef] [Green Version]
  67. Becker, A.; Marcon, M.; Ghafoor, S.; Wurnig, M.C.; Frauenfelder, T.; Boss, A. Deep Learning in Mammography. Investig. Radiol. 2017, 52, 434–440. [Google Scholar] [CrossRef]
  68. Kim, E.-K.; Kim, H.-E.; Han, K.; Kang, B.J.; Sohn, Y.-M.; Woo, O.H.; Lee, C.W. Applying Data-driven Imaging Biomarker in Mammography for Breast Cancer Screening: Preliminary Study. Sci. Rep. 2018, 8, 2762. [Google Scholar] [CrossRef] [Green Version]
  69. Sun, D.; Wang, M.; Li, A. A Multimodal Deep Neural Network for Human Breast Cancer Prognosis Prediction by Integrating Multi-Dimensional Data. IEEE/ACM Trans. Comput. Biol. Bioinform. 2018, 16, 841–850. [Google Scholar] [CrossRef]
  70. Muhammad, M.; Zeebaree, D.; Brifcani, A.M.A.; Saeed, J.; Zebari, D.A. Region of Interest Segmentation Based on Clustering Techniques for Breast Cancer Ultrasound Images: A Review. J. Appl. Sci. Technol. Trends 2020, 1, 78–91. [Google Scholar] [CrossRef]
  71. Ribli, D.; Horváth, A.; Unger, Z.; Pollner, P.; Csabai, I. Detecting and classifying lesions in mammograms with Deep Learning. Sci. Rep. 2018, 8, 4165. [Google Scholar] [CrossRef] [PubMed] [Green Version]
  72. Jiao, Z.; Gao, X.; Wang, Y.; Li, J. A parasitic metric learning net for breast mass classification based on mammography. Pattern Recognit. 2018, 75, 292–301. [Google Scholar] [CrossRef]
  73. Chougrad, H.; Zouaki, H.; Alheyane, O. Deep Convolutional Neural Networks for breast cancer screening. Comput. Methods Programs Biomed. 2018, 157, 19–30. [Google Scholar] [CrossRef]
  74. Wang, H.; Feng, J.; Zhang, Z.; Su, H.; Cui, L.; He, H.; Liu, L. Breast mass classification via deeply integrating the contextual information from multi-view data. Pattern Recognit. 2018, 80, 42–52. [Google Scholar] [CrossRef]
  75. Lehman, C.D.; Yala, A.; Schuster, T.; Dontchos, B.; Bahl, M.; Swanson, K.; Barzilay, R. Mammographic Breast Density Assessment Using Deep Learning: Clinical Implementation. Radiology 2019, 290, 52–58. [Google Scholar] [CrossRef] [Green Version]
  76. Byra, M.; Galperin, M.; Ojeda-Fournier, H.; Olson, L.; O’Boyle, M.; Comstock, C.; Andre, M. Breast mass classification in sonography with transfer learning using a deep convolutional neural network and color conversion. Med. Phys. 2018, 46, 746–755. [Google Scholar] [CrossRef]
  77. Fujioka, T.; Kubota, K.; Mori, M.; Kikuchi, Y.; Katsuta, L.; Kasahara, M.; Oda, G.; Ishiba, T.; Nakagawa, T.; Tateishi, U. Distinction between benign and malignant breast masses at breast ultrasound using deep learning method with convolutional neural network. Jpn. J. Radiol. 2019, 37, 466–472. [Google Scholar] [CrossRef]
  78. Hickman, S.E.; Woitek, R.; Le, E.P.V.; Im, Y.R.; Luxhøj, C.M.; Aviles-Rivero, A.I.; Baxter, G.C.; MacKay, J.W.; Gilbert, F.J. Machine Learning for Workflow Applications in Screening Mammography: Systematic Review and Meta-Analysis. Radiology 2022, 302, 88–104. [Google Scholar] [CrossRef]
  79. Myers, E.R.; Moorman, P.G.; Gierisch, J.M.; Havrilesky, L.J.; Grimm, L.; Ghate, S.V.; Davidson, B.; Mongtomery, R.C.; Crowley, M.J.; McCrory, D.C.; et al. Benefits and Harms of Breast Cancer Screening. JAMA 2015, 314, 1615–1634. [Google Scholar] [CrossRef]
  80. Kayode, A.A.; Akande, N.O.; Adegun, A.A.; Adebiyi, M.O. An automated mammogram classification system using modified support vector machine. Med. Devices: Évid. Res. 2019, 12, 275–284. [Google Scholar] [CrossRef] [PubMed]
  81. Wei, M.; Du, Y.; Wu, X.; Su, Q.; Zhu, J.; Zheng, L.; Lv, G.; Zhuang, J. A Benign and Malignant Breast Tumor Classification Method via Efficiently Combining Texture and Morphological Features on Ultrasound Images. Comput. Math. Methods Med. 2020, 2020, 5894010. [Google Scholar] [CrossRef] [PubMed]
  82. Dutta, K.; Roy, S.; Whitehead, T.; Luo, J.; Jha, A.; Li, S.; Quirk, J.; Shoghi, K. Deep Learning Segmentation of Triple-Negative Breast Cancer (TNBC) Patient Derived Tumor Xenograft (PDX) and Sensitivity of Radiomic Pipeline to Tumor Probability Boundary. Cancers 2021, 13, 3795. [Google Scholar] [CrossRef] [PubMed]
  83. Zhang, Y.; Weng, Y.; Lund, J. Applications of Explainable Artificial Intelligence in Diagnosis and Surgery. Diagnostics 2022, 12, 237. [Google Scholar] [CrossRef]
  84. Roy, S.; Whitehead, T.D.; Quirk, J.D.; Salter, A.; Ademuyiwa, F.O.; Li, S.; An, H.; Shoghi, K.I. Optimal co-clinical radiomics: Sensitivity of radiomic features to tumour volume, image noise and resolution in co-clinical T1-weighted and T2-weighted magnetic resonance imaging. Ebiomedicine 2020, 59, 102963. [Google Scholar] [CrossRef]
  85. Roy, S.; Whitehead, T.D.; Li, S.; Ademuyiwa, F.O.; Wahl, R.L.; Dehdashti, F.; Shoghi, K.I. Co-clinical FDG-PET radiomic signature in predicting response to neoadjuvant chemotherapy in triple-negative breast cancer. Eur. J. Nucl. Med. 2021, 49, 550–562. [Google Scholar] [CrossRef]
  86. Lambin, P.; Leijenaar, R.T.H.; Deist, T.M.; Peerlings, J.; de Jong, E.E.C.; van Timmeren, J.; Sanduleanu, S.; Larue, R.T.H.M.; Even, A.J.G.; Jochems, A.; et al. Radiomics: The bridge between medical imaging and personalized medicine. Nat. Rev. Clin. Oncol. 2017, 14, 749–762. [Google Scholar] [CrossRef] [Green Version]
  87. Liang, W.; Yang, P.; Huang, R.; Xu, L.; Wang, J.; Liu, W.; Zhang, L.; Wan, D.; Huang, Q.; Lu, Y.; et al. A Combined Nomogram Model to Preoperatively Predict Histologic Grade in Pancreatic Neuroendocrine Tumors. Clin. Cancer Res. 2019, 25, 584–594. [Google Scholar] [CrossRef] [Green Version]
  88. Park, H.; Lim, Y.; Ko, E.S.; Cho, H.-H.; Lee, J.E.; Han, B.-K.; Choi, J.S.; Park, K.W. Radiomics Signature on Magnetic Resonance Imaging: Association with Disease-Free Survival in Patients with Invasive Breast Cancer. Clin. Cancer Res. 2018, 24, 4705–4714. [Google Scholar] [CrossRef] [Green Version]
  89. Understanding Reproducibility and Replicability—Reproducibility, and Replicability in Science—NCBI Bookshelf. Available online: (accessed on 13 December 2022).
  90. Kukull, W.A.; Ganguli, M. Generalizability: The trees, the forest, and the low-hanging fruit. Neurology 2012, 78, 1886–1891. [Google Scholar] [CrossRef] [Green Version]
  91. Duwa, B.B.; Ozsoz, M.; Al-Turjman, F. Applications of AI, IoT, IoMT, and Biosensing Devices in Curbing COVID-19. In AI-Powered IoT for COVID-19; CRC Press: Boca Raton, FL, USA, 2020; pp. 141–158. [Google Scholar] [CrossRef]
  92. Ibrahim, A.U.; Ozsoz, M.; Al-Turjman, F.; Coston, P.P.; Duwa, B.B. How Artificial Intelligence and IoT Aid in Fighting COVID-19. In AI-Powered IoT for COVID-19; CRC Press: Boca Raton, FL, USA, 2020; pp. 159–168. [Google Scholar] [CrossRef]
  93. World Health Organization. WHO Position Paper on Mammography Screening; World Health Organization: Geneva, Switzerland, 2014. [Google Scholar]
  94. Lei, Y.-M.; Yin, M.; Yu, M.-H.; Yu, J.; Zeng, S.-E.; Lv, W.-Z.; Li, J.; Ye, H.-R.; Cui, X.-W.; Dietrich, C.F. Artificial Intelligence in Medical Imaging of the Breast. Front. Oncol. 2021, 11, 2892. [Google Scholar] [CrossRef] [PubMed]
  95. Tsantekidis, A.; Passalis, N.; Tefas, A. Deep Reinforcement Learning; McGraw-Hill: New York, NY, USA, 2022; pp. 117–129. [Google Scholar] [CrossRef]
  96. LeCun, Y.; Bengio, Y.; Hinton, G. Deep learning. Nature 2015, 521, 436–444. [Google Scholar] [CrossRef] [PubMed]
  97. Wang, S.; Summers, R.M. Machine learning and radiology. Med. Image Anal. 2012, 16, 933–951. [Google Scholar] [CrossRef] [Green Version]
  98. Zhavoronkov, A. Artificial Intelligence for Drug Discovery, Biomarker Development, and Generation of Novel Chemistry. Mol. Pharm. 2018, 15, 4311–4313. [Google Scholar] [CrossRef] [PubMed] [Green Version]
  99. Roy, S.; Meena, T.; Lim, S.-J. Demystifying Supervised Learning in Healthcare 4.0: A New Reality of Transforming Diagnostic Medicine. Diagnostics 2022, 12, 2549. [Google Scholar] [CrossRef] [PubMed]
  100. Schölkopf, B.; Platt, J.; Hofmann, T. Greedy layer-wise training of deep networks. ” Adv. Neural. Inf. Process Syst. 2007, 19, 153–160. [Google Scholar]
  101. Ligthart, A.; Catal, C.; Tekinerdogan, B. Analyzing the effectiveness of semi-supervised learning approaches for opinion spam classification. Appl. Soft Comput. 2020, 101, 107023. [Google Scholar] [CrossRef]
  102. Ozsahin, D.U.; Sheshakli, S.; Kibarer, A.G.; Denker, A.; Duwa, B.B. Analysis of early stage breast cancer treatment techniques. Appl. Multi-Criteria Decis.-Mak. Theor. Healthc. Biomed. Eng. 2021, 71–80. [Google Scholar] [CrossRef]
  103. George, K.; Faziludeen, S.; Sankaran, P. Breast cancer detection from biopsy images using nucleus guided transfer learning and belief based fusion. Comput. Biol. Med. 2020, 124, 103954. [Google Scholar] [CrossRef]
  104. Talo, M. Automated classification of histopathology images using transfer learning. Artif. Intell. Med. 2019, 101, 101743. [Google Scholar] [CrossRef] [Green Version]
  105. Yuan, J.; Hu, Z.; Mahal, B.A.; Zhao, S.D.; Kensler, K.H.; Pi, J.; Hu, X.; Zhang, Y.; Wang, Y.; Jiang, J.; et al. Integrated Analysis of Genetic Ancestry and Genomic Alterations across Cancers. Cancer Cell 2018, 34, 549–560. [Google Scholar] [CrossRef] [PubMed] [Green Version]
  106. Tagliafico, A.S.; Piana, M.; Schenone, D.; Lai, R.; Massone, A.M.; Houssami, N. Overview of radiomics in breast cancer diagnosis and prognostication. Breast 2019, 49, 74–80. [Google Scholar] [CrossRef] [PubMed]
Figure 1. Block diagram of AI in BC diagnosis.
Figure 1. Block diagram of AI in BC diagnosis.
Diagnostics 13 00045 g001
Figure 2. Showing publications on AI in breast cancer diagnosis between 2012–2022.
Figure 2. Showing publications on AI in breast cancer diagnosis between 2012–2022.
Diagnostics 13 00045 g002
Figure 3. Size of input or training data used by the different countries where the studies were carried out.
Figure 3. Size of input or training data used by the different countries where the studies were carried out.
Diagnostics 13 00045 g003
Figure 4. Summary of task, date, and number of images or studies used in figure.
Figure 4. Summary of task, date, and number of images or studies used in figure.
Diagnostics 13 00045 g004
Figure 5. Relationship between AI, ML, and DL.
Figure 5. Relationship between AI, ML, and DL.
Diagnostics 13 00045 g005
Table 1. Summary of the results of 36 studies.
Table 1. Summary of the results of 36 studies.
Author’s NameYearTitleSummary of Result
Kooi et al. [52] 2017The advantages of Large-scale DL for CAD of breast lesionsThe result showed that CNN performs better than CAD in a low sensitivity when compared.
Pandey et al. [57] 2018The systematic and rapid segmentation of the region of interest (ROI) and breast density in MRIThe pictorial-based analysis displayed excellent segmentation in terms of accuracy, specificity, AUC, and sensitivity
Sadoughi et al. [60]2018AI methodology in the diagnosis of BCThe result showed that SVM has the highest accuracy (100%) when compared with various images from thermographs (99.85%), mammograms (93.69%), and ultrasound (US) (95.85%)
Le et al. [61] 2019The use of AI in breast imaging The transition from the traditional rule-based CAD to an enhanced embedded knowledge that reduces diagnostic error and improves radiologist accuracy.
Huang et al. [32]2020AI in BC, prognosis, and diagnosisThe challenges of cancer prognosis and diagnosis are dealt with using advances in AI
H-p et al. [59] 2020Combination of AI and CAD in BC. Recent advancements and difficultiesDL employed in clinical image datasets combined with CAD provides a novelty workflow in clinical practice
Shah et al. [44] 2021AI in detecting BCAI showed higher diagnostic accuracy than CAD in the detection of BC in mammogram
Freeman et al. [54]2021Utilization of AI for image analysis in BC screening initiatives: a comprehensive assessment of the diagnostic validityIn the 3 studies, AI detects 53%, 45%, and 50% of low-risk cancer. In addition, AI detects 10%, 4%, and 0% BC data from the already screened data set used by radiologists.
Utilization of AI for image analysis in BC screening initiatives: a comprehensive assessment of the diagnostic validity
Shah et al. [44] 2022Trend and direction in the application of AI in the diagnosis of BCThe availability of the data set makes it possible for AI to be applied using image modalities to the mammogram.
Nassif et al. [58] 2022The use of AI BC detectionImages from histopathological BC examinations are less expensive and very common, therefor they are used in DL BC detection
Dileep and Gianchandani Gyani [45] 2022AI application in BC screening and diagnosisIncorporation of AI into different screening methodologies to detect breast mass, density, and segmentation.
Shao et al. [62] 2022The application of AI to the clinical study of BCThe AI was able to classify BC according to the many kinds of data, such as plain radiographs, cancer genes, health records, pharmacological information, and biological works of literature.
Mendes et al. [55]2022An Overview of the Role of AI in Imaging analysis of BC and Its Various ApplicationsAI application in BC images using ML, DL to predict the risk of cancer, mammogram evaluation, and labeling of data set, AI showed an overview of the advantages and challenges including the potential solution during the BC imaging survey
Table 2. Various algorithms of AI applications and performances in breast cancer diagnosis.
Table 2. Various algorithms of AI applications and performances in breast cancer diagnosis.
ReferenceDateTaskCountryTumor TypeModel SourceNo Images/StudiesType of ModelSize of Input/TrainingIndependent Test SetPerformance ValidationThe Area under the Curve (AUC) Sensitivity
Tan et al. [63]2013SVM classifierUSABreast mammogramGaussian kernel994 womenGaussian kernel349 benign, 362 cancerousNANA0.73NANANA
Qiu et al. [64]2016ClassificationUSAbreastCNN 270CNN, CAD200 cases, 70 casesNANA0.70706071.4
Ayelet Akselrod-Ballin [65]2016DLCanadaHistopathologica + mammographyR-CNN52,936In house Lab9611 NA10550.7287.077.3 91.0
Samala et al. [66]2016mass detection and classificationUSAMammography + X-ray SFM and DM28,330Cuda- convNet28,330 94NA0.8183.091.0NA
Becker A, [67]2017ClassificationSwitzerlandFibro glandular
volumetric breast density.
Quantra 2.2 510,000ANN + Volpara DensityGroup 1 = 95/95, Group 2 = 83/513NANA0.7978.084.0 81.0
Kim al. [68]2018Testing ANNsSouth KoreaBreast DIB-MG
29,107ANN18 cancers and 233 controlsNANA0.9973.7 versus 66.672.0 versus 92.7NA
Sun et al. [69]2018Multimodal DNNChinaBreast cancerMetabric1980DNN1054 NANA8.4095992.4
Mohammed et al. [70]2018classificationIraqUltrasound Breast ImagesOI + ANN classifier1393ANN classifier900. 300NANANA79.484.7682.0
Ribli D. et al. [71]2018classificationHungaryMammogramsCNN86,000CNN7700, 847NANA0.8598.799.698.7
Jiao et al. [72]2018Alex Net + ImageNetChinaDL ClassificationDDSM 300 images300 DDSM + MIAS set300 150150NANANADDSM = 97.4 MIAS = 96.7
Chougrad et al. [73]2018Comparative image classificationMorrocoMammography + X-rayDDSM, INbreast, BCDR 6116DDSM, INbreast, BCDR 641, 688DDSM
Wang H et al. [74]2018classificationChinaMammographyMV DCNN736MV-DNN, + MAP36829574DNN = 0.828, MAP = 0.846NANANA
Lehman et al. [75]2019Breast segment classificationUSABreast densityDMs + ResNet1827,684Pretrained ResNet1827,684 5741 + 10768738 Kappa = 0.67NANA77
Byra et al. [76]2019Classification of Breast USUSABreast USINbreastes582 AUC (VGG19 + FT + ML) 582150 1500.936NANANA
Fujioka et al. [77]2019Breast mass segmentationJapanBreast USDCNN947DCNN480 benign, 467 cancerousNANA0.913NANANA
E.P.V Le [78] 2019classificationUkBreast imagingiCAD’s Detection + ScreenPoint MedicalTranspara640,000 CNN + DREAM Challenge318,000NANA0.91
84.091 [79]NA
Kayode et al. [80]2019SVM classifierNigeriaBreast GLCM322Texture Feature + SVM126 Normal, 60 benign, 48 malignantNANANA94.591.3NA
Tomoyuki Fujioka,
Kazunori Kubota
et al. [77]
2019ClassificationJapanBreast tumorCNN1536CNN + Res152897 malig, 639 benignNANA0.95190.987.0 NA
Wei M [81]2020ClassificationChinaBreast tumorRadial Range Spectrum + GLCM1061SVM classifier589 malignant, 472 benignNANA0.9387.087.687.3
Tomoyuki Fujioka,
Kazunori Kubota
et al. [77]
2020DifferentiationJapanBreast cancerCNN576.6CNN 48 benign, 72 malignantNANA0.93,
95.8, 58.3, 91.792.5 60.4
92.5 65.8
Table 3. Comparison of the latest related studies.
Table 3. Comparison of the latest related studies.
ReferenceCountryTopicAimMethods UsedResultLimitation
Dutta et al., 2021 [82]USAThe classification of Triple-Negative Breast Cancer (TNBC) using DL methodology on Patient Derived Tumor Xenograft (PDX) and Tumor Probability Boundary Sensitivity of Radiomic PipelineDL is used in the automated pathway for the detection and quantification of TNBC PDX tumors from nonclinical weighted (T1W) images and weighted images (T2WI) of M RIManual comparison of delineation with U-Net, dense U-Net, Res-Net, recurrent residual U-Net (R2UNet), and dense R2U-Net The compared networks had a score ranging between 1% and 3% Time intensive. Variations in the results of repeated measurements cannot be reproduced.
Zhang et al., 2022 [83] UKDiagnostic and Surgical Applications of AI that Can Be interpreted or explainedOvercome AI’s explainable or interpreted (XAI) black box identityLiterary search from databases between 2019 to 2021XAI is heading in several interesting directions.Black box identity of some of the AI models
Roy et al., 2020 [84]USARadiomic characteristics and resolution clinical weighted (T1W) image and weighted image (T2WI) from MRIRadiomics characteristics were used in a combined clinical trial.The study uses TNBC patients Sixteen volume-dependent characteristics of radiomics were identified Models used are genetically engineered PDXs are also used.
Roy et al., 2022 [85] USAPredicting response to neoadjuvant in TNBC by using a co-clinical FDG-PET (RadSig) radiomic signatureOptimization and identification of radiometric characteristics and therapeutic responseThe study uses TNBC patients as well as PDXsNB-RadSig has the highest prediction and therapeutic response.SVM-RadSig NB-RadSig superior to Standardized uptake values of mean, maximum, and peak
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

Share and Cite

MDPI and ACS Style

Uzun Ozsahin, D.; Ikechukwu Emegano, D.; Uzun, B.; Ozsahin, I. The Systematic Review of Artificial Intelligence Applications in Breast Cancer Diagnosis. Diagnostics 2023, 13, 45.

AMA Style

Uzun Ozsahin D, Ikechukwu Emegano D, Uzun B, Ozsahin I. The Systematic Review of Artificial Intelligence Applications in Breast Cancer Diagnosis. Diagnostics. 2023; 13(1):45.

Chicago/Turabian Style

Uzun Ozsahin, Dilber, Declan Ikechukwu Emegano, Berna Uzun, and Ilker Ozsahin. 2023. "The Systematic Review of Artificial Intelligence Applications in Breast Cancer Diagnosis" Diagnostics 13, no. 1: 45.

Note that from the first issue of 2016, this journal uses article numbers instead of page numbers. See further details here.

Article Metrics

Back to TopTop