Abstract

Breast cancer threatens the lives of millions of women every year and causes death around the world. With the development of Artificial Intelligence (AI), its implementation in the diagnosis and treatment of breast cancer has become prominent. On account of this, there are numerous studies that focus on different aspects of the incorporation of AI in breast cancer diagnosis and treatment. This research paper aims to give an overview of the developments of AI in breast cancer diagnosis and chemotherapy treatment, as well as the incorporation of AI in apps and devices, and address the challenges and ethical issues of the implementation of AI. There have been many developments of AI in breast cancer diagnosis, including the usage of Explainable Artificial Intelligence (XAI) to understand the internal reasoning behind black box AI models, numerous papers that show that AI-supported breast cancer screenings are proficient and the implementation of deep learning in the risk predictions of breast cancer. The usage of AI also plays an essential role in chemotherapy in which AI can be an effective tool because of its ability to predict treatment responses and minimise toxicity while selecting optimal treatment plans. This paper also discusses the integration of AI to the real world and its use in improving diagnosis, surveillance and treatment by incorporating wearable detection gadgets. By enabling constant, personalised monitoring, a preventive element can be added to oncology therapy. AI can also be implemented in mobile apps which allows for the personalisation of treatment. In addition, there are also many challenges and ethical issues such as AI biases that must be dealt with. The incomprehension of black box AI models is still a prominent issue despite the development of XAI. Incorrect readings made by AI can lead to the same outcomes as if a radiologist were to incorrectly assess a mammogram, leading to the suffering of the patient. Additionally, trust in AI is needed by both patients and doctors. In short, this study provides an overview of the implementation of AI in the diagnosis and treatment of breast cancer, revealing not only its proficiency and reliability but also the issues that must be solved.

Introduction

What if artificial intelligence (AI) could not only detect breast cancer earlier than doctors but also guide personalised treatment plans tailored to each individual patient? Breast cancer is the most commonly diagnosed and life-threatening cancer among women worldwide, presenting a major public health challenge that affects millions every year. In the United States alone, approximately 316,950 women are expected to be diagnosed with invasive breast cancer in 2025 (Han et al., 2011). This disease impacts about 1 in 8 women during their lifetime, with many diagnosed at a localised stage when treatment tends to be more effective (Han et al., 2011). Early detection is crucial for improving survival rates and quality of life, highlighting the urgent need for enhanced diagnostic strategies that can also address disparities in rehabilitation globally.

This study aims to analyse the developments of AI in breast cancer diagnosis and treatment, propose the authors’ new idea about AI integration and examine the ethical perspectives of incorporating AI in breast cancer care (Carter et al., 2019). Artificial intelligence is revolutionising breast cancer by improving both detection and treatment. Beyond imaging, AI-powered wearables and devices, such as smart patches and health trackers, allow continuous monitoring of patient symptoms, activity levels and treatment responses. By analysing this longitudinal data, AI can identify early warning signs, track side effects and support timely interventions, creating a more proactive approach to breast cancer management (Royea et al., 2021). In chemotherapy, AI enables personalised treatment strategies by analysing patient data, tumour biology and treatment outcomes. Clinical decision support systems and computational oncology tools can predict chemotherapy responses, optimal drug sequencing and short-term risks, helping clinicians tailor interventions to individual patients. By integrating AI into both screening and treatment, breast cancer care moves toward a patient-centred approach where early detection and precision therapy work together to improve outcomes, minimise unnecessary treatments and enhance quality of life (Kurz & Axenie, 2020).

The Developments of AI in Breast Cancer Diagnosis (JJ)

The use of artificial intelligence in the detection and diagnosis of breast cancer has seen significant advancements in recent years. This has involved navigating key challenges, including the black box model, and evaluating proficiency, accuracy and overall potential in real-world clinical settings.

THE BLACK BOX MODEL

AI systems are trained on available datasets and generate desired outputs based on these inputs. However, the reasoning and decision-making process behind these outputs is often unclear (Bai et al., 2024). This is what is known as a black box model. Without being able to comprehend the internal evaluation process of an AI system, radiologists cannot interpret and understand diagnostic outputs.

Bai et al. (2024) state that Explainable Artificial Intelligence (XAI) can be used to uncover the internal reasoning behind decisions made by AI, increasing the transparency and comprehensibility of AI in breast cancer diagnosis (Bai et al., 2024). There are two main types of XAI models: intrinsic and post-hoc. While intrinsic models, also known as transparent AI models, are inherently explainable and comprehendible, post-hoc models are not and must be explained using an external method (Bai et al., 2024). Explanations of XAI can be divided into two types: local and global. Global explanations provide an overview of how the entire model operates, while local explanations interpret the reasoning for singular outputs (Bai et al., 2024). XAI models include: SHapley Additive exPlanations (SHAP), Gradient-weighted Class Activation Map (Grad-CAM) and Local Interpretable Model-agnostic Explanations (LIME) (Bai et al., 2024).

• SHAP is a post-hoc explainable model. It explains the internal evaluation process within the AI by computing the contributions to the reasoning process of each part of the input given to the black box AI model (Bai et al., 2024).
• Grad-CAM is also a post-hoc explainable model. It identifies regions of an image that impact the output of the AI, using global average pooling as a heatmap, and gives these regions values based on importance (Bai et al., 2024).
• LIME is also a post-hoc model. It creates locally accurate explanations of outputs by providing slightly altered inputs. It then obtains the new outputs from the AI and gives the adjusted inputs a value based on their impact. It then builds an interpretable model based on these inputs and their corresponding values (Bai et al., 2024).

Overall, XAI enables radiologists to understand the decision-making processes behind AI in breast cancer evaluation and diagnosis, increasing its transparency, reliability and understandability.

AI-SUPPORTED SCREENINGS

Despite the large quantity of data used to train them, AI models do not always provide accurate evaluations of mammograms. However, numerous studies have compared the proficiency of breast cancer screenings that were AI-supported to those that were not. Through these comparisons, it was concluded that the proficiency of AI-supported screening surpassed the proficiency of screenings without the support of AI (Chang et al., 2025; Watanabe et al., 2019; Rodríguez-Ruiz et al., 2019).

Moreover, Leibig et al. (2022) have proposed a new AI-supported approach to breast cancer screenings: a two-part decision-referral AI system. In this system, an AI evaluates a mammogram and refers the case for radiologist assessment if it is unsure or detects breast cancer. When Leibig et al. (2022) evaluated the proficiency of this system, it was revealed to be more accurate than the assessments made by either an AI or radiologist alone.

In short, these studies demonstrated that the implementation of AI increased the overall accuracy and effectiveness of breast cancer screenings.

DEEP LEARNING MODELS

The assessment of patient risk is crucial in the diagnostic process of cancer. The Tyrer-Cuzick Model, a popular and routinely used clinical practice, uses breast density as a factor to evaluate these risks (Yala et al., 2019). However, Yala et al. (2019) argue that the assessments made by radiologists are subjective and different possible evaluations could be made. To resolve this issue, they suggest the integration of deep learning (DL) models in evaluating the risks of breast cancer (Yala et al., 2019). In particular, when a hybrid deep learning model was compared with the Tyrer-Cuzick model, the hybrid DL model was more accurate (Yala et al., 2019), revealing the strong potential of deep learning in clinical settings.

AI in Treatment Response Prediction Models and Enhancing Chemosensitivity (HN)

AI models play an essential role in treatment outcome prediction and chemotherapy regimen enhancement. The following analysis explores the implementation of AI in chemotherapy applications. The discussion will outline current AI applications and assess its current status and future prospects. 

The accuracy of pharmacokinetic measurements, which indicates how drugs are handled in the body, is improved through integrative computational and mechanistic modelling that uses biological information to predict resistance based on tumour traits (Frieboes et al., 2009). The CHIMERA system uses machine learning to enhance personalised chemotherapy and surgery schedules through its mechanistic modelling approach which connects tumour growth with chemotherapy responses (Axenie & Kurz, 2020a). The PERFECTO model combines tumour growth functions and chemotherapy response data from clinical records to forecast treatment outcomes with strong patient-specific accuracy. These models help to determine intervention timing while improving treatment strategies which might eliminate the requirement for additional clinical trials or delayed patient care (Axenie & Kurz, 2020b). 

Through multi-omics data integration, AI technologies can be utilised to merge genomic and transcriptomic and clinical data into one comprehensive system for better treatment response predictions (Ruffalo et al., 2019). The network smoothing algorithm extracts new features from high-throughput data which boosts the capability to forecast aromatase inhibitor responses in estrogen receptor-positive (ER+) post-menopausal breast cancer tumours (Ruffalo et al., 2019). In addition, research has identified 17p12 deletion as a genomic change which serves as a tumour resistance predictor against neoadjuvant chemotherapy (Han et al., 2011). The epigenetic silencing of BRCA1 shows predictive value for increased platinum-based agent sensitivity. These biomarkers demonstrate their effectiveness as decision tools for chemotherapy selection (Stefansson et al., 2012). AI-based models can integrate these genomic and epigenetic biomarkers to enhance chemotherapy response predictions which will lead to better patient survival outcomes.

AI technology enables cancer therapy through two main functions which involve predicting treatment results and developing methods to increase drug sensitivity. The absence of Fanconi anemia/BRCA (FA/BRCA) pathway leads tumours to become more responsive to DNA-damaging drugs (Mulligan et al., 2013). AI tools have enabled scientists to build a 44-gene microarray system that successfully detects tumours with DDR deficiency. The assessment uses AI to forecast pathological complete response (pCR) of neoadjuvant DNA-damaging chemotherapy which leads to enhanced relapse-free survival during adjuvant treatment (Mulligan et al., 2013). Additionally, deep learning algorithms have been developed to analyse pathological complete response (pCR) outcomes from neoadjuvant chemotherapy treatments in breast cancer patients (Qu et al., 2020). Scientists are conducting studies to evaluate chemosensitisers, including oxygen-carbon nanotubes, for boosting the effectiveness of paclitaxel drugs against breast cancer cells. Traditional Chinese medicine formulas undergo network pharmacology and bioinformatics analysis as researchers explore these approaches. This research demonstrates how AI helps understand pathway influences that could enhance chemosensitivity (Wang et al., 2018). 

AI demonstrates high effectiveness in optimising chemotherapy through its ability to predict treatment response and minimise toxicity while selecting optimal treatment plans. To achieve the maximum potential of AI in personalised cancer therapy, the future will depend on continuous utilisation of computational models, multi-omics and bioinformatics approaches.

AI-Enabled Wearables for Early Detection, Patient Monitoring and Treatment Support (JB)

AI-powered wearables and devices are transforming the diagnosis, surveillance and treatment of breast cancer, ranging from wearable ultrasound patches and thermal imaging to real-time biomarker sensors and side-effect prediction algorithms. 

EARLY DETECTION

A new wearable cancer detection technology has been released in the U.S. (Du et al., 2023), which is a wearable ultrasound honeycomb-structured, flexible patch worn in the bra. It is capable of scanning breast tissue throughout the day and can transmit data to a smartphone for analysis (Du et al., 2023). Research has shown  that this device can detect cysts as small as 0.3 cm, offering frequent, user-friendly monitoring of breast tissue. This device potentially transforms breast cancer detection, promising to identify tumours that appear between routine mammograms. These account for 20-30% of cases and are often more aggressive (Du et al., 2023).

Similarly, the Cyrcadia Breast Monitor (iTBra) uses thermal sensors embedded in patches worn under a bra. By capturing circadian temperature fluctuations from breast tissue, the iTBra uses machine learning to distinguish malignant from benign lesions based on clinical data, with 78% accuracy, 83.6% sensitivity and 71.5% specificity (Royea et al., 2021). Even though it is still under validation, which requires larger-scale trials and regulatory approval, pilot trials involving over 500 participants show promise – particularly for dense breast tissue where mammography is less effective (Royea et al., 2021; El Camino Health, 2014).

Additionally, AI-powered thermal imaging provides a non-invasive approach to breast screening. For instance, Niramai’s Thermalytix system is a portable, touchless breast cancer detection technology which uses AI to interpret thermographic patterns. This offers a better option for younger women (under 45) and in settings where mammography is less accessible. This helps promote earlier detection and greater equity in underserved populations (Niramai, 2025).

PATIENT MONITORING

There is significant promise of AI-enabled wearable technology in oncology (Birla et al., 2025). Continuous monitoring of vitals and physical activity allows continuous patient health assessment and early intervention where needed. AI-enabled wearables such as fitness trackers (e.g., FitBit) and physiological sensors provide continuous data on patients’ activity, sleep and vital signs. Studies show that lower physical activity levels, detectable through these wearables, corresponds with reduced treatment tolerance and poorer prognosis (Birla et al., 2025). AI models analysing this data can identify trends and predict outcomes, improving prognosis and enabling proactive care adjustments.

There are different technologies that also include wearable biosensors capable of monitoring biomarker changes. For instance, flexible patch sensors that measure inflammatory markers like C-reactive protein (CRP) via sweat have been studied. AI algorithms interpret the electrochemical signals to track treatment response and detect complications earlier (Ghaffari et al., 2021). Similarly, strain sensors embedded in films detect electrical impedance changes tied to tumour volume shifts, allowing AI to monitor tumour growth or regression in real time (Abramson et al., 2023).

TREATMENT SUPPORT

AI is not only assisting in detection and monitoring, but also in predicting post-treatment complications. A model trained on over 6,300 patient cases can predict lymphedema, which is common after breast surgery and radiation, with an overall accuracy of 73.4%, 81.6% sensitivity and 72.9% specificity. Though it is not yet embedded in wearable hardware, these prediction tools can be combined with different wearables (e.g., tracking arm circumference or activity levels) to enable earlier intervention (Ozmen & Schwarz, 2024). 

Additionally, clinical decision support systems (CDSS) using AI are reshaping how oncologists select and adapt treatments. In one study involving nearly 2,000 cases, oncologist agreement with treatment recommendations increased from 56% to 61% when guided by CDSS, especially for hormone receptor-positive and stage four breast cancers (Akkus et al., 2025). However, excessive confidence in AI can lead to overeliance and reduce diagnostic accuracy, stressing the need for explainable and calibrated AI interfaces (Rosbach et al., 2024). 

By enabling personalised monitoring, AI-enabled wearables add a preventive element to oncology therapy. When combined with explainable AI-assisted decision tools, the result is an improved, responsive and patient-centred management of breast cancer. As science and clinical trials go further, this technology-based approach holds a positive direction for improved outcomes and better quality of life for patients.

Using AI to Create Personalised Lifestyle Changes in Breast Cancer Treatment (GR)

Artificial intelligence is changing the way breast cancer diagnosis is taking place in clinical settings. One promising idea is the usage of apps and wearable devices that track the individual treatment process for patients. These devices can include watches, ipads and patches which track symptoms, medications, physical activity and mammogram results over time.

Mobile apps offer patients the opportunity to stay involved in their own health instead of relying solely on healthcare professionals. Unlike traditional monitoring apps, these enable patients to actively track symptoms, side effects and lifestyle factors such as exercise and diet, providing tailored feedback. For example, AI-assisted apps can analyse how patients are feeling, such as fatigue, pain and mood, to predict symptoms and provide reminders and recommendations (Zhou et al., 2024). Mogharab et al. (2025) have discovered that mobile apps for breast cancer treatment have significantly helped patients recover in several ways, including supporting weight management, enhancing quality of life, reducing stress and anxiety, improving coping with symptoms and increasing the overall wellbeing of patients. These apps are able to provide information to doctors and transforms the patient into an active partner in their care, helping to bridge gaps between hospital and home treatment. 

Lifestyle and nutrition factors play a significant role in influencing breast cancer prognosis, risk of recurrence and treatment tolerance (Hamer & Warner, 2017). Through AI-driven diet and lifestyle management, these apps generate dietary plans and provide adaptive exercise recommendations. A recent review highlighted that AI-powered nutrition apps improved the dietary quality and conditions of breast cancer survivors, supporting weight management and metabolic health (Pan et al., 2024). Integration of AI chatbots in apps and devices can offer continuous patient support by answering questions about various concerns, particularly in medications, diet, symptoms and side effects, increasing overall patient engagement and reducing stress and anxiety amongst breast cancer patients (Shaban et al., 2025). Merging lifestyle coaching with capturing health data in wearable devices can create tools that create a holistic system for cancer survivors based on personal requirements. 

Wearable devices with built-in AI can go above and beyond detection and tracking symptoms: they can create personalised and patient-centred exercise plans critical in efficient recovery (Orhan et al., 2025). AI-enabled fitness wearables collect continuous movement and exercise data to assess patient responses, monitoring daily step counts and improving activity levels and quality of life in survivors (Birla et al., 2024). Depending on the patient’s tolerance, these AI wearable devices can help build routines and recommend exercises for patients. If these devices detect low activity and increased fatigue, the AI can adjust rehabilitation in real time without the assistance of doctors. This personalised approach enhances safety and reduces any arising complications such as injuries. Moreover, rehabilitation through wearable devices with AI can reduce the burden on healthcare workers. 

Research has proved that AI can also improve personalisation in chemotherapy by suggesting amended doses and anticipating tumour responses (Blasiak et al., 2025). Instead of using the one-size-fits-all approach, doctors can use AI insights to design accurate treatment plans that are safer for each individual. Additionally, AI also has the capacity to identify patients with higher risk factors for experiencing complications or side effects like reduced tolerance to certain drugs, allowing care teams the opportunity to make meaningful changes early (Manikis et al., 2023).

By utilising AI sources in this manner, cancer care can become less reactive and more efficient. These models shift the subject from treating problems after they appear to designing care plans which can alter and adapt to certain patients’ individual needs. This strengthens the overall personalisation of breast cancer treatment. 

Ethics of Using AI in Diagnosis and Treatment (AD)

AI is an innovative new technology that is being explored as a path to improve the way that patients with breast cancer are cared for. These AI systems can process thousands of images and patient records faster than any doctor while maintaining accuracy, reducing the workload of doctors involved. It is difficult to look beyond the benefits of this technology, however there are some challenges and ethical concerns that must be tackled. This section will explain the most important of these issues, as well as possible solutions.

For breast cancer detection, AI is trained on existing data such as mammograms. If the training data is biased towards one group of people, results may be less accurate for individuals who fall outside of that group; for instance, if the majority of images are from white, middle-aged women, then the AI output may be less accurate for women of other ethnicities or ages (Miyawaki et al., 2025). This algorithmic bias is found in all types of AI (Corti et al., 2022), however due to the sensitive nature of people’s health, it is very important that this type of bias is minimised. Many breast cancer AI systems are developed in high-income countries that have Caucasian populations (Miyawaki et al., 2025). As a result, they may make mistakes when diagnosing other ethnic groups, such as Black or Asian people (Nguyen et al., 2025). This creates a gap in the quality of care between Caucasian women and other ethnic groups. Unless datasets are deliberately made more diverse, AI could reinforce, rather than reduce, existing gaps (Goh et al., 2025). A solution to this could include sharing data between hospitals from different areas or ensuring that there is a suitable spread in different types of diversity (Goh et al., 2025).

As previously mentioned, AI often acts as a “black box”, meaning it comes to conclusions without clearly explaining how that decision was reached (Carter et al., 2019). The process AI uses to create answers involves converting an input into an array of numbers, which is then passed through a series of layers made up of artificial neurons (Ansari et al., 2025). Each neuron adjusts the values in the array using weights and biases, eventually producing another set of numbers at the end of the neural network that can be interpreted as an output (Bai et al., 2024). The AI is trained by being given an input array and an expected output array, and through trial and error, it adjusts its weights and biases to connect the two (Xu et al., 2020). This creates a situation where the AI does not have reasoning behind why it made a decision, only that it has mathematically learnt a pattern (Rosbach et al., 2025). Doctors may not want to trust an answer they cannot verify, leading to an issue where even the most accurate AI has no merit in clinical settings (Pesapane et al., 2024). To solve this, scientists are developing explainable AI (XAI). These systems show which parts of an image influenced the decision, or provide reasons that a human can interpret (Ansari et al., 2025). For example, an XAI model might highlight the area of the breast image where it spotted abnormal tissue. This would help verify whether the AI is looking at the right things or whether it is coming to its conclusions through inaccurate readings.

While these AI systems iare often correct, there have been (and will be) cases where mistakes are made. If an AI makes an error, the consequences are the same as if a human made the same error. The challenge here lies in pinpointing accountability; the blame could lie with the developer, the hospital or the doctor. Cestonaro, et al. explored medical liability and found no clear conclusion on how to handle AI-related mistakes. Typically, doctors are currently held responsible, even though they cannot always verify the logic behind an AI decision (Goh et al., 2025), yet developers also need to be held accountable. Therefore, malpractice law should be appropriately updated to state clearly what to do in the event of an AI mistake (Carter et al., 2025).

Finally, for AI to succeed in breast cancer care, patients must trust it. Pesapane et al. (2024) show that most patients prefer AI to assist rather than replace human doctors; people feel more comfortable when a radiologist interprets their mammogram and uses AI as a second opinion. Additionally, patients might have reservations about trusting AI with highly sensitive data, considering it is vulnerable to cyber attacks. Furthermore, doctors may not welcome AI, with concerns regarding overreliance, anchor bias and automation bias. To allow the seamless implementation of AI in healthcare, there should be clear communication, education and consent present for all involved.

Conclusion

This paper gave an overview of the developments of AI in the diagnosis and treatment of breast cancer through the demonstration of the improvements of AI in breast cancer diagnosis, its potential in chemotherapy and the benefits of its implementation in apps and devices to allow personal care and treatment. In addition, it explored the various challenges and ethical issues that occur due to the nature of AI, as well as potential solutions.

Firstly, the paper: explored the use of XAI to enable the interpretation of internal decision-making processes behind black box AI models; reviewed numerous papers revealing the proficiency of AI when supporting radiologists in breast cancer screenings; and emphasised the potential of deep learning in predicting the risks of breast cancer in clinical settings. Secondly, it highlighted AI’s potential in chemotherapy due to its ability to be an effective decision-making tool that predicts treatment responses, minimises toxicity and selects optimal treatment plans. Thirdly, it explained the integration of AI in wearable gadgets and other devices for the diagnosis, surveillance and treatment of breast cancer, personalising care and encouraging patient involvement. Lastly, it analysed the many challenges and ethical issues of integrating AI into the diagnosis and treatment of breast cancer – including AI biases, black box models, accountability and lack of trust – and proposed potential future solutions.

Overall, this paper has given an overview of the development of AI in the diagnosis and treatment of breast cancer, its effectiveness, proficiency and practicality, and its additional challenges and ethical issues.

Abramson, A., Chan, C.T., Khan, Y., Mermin-Bunnell, A., Matsuhisa, N., Fong, R., Shad, R., Hiesinger, W., Mallick, P., Gambhir, S.S. & Bao, Z. (2022). A flexible electronic strain sensor for the real-time monitoring of tumor regression. Science Advances, 8(37).

Ansari, Z.A., Tripathi, M.M. & Ahmed, R. (2025). The role of explainable AI in enhancing breast cancer diagnosis using machine learning and deep learning models, Discover Artificial Intelligence, 5(75).

Axenie, C. & Kurz, D. (2020a). Combining Mechanistic Models and Machine Learning for Personalized Chemotherapy and Surgery Sequencing in Breast Cancer, bioRxiv [preprint], 140756.

Axenie, C. & Kurz, D. (2020b). PERFECTO: Prediction of Extended Response and Growth Functions for Estimating Chemotherapy Outcomes in Breast Cancer, bioRxiv [preprint], 424759.

Bai, S., Nasir, S., Kahn, R.A., Arif, S., Meyer, A. & Konik, H. (2024). Breast cancer diagnosis: A comprehensive exploration of explainable artificial intelligence (XAI) techniques, arXiv, 2406.00532. 

Birla, M., Rajan, Roy, P.G., Gupta, I. & Malik, P.S. (2024). Integrating Artificial Intelligence-Driven Wearable Technology in Oncology Decision-Making: A Narrative Review, Oncology, 103(1), pp. 69-81.

Björnsson, B., Borrebaeck, C., Elander, N., Gasslander, T., et al. (2020). Digital twins to personalize medicine, Genome Medicine, 12(4).

Blasiak, A., Tan, W.S. & Ho, D. (2025). Personalized dose selection platform for patients with solid tumors in the PRECISE CURATE.AI feasibility trial, NPJ Precision Oncology, 9(1), 49.

Carter, S., Rogers, W., Win, K.T., Frazer, H., Richards, B. & Houssami, N. (2019). The ethical, legal and social implications of using artificial intelligence systems in breast cancer care, The Breast, 49, pp.25–32.

Cestonaro, C., Delicati, A., Marcante, B., Caenazzo, L. & Tozzo, P. (2023). Defining medical liability when artificial intelligence is applied on diagnostic algorithms: a systematic review, Frontiers in Medicine, 10(3389).

Chang, Y.W., Ryu, J.K., An, J.K., Choi, N., Park, Y.M., Ko, K.H. & Han, K. (2025). Artificial intelligence for breast cancer screening in mammography (AI-STREAM): preliminary analysis of a prospective multicenter cohort study, Nature Communications, 16(2248).

Corti, C., Cobanaj, M., Marian, F., Dee, E.C., Lloyd, M.R., Marcu, S., Dombrovschi, A., Biondetti, G.P., Batalini, F., Celi, L.A. & Curigliano, G. (2022). Artificial intelligence for prediction of treatment outcomes in breast cancer: Systematic review of design, reporting standards, and bias, Cancer Treatment Reviews, 108(102410). 

Du, W., Lin, Z., Suh, E., Lin, D., Marcus, C., Ozkan, L., Ahuja, A.T., Fernandez, S., Shuvo, I.I., Sadat, D., Liu, W., Li, F., Chandrakasan, A.P., Ozmen, T. & Dagdeviren, C. (2023). Conformable ultrasound breast patch for deep tissue scanning and imaging, Science Advances, 9(30).

El Camino Health (2014). Cyrcadia Health enters clinical trials for wearable breast cancer detection device, El Camino Health [online]. <https://www.elcaminohealth.org/newsroom/cyrcadia-health-enters-clinical-trials-wearable-breast-cancer-detection-device>

Frieboes, H., Edgerton, M., Fruehauf, J., Rose, F., Worrall, L.K., Gatenby, R., Ferrari, M. & Cristini, V. (2009). Prediction of drug response in breast cancer using integrative experimental/computational modeling, Cancer Research, 69(10), pp. 4484-4492.

Ghaffari, R., Rogers, J. A. & Ray, T. R. (2021). Recent progress, challenges, and opportunities for wearable biochemical sensors for sweat analysis. Sensors and Actuators B Chemical, 332(129447). 

Goh, S., Goh, R.S.J., Chong, B., Ng, Q.X., Koh, G.C.H., Ngiam, K.Y. & Hartman, M. (2025). Challenges in implementing artificial intelligence in breast cancer screening programs: systematic review and framework for safe adoption, Journal of Medical Internet Research, 27, e62941.

Hakim, A. & Awale, R.N. (2020). Thermal Imaging — An Emerging Modality for Breast Cancer Detection: A Comprehensive Review, Journal of Medical Systems, 44(8), 136.

Hamer, J. & Warner, E. (2017). Lifestyle modifications for patients with breast cancer to reduce recurrence and improve prognosis: a review, Canadian Medical Association Journal, 189(7), E455–E461.

Han, W., Woo, J., Jeon, Y., Yang, S.-J., Cho, J., Ko, E., Kim, T.-Y., Im, S., Oh, D., Park, I., Hwang, K., Moon, H., Yang, K.-S. & Noh, D. (2011). 17p12 deletion in breast cancer predicts resistance to neoadjuvant chemotherapy, Experimental and Therapeutic Medicine, 2(5), pp. 799-804.

Hiesinger, W., Mallick, P., Gambhir, S.S. & Bao, Z. (2022). A flexible electronic strain sensor for the real-time monitoring of tumor regression, Science Advances, 8(37).

Leibig, C., Brehmer, M., Bunk, S., Byng, D., Pinker, K. & Umutlu, L. (2022). Combining the strengths of radiologists and AI for breast cancer screening: A retrospective analysis, The Lancet Digital Health, 4(7), E507–E519.

Manikis, G.C., Souliotis, K. & Bamia, C. (2023). Personalized risk analysis to improve psychological outcomes in breast cancer survivors: machine learning models from the BOUNCE Back project, Journal of Medical Internet Research, 25, e45678. 

McKinney, S.M., Sieniek, M., Godbole, V., et al. (2020). International evaluation of an AI system for breast cancer screening, Nature, 577, pp. 89–94. 

Miyawaki, I.A., Banerjee, I., Batalini, F., Campello Jorge, C.A., Celi, L.A., Cobanaj, M., Dee, E.C., Gichoya, J.W., Kaffey, Z., Lloyd, M.R., McCullum, L., Ranganathan, S., & Corti, C. (2025). Global disparities in artificial intelligence-based mammogram interpretation for breast cancer: A scientometric analysis of representation, trends, and equity, European Journal of Cancer, 220(115394). 

Mogharab, G., Mousavi, M., Haseli, S., Gharaati, F., Salehi, T., Mohammadi, A. & Zarghami, M. (2025). Expert evaluation of mobile health apps for breast cancer: a scoping review, BMC Health Services Research, 25(692).

Mulligan, J., Hill, L., Deharo, S., Irwin, G., et al. (2014). Identification and Validation of an Anthracycline/Cyclophosphamide–Based Chemotherapy Response Assay in Breast Cancer, Journal of the National Cancer Institute, 106. 

Nguyen, D.L., Ren, Y., Jones, T.M., Thomas, S.M., Lo, J.Y. & Grimm, L.J. (2024). Patient characteristics impact performance of AI algorithm analyzing true-negative screening digital breast tomosynthesis examinations, Radiology, 311(2), e232286. 

Niramai (2025). Thermalytix: AI-Powered Breast Cancer Screening Test, Niramai [online]. <https://niramai.com/about/thermalytix/>

Orhan, S. A. & Özşaker, E., (2025). Lymphedema, fatigue, and quality of life in breast cancer survivors with axillary lymph node dissection, Lymphatic Research and Biology, 23(2), pp.107–114. 

Ormel, H.L., van der Schoot, G.G.F., Sluiter, W.J., Jalving, M., Gietema, J.A. & Walenkamp, A.M.E. (2022). Physical activity monitoring using wearable devices in breast cancer patients: A randomized controlled trial, Breast Cancer Research and Treatment, 194(3), pp. 593–604. 

Ozmen, B.B. & Schwarz, G. (2024). Predicting Breast Cancer Related Lymphedema after Immediate Lymphatic Reconstruction: An Artificial Intelligence Approach with Synthetic Data, Plastic and Reconstructive Surgery Global Open, 12(1S1), p. 62.

Pan, L.C., Wright, E. & Lee, Y.Y. (2024). Applications of AI in digital nutrition for cancer survivors: weight prediction and personalized diet plans, Digital Health, 10, pp. 1–12. 

Pesapane, F., Giambersio, E., Capetti, B., Monzani, D., Grasso, R., Nicosia, L., Rotili, A., Sorce, A., Meneghetti, L., Carriero, S., Santicchia, S., Carrafiello, G., Pravettoni, G. & Cassano, E. (2024). Patients’ Perceptions and Attitudes to the Use of Artificial Intelligence in Breast Cancer Diagnosis: A Narrative Review, Life, 14(4), 454.

Ruffalo, M., Thomas, R., Chen, J., Lee, A.V., Oesterreich, S. & Bar-Joseph, Z. (2019). Network-guided prediction of aromatase inhibitor response in breast cancer, PLoS Computational Biology, 15.

Rodríguez-Ruiz, A., Krupinski, E., Mordang, J.J., Schilling, K., Heywang-Köbrunner, S.H., Sechopoulos, I. & Mann, R.M. (2019). Detection of Breast Cancer with Mammography: Effect of an Artificial Intelligence Support System, Radiology, 290(2), pp. 1–10. 

Rosbach, E., Ganz, J., Ammeling, J., Riener, A. & Aubreville, M. (2025). Automation bias in AI-assisted medical decision-making under time pressure in computational pathology, Informatik Aktuell, pp. 129–134.

Royea, R., Buckman, K.J., Benardis, M., Holmes, J., Fletcher, R.L., Eyk, N., Rajendra Acharya, U. & Ellenhorn, J.D.I. (2021). An introduction to the Cyrcadia Breast Monitor: A wearable breast health monitoring device, Computer Methods and Programs in Biomedicine, 197, 105758. 

Shaban, M., Khan, F. & Al-Mutairi, H. (2025). Empowering breast cancer patients through AI chatbots: improving knowledge and engagement in oncology care, BMC Nurs, 24(1), 994.

Stefansson, O.A., Villanueva, A., Vidal, A., Martí, L. & Esteller, M. (2012). BRCA1 epigenetic inactivation predicts sensitivity to platinum-based chemotherapy in breast and ovarian cancer, Epigenetics, 7(11), pp. 1225-1229.

Wang, J., Li, Y., Luo, X., Wang, L. & Xu, L. (2023). The effectiveness of mobile health (mHealth) interventions on improving adherence and outcomes among breast cancer patients: A systematic review, Supportive Care in Cancer, 31(7), pp. 3335–3348. 

Wang, N., Yang, B., Zhang, X., Wang, S., Zheng, Y., Li, X., Liu, S.-S., Pan, H., Li, Y., Huang, Z., Zhang, F. & Wang, Z. (2018). Network pharmacology-based validation of Caveolin-1 as a key mediator of Ai Du Qing inhibition of drug resistance in breast cancer, Frontiers in Pharmacology, 9.

Watanabe, A.T., Lim, V., Vu, H.X., Chim, R., Weise, E., Liu, J., Bradley, W.G. & Comstock, C.E. (2019). ‘Improved cancer detection using artificial intelligence: a retrospective evaluation of missed cancers on mammography’, Journal of Digital Imaging, 32(4), pp. 625–637.

Xu, F., Sepúlveda, M.-J., Jiang, Z., Wang, H., Li, J., Liu, Z., Yin, Y., Roebuck, M.C., Shortliffe, E.H., Yan, M., Song, Y., Geng, C., Tang, J., Purcell Jackson, G., Preininger, A.M. & Rhee, K. (2020). Effect of an artificial intelligence clinical decision support system on treatment decisions for complex breast cancer, JCO Clinical Cancer Informatics, 4, pp. 824–838.

Yala, A., Lehman, C., Schuster, T., Portnoi, T. & Barzilay, R. (2019). A deep learning mammography-based model for improved breast cancer risk prediction, Radiology, 292, pp. 60–66.

Yu-Hong, Q., Zhu, H.-T., Cao, K., Li, X., Ye, M. & Sun, Y. (2020). Prediction of pathological complete response to neoadjuvant chemotherapy in breast cancer using a deep learning (DL) method, Thoracic Cancer, 11(3), pp. 651-658.

Zhou, Y., Chen, Y., Xu, X., Zhao, W., Xu, Y. & Sun, X. (2024). Effectiveness of a mobile app with intelligent design on symptoms and quality of life among young breast cancer survivors: A randomized controlled trial, PLOS ONE, 19(5), e0301897.