1 Introduction

The SDGs, established by the United Nations, serve as a framework for creating a more equitable and sustainable future for everyone. Among the SDGs, those addressing health and well-being (SDG 3) (Fig. 1), gender parity (SDG 5), and reducing disparities (SDG 10) are particularly relevant to female cancer patients [1, 2]. Cancer continues to be a major contributor to mortality among women globally, necessitating innovative approaches to improve outcomes [3, 4]. Addressing the financial and societal aspects of cancer care is also crucial, aligning with SDGs focused on poverty eradication (SDG 1), education (SDG 4), sustainable economic development (SDG 8), and technological and industrial advancement (SDG 9) [5, 6]. Furthermore, the global nature of cancer care requires international collaboration and partnership, emphasizing the importance of global partnerships (SDG 17) [7, 8].

Fig. 1
figure 1

SDG 3, SDG 5 and SDG 10 associated with other SDGs related to health. The illustrations were created with BioRender.com

Full size image

CI encompasses a variety of advanced computational methods including artificial intelligence (AI) [9], machine learning (ML) [10], neural networks [11], and data mining [12], which have shown significant promise in revolutionizing cancer healthcare. These technologies offer potential solutions for early diagnosis [13, 14], personalized treatment [15], and efficient management [16] of cancer, thereby contributing to the achievement of relevant SDGs [17].

This narrative review aims to offer a comprehensive overview of how CI methods are applied to address the challenges faced by female cancer patients and to achieve the SDGs. It will delve into specific applications of CI in oncology, discuss the benefits and limitations of these technologies, and suggest pathways for future research. The review emphasizes the importance of interdisciplinary collaboration and innovation in fully leveraging CI's potential to improve female cancer care and promote sustainable development.

2 Computational intelligence techniques in oncology

2.1 Machine learning and predictive analytics

ML algorithms have become integral in cancer research, offering robust methods for predictive modeling and early diagnosis. These algorithms can analyze extensive datasets to detect patterns and make predictions that may elude human experts [18].

Support vector machines (SVMs): SVMs have demonstrated high accuracy in classifying cancerous tissues in breast cancer patients. Studies by Cruz and Wishart, for instance, have shown that SVMs effectively distinguish between malignant and benign tumors using mammographic data, supporting SDG 3 [19].

In addition, Chen et al. proposed and implemented a swarm intelligence-based SVM classifier (PSO_SVM) for breast cancer diagnosis. Their approach required a smaller set of support vectors for training while achieving a high predictive accuracy of 99.3% through tenfold cross-validation. Notably, they identified five informative features that could provide valuable insights into breast cancer, aiding physicians in making precise diagnostic decisions [20].

Moreover, Bilal et al. recently demonstrated that hybrid SVM models, such as the improved quantum-inspired binary Grey Wolf Optimizer (IQI-BGWO-SVM) technique, outperform contemporary classification approaches on the Mammographic Image Analysis Society (MIAS) dataset. In tenfold cross-validation, this model exhibited significant accuracy, sensitivity, and specificity [21]. These advancements underscore the potential of new SVM techniques in improving breast cancer diagnosis and treatment, further aligning with SDG 3 (Fig. 2).

Fig. 2
figure 2

SDG 3 associated with other SDGs related to health

Full size image

Decision trees and random forests: These models have been employed to predict breast cancer (BC) recurrence. For instance, a study by Delen, Walker, and Kadam demonstrated that decision trees could accurately predict cancer survival outcomes by analyzing patient demographics and clinical data, contributing to SDG 3 [22].

Similarly, Khan et al. developed a hybrid prognostic scheme based on fuzzy decision trees to predict BC survivability. They found this method to be an efficient alternative to traditional classifiers, offering greater robustness and balance, with the potential for significant performance enhancement [23].

Furthermore, Jin et al. developed and evaluated a machine learning model based on random forest for anticipating outcomes among BC patients experiencing a suboptimal response to neoadjuvant chemotherapy. This model showed good performance in predicting outcomes for BC patients with stable or progressive disease post-neoadjuvant chemotherapy, potentially aiding in detecting tumor recurrence and further supporting SDG 3 [24].

Neural networks: Deep learning, a subset of neural networks, has been particularly successful in image recognition tasks related to cancer detection. For example, Esteva et al. utilized convolutional neural networks (CNNs) to achieve dermatologist-level accuracy in classifying skin cancer from images, promoting SDG 3 [25].

In another study, Teng et al. designed a neural network robotic system to assess the impact of music therapy on mitigating adverse reactions following chemotherapy in BC patients. This robotic system-assisted music therapy showed a significant improvement, increasing the cure rate by 7.84% [26].

Additionally, Zhao et al. developed a diagnostic model for endometrial cancer (EC) employing an artificial neural network. This model highlighted the increased expression of certain up-regulated genes, such as matrix metalloproteinases MMP12 and MMP9, and decreased expression of down-regulated oncogenes, facilitating the development of diagnosis strategies for EC [27].

Hutt et al. established a neural network computer model with 98.6% accuracy in predicting changes in a patient’s overall cancer risk. This model aims to guide the need for testing, predict diagnoses, and advise on preventative actions, effectively minimizing unnecessary invasive procedures for EC patients. This tool can be valuable resource for physicians, complementing other indicators, to strengthen preventive measures before potential EC development [28].

2.2 Artificial intelligence in personalized medicine

AI technologies are spearheading the move towards personalized medicine, shaping treatments to match the unique characteristics of each patient. This approach holds significant promise in cancer care, where variability in genetic profiles and tumor behavior necessitates customized treatment plans [29].

Genomic analysis: AI algorithms analyze genomic data to identify mutations and alterations associated with specific types of cancer. For instance, Cheng et al. reported that using AI to interpret next-generation sequencing (NGS) data has facilitated the identification of actionable mutations in breast cancer, aiding in the realization of SDG 3 [30].

Similarly, Zhang et al.'s research identified several novel genetic alterations and pathways in ovarian cancer (OC) through NGS. This provided valuable insights into the molecular mechanisms of OC and contributed to early clinical diagnosis [31].

Drug discovery: AI models are revolutionizing drug discovery by predicting the efficacy of new compounds. Companies like Atomwise use the AtomNet algorithm to screen billions of chemical compounds, identifying potential drug candidates for cancer treatment, aligning with SDG 3. For instance, Sobhani et al. screened 10 million compounds targeting cytotoxic T-lymphocyte-associated antigen 4 (CTLA-4) using virtual screening techniques. Several of these compounds demonstrated the ability to inhibit tumor growth in both preventive and therapeutic contexts in syngeneic and CTLA-4-humanized mouse models. These findings underscore the potential of AI-driven approaches in designing small molecules for cancer immunotherapy [32].

In a similar vein, Raies et al. introduced Drugnome AI, a tool designed to assess the druggability potential of every protein-coding gene in the human exome. This tool integrates data at the gene level from 15 different sources to produce exome-wide predictions, leveraging labeled sets of known drug targets. Notably, features from protein–protein interaction networks emerged as the top predictors. Employing a stochastic semi-supervised machine learning method, DrugnomeAI displays significant promise in identifying druggable targets for oncology [33].

Furthermore, Wang et al. developed KG4SL, a novel model utilizing a graph neural network (GNN) that incorporates messaging from knowledge graphs (KG) to enhance its predictions. Experimental results demonstrated a marked improvement in synthetic lethality predictions when the KG was integrated into the GNN [34].

Predictive analytics for treatment response: AI tools empower oncologists to predict patient responses to therapies, aiding in the selection of optimal treatments. Integration of AI with electronic health records (EHRs) enables continuous monitoring and adjustment of treatment plans based on real-time patient data, supporting SDG 3 and SDG 9 [35].

For instance, Moghadas-Dastjerdi et al. developed a predictive model for chemotherapy response in BC patients using second derivative texture analysis of quantitative CT images combined with machine learning. This model categorizes patients into responders and non-responders before treatment, significantly advancing the precision oncology paradigm for BC [36].

Similarly, Sannachi et al. demonstrated the integration of quantitative ultrasound, texture analysis, texture derivative, and molecular subtype techniques to forecast treatment response in BC patients before neoadjuvant chemotherapy. Their models effectively differentiate between responders and non-responders, correlating closely with histopathology and post-surgery outcomes. This approach suggests that integrating quantitative ultrasound-texture biomarkers and molecular subtypes as early treatment response indicators could optimize personalized treatment regimens [37].

Moreover, Huang et al. employed SVM-based algorithms to predict the collective response of 273 ovarian cancer patients to various chemotherapeutic agents with 80% accuracy. These algorithms utilize patient-specific gene expression profiles, providing tailored predictions of drug responses [9].

2.3 Data mining and big data analytics

The explosion of healthcare data, including clinical records, imaging data, and genomic information, presents an opportunity for data mining techniques to uncover insights that can improve cancer care [38].

Pattern recognition: Data mining algorithms play a crucial role in identifying patterns and correlations within large datasets, aiding in the understanding of disease progression and treatment outcomes. For example, these techniques have significantly advanced mammography data analysis, leading to enhanced breast cancer screening protocols and supporting SDG 3.

In their study, Diz et al. utilized data mining to assist oncologists in classifying and diagnosing breast cancer. They utilized a database of 410 images containing microcalcifications, masses, and normal tissue samples. Their approach included applying the gray level co-occurrence matrix and gray level run length matrix for feature extraction, alongside various data mining classifiers. The findings demonstrated high positive predictive values (70%) and strong accuracy (> 65%) in distinguishing mammographic findings and classifying the Breast Imaging Reporting and Data System (BI-RADS) scale [39].

Lee et al. introduced CBIS-DDSM (Curated Breast Imaging Subset of DDSM), an updated version of the Digital Database for Screening Mammography (DDSM) standardized for evaluating CADx and CADe systems in mammography. This dataset includes decompressed images, curated data selected by trained mammographers, and updated mass segmentation with bounding boxes, alongside pathologic diagnoses for training. CBIS-DDSM, formatted similarly to modern computer vision datasets, encompasses 753 cases of calcifications and 891 cases of masses, facilitating advanced analysis of decision support systems [40].

In their comprehensive review, Oskouei et al. examined diverse research on breast cancer diagnosis, treatment, and prognosis. They emphasized that many studies focus extensively on comparing the accuracy of different data mining algorithms and techniques in diagnosing breast cancer [12].

Integration of multimodal data: Integration of multimodal data, which combines clinical, imaging, and genomic information, provides a comprehensive patient view, enhancing both the accuracy of diagnoses and the personalization of treatment plans. This approach has been shown to significantly boost the predictive capabilities of computational intelligence models in oncology, thereby contributing to advancements in both SDG 3 and SDG 9 [41, 42].

Boehm et al. conducted a study involving 444 patients predominantly diagnosed with late-stage high-grade serous ovarian cancer. They identified quantitative features such as tumor nuclear size from hematoxylin and eosin staining, and omental texture from contrast-enhanced computed tomography, both of which were found to correlate with prognosis. Their findings underscored the complementary prognostic insights provided by these features alongside clinicogenomic factors. Integrating histopathological, radiological, and clinicogenomic machine-learning models presented a promising approach for refining risk stratification among cancer patients through multimodal data fusion [43].

In their survey of current practices in breast cancer detection and prognosis, Archana Mathur et al. explored the evolution of AI-based approaches. They highlighted a significant shift from uni-modal to multi-modal data integration for improved detection efficacy, aligning closely with clinical observations [44].

Epidemiological insights: Big data analytics have revolutionized cancer epidemiology, facilitating large-scale studies to discern risk factors and trends across diverse populations. This data-driven approach is pivotal for guiding public health initiatives aimed at cancer prevention and early detection, aligning with SDG 3 and SDG 17 [45,46,47].

Sun et al. analyzed the 2022 incidence and mortality rates of five types of female genital cancers, focusing on their evolutionary trends. Using data from 2018 and trends observed from 2010 to 2018, they projected incidences and mortalities in China, calculating age-standardized rates (ASIR and ASMR) with Segi’s world standard population. Most female genital organ cancers exhibited increased ASIRs and ASMRs over the study period, although there has been a recent slowdown in rural areas for vulvar and cervical cancers [48].

In their study, Zhang et al. compared spatial clustering and hotspot patterns of 11 common cancers across mainland China to identify associated environmental and behavioral risk factors. Employing the global Moran’s Index, they identified significantly higher spatial clustering degrees (p < 0.05) for esophageal, stomach, and liver cancers compared to others. Stratifying by gender, they found notably higher spatial clustering degrees (p < 0.001) for female esophageal cancer, male stomach cancer, male liver cancer, and female lung cancer. Moreover, male liver cancer exhibited a significantly higher clustering degree than female liver cancer (p < 0.001), while female lung cancer showed higher clustering than male lung cancer (p < 0.001) [49].

3 Achieving SDGs through CI in female cancer care

3.1 Health and well-being (SDG 3)

Achieving SDG 3 involves ensuring health and well-being for all ages, with CI techniques playing a pivotal role in improving life and care quality for female cancer patients.

Early detection and diagnosis: Machine learning models significantly improve the accuracy and speed of cancer diagnosis, facilitating early intervention. AI-driven imaging tools, utilized in mammography and radiology, detect tumors at earlier stages, thereby improving survival rates [50].

Xu et al. developed a dynamic multimodal variational autoencoder (DMVAE) to exploit intrinsic correlations between different data modalities, enabling feature imputation when modalities are missing. This AI diagnostic system predicts responses to the gonadotropin-releasing hormone (GnRH) stimulation test by integrating data from laboratory tests, electronic health records (EHRs), ultrasonography, and radiography reports. The DMVAE integrates features from all modalities to forecast outcomes for diagnosing Central Precocious Puberty. Experimental results demonstrate the DMVAE's superiority over standard methods, enabling confident diagnoses in about two-thirds of patients with 1.0 specificity, while confirming diagnoses in about one-third of patients using GnRH or GnRH analog stimulation tests [51].

Personalized treatment: Advanced computing techniques enable the development of customized treatment plans, enhancing effectiveness and minimizing side effects. For instance, AI models can predict which patients are likely to benefit from specific chemotherapy protocols, thereby optimizing treatment outcomes [52, 53].

A notable example is the work by Coudray et al., who developed a deep convolutional neural network (Inception v3) trained on whole-slide images from The Cancer Genome Atlas. This model can classify images as adenocarcinoma (LUAD), squamous cell carcinoma (LUSC), or normal lung tissue with high accuracy. Furthermore, it was trained to predict the ten most commonly mutated genes in LUAD, and it identified six of these genes—STK11, EGFR, FAT1, SETBP1, KRAS, and TP53—that could be inferred from pathology images, thus aiding physicians in creating personalized treatment plans [54].

In another study, Jiang et al. introduced several ML algorithms to predict axillary lymph node metastasis (LNM) in breast invasive micropapillary carcinoma (IMPC) patients and to identify LNM risk factors. Among these algorithms, the extreme gradient boosting (XGBoost) model outperformed the traditional logistic regression-based nomogram. When combined with the SHapley Additive exPlanations (SHAP) framework, it more intuitively highlighted the impact of various variables on LNM, identifying tumor size as the most significant risk factor. Consequently, the prediction score generated by the XG Boost model serves as a valuable indicator for LNM, significantly aiding in precision medicine [55].

Ongoing monitoring and management: Besides advancing personalized treatment, wearable devices and AI-driven remote monitoring systems enable continuous tracking of patient health, facilitating timely interventions. These technologies monitor vital signs, medication adherence, and symptom progression, offering a comprehensive approach to managing cancer care [56].

Wahidi et al. developed and tested an automated platform using CareSignal for remote symptom monitoring in female cancer patients undergoing radiotherapy. At the start of treatment, patients received a weekly SMS survey to assess treatment-related side effects' severity. The study evaluated patient perceptions of the system, including communication with healthcare providers, feasibility, and overall satisfaction. Despite their illness, patients demonstrated high compliance and expressed willingness to continue using the software in the future [57].

Additionally, Tan et al. examined the impact of an automated SMS platform on patient compliance with scheduled radiation therapy appointments. From July 2016 to January 2017, they implemented a program that sent daily appointment reminders to enrolled patients. Their analysis compared compliance rates between recipients of SMS reminders and those without reminders. The findings highlighted significant improvement in appointment adherence among patients receiving SMS reminders, affirming SMS as an effective intervention for enhancing compliance in populations at risk of missing appointments [58].

3.2 Gender equality (SDG 5)

Promoting gender equality involves addressing disparities in healthcare access and outcomes between men and women. CI technologies can empower women by providing them with better healthcare solutions.

Access to advanced diagnostics: Enhancing access to advanced diagnostic tools can significantly narrow disparities in cancer detection rates among women. Mobile health (mHealth) applications integrated with AI are pivotal in providing diagnostic services to remote and underserved areas, thereby improving healthcare accessibility [59, 60].

In their facility-based cross-sectional study, Shimels et al. examined 391 clinically confirmed cervical cancer patients. Data collection was facilitated by the Kobo Collect application, and analysis was conducted using SPSS version 26.0. The study underscored that nearly two-thirds of the patients perceived their healthcare access positively. Moreover, it revealed that almost half of the cervical cancer patients in Addis Ababa had optimal access to essential healthcare services required for their treatment [61].

Volynskaya et al. reviewed tools designed to enhance patient access to subspecialty pathology services across diverse geographic locations. Their focus was on ensuring timely reporting by appropriate pathologists and developing a systematic approach to deliver high-quality pathology services across a broad healthcare network. Their study emphasized the creation of a multisite pathology informatics platform to improve surgical pathology and hematopathology using advanced laboratory information systems [62].

Awareness and education: AI-driven platforms are pivotal in disseminating information about female-specific cancers, fostering awareness, and promoting early screening. These platforms utilize AI-powered educational tools tailored to various demographic groups, thereby enhancing their effectiveness [63, 64].

In their study, Griewing et al. explored the awareness and adoption intentions of healthcare professionals specializing in breast cancer towards digital health applications (DHA), AI, and blockchain technology (BCT). They found that awareness levels were highest for DHA (68.9%), followed by AI (66.7%) and BCT (24.4%). The study underscored the critical role of professional education in bridging the gap between the clinical benefits of emerging health technologies and their integration into breast cancer care. Griewing et al. demonstrated that brief educational interventions could significantly increase HCPs' willingness to adopt these technologies, suggesting targeted educational programs could effectively integrate them into clinical practice for female cancers [65].

Furthermore, Pesapane et al. highlighted a strong preference among breast cancer patients for AI to complement rather than replace traditional diagnostic processes. Their research emphasized the collaborative role of AI alongside radiologists' expertise in enhancing diagnostic accuracy [66].

Empowering women in research: Encouraging women's participation in cancer research and AI development ensures that their perspectives and needs are effectively addressed, leading to more equitable healthcare solutions [67,68,69,70].

Perera et al. conducted a cross-sectional study using data from ClinicalTrials.gov to assess female participation in US oncology trials. Their findings reveal significant underrepresentation of women relative to their disease burden in various oncology trials, particularly in fields with sex-specific outcome disparities such as surgical oncology, bladder, head/neck, and stomach cancer trials [71].

In another study, Varma et al. examined the representation of individuals by sex, age, and race in premarketing and postmarketing studies for novel cancer therapeutics approved by the FDA from 2012 to 2016. They found that women were adequately represented in both types of studies. However, the study underscores the need for the FDA to enhance transparency in clinical trial demographic data and further promote women's participation across these studies [72].

3.3 Reduced inequalities (SDG 10)

Reducing inequalities involves ensuring that everyone, regardless of their background, has access to quality healthcare. CI technologies can help bridge the gap between different socio-economic groups [73, 74].

Equitable access to healthcare: Telemedicine and remote diagnostic tools can provide quality care to patients in low-resource settings. AI-based platforms can deliver diagnostic and treatment services to underserved populations, reducing healthcare disparities [75, 76].

Rajan et al. explored the impact of telemedicine in Nepal, focusing on reducing gender-based barriers for rural women and girls accessing healthcare. Their study highlighted significant reductions in travel restrictions, treatment costs, and concerns related to sexual and reproductive health consultations. It also improved access to timely healthcare services and facilitated better time management for household responsibilities and activities. The study concludes that rural telemedicine plays a crucial role in overcoming gender-specific barriers to essential healthcare services [77].

Paladino et al. assessed the THRIVE app's impact, a web-enabled tool transmitting patient-reported symptoms to healthcare teams. Their study aimed to improve timely symptom management in a region with significant racial disparities in breast cancer outcomes. They evaluated outcomes such as adherence to adjuvant endocrine therapy, symptom burden, quality of life, self-efficacy in symptom management, and healthcare costs facilitated by the THRIVE app [78].

Affordable healthcare solutions: AI technologies offer scalable solutions that can significantly reduce the cost of healthcare services by automating routine tasks and optimizing resource allocation. This potential can enhance the affordability and accessibility of cancer care [79].

AI emerges as a pivotal tool in addressing disparities in cancer treatment effectiveness and affordability. By integrating and processing data from diverse databases, AI facilitates drug repurposing and expedites the discovery of new drugs cost-effectively and within shorter timeframes [80, 81].

In their study, Zhou et al. examined the concordance between treatment recommendations from IBM Watson for Oncology (WFO) and clinical decisions made by oncologists at a cancer care center in China. The research aimed to highlight treatment approach disparities between China and the U.S., revealing varying levels of concordance across different cancer types, notably lower rates for gastric cancers [82].

Global collaborations: International collaborations and knowledge-sharing initiatives can leverage CI technologies to address global health disparities. By sharing data, resources, and expertise, organizations can significantly improve cancer care in developing countries [83, 84].

For instance, Bhatt et al. developed a mHealth prototype and implemented a mHealth-supported cervical and oral cancer screening intervention at three sites serving underserved, low-health-literacy communities. This study screened 8,686 individuals, with the majority being screened for oral cancer. The positivity rates were 28% for cervical cancer screening, with 37% of those individuals attending follow-up appointments, and 5% for oral cancer screening, with 31% attending follow-up. Funded by a Global Innovation Initiative grant—a collaborative program between the British Council and the Institute of International Education—this project underscores the significant impact of international cooperation and support [85].

Complementing these efforts, a review by Chuang et al. offers a comprehensive overview of the educational and research programs provided by international organizations. These organizations' collaborative efforts with centers in low- and middle-income countries aim to build sustainable programs that enhance education, training, and research in gynecologic cancers globally [86].

3.4 No poverty (SDG 1)

Addressing the financial burden of cancer care is crucial for reducing poverty levels among female cancer patients and their families.

Financial risk protection: AI can optimize healthcare resource allocation and reduce unnecessary expenditures, making cancer care more cost-effective and alleviating the financial burden on patients [87, 88].

For example, Ishii-Rousseau et al. established practical metrics and benchmark data to demonstrate the clinical and operational utility of digital pathology (DP) in a pathology laboratory. By reviewing metrics such as ancillary test requests for recurrent/metastatic disease, cost analysis, turnaround time (TAT), and a DP experience survey, they conducted a comprehensive cost analysis, revealing a 5-year savings of $1.3 million. This study illustrates the benefits of applying clinical AI and implementing DP operations, including reduced confirmatory testing for patients with metastatic/recurrent disease and long-term decreases in off-site pathology asset costs. These advancements not only enhance cancer patient care but also demonstrate a significant return on investment, justifying the adoption of DP [89].

Similarly, Khanna et al. evaluated the impact of AI technology on healthcare costs, particularly in diagnosis and treatment, and compared it to traditional, non-AI-based approaches. Utilizing the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) methodology, they selected the top 200 studies focusing on AI in healthcare with an emphasis on cost reduction. Their study model demonstrates substantial cost savings through the use of AI tools in diagnosis and treatment. Furthermore, they suggest that the economics of AI can be improved by incorporating pruning, reducing AI bias, enhancing explainability, and securing regulatory approvals [90].

Social protection systems: AI can support the design and implementation of social protection systems that provide financial assistance and healthcare services to low-income female cancer patients, thereby reducing poverty [91,92,93].

In their comprehensive review, Bajwa et al. explore recent advancements in AI within healthcare, providing a roadmap for the development of effective, reliable, and secure AI systems. They also discuss future directions for integrating AI into healthcare practices [94].

In a related review, Abbod et al. delve into the foundational principles of AI techniques and their application in managing urological cancers, highlighting AI's transformative potential in optimizing treatment strategies [95].

Wahl et al. offer insights into how AI can enhance health outcomes in resource-limited settings, addressing ethical concerns surrounding patient safety and privacy within AI-driven healthcare interventions [96].

Finally, Reddy S reviews generative AI in healthcare, emphasizing its practical utility while advocating for careful planning and ethical management in implementation. The review underscores the importance of social factors, data privacy, security measures, and the integration of clinician expertise alongside AI technologies [97].

3.5 Quality education (SDG 4)

Quality education is essential for empowering female cancer patients and healthcare professionals.

Education and training for healthcare providers: AI-driven educational platforms offer up-to-date training on the latest cancer treatment protocols and technologies, thus enhancing the quality of care [98,99,100].

Building on this, Weerarathna et al., in their review of AI applications across the cancer care spectrum, discussed the challenges of translating AI from theoretical research to practical use in real-world cancer treatment. They particularly emphasized the emerging benefits of AI in female cancer care [101].

Additionally, Pan et al. outlined the features and specific application scenarios for AI-driven digital health technologies (DHTs) in cancer survivorship care services. They underscored the critical role of involving both patients and health practitioners in the development and implementation of these technologies [102].

Patient education: AI tools empower cancer patients by providing personalized information on their diagnosis, treatment options, and necessary lifestyle adjustments, facilitating informed health decisions [103, 104].

In a study by Shin et al., they investigated the impact of educating and counseling young female breast cancer patients about ovarian function and fertility after cancer treatment. Analyzing responses from 109 participants, the study aimed to assess how this education influenced their understanding of reproductive health issues. While the research highlighted challenges in educational effectiveness and satisfaction, it underscored the crucial need to educate patients about the impact of cancer treatments on ovarian function and fertility preservation [105].

Additionally, Amgad et al. developed the Histomic Prognostic Signature (HiPS) risk score aligned with American Joint Committee on Cancer (AJCC) staging criteria. This AI-driven system significantly enhances prognostic accuracy for patients with nonmetastatic invasive breast carcinomas. Their findings demonstrate that the HiPS score serves as a robust predictor of survival outcomes in estrogen receptor positive (ER+) and human epidermal growth factor receptor 2 positive (HER2+) breast cancers. Moreover, HiPS shows strong correlations with established epidemiologic and genomic risk profiles and validates well against widely used gene expression assays like Oncotype DX and MammaPrint, crucial for predicting recurrence and metastasis risks [106].

3.6 Decent work and economic growth (SDG 8)

Improving the economic conditions of female cancer patients is crucial for their well-being.

Employment opportunities: AI technologies not only support the creation of flexible employment opportunities for female cancer patients, enabling them to work while undergoing treatment [107,108,109,110], but they also demonstrate significant benefits across various industries discussed by Shen et al. These include healthcare, transportation, and production environment control, where AI integration has proven transformative [111]. Furthermore, Yabroff et al. underscore the financial challenges faced by cancer survivors in the US and propose practical strategies to mitigate financial toxicity, emphasizing the importance of supportive work environments and stable income sources for employees affected by cancer [112].

Safe work environments: AI applications can effectively monitor and ensure safe working conditions for cancer patients who continue to work during their treatment [113].

In their survey of 157 Italian occupational physicians, Rondinone et al. highlighted the need to equip these professionals with updated strategies and practical tools. Their findings underscore the importance of collaboration within occupational health and safety systems to identify suitable accommodations that support the resilience of workers affected by cancer in their workplaces [114].

Tamminga et al. explored the role of employers in supporting employees diagnosed with cancer to maintain or return to work. They identified a lack of effective interventions and tools to assist employers in this crucial aspect. Integrating supportive measures in workplaces can significantly benefit the retention and successful return to work of employees with cancer, enhancing their quality of life and societal contributions. Future research should focus on developing and evaluating specific tools and interventions that empower employers to effectively support their employees [115].

3.7 Industry, innovation, and infrastructure (SDG 9)

Innovation in healthcare infrastructure is essential for improving cancer care.

Healthcare infrastructure: Investment in AI-driven healthcare infrastructure holds significant promise for enhancing the efficiency and accessibility of cancer care services [116, 117]. Koh et al. delved into the challenges and opportunities of AI and machine learning (ML) in cancer imaging, discussing the evolution of algorithms into widely deployable tools. They stressed the need for interdisciplinary collaboration to expedite the development of practical AI and ML clinical tools aimed at improving cancer patient care and outcomes [118]. Meanwhile, Combi et al. conducted an extensive review of major conferences and journals, focusing on telemedicine projects and software systems. Their findings underscored the relevance of telehealth technology not only in developed nations but also in developing countries, where healthcare systems are adapting to meet current demands, thereby highlighting its global impact [119].

Cyber risks and AI/ML in healthcare: Amidst the rapid advancement in AI and ML technologies for cancer care, addressing cyber risks has become increasingly crucial. During infectious disease outbreaks, the focus on developing new healthcare resources can inadvertently expose vulnerabilities in digital infrastructure. Cyberattacks targeting healthcare systems can disrupt patient care, compromise sensitive data, and hinder the overall efficacy of cancer treatment technologies [120].

Integrating cybersecurity measures into the development of AI-driven healthcare systems is essential for safeguarding these innovations. Recent studies, including the work of Radanliev and De Roure, D, highlight the role of AI/ML in mitigating these cyber risks by enhancing the security and resilience of healthcare infrastructures during pandemics [121]. By incorporating advanced cybersecurity protocols and developing robust systems to protect against cyber threats, the healthcare sector can ensure that the benefits of AI/ML technologies are realized without compromising data integrity or patient safety [29]. Ensuring that healthcare systems are not only innovative but also secure aligns with SDG 9, which emphasizes the importance of resilient and secure infrastructure. The integration of cybersecurity measures into healthcare innovation helps maintain the trust and reliability of AI-driven solutions, ultimately contributing to more effective and sustainable cancer care.

Research and development: Supporting AI research in cancer care holds promise for groundbreaking innovations, including immunotherapies and enhanced treatment options for female cancer patients [122,123,124].

Achar et al. employed robotic sampling to investigate longitudinal T cell responses in both mouse and human models, integrating machine learning and modeling to understand how T cells interpret information from specific antigens. Their analysis revealed that cytokine release patterns carry information about the type of antigen encountered, identifying six distinct cellular responses—expanding beyond the traditional three recognized types. This deeper understanding is set to advance strategies in cancer immunotherapy, particularly those involving engineered and tailored T cell responses [125].

In a complementary effort to revolutionize cancer treatment, This et al. pioneered the use of ML algorithms to classify T cells using complex datasets, including signals from polyclonal T cells. Utilizing deep learning techniques, they accurately predicted the activation of T cell receptor (TCR)-transgenic CD8 + T cells based on calcium fluctuations, validating their approach across various TCRs and polyclonal populations. This study establishes a foundational method for identifying antigen-specific TCR sequences crucial for advancing T cell therapies [126].

3.8 Partnerships for the goals (SDG 17)

Collaborative efforts are crucial for leveraging AI in cancer care.

International collaborations: Collaborations among countries, institutions, and organizations play a crucial role in advancing AI technologies and best practices in cancer care [127,128,129].

The International Collaboration on Cancer Reporting (ICCR) has standardized datasets for histological reporting of breast cancer to ensure global consistency. Laokulrath et al. reviewed ICCR's recommendations for reporting invasive breast carcinoma, comparing them with guidelines from regional organizations such as RCPath, CAP, and RCPA. Their comprehensive analysis provided practical insights, highlighted divergent aspects, and proposed pathways for establishing a unified global framework for breast cancer pathology reporting [130].

In parallel, Tzelves et al. evaluated the implementation of an AI-driven follow-up strategy via the Artificial Intelligence Supporting Cancer Patients across Europe (ASCAPE) platform. This collaborative initiative involves 15 partners across 7 countries, comprising academic medical centers and research institutions. Leveraging advancements in Big Data and AI, ASCAPE offers personalized predictions for quality of life issues among breast and prostate cancer patients, alongside actionable insights for physicians. The study underscores ASCAPE's potential to democratize access to AI tools, bridging healthcare provider disparities and benefiting cancer patients across Europe and beyond [131].

Resource sharing: Global partnerships can support the sharing of data, resources, and expertise to improve cancer care for women worldwide [132,133,134].

Girdler et al. investigated collaborative efforts within the global retinoblastoma community (1RBW) aimed at improving patient outcomes. Their survey of 170 treatment centers identified 112 partner institutions, including 80 treatment centers and 32 organizations. These collaborations primarily focused on patient referrals, consultations, and capacity-building initiatives, reflecting extensive global efforts to reduce retinoblastoma mortality. The study also highlighted opportunities to strengthen connections with peripheral treatment centers for broader impact [135].

In West Africa, Aidam and Sombié documented The West African Health Organization's (WAHO) research development program from 2009 to 2013. Using the Knowledge for Better Health Research Capacity Development Framework, this initiative supported 24 health research projects across all 15 ECOWAS (Economic Community of West African States) member countries. Emphasizing stewardship, financing, sustainable resourcing, and research utilization, the program significantly bolstered regional health research capacity [136].

4 Challenges and barriers

4.1 Technical challenges

Data quality and integration: Effective AI models depend on high-quality, integrated data, necessitating collaborative efforts to standardize data collection and sharing practices [137].

Addressing the challenges of integrating AI into clinical practice, particularly with multimodal data, involves managing issues such as missing data, fairness in dataset shifts, interpretability limitations, and navigating regulatory guidelines. Lipkova et al. proposed strategies to enhance AI interpretability and explore the potential of multimodal data connections in clinical settings. Their study critically assessed barriers to clinical adoption and offered solutions to overcome these challenges [138].

Furthermore, Bernardi et al. evaluated digital health technology interventions aimed at improving healthcare research data quality. They identified a spectrum of barriers, including technical, motivational, economic, political, legal, ethical, organizational, human resources, and methodological factors. Their findings underscore the importance of standardized practices and collaborative initiatives to enhance data quality in healthcare research [139].

Algorithm interpretability: Developing interpretable AI algorithms is crucial for their acceptance and effective use in clinical settings [140, 141]. Ennab et al. contributed to this area by designing an interpretability-based model that employs algorithms simulating human reasoning abilities. Through statistical analysis of datasets containing variables from medical images and patient symptoms, they identified relative weights crucial for healthcare conditions. Their model integrates two interpretability algorithms, blending deep learning with rules-based approaches based on these variable weights [142].

Meanwhile, Metta et al. explored the practical application of local Explainable Artificial Intelligence (XAI) methods within healthcare, particularly highlighting the effectiveness of the Local Rule-Based Explanations (LORE) technique. Their review underscores the pivotal role of interpretability and transparency in AI systems for disease diagnosis, predicting patient outcomes, and customizing treatment plans. Despite the inherent challenges and trade-offs between interpretability and model performance, their findings emphasize how local XAI methods like LORE can significantly enhance decision-making in healthcare. By offering detailed, case-specific insights, these methods contribute to improving both physicians' and patients' understanding of machine learning models and their implications [143].

4.2 Ethical and social considerations

Medical decision-making: The integration of AI into medical decision-making raises ethical concerns regarding transparency and accountability, emphasizing the need for human oversight and preserving patient autonomy [144, 145]. Xu et al. conducted an extensive review on interpretability within Clinical Decision Support Systems (CDSS), providing insights into methods aimed at enhancing interpretability from both technological and medical perspectives. Their study explores various strategies, including ante-hoc methods such as fuzzy logic, decision rules, logistic regression, and decision trees for knowledge-based AI. They also examine the utility of white box models and post hoc techniques like feature importance analysis, sensitivity analysis, visualization tools, and activation maximization for black box models. Key factors influencing CDSS interpretability identified include data characteristics, biomarkers, dynamics of human-AI interaction, and specific needs of clinicians and patients [146].

Addressing critical legal and ethical issues related to AI in healthcare, Naik et al. emphasize the necessity for algorithmic transparency, privacy protection, and safeguarding the interests of all stakeholders [147]. They also stress the importance of implementing robust cybersecurity measures to mitigate vulnerabilities associated with AI technologies in healthcare settings. These considerations are crucial as AI continues to evolve as a tool in medical decision-making, balancing innovation with ethical responsibilities to ensure patient trust and safety.

Addressing bias in algorithms: Ensuring fairness in AI algorithms is crucial to avoid exacerbating healthcare disparities. Developing unbiased AI models is essential for delivering equitable care to all patients [148, 149].

In the realm of medical education, Franco D'Souza et al. provide a comprehensive guide outlining twelve essential recommendations for navigating ethical challenges in integrating AI. Their suggestions include advocating for transparency, mitigating bias, rigorously validating content, securing data, obtaining informed consent, fostering collaboration among stakeholders, enhancing educator training, empowering students, implementing continuous monitoring practices, establishing clear accountability measures, adhering to standardized guidelines, and forming dedicated ethics committees. Embracing these principles enables medical educators to elevate educational standards and prepare a healthcare workforce proficient in both technology and ethics, ensuring effective AI integration [150].

Regarding cancer healthcare, Dankwa-Mullan and Weeraratne explore the applications of AI and ML tools, emphasizing strategies to mitigate bias and promote equity. Their recommendations underscore the necessity of a coordinated and collaborative approach to effectively address embedded biases across these technologies, aiming to optimize their contribution to equitable cancer care [151].

Patient acceptance and trust: Educating patients about the benefits and limitations of AI can enhance their acceptance and trust in these technologies. It's also crucial to involve patients in the decision-making process [152,153,154].

Mr. Esmaeilzadeh argues in his paper that rectifying flawed business models and inefficient workflows in healthcare requires more than just deploying AI-driven tools. Healthcare organizations must tackle root problems such as misaligned financial incentives, dysfunctional medical workflows, and fragmented electronic health records systems. Alongside AI adoption, improving care coordination among providers and enhancing patient education and engagement models are essential. This holistic approach recognizes the need for investments beyond financial resources, emphasizing the importance of nurturing human capital, promoting continuous learning, and fostering a supportive environment for AI integration.

The paper underscores the critical role of clear regulations in establishing trust, ensuring safety, and guiding the ethical use of AI. It advocates for coherent frameworks addressing transparency, model accuracy, data quality control, liability, and ethics. Furthermore, Esmaeilzadeh stresses the importance of advancing AI literacy within academia to prepare future healthcare professionals for an increasingly AI-driven landscape [155].

Concluding with a vision for the future of healthcare with AI, the paper offers strategic suggestions for responsible engagement. It emphasizes thoughtful innovation to unlock significant benefits for healthcare organizations, physicians, nurses, and patients while proactively managing risks.

Shevtsova et al. identified factors influencing trust and acceptance of AI-powered medical technologies among stakeholders, including patients. Their review of existing literature pinpointed various determinants of trust in AI applications in medicine. Subsequently, they conducted an electronic survey among key stakeholder groups based on these factors. Results indicated that variables such as patient gender, age, and education level had minimal relevance [156]. Nevertheless, this study offers valuable insights for medical AI implementers, highlighting key drivers of trust and acceptance among stakeholders and providing guidance for fostering acceptance of AI applications in medical settings.

4.3 Economic and regulatory barriers

Cost of implementation: The high initial cost associated with AI solutions poses a significant barrier for healthcare providers aiming to integrate these technologies. To foster widespread adoption, it is crucial to develop cost-effective strategies and secure adequate funding support [157,158,159].

Challenges in AI implementation: Khan et al. outlined several obstacles in the adoption of AI within the healthcare sector. These include the lack of price transparency, escalating costs, data privacy concerns, socio-economic implications, ethical considerations, cybersecurity risks, and issues related to developer capabilities. Addressing these challenges is essential for overcoming barriers to successful AI integration in medicine [160].

Navigating regulatory approvals: The regulatory landscape for AI in healthcare is complex and navigating it can be challenging. Ensuring compliance with regulations and obtaining necessary approvals are crucial steps that can potentially delay the deployment of AI solutions [161, 162].

Auditory aspects of ML for health algorithms: Oala et al. explored the technical validation, clinical evaluation, and regulatory assessment of machine learning algorithms for health applications (ML4H). They proposed auditing tools and recommended interactive regulatory checklists tailored to specific use-cases. Their suggestions include automated audit report generation, standardized test cases, and educational resources for auditors to enhance regulatory oversight [163].

Sustainability: To ensure the sustained effectiveness of AI interventions, ongoing investment in maintenance and updates is essential [164,165,166]. Goh HH and Vinuesa R discuss a range of AI applications poised to accelerate progress towards achieving the SDGs, alongside challenges that may hinder these advancements. They propose the establishment of an additional SDG dedicated to digital technologies, underscoring its pivotal role in global development alongside existing SDGs [167].

5 Future directions and recommendations

Research priorities: Future research should prioritize the development of advanced AI models capable of effectively managing intricate and heterogeneous datasets. Exploring AI applications in less-explored realms of female cancer care holds promise for uncovering novel insights and enhancing outcomes [168].

Interdisciplinary collaboration: Fostering collaboration between computer scientists, healthcare professionals, and policymakers can accelerate the development and implementation of AI technologies. Interdisciplinary teams can address the multifaceted challenges in cancer care more effectively [169].

Policy recommendations: Policymakers should facilitate the integration of AI in healthcare by offering funding, establishing supportive regulatory frameworks, and promoting education and training programs. Encouraging the implementation of industry standards can significantly enhance the effectiveness and safety of AI applications [170, 171].

6 Conclusion

The integration of computational intelligence techniques holds significant potential for transforming female cancer care and advancing the Sustainable Development Goals. By addressing the specific challenges faced by female cancer patients, CI can contribute to achieving health and well-being (SDG 3), promoting gender equality (SDG 5), and reducing inequalities (SDG 10). The application of these advanced technologies, including artificial intelligence, machine learning, and data mining, enables earlier diagnosis, more personalized treatment plans, and more efficient management of cancer, which are crucial in improving patient outcomes and quality of life.

However, the successful implementation of CI in female cancer care requires overcoming several challenges. These include ensuring data privacy and security, addressing biases in AI algorithms, and ensuring equitable access to these technologies across different regions and populations. Additionally, the integration of CI with existing healthcare systems necessitates significant investment in infrastructure, education, and training for healthcare professionals.

Continued research, innovation, and collaboration among stakeholders—including researchers, healthcare providers, policymakers, and technology developers—are essential for developing more effective, equitable, and sustainable healthcare solutions. Interdisciplinary efforts that combine the expertise of oncologists, data scientists, and engineers are particularly important in fully realizing the potential of CI to advance female cancer care.

This review underscores the importance of leveraging technology to not only improve health outcomes for female cancer patients but also to support broader societal goals such as gender equality and reducing healthcare disparities. By fostering global partnerships and prioritizing ethical considerations in the development and deployment of CI technologies, the global healthcare community can work towards a future where the benefits of advanced computational methods are accessible to all, contributing to a more sustainable and equitable world.

In conclusion, the promise of computational intelligence in achieving the SDGs is vast, but it requires a concerted effort to ensure that the advancements in technology are translated into real-world benefits for female cancer patients. By addressing the challenges and leveraging the opportunities presented by CI, we can make significant strides toward a future where female cancer care is not only more effective but also more inclusive and just.