Artificial intelligence (AI) systems are increasingly deployed in clinical practice, particularly in radiology, pathology, endoscopy, and decision support. While these tools improve efficiency and accuracy, concerns have arisen about deskilling—the erosion of physicians’ expertise due to reliance on automation. We conducted a narrative review of empirical studies, randomized trials, and theoretical analyses published up to August 2025. The focus was on quantitative evidence of decreased performance following AI exposure, automation bias, and structural changes in training environments. Sources included PubMed, Embase, and gray literature. Evidence of clinical deskilling, though scarce, is consistent across specialties. In a multicenter randomized trial in colonoscopy, the adenoma detection rate (ADR) dropped significantly from 28.4% to 22.4% when endoscopists reverted to non-AI procedures after repeated AI use, while ADR remained stable with AI assistance (25.3%). In radiology, a controlled study of 27 breast imaging radiologists showed that erroneous AI prompts increased false-positive recalls by up to 12%, even among experienced readers. In computational pathology, experimental web-based tasks revealed that over 30% of participants reversed correct initial diagnoses when exposed to incorrect AI suggestions under time constraints. Structural deskilling has been reported in cytology following the UK’s transition to human papillomavirus primary screening, leading to an 80%-85% reduction in case volumes and consolidation of laboratories from 45 to 8 centers, with major implications for training capacity. Across domains, analyses confirm the presence of automation bias and highlight risks of diminished independent diagnostic reasoning. Although limited in number, empirical studies consistently demonstrate that AI can inadvertently impair physicians’ performance or reduce opportunities for skill maintenance. Quantitative evidence of decreased diagnostic accuracy, error propagation, and training erosion underscores the need for longitudinal monitoring, adaptive curricula, and regulatory frameworks to mitigate deskilling. Safeguarding clinical expertise should be considered a central component of AI safety and resilience in medicine.

PURPOSE: Digitalization of the health care system is transforming cancer patient care. Although many studies have investigated the determinants of a limited digital health literacy, the association between frailty factors and overall survival (OS) of these patients has never been assessed. METHODS: A retrospective noninterventional study included 15,244 adult patients with cancer diagnosed between January 1, 2015, and December 31, 2017, and treated at the Centre Léon Bérard. Limited e-health literacy was defined as the absence of an e-mail address in the electronic patient record. An Inverse Probability of Treatment-Weighted Kaplan-Meier estimate and a multivariate Cox proportional hazards model including interaction terms were used to adjust for confounding on measured covariates. RESULTS: In total, 15,244 adults with cancer were included: 55% women, with a median age of 62 years (19-103), and 35.5% had a metastatic disease. More than half (n = 8,771, 57.5%) had entered their e-mail address in their electronic patient record, and 4,020 (26.4%) opened their own patient portal. The median follow-up was 3.6 years (range: 0-6.8). Inverse Probability of Treatment-weighted Kaplan-Meier estimates showed a significantly better OS for patients with an e-mail address ( P < .001). In multivariate analysis integrating interaction terms, male gender (hazard ratio [HR] = 1.27; 95% CI, 1.15 to 1.41; P < .001), older age (HR = 1.02; 95% CI, 1.02 to 1.03; P < .001), de novo metastatic setting (HR = 2.63; 95% CI, 2.47 to 2.79; P < .001), and no e-mail address (HR = 1.63; 95% CI, 1.33 to 2.00; P < .001) were significantly associated with worse OS. CONCLUSION: Our results support a strong association between the limited level of literacy and OS. A more in-depth study integrating variables such as socioeconomic level and location of residence would enrich these results.

The response to immunotherapy has been little investigated in overweight and obese cancer patients. We evaluated the relationships between BMI, toxicity, and survival in patients treated by immunotherapy for metastatic cancer. We included metastatic cancer patients treated by immunotherapy between January 2017 and June 2020 at the Centre Léon Bérard. In total, 272 patients were included: 64% men and 36% women, with a median age of 61.4 years. BMI ≥ 25 in 34.2% and 50% had non-small cell lung cancer (n = 136). Most received monotherapy, with nivolumab in 41.9% and pembrolizumab in 37.9%. Toxicity, mostly dysthyroiditis, occurred in 41%. Median overall survival (OS), estimated by Kaplan–Meier analysis, was significantly longer for patients with a BMI ≥ 25 than for those with a BMI < 25 (24.8 versus 13.7 months HR = 0.63; 95% CI 0.44–0.92, p = 0.015), and for patients experiencing toxicity than for those without toxicity (NR versus 7.8 months, HR = 0.22; 95% CI 0.15–0.33, p < 0.001). Adjusted OS was associated with toxicity, and the occurrence of toxicity was associated with sex and histological features but not with BMI. Thus, being overweight and experiencing toxicity was associated with longer overall survival in patients treated by immunotherapy. More attention should be paid to body composition in the care of cancer patients.

Background: Real-world data (RWD) related to the health status and care of cancer patients reflect the ongoing medical practice, and their analysis yields essential real-world evidence. Advanced information technologies are vital for their collection, qualification, and reuse in research projects. Methods: UNICANCER, the French federation of comprehensive cancer centres, has innovated a unique research network: Consore. This potent federated tool enables the analysis of data from millions of cancer patients across eleven French hospitals. Results: Currently operational within eleven French cancer centres, Consore employs natural language processing to structure the therapeutic management data of approximately 1.3 million cancer patients. These data originate from their electronic medical records, encompassing about 65 million medical records. Thanks to the structured data, which are harmonized within a common data model, and its federated search tool, Consore can create patient cohorts based on patient or tumor characteristics, and treatment modalities. This ability to derive larger cohorts is particularly attractive when studying rare cancers. Conclusions: Consore serves as a tremendous data mining instrument that propels French cancer centres into the big data era. With its federated technical architecture and unique shared data model, Consore facilitates compliance with regulations and acceleration of cancer research projects.

PURPOSE: Therapeutic compliance, or adherence, is critical in oncology because of the complexity and duration of cancer treatment regimens. Nonadherence can lead to suboptimal therapeutic outcomes, increased disease progression, higher mortality rates, and elevated health care costs. Traditional methods to enhance compliance, such as patient education and regular follow-ups, have shown limited success. MATERIALS AND METHODS: This review examines the potential of digital health technologies to improve adherence in oncology. Various studies and trials are analyzed to assess the effectiveness of these technologies in supporting patients with cancer. RESULTS: mHealth applications have been shown to improve medication adherence through features like medication reminders and symptom tracking. Telemedicine facilitates continuous care and reduces the need for travel, significantly improving adherence and patient satisfaction. Patient-reported outcome measures enhance clinical decision making and personalized treatment plans by incorporating patient feedback. Electronic medical records and patient portals improve compliance by providing easy access to medical information and fostering better patient-provider communication. Connected pillboxes aid in consistent medication intake and reduce dispensing errors. CONCLUSION: Digital health technologies offer significant benefits in oncology by enhancing patient engagement, improving adherence to treatment protocols, and enabling comprehensive cancer care management. However, challenges such as the digital divide, data privacy concerns, and the need for tailored interventions must be addressed. Future research should focus on evaluating the effectiveness of digital interventions and developing personalized digital health tools to maximize therapeutic compliance.

Real-world data (RWD) provide essential insights into the effectiveness and safety of breast cancer treatments, particularly in diverse patient populations, where traditional clinical trials may have limitations. Integrating RWD into breast cancer research enhances the understanding of treatment outcomes and supports clinical decision-making, complementing the findings from controlled clinical studies. This article reviews the integration of RWD into breast cancer research, highlighting the benefits and challenges. Various sources of RWD, including electronic health records (EHRs), insurance claims, and patient registries, are examined, with a focus on their application in studies of triple-negative breast cancer. The article also explores the role of artificial intelligence (AI) in managing RWD, particularly through technologies like natural language processing (NLP) and predictive analytics, which enhance data collection, storage, and analysis. RWD has demonstrated significant value in informing clinical decision-making and improving patient outcomes in breast cancer treatment. The integration of AI into the management of RWD has provided deeper insights into patient outcomes and supported personalized treatment strategies. Specific studies leveraging RWD have shown improved understanding of breast cancer subtypes, such as triple-negative breast cancer, and enhanced the effectiveness of treatment protocols. Despite the benefits, challenges remain in integrating RWD and AI into clinical practice, particularly regarding transparency, interpretability, and ethical considerations. Addressing these challenges requires robust data governance frameworks, interdisciplinary collaboration, and investment in advanced analytical tools. The potential for RWD and AI to transform breast cancer treatment and improve patient care is significant, underscoring the need for ongoing research and collaboration.