AI-enabled approaches may predict cutaneous toxicities in cancer immunoradiotherapyResearchers explore AI to predict skin side effects from combined cancer treatments

This systematic mini-review synthesizes emerging AI-enabled approaches for the precision prediction and management of cutaneous toxicities—specifically radiation dermatitis and immune-related skin adverse events—in patients with advanced malignant tumors receiving immunoradiotherapy (ICI+RT). The review describes these toxicities as arising from the convergence of radiation-induced tissue injury and immune checkpoint blockade–driven immune amplification, involving interconnected pathways like DNA damage–associated danger signaling and dysregulated T-cell responses. The authors note that toxicities often emerge early, exhibit substantial inter-patient heterogeneity, and can compromise treatment continuity and quality of life, with broader application of immunoradiotherapy accompanied by a high incidence, though exact incidence rates are not reported.

The proposed AI-enabled approaches aim to move beyond conventional reactive, grading-based management by integrating clinicodosimetric variables, spatial dose topology, imaging-/radiomics-derived tissue susceptibility, and immune–inflammatory surrogates into predictive models. The review positions cutaneous toxicity as a tractable and clinically meaningful endpoint for precision management aligned with data-driven oncology goals. However, the safety and tolerability profile of these AI-guided strategies, as well as their impact on serious adverse events, are not reported.

Key limitations include the nature of the publication as a review of emerging approaches without reporting primary trial results, sample sizes, follow-up, or specific outcome measures. The certainty of the evidence is low, as the review does not report on the predictive utility, clinical efficacy, or validated performance of the AI models discussed. Practice relevance is therefore restrained; while the conceptual framework for AI integration is presented, its clinical implementation awaits prospective validation and demonstration of improved patient outcomes compared to standard management.