Shaping AI-Based Decision Support in Kidney Cancer: Multidisciplinary Consensus from the IKCSEU25 ART Think Tank

Abstract

Background: Artificial intelligence (AI) has the potential to significantly enhance clinical decision-making in oncology. However, its application in renal cell carcinoma (RCC) remains limited. The ART (Artificial Intelligence in Renal Tumors) project is a Spanish, multi-institutional initiative aimed at developing a dynamic, transcriptomics-based AI model to guide systemic treatment decisions for patients with metastatic RCC (mRCC). Objective: The aim of this paper is to present the rationale, methodology, and early implementation challenges of the ART project, as discussed during a dedicated Think Tank session at the 2025 International Kidney Cancer Symposium Europe (IKCSEU25), and to gather expert insights on its clinical and regulatory viability. Design, Setting, and Participants: The ART project includes three phases: (1) retrospective algorithm training using clinical and transcriptomic data from completed trials; (2) a prospective, non-interventional study collecting multi-omic and clinical data from 500 patients across 30 centers; and (3) a future comparative analysis of ART-guided versus standard clinical decisions. The AI model is designed to evolve continuously through ongoing data integration. Results and Limitations: Experts underscored the importance of integrating multimodal data—including circulating biomarkers and immune profiling—while expressing concerns about the reliance on short-term endpoints. Key barriers identified included data harmonization, external validation, and regulatory uncertainty regarding adaptive algorithms. The absence of a clear approval pathway for non-static clinical decision support systems also poses a challenge. Despite limited initial funding, the ART platform has generated strong institutional engagement and may serve as a scalable model for clinician-oriented AI tools. Conclusions: The ART project represents an innovative approach to AI-driven personalization of kidney cancer treatment. Expert feedback from IKCSEU25 highlighted the scientific robustness of the initiative, while also emphasizing the need for broader validation, regulatory clarity, and the use of clinically meaningful endpoints to support real-world implementation. Patient Summary: Experts reviewed a new AI-based tool being developed in Spain to help doctors choose the best treatments for kidney cancer. The tool shows promise but needs further testing and must meet regulatory standards before it can be used in routine clinical care.

Ähnliche Arbeiten