Abstract A083: A Conceptual Framework for Communicating AI-Based Prostate Cancer Risk Predictions

Abstract

Abstract Background In prostate cancer care, artificial intelligence (AI) models are increasingly used to generate personalized risk predictions by analyzing diverse data inputs, including clinical variables, imaging findings, and genomic indicators such as polygenic risk scores. However, effectively communicating these complex AI-derived risk estimates to patients presents significant challenges. Patients often struggle to understand probabilistic risk information, particularly when advanced algorithms generate it. Traditional cancer risk communication methods might not consider the complexities of AI outputs or differences in patient health literacy and numeracy. We propose a conceptual framework to guide patient-centered communication of AI-driven risk predictions. Methods We reviewed the existing literature on AI-based risk modeling, health risk communication, numeracy, dual-process information processing, and related theories to develop a communication framework designed explicitly for AI-based prostate cancer risk prediction tools. Results The proposed framework emphasizes clear, patient-specific risk communication strategies to enhance understanding and foster trust. It highlights a dual-process approach: offering intuitive visual aids and narrative explanations for quick, instinctive understanding, while also providing accurate statistics and explanations of the AI model’s reasoning for more deliberate processing. By addressing various patient numeracy levels and clearly explaining the AI model’s inputs, assumptions, and uncertainty in plain language, the framework aims to improve understanding and trust in AI-generated risk estimates. Additionally, to promote equity, it includes culturally responsive and accessible communication strategies tailored to patients’ linguistic, literacy, and cultural backgrounds. Conclusion This conceptual framework offers guidance on integrating AI-based risk prediction tools into prostate cancer consultations in a manner that is patient-friendly. By enhancing patients’ understanding of personalized risk information, building greater trust in AI-driven predictions, and addressing various communication needs, the framework helps facilitate more informed decision-making. Ultimately, patient-focused communication can help reduce disparities and promote more equity in prostate cancer care. Citation Format: Tingyu Zou, Folakemi Odedina. A Conceptual Framework for Communicating AI-Based Prostate Cancer Risk Predictions [abstract]. In: Proceedings of the AACR Special Conference in Cancer Research: Innovations in Prostate Cancer Research and Treatment; 2026 Jan 20-22; Philadelphia PA. Philadelphia (PA): AACR; Cancer Res 2026;86(2_Suppl):Abstract nr A083.

Ähnliche Arbeiten