AI is rapidly transforming oncology, but many clinicians lack practical tools to evaluate its real-world reliability. A new study introduces a clinician-focused checklist and questionnaire designed to help oncologists critically assess artificial intelligence and machine learning models for cancer care.
Unlike existing guidelines, which are often written for developers, these tools are tailored specifically for physicians who must make clinical judgments based on complex AI outputs. The authors emphasize that oncology’s multidisciplinary and data-intensive nature requires rigorous scrutiny to ensure that AI applications are both safe and clinically meaningful.
The initiative produced two complementary resources. The first is a yes or no checklist that allows clinicians to quickly scan whether a model adheres to widely recognized standards in the field. The second is an open-ended questionnaire intended for deeper evaluation, prompting physicians to explore critical aspects of model development and potential clinical impact. Both tools were developed through collaboration between oncologists and artificial intelligence experts, ensuring they address the realities of clinical practice rather than abstract technical theory.
The development process drew on insights from 24 published articles, with iterative refinement to capture domains most relevant to oncology practice. Four case examples of AI applications in cancer care were analyzed to demonstrate how the checklist and questionnaire can be applied in real-world settings. These examples highlight how the tools facilitate structured, case-based learning and support informed clinical decision-making.
By equipping oncologists with a systematic approach, the authors argue that these resources will help bridge the gap between rapidly advancing artificial intelligence technologies and the practical needs of frontline clinicians. Importantly, the tools also underscore the interdisciplinary collaboration required for safe and effective AI integration into oncology.
As the use of artificial intelligence expands across cancer diagnosis, treatment planning, and patient monitoring, structured evaluation will be essential to safeguard patient outcomes while enabling innovation. These new resources represent a step toward empowering clinicians to take an active role in shaping how AI enters oncology practice.
Reference: Siddiqui NS et al. Clinician’s Artificial Intelligence Checklist and Evaluation Questionnaire: Tools for Oncologists to Assess Artificial Intelligence and Machine Learning Models. JCO Clin Cancer Inform. 2025:9:e2500067.
