The Oesophageal Cancer Multi-Disciplinary Tool: A Responsibly Co-Designed, Externally Validated, Machine Learning Clinical Decision-Support System

Background The oesophageal cancer (OC) multi-disciplinary team (MDT) operates under significant pressures, handling complex decision-making. Machine learning (ML) has the potential to enhance efficiency, consistency, and cost-effectiveness by identifying intricate decision-making patterns provided it is trained and implemented responsibly. We present an externally validated ML-based clinical decision support system (CDSS) designed to predict OC MDT treatment decisions and prognosticate palliative scenarios, co-designed using Responsible Research and Innovation (RRI) principles. Methods Clinicopathological data collected from 1931 patients over 13 years were used to develop and validate ML models trained to predict curative and palliative treatment pathways and palliative prognosis. 953 OC cases from a leading tertiary referral centre trained models which were externally validated on 978 OC cases from a unit in another geographical region. Model performance was evaluated using Area Under Curve (AUC) for treatment classifiers and calibration curves for survival models. A parallel RRI program combining interviews and inter-disciplinary workshops with computer scientists, clinicians and patient representatives allowed end-users to contribute towards the co-design of the CDSS user interface. Results The ML models performed well with AUCs between 0.873–0.909 for the primary model, and 0.711–0.815 for the palliative classifier. Predicted survival probability from the palliative survival model was well calibrated over the first 12 months post-diagnosis in both cohorts. The RRI program provided a collaborative, multidisciplinary environment in which to develop the CDSS which is user-friendly and quick to use. Conclusion We present a novel, responsibly developed, externally validated AI CDSS trained to predict oesophageal cancer MDT decisions. It represents the foundations of a transformative application of ML, personalised, consistent and efficient MDT decision-support within OC which aligns to RRI principles.