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

Abstract Aim The oesophageal cancer (OC) multi-disciplinary team (MDT) operates under significant pressures. Machine learning (ML) may 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 clinical decision support system (CDSS) aimed at predicting OC MDT treatment decisions and prognosticating palliative scenarios, co-designed using Responsible Research and Innovation (RRI) principles. Method Clinicopathological data collected from 1931 patients over 13 years were used to train and validate ML models predicting initial treatment pathways and palliative prognosis. 953 cases from a leading tertiary unit trained models which were externally validated on 978 cases from another geographical catchment. 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 workshops with computer scientists, clinicians and patient representatives allowed end-users to co-design 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 to develop the CDSS which is user-friendly and quick to use. Conclusions 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.