Clinical heterogeneity across oropharyngeal datasets and implications for outcome prediction

Multi-institutional oropharyngeal cancer (OPC) datasets are increasingly used for research, yet systematic evaluation of data completenees and in clinical impact remains limited, Unrecognized institutional heterogeneity in variable availability, coding conventions, and patient characteristics can introduce substantial bias into combined analysis. A cross-sectional analysis and retrospective cohort study was carried out to systematically characterize data availability patterns and clinical heterogeneity across two major OPC datasets: the RADCURE dataset (2005-2017) from Princess Margaret Cancer Centre (Toronto, Canada) and the HNSCC-Oropharyngeal dataset (2005-2012) from MD Anderson Cancer Center (Houston, TX, USA). The final analytic cohort comprised 1972 participants; 1482 from Toronto and 490 from Houston. Substantial institutional differences were observed across demographic characteristics, HPV distribution, and treatment modality, After comprehensive adjustment for patient and disease characteristics, the survival hazard ratio for treating institution was 1.19 (95% CI 0.93-1.52), with treating institution consistently ranking as a low-importance predictor across all machine learning algorithms and time points. The authors conclude that systematic harmonization and rigorous assumption testing are prerequisites for multi-institutional OPC analyses; after appropriate adjustment, treating institution does not independently predict survival.