Variability in RECIST 1.1 Progressions in Oncology Trials

This technical report investigates the variability in double-read oncology trials, specifically focusing on the discrepancies in assessing disease progression using RECIST 1.1 criteria. The study emphasizes the importance of Blinded Independent Central Review (BICR) with double reads and adjudication in clinical trials to ensure data quality. It analyzes data from five clinical trials involving 1932 lung cancer patients treated with immunotherapy or targeted therapy. The findings reveal a 39.3% discordance rate in the date of progressive disease (DoPD) assessments, with various RECIST components contributing to this variability. New lesion detection was identified as a significant factor in discordance. The report discusses the implications of these discrepancies on survival curves and the potential for improving trial precision through better understanding of RECIST component reliability. The aim is to enhance the accuracy of progression assessments, thereby supporting clinical decision-making in oncology.