Development of a novel methodology to assess response to lymphoma treatment in real-world data.

Abstract

e19081 Background: Innovation in the use of real-world data (RWD) for regulatory purposes is an important provision of 21st Century Cures Act. Numerous initiatives to enhance RWD methodologies in oncology drug development are ongoing, including evaluation of treatment response assessment to support effectiveness research. Blinded independent central review (BICR) standardizes response assessment in clinical trials, and corresponding RWD methodologies are needed. This study assess feasibility of BICR using RWD, and evaluates a novel, standardized treatment response methodology among patients with diffuse large B-cell lymphoma (DLBCL). Methods: A retrospective multisite chart review identified adult patients with DLBCL treated with first-line chemoimmunotherapy in US (01JAN2015-31DEC2022). Clinical data, including physician-reported response, were abstracted from medical charts and positron emission tomography-computed tomography (PET-CT) reports at treatment initiation and first response assessment. Lugano 2014 criteria, such as standardized uptake value (SUV) and Deauville score, were extracted and converted to an algorithm (rwLugano). Absence of normal tissue SUV in PET-CT reports was addressed using published values for background, mediastinum, and liver to facilitate a real-world Deauville (rwDeauville) score for rwLugano classification. PET-CTs underwent deidentified digital image transfer for BICR using Lugano 2014 criteria by two lymphoma radiologists. Descriptive analyses assessed BICR feasibility and completeness of rwLugano data elements. Results: Six oncology practices, comprising 77 physicians, abstracted medical records and PET-CT reports for 185 patients. PET-CT scans were available for BICR for 174/185 (94%) patients (41% female, mean age 66 years). BICR discordance for Deauville score at baseline and initial response occurred in 7% (12/174) and 29% (50/174), respectively. For response assessment per Lugano, BICR discordance occurred in 9% (16/174), and were further adjudicated by a third party. Lugano discordance was often related to interpretation of marrow activity and/or new disease sites, highlighting variability in lymphoma treatment response assessment. Medical record review found Deauville score charted for 97 (56%) patients at baseline and 145 (83%) at first response. Tumor SUV was complete (174/174, 100%) at baseline and for all patients (63/63, 100%) not charted as CR at first response. Thus, 29 patients (17%) required calculation of rwDeauville at first response, providing data to calculate rwLugano for all patients in the final cohort (174/174, 100%). Conclusions: This study suggests BICR, though resource intensive, can be performed, and the component measures necessary to assess response per Lugano criteria can be obtained using RWD. Accordingly, rwLugano warrants further analysis for concordance with BICR- and physician-reported response.