(017) Feasibility of Using Deidentified Positron Emission Tomography–Computed Tomography (PET-CT) Scans to Support a Blinded Independent Central Review for Lymphoma Treatment Response Assessment using Real-World Data

Background: Clinical outcome evaluation criteria, such as RECIST v1.1 and Lugano 2014, provide a standardized approach to assess treatment efficacy via tumor response in oncology clinical trials. Application of standardized methodologies for outcome classification using real world data are currently lacking.

Objectives: To explore the feasibility of performing a blinded independent central review (BICR) of positron emission tomography–computed tomography (PET-CT) scans in a real-world data (RWD) cohort of adult patients with diffuse large B-cell lymphoma (DLBCL).

Methods: We conducted a retrospective multisite RWD study using Cardinal Health’s Practice Research Network (PRN) to identify adults with DLBCL treated with first line chemoimmunotherapy in US clinical practice (01JAN2015-31DEC2022). Participating PRN sites abstracted clinical data derived from medical records and PET-CT imaging reports recorded at treatment initiation and first response assessment. Deidentified PET-CT digital images were evaluated via BICR using Lugano 2014 criteria by two experienced lymphoma radiologists. BICR discordances were resolved by a 3rd reviewer.

Results: The study included 185 patients with DLBCL across 7 PRN sites. PET-CT reports were readily available for all patients (n=185, 100%); however, procurement of digital PET-CT images for transfer to data-capture software, de-identification, and transmission for BICR posed significant logistical challenges including submitting a request, frequently via paper form, and delays in obtaining scans. Baseline and follow-up PET-CT scans were available for 174 (94%) patients [female n=71 (41%); mean age 66 years] across 6 PRN sites. For the 174 patients in the RWD cohort, BICR classification of initial response per Lugano was discordant for 16 (9%), with all cases achieving concordance after adjudication with a 3rd party. Per BICR, the complete response rate was 82%. Discordances in final Lugano classification were often related to interpretations of bone marrow activity and/or new sites of disease, highlighting the variability within lymphoma treatment response assessment.

Conclusions: This study demonstrated that pairing manual chart abstraction with deidentified digital image transfer for BICR is challenging, yet feasible. Such approaches may reduce outcome misclassification with use of RWD.