Imaging Biomarkers to Predict Outcomes in Patients with Large B-Cell Lymphoma with a Day 28 Partial Response By PET/CT Imaging Following CAR-T Therapy

Introduction: CD19 Chimeric Antigen Receptor T-cell (CAR-T) therapy is now a commonly used treatment for relapsed/refractory (R/R) Large B-cell Lymphoma (LBCL). However, predictors of long-term response remain poorly defined. In particular, partial response (PR) at first tumor assessment at Day 28 (D28) is a source of uncertainty both for clinicians and patients. In the pivotal CAR-T trials for LBCL, approximately half of these patients eventually achieved a complete remission (CR), while the other half experienced progressive disease (PD) (Neelapu et al, NEJM 2017). Herein, we present real-world data on 24 patients achieving a PR on D28 by PET/CT imaging following CAR-T therapy for R/R LBCL. We explore whether differences between baseline and D28 PET/CT imaging might predict progression free survival (PFS), overall survival (OS), best overall response rate (B-ORR), or last overall response rate (L-ORR). Methods: Out of 75 patients receiving CAR-T therapy at a single institution, we retrospectively identified and reviewed 24 (32%) as achieving a PR on D28. Two independent radiologists collected baseline (pre-CAR-T therapy) and D28 PET/CT Standard Uptake Value (SUV) max and Total Tumor Metabolic Volume (TMV, in cm 3) using ROVER software. The Intraclass Correlation Coefficients (ICC) as a measure of absolute agreement between two readers was 0.99 for SUV the 0.97 for TMV. There was a strong absolute agreement between the two radiologists. For simplicity of data interpretation and given this concordance we present the results of one reviewer. Univariable Cox regression model was used to calculate PFS and OS. All statistical tests were 2-sided and conducted at the 0.05 level of significance. Results: Of the 24 patients with PR on D28 PET/CT, median follow-up time was 1.9 years with 17 patients (71%) still alive at last follow-up (see Fig 1a). Median age was 51 years-old, 46% were female, 66.7% had stage III/IV disease, all patients had ECOG ≤2, 58% received bridging therapy, and half had ≥3 lines of prior therapy (see Table 1A). Results of the univariable Cox regression model revealed that a lower D28 SUV max (p=0.004), lower TMV at both baseline (p=0.03), and at D28 (p=0.01) may be predictive of better OS. Longer PFS was found with lower D28 SUV max (p=0.002) and lower TMV at both baseline (p=0.01), and D28 (p=0.04). In analysis of B-ORR achieved by PET/CT, half of patients in PR at D28 ultimately achieved a CR (see Table 1B). OS was significantly lower in t