Prognostic and theranostic 18F-FDG PET biomarkers for anti-PD1 immunotherapy in metastatic melanoma: association with outcome and transcriptomics

• DOI: 10.1007/s00259-019-04411-7
• Publication Date: 2019-07-25T07:24:13Z
• Authors: Romain-David Seban; John S. Nemer; Aurélien Marabelle; Randy Yeh; Eric Deutsch; Samy Ammari; Antoine Moya-Plana; Fatima-Zohra Mokrane; Robyn D. Gartrell; Grace Finkel; Luke Barker; Amélie E. Bigorgne; Lawrence H. Schwartz; Yvonne Saenger; Caroline Robert; Laurent Dercle
• Keywords: Adult; Aged, 80 and over; Biopsy; Programmed Cell Death 1 Receptor; Middle Aged; Prognosis; Magnetic Resonance Imaging; 3. Good health; Young Adult; 03 medical and health sciences; Treatment Outcome; 0302 clinical medicine; Fluorodeoxyglucose F18; Positron-Emission Tomography; Biomarkers, Tumor; Humans; Immunotherapy; Neoplasm Metastasis; Transcriptome; Melanoma; Aged; Follow-Up Studies; Retrospective Studies

An imaging-based stratification tool is needed to identify melanoma patients who will benefit from anti Programmed Death-1 antibody (anti-PD1). We aimed at identifying biomarkers for survival and response evaluated in lymphoid tissue metabolism in spleen and bone marrow before initiation of therapy.This retrospective study included 55 patients from two institutions who underwent 18F-FDG PET/CT before anti-PD1. Parameters extracted were SUVmax, SUVmean, HISUV (SUV-based Heterogeneity Index), TMTV (total metabolic tumor volume), TLG (total lesion glycolysis), BLR (Bone marrow-to-Liver SUVmax ratio), and SLR (Spleen-to-Liver SUVmax ratio). Each parameter was dichotomized using the median as a threshold. Association with survival, best overall response (BOR), and transcriptomic analyses (NanoString assay) were evaluated using Cox prediction models, Wilcoxon tests, and Spearman's correlation, respectively.At 20.7 months median follow-up, 33 patients had responded, and 29 patients died. Median PFS and OS were 11.4 (95%CI 2.7-20.2) and 28.5 (95%CI 13.4-43.8) months. TMTV (>25cm3), SLR (>0.77), and BLR (>0.79) correlated with shorter survival. High TMTV (>25 cm3), SLR (>0.77), and BLR (>0.79) correlated with shorter survival, with TMTV (HR PFS 2.2, p = 0.02, and HR OS 2.5, p = 0.02) and BLR (HR OS 2.3, p = 0.04) remaining significant in a multivariable analysis. Low TMTV and TLG correlated with BOR (p = 0.03). Increased glucose metabolism in bone marrow (BLR) was associated with transcriptomic profiles including regulatory T cell markers (p < 0.05).Low tumor burden correlates with survival and objective response while hematopoietic tissue metabolism correlates inversely with survival. These biomarkers should be further evaluated for potential clinical application.