18F-FDG PET/CT Radiomics Features as Predictive Biomarkers in Diffuse Large B-Cell Lymphoma

Abstract Background/Aim: The aim of this study was to evaluate the association between radiomic features (RFs) extracted from baseline 18F-FDG PET/CT (FDG-PET) with progression-free survival (PFS) and overall survival (OS) in diffuse large-B-cell lymphoma (DLBCL) patients eligible for standard-of-care first line chemotherapy.Materials and Methods: Records of 376 newly diagnosed DLBCL patients who underwent baseline FDG-PET at our institution were retrospectively analyzed. Inclusion criteria were: 1) proven DLBCL; 2) FDG-PET as baseline procedure prior to first-line chemotherapy; 3) no lost follow-up. Demographic data, International Prognostic Index (IPI) and conventional PET imaging parameters were collected. FDG-PET RFs were extracted from the lesion showing the highest uptake (target lesion). A radiomic score to predict PFS and OS was obtained by multivariable Elastic Net Cox model. PFS and OS were explored with the Kaplan-Meier method and Cox proportional hazards regression methods. Radiomic univariate model, as well as clinical and clinical-radiomic multivariable models for PFS and OS were obtained. Prediction accuracy was evaluated by C-index.Results: One-hundred and twelve (n=112) patients were included (55 females, 57 males, median age 62.3 years). 9/112(8%) were stage-I, 31/112(27.7%) stage-II, 8/112(7.2%) stage-III and 64/112(57.1%) stage-IV. Median follow-up was 34.7 months (IQR 11.3-66.3). Radiomic score resulted significantly associated with both PFS and OS at Cox multivariable analysis (p<0.001), outperforming conventional PET parameters. C-index (95%CI) for PFS prediction were 0.67(0.58-0.76), 0.81(0.75-0.88) and 0.84(0.77-0.91) for clinical, radiomic and clinical-radiomic model, respectively. C-index for OS were 0.77(0.66-0.89), 0.84(0.76-0.91) and 0.90(0.81-0.98). In the Kaplan-Meier analysis on population stratified according to IPI (low-IPI vs high-IPI), the radiomic score was significant predictor of PFS (p<0.001). Conclusion: The radiomic score was significant predictive biomarker of PFS and OS in DLBCL patients. The extraction of RFs from baseline FDG-PET might be proposed in DLBCL to stratify high-risk vs. low-risk patients of relapse after first-line therapy, especially in low-IPI patients.