[¹⁸F]FDG PET/CT 用于接受 CAR-T 细胞治疗的弥漫大 B 细胞淋巴瘤患者的预后与毒性预测

Chimeric antigen receptor (CAR) T-cell therapy has been confirmed to benefit patients with relapsed and/or refractory diffuse large B-cell lymphoma (DLBCL). It is important to provide precise and timely predictions of the efficacy and toxicity of CAR T-cell therapy. In this study, we evaluated the value of [ 18 F]fluorodeoxyglucose positron emission tomography/computed tomography ([ 18 F]FDG PET/CT) combining with clinical indices and laboratory indicators in predicting outcomes and toxicity of anti-CD19 CAR T-cell therapy for DLBCL patients.

Thirty-eight DLBCL patients who received CAR T-cell therapy and underwent [ 18 F]FDG PET/CT within 3 months before (pre-infusion) and 1 month after CAR T-cell infusion (M1) were retrospectively reviewed and regularly followed up. Maximum standardized uptake value (SUV max ), total lesion glycolysis (TLG), metabolic tumor volume (MTV), clinical indices, and laboratory indicators were recorded at pre-infusion and M1 time points, and changes in these indices were calculated. Progression-free survival (PFS) and overall survival (OS) were as endpoints. Based on the multivariate Cox regression analysis, two predictive models for PFS and OS were developed and evaluated the efficiency. Pre-infusion indices were subjected to predict the grade of cytokine release syndrome (CRS) resulting from toxic reactions.

For survival analysis at a median follow-up time of 18.2 months, patients with values of international prognostic index (IPI), SUV max at M1, and TLG at M1 above their optimal thresholds had a shorter PFS (median PFS: 8.1 months [IPI 2] vs. 26.2 months [IPI < 2], P = 0.025; 3.1 months [SUV max 5.69] vs. 26.8 months [SUV max < 5.69], P < 0.001; and 3.1 months [TLG 23.79] vs. 26.8 months [TLG < 23.79], P < 0.001). In addition, patients with values of SUV max at M1 and SUV max % above their optimal thresholds had a shorter OS (median OS: 12.6 months [SUV max 15.93] vs. 'not reached' [SUV max < 15.93], P < 0.001; 32.5 months [ SUV max % -46.76] vs. 'not reached' [ SUV max % < -46.76], P = 0.012). Two novel predictive models for PFS and OS were visualized using nomogram. The calibration analysis and the decision curves demonstrated good performance of the models. Spearman's rank correlation (r s ) analysis revealed that the CRS grade correlated strongly with the pre-infusion SUV max (r s = 0.806, P < 0.001) and moderately with the pre-infusion TLG (r s = 0.534, P < 0.001). Multinomial logistic regression analysis revealed that the pre-infusion value of SUV max correlated with the risk of developing a higher grade of CRS (P < 0.001).

In this group of DLBCL patients who underwent CAR T-cell therapy, SUV max at M1, TLG at M1, and IPI were independent risk factors for PFS, and SUV max at M1 and SUV max % for OS. Based on these indicators, two novel predictive models were established and verified the efficiency for evaluating PFS and OS. Moreover, pre-infusion SUV max correlated with the severity of any subsequent CRS. We conclude that metabolic parameters measured using [ 18 F]FDG PET/CT can identify DLBCL patients who will benefit most from CAR T-cell therapy, and the value before CAR T-cell infusion may predict its toxicity in advance.