Semiquantitative PET Parameters Refine Prognosis in CAR T–Treated Lymphoma After 1 and 3 Months: A Prospective Single-Center Study

Chimeric antigen receptor T-cell (CAR T) therapy has shown remarkable efficacy in treating relapsed or refractory large B-cell lymphoma. However, for nearly half of these patients, the therapy eventually does not achieve durable remission. We investigated whether semiquantitative PET parameters (namely, SUV_max, metabolic tumor volume [MTV], and total lesion glycolysis [TLG]) could improve risk stratification 1 mo (PET1m) and 3 mo (PET3m) after CAR T infusion. <b>Methods:</b> In this prospective, single-center cohort study, patients with large B-cell lymphoma received axicabtagene ciloleucel or tisagenlecleucel. [¹⁸F]FDG PET/CT scans were acquired at baseline, 1 mo, and 3 mo after infusion. MTV and TLG were calculated using a threshold SUV_max of 4 or greater. Patients were followed for overall survival (OS), progression-free survival (PFS), and duration of response (DoR). The imaging assessment was based on the Lugano recommendation for response assessment. Prognostic factors were identified using univariate and multivariate Cox regression. <b>Results:</b> Sixty-one patients were enrolled, with a median follow-up of 18 mo. Twenty-eight (46%) patients died. Kaplan-Meier analysis with log-rank tests indicated a significant association of elevated Deauville score (DS), SUV_max, MTV, and TLG with OS (all <i>P</i> < 0.05). DS cutoff was arbitrarily fixed at 4. The optimal SUV_max, MTV, and TLG cutoffs at PET1m were 9.1, 60.8, and 97.0, respectively; whereas at PET3m, they were 6.3, 120.1, and 436.9, respectively. Patients with an SUV_max of 6.3 or greater at PET3m had an 8-fold increase in risk of death (hazard ratio [HR], 8.15; 95% CI, 2.81-23.6; <i>P</i> < 0.01) compared with those below this cutoff. Similarly, higher MTV (≥120.1) at PET3m yielded a nearly 10-fold risk (HR, 9.87; 95% CI, 3.65-26.7; <i>P</i> < 0.01). DS, SUV_max, MTV, and TLG at both PET1m and PET3m were associated with OS and PFS (all <i>P</i> < 0.05), whereas PET3m parameters also correlated with DoR (<i>P</i> < 0.05). Harrell C-index values were higher for PET3m measures than for PET1m, though differences were not statistically significant (<i>P</i> > 0.05). On multivariable analysis, older age (HR, 1.10), bridging therapy (HR, 10.91), elevated lactate dehydrogenase (HR, 6.43), increased fibrinogen (HR, 5.27), and higher SUV_max at PET3m (HR, 11.03) independently predicted poorer OS. There were no significant associations between SUV_max, MTV, and TLG with CAR T-related toxicities. <b>Conclusion:</b> Semiquantitative PET parameters, such as SUV_max, MTV, and TLG, at 1 mo and 3 mo after CAR T-cell therapy correlate significantly with OS, PFS, and DoR. [¹⁸F]FDG PET/CT at 3 mo may offer slightly stronger prognostic discrimination, but both time points can be used for early risk stratification.