Baseline [18F] FDG PET biomarkers of host and tumor predict efficacy and CRS in multiple myeloma patients treated with CAR-T cell therapy

Abstract

Abstract Background: Chimeric antigen receptor T-cell (CAR-T) therapy has transformed the treatment landscape for Relapsed/Refractory multiple myeloma (R/R MM), yet clinical outcomes and toxicity remain heterogeneous. Reliable pre-treatment predictors of CAR-T efficacy are limited. While tumor burden is commonly evaluated, the metabolic activity of immune-related organs may reflect host immune status and influence therapeutic outcomes. Methods We retrospectively analyzed baseline [¹⁸F] FDG PET/CT scans from 120 patients with R/R MM treated with BCMA-directed CAR-T therapy. We quantified metabolic activity in both tumor lesions and immune-related organs, including the spleen, liver, and bone marrow. Treatment response was defined by International Myeloma Working Group (IMWG) criteria, and cytokine release syndrome (CRS) was graded according to ASTCT guidelines. A machine learning model was developed to integrate imaging and clinical features for predicting therapeutic response and CRS risk. Results Patients with lower pre-treatment spleen metabolic activity were significantly more likely to achieve deep response (≥VGPR) (OR 3.8, 95% CI: 1.7–8.3; P < 0.001). High metabolic activity in extramedullary lesions-especially those with elevated SUVmax and metabolic tumor volume-was strongly predictive of poor response (OR 0.4, 95% CI: 0.2–0.8; P = 0.007). Higher liver metabolic activity was independently associated with grade ≥2 CRS (OR 4.5, 95% CI: 2.0–10.2; P < 0.001). For CRS, elevated liver glycolysis was associated with increased risk of grade ≥2 CRS (OR 3.2, 95% CI: 1.5–6.9; P < 0.001). The combined model incorporating imaging and clinical features achieved an area under the ROC curve (AUC) of 0.83 for predicting response and 0.79 for predicting CRS severity. Conclusions Metabolic activity in immune-related organs—particularly the spleen and liver—alongside high-risk tumor profiles such as intense extramedullary disease, provides noninvasive insight into both therapeutic efficacy and toxicity in CAR-T-treated multiple myeloma. These findings support the integration of immune-metabolic imaging into pre-treatment assessment to improve risk stratification and guide clinical decision-making. Future studies will validate these biomarkers through correlation with tumor and peripheral immune profiles, including flow cytometric analysis of blood and tissue samples to elucidate microenvironmental mechanisms.