PET-INTEGRATED NOMOGRAMS FOR SURVIVAL PREDICTION IN PATIENTS WITH LARGE B-CELL LYMPHOMA TREATED WITH CAR T-CELL THERAPY
Beatrice Casadei et al
IRCSS Azienda Ospedaliero-Universitaria di Bologna, Bologna, Italy
Background
- CD19-directed chimeric antigen receptor T-cell (CART) therapy is effective in relapsed/refractory large B-cell lymphoma (R/R LBCL), however clinical outcomes remain heterogeneous. Reliable prognostic stratification prior to CART infusion is therefore crucial to optimise patient selection and management.
- ^18F-FDG PET/CT is the standard imaging modality for staging and response assessment in LBCL, providing quantitative metabolic parameters that reflect tumour burden and disease biology. When integrated with clinical and laboratory variables, these parameters may improve prognostic modelling.
- The aim of the study is to develop PET-integrated nomograms for predicting progression-free survival (PFS) and overall survival (OS) in R/R LBCL patients (pts) treated with CART.
Methods
- This post-hoc analysis of a retro-prospective, single-center study included R/R LBCL pts treated with CART in routine clinical practice between August 2019 and November 2023.
- Baseline variables (n=35) included demographics, histology (DLBCL NOS vs high grade B cell lymphoma [HGBCL]), number and response to prior lines of treatment, bridging therapy (BT, yes or no), laboratory parameters (e.g. lactate dehydrogenase [LDH], C-reactive protein [CRP], hemoglobin [Hb], white blood cell [WBC] and lymphocyte [lym] count), and PET metrics (maximum and mean standardized uptake value, metabolic tumor volume [MTV], total lesion glycolysis [TLG], extranodal/bone involvement).
- Variable selection used adaptive LASSO with bootstrap stability selection (500 resamples). Multivariable Cox models were developed for PFS and OS. Model performance was assessed using Harrell’s concordance index (C-index), time-dependent area under the curve (AUC), and calibration plots.
Results
- A total of 141 pts were analysed: 111 (78.7%) had DLBCL NOS and 30 HGBCL. Most pts had stage III/IV disease (73%), received a maximum of 2 previous line of treatment (73.7%), and were refractory to the last therapy (93%). BT was performed in 84% pts. CART products included axi-cel (n=109), tisa-cel (n=30), and liso-cel (n=2). With a median follow-up of 16.6 months, the median PFS and OS were 13.3 and 40.2 months, respectively.
- On univariate analysis, PET-derived parameters (MTV, TLG, bone involvement) and laboratory variables (LDH, CRP, ferritin, Hb, WBC and lym count) were associated with outcomes (p<0.05).
- In the multivariable OS model (C-index of 0.798, 95%CI 0.731–0.856), four independent predictors emerged: Hb (hazard ratio [HR] 0.72, 95%CI 061-084, p<0.001), WBC (HR 1.17, 95%CI 1.09-1.26, p<0.001) and lym_count (HR 0.41, 95%CI 0.22-0.75, p=0.004) as a protective factors, while bone involvement on PET (HR 2.36, 95%CI 1.28-4.36, p=0.006) as an adverse factor. In the PFS model (C-index 0.689, 95%CI 0.623–0.756), Hb remained protective (HR 0.78, 95%CI 0.68-0.89, p<0.001), while receiving BT (HR 2.53, 95%CI 1.14-5.63, p=0.02) was adverse. Time-dependent ROC showed AUCs at 12 months of 0.86 (OS) and 0.71 (PFS).
Conclusion
- We developed PET-integrated nomograms that predict outcomes after CART in R/R LBCL. These models underscore the interplay between host marrow reserve (Hb and lym count) and metabolic tumour distribution, including bone involvement.
- The proposed tools may support trial design and individualized decision-making and warrant prospective multicentre validation.
