Personalized baseline and residual TMTV influence treatment response and outcomes in relapsed/refractory lymphomas: results from the GATA study

Abstract

Purpose

Total metabolic tumor volume (TMTV) at baseline becomes a key biomarker in several lymphoma subtypes. Variability in segmentation methods such as 41%SUVmax and SUVmax > 4 has limited its clinical application. Additionally, immune-checkpoint-inhibitors introduced challenges in response assessment due to pseudoprogression, complicating the use of traditional metrics. This study investigates the prognostic impact of baseline- and residual-TMTV and introduces a novel personalized-liver-based-threshold (pTMTV_liver) to enhance precision in patient stratification.

Methods

We analyzed 91 patients with relapsed/refractory diffuse-large-B-cell lymphoma and follicular lymphoma from the GATA trial, comparing patient’s outcome according to three segmentation methods: TMTV_41%, TMTV₄, and pTMTV_liver. pTMTV_liver used a threshold of 200%SUVmean_liver aligning with 125%SUVmax_liver to enhance standardization and reduce variability.

Results

Baseline-TMTV significantly influenced prognosis with higher TMTV correlating with shorter PFS and OS (p < 0.0001 for all methods). Residual-TMTV, particularly with pTMTV_liver and TMTV₄, stratified no-CMR patients with the lowest predictive errors and better predictive accuracy compared to TMTV_41% Multivariate analyses confirmed residual-pTMTV_liver as superior for prognostic performance for PFS (HR:5.10; C-index:0.724) and OS (HR:4.00; C-index:0.853) compared to TMTV₄ and Deauville Score (DS). The DS alone did not fully capture the heterogeneity of outcomes of DS4-5 patients.

Conclusion

Baseline- and residual-TMTV strongly influence prognosis and response in lymphoma patients. The novel personalized pTMTV_liver method offers improved accuracy of patient stratification, particularly for those with DS4-5, providing more reliable risk assessment. Larger cohorts are needed to validate these findings and optimize residual-TMTV-based clinical applications.