Beyond target lesions: Prognostic value of longitudinal AI-derived whole-body [<sup>1</sup>⁸F]FDG PET/CT metrics in metastatic melanoma.
Authors
Affiliations (7)
Affiliations (7)
- Clinical Cooperation Unit Nuclear Medicine, German Cancer Research Center - Deutsches Krebsforschungszentrum (DKFZ), Im Neuenheimer Feld 280, 69210, Heidelberg, Germany. [email protected].
- School of Electrical and Computer Engineering, National Technical University of Athens, Athens, Greece.
- Department of Biostatistics, German Cancer Research Center - Deutsches Krebsforschungszentrum (DKFZ), Heidelberg, Germany.
- Clinical Cooperation Unit Nuclear Medicine, German Cancer Research Center - Deutsches Krebsforschungszentrum (DKFZ), Im Neuenheimer Feld 280, 69210, Heidelberg, Germany.
- Department of Diagnostic and Interventional Radiology, University Clinic Heidelberg, Heidelberg, Germany.
- Medical Faculty Heidelberg, Department of Dermatology, Heidelberg University, Heidelberg, Germany.
- National Center for Tumor Diseases (NCT), NCT Heidelberg, a partnership between, DKFZ and University Hospital Heidelberg, Heidelberg, Germany.
Abstract
To investigate the prognostic value of an artificial intelligence (AI)-based semi-automated tool for longitudinal whole-body quantification of total metabolic tumor volume (TMTV) and total lesion glycolysis (TLG) on [<sup>1</sup>⁸F]FDG PET/CT in patients with metastatic melanoma undergoing immune checkpoint inhibitor (ICI) therapy, and to assess its prognostic relevance alongside established PET-based metabolic response criteria. Forty-three patients with unresectable metastatic melanoma treated with ICIs underwent [<sup>1</sup>⁸F]FDG PET/CT at baseline (n = 43), after two cycles (interim; n = 41), and after four cycles of therapy (late; n = 43). Whole-body tumor segmentation was performed using a previously validated AI-based framework combining ensemble unsupervised segmentation and deep representation learning, followed by expert review. TMTV and TLG were calculated for each time point. Overall survival (OS) was analyzed using Kaplan-Meier estimates, log-rank tests, and Cox proportional hazards regression. Multivariable models included LDH, AJCC stage, and ECOG performance status. Metabolic response was additionally assessed using EORTC, PERCIST, PERCIMT, imPERCIST5, and iPERCIST criteria. Median follow up [95% CI] of the patient cohort from the date of baseline PET/CT was 97.8 months [90.1-134.5 months]. Automated whole-body segmentation and volumetric quantification were feasible in all examinations. In univariable Cox analysis, baseline TMTV and TLG showed non-significant associations with OS. At interim and late PET/CT, higher TMTV and TLG were significantly associated with worse OS (interim: p = 0.04 for both; late: p = 0.03 for TMTV, p = 0.02 for TLG). Median-based dichotomization confirmed shorter OS in patients with elevated volumetric parameters with a statistically significant association for TLG (logrank p = 0.021) and a trend toward significance for TMTV (logrank p = 0.10) at baseline. At interim imaging, both TMTV (logrank p = 0.01) and TLG (logrank p = 0.04) were significantly associated with worse OS, with the strongest associations observed at late follow-up (both logrank p = 0.009). In multivariable analysis, elevated late TMTV and TLG remained independently associated with poorer OS, alongside increased LDH, whereas lower ECOG performance status independently predicted improved survival. Among the applied metabolic response criteria, at both interim and late PET/CT, survival curves demonstrated a non-significant trend toward longer OS in responders (CMR + PMR) compared with non-responders (SMD + PMD), with this trend being most pronounced for PERCIMT. AI-based longitudinal whole-body quantification of TMTV and TLG on [<sup>1</sup>⁸F]FDG PET/CT appears to provide independent prognostic information in metastatic melanoma patients undergoing immunotherapy. In particular, volumetric PET metrics derived from late follow-up imaging demonstrate robust association with OS and may outperform conventional response criteria. AI-driven whole-body tumor burden assessment may enhance objective risk stratification and support treatment monitoring in the era of immunotherapy.