The predictive power of 18F-FDG PET/CT two-lesions radiomics and conventional models in classical Hodgkin’s Lymphoma: a comparative retrospectively-validated study

• DOI: 10.1007/s00277-025-06190-8
• Publication Date: 2025-01-14T04:53:05Z
• Authors: Elizabeth Katheri...; David Morland; Roberto Gatta; Luca Boldrini; Marco De Summa; Silvia Chiesa; Annarosa Cuccaro; Elena Maiolo; Stefan Hohaus; Salvatore Annunziata
• Keywords: Classical Hodgkin lymphoma; PET-CT

In a previous preliminary study, radiomic features from the largest and hottest lesion in baseline 18F-FDG PET/CT (bPET/CT) of classical Hodgkin's Lymphoma (cHL) predicted early response-to-treatment prognosis. Aim this large retrospectively-validated study is to evaluate predictive role two-lesions radiomics comparison with other clinical conventional models. cHL patients bPET/CT between 2010 2020 were retrospectively included randomized into training-validation sets. Target lesions were: Lesion_A, axial diameter (Dmax); Lesion_B, highest SUVmax. Total-metabolic-tumor-volume (TMTV) was calculated 212 extracted. harmonized using ComBat across two scanners. Outcomes progression-free-survival (PFS) Deauville Score at interim (DS). For each outcome, three models their combinations trained validated: - model "R"; "P"; "C". 197 (training = 118; validation 79): 38/197 (19%) had adverse events 42/193 (22%) DS ≥ 4. training phase, only one feature selected for PFS prediction "R" (Lesion_B F_cm.corr, C-index 66.9%). Best "C" combined stage IPS (C-index 74.8%), while optimal "P" TMTV Dmax 63.3%). After internal validation, "C", "C + R", "R P" R significantly PFS. The best validated R" 66.3%). No prediction. combination showed an added prognostic power cHL. As single models, parameters maintain power, or alone seem be sub-optimal predict survival.