TOPLINE

A composite score derived from baseline FDG PET-CT imaging — combining total metabolic tumor volume (TMTV), maximal interlesion distance (Dmax), and median peripheral centroid distance (medPCD) — identified patients with primary mediastinal B-cell lymphoma (PMBL) at higher risk for death.

METHODOLOGY

• Researchers conducted a post hoc analysis of baseline 18F-FDG PET-CT scans from the LYSA cohort to assess whether radiomic features predicted survival in patients with newly diagnosed primary mediastinal B-cell lymphoma.
• They included 180 treatment-naive patients (median age, 35.7 years; 103 women) across multiple centers in France who were treated between 2007 and 2017.
• Researchers extracted 12 three-dimensional PET features and selected three nonredundant ones for prognosis: TMTV (overall tumor burden), Dmax (reflecting disease spread), and medPCD (reflecting lesion compactness/massiveness).
• Progression-free and overall survival were assessed over a median follow-up duration of 4.41 years.
• A composite score was created by combining TMTV, Dmax, and medPCD, classifying each feature as high or low and counting high-risk features.

TAKEAWAY

• Higher TMTV and higher Dmax were each associated with shorter progression-free survival and shorter overall survival. Higher medPCD was associated with shorter overall survival (P < .05 for all).
• The composite radiomic score was an independent predictor of overall survival (adjusted hazard ratio [HR], 7.76; P = .011).
• Patients with 2-3 high-risk radiomic features had significantly worse 5-year progression-free and overall survival than those with 0-1 features (P < .01).
• The composite radiomic score showed superior prognostic discrimination compared with individual parameters or traditional clinical factors.

IN PRACTICE

"By integrating measures of tumor burden, spatial dissemination, and structural compactness, this model identifies a subgroup with inferior outcomes at diagnosis and may help inform future trial-based risk-adapted strategies, while also supporting the exploration of deescalation approaches in low-risk patients," the authors of the study wrote.

SOURCE

The study was led by Pierre Decazes, Centre Henri Becquerel in Rouen, France. It was published online on July 15 in Blood Advances.

LIMITATIONS

The retrospective design may have introduced selection bias. Treatment heterogeneity may have affected the results. The study lacked an external validation cohort to assess generalizability.

DISCLOSURES

The study was supported by grants from Ligue Contre le Cancer, groupement d'intérêt public Cancéropôle Nord-Ouest, Force Hémato, and Institut Carnot Consortium for Lymphoma Research. The authors reported no relevant conflicts of interest.

This article was created using several editorial tools, including AI, as part of the process. Human editors reviewed this content before publication.