Risk prediction in diffuse large B-cell lymphoma improves when combining baseline PET features with interim PET...

Area of Science:

• Oncology
• Radiology
• Medical Imaging

Background:

• Accurate risk stratification in diffuse large B-cell lymphoma (DLBCL) is vital for timely treatment escalation.
• The International Metabolic Prognostic Index (IMPI) shows promise but may be enhanced by additional predictive factors.
• Radiomic features and early treatment response could refine prognostic models.

Purpose of the Study:

• To develop and evaluate dynamic risk prediction models for DLBCL treatment failure.
• To assess the added value of radiomic features and interim PET (iPET) response beyond existing prognostic indices.
• To compare novel models against the IMPI and ClinicalPET models.

Main Methods:

• Utilized data from the PETRA database of newly diagnosed DLBCL patients treated with R-CHOP.
• Developed Cox regression models to identify optimal transformations of Dmaxbulk, SUVpeak, and ΔSUVmax.
• Validated models using cross-validated c-index for discrimination and Akaike Information Criterion (AIC) for model fit.
• Compared baseline and interim PET-based risk models.

Main Results:

• The best baseline model included age, metabolic tumor volume (MTV), and Dmaxbulk (c-index 0.70).
• Incorporating iPET response (ΔSUVmax) significantly improved outcome prediction (c-index 0.74).
• The dynamic model demonstrated wider segregation of Kaplan-Meier curves and better risk classification accuracy.

Conclusions:

• A dynamic risk assessment incorporating baseline and interim PET features offers superior outcome prediction in DLBCL.
• Radiomic features and early treatment response provide valuable prognostic information.
• These findings support the development of more personalized treatment strategies for DLBCL.