Radiomic Features Prognosticate Treatment Response in CAR-T Cell Therapy

Area of Science:

• Oncology
• Radiology
• Immunotherapy

Background:

• Diffuse large B-cell lymphoma (DLBCL) is the most common aggressive non-Hodgkin lymphoma.
• Predicting response to advanced therapies like axicabtagene ciloleucel (axi-cel) remains a challenge.
• Chimeric antigen receptor (CAR) T-cell therapy offers a promising treatment avenue for relapsed/refractory DLBCL.

Purpose of the Study:

• To develop a radiomics signature for prognostication of axi-cel treatment response in DLBCL patients.
• To evaluate the prognostic ability of radiomic features derived from baseline PET/CT scans.
• To correlate radiomic metrics with overall survival (OS) and progression-free survival (PFS).

Main Methods:

• A cohort of 155 relapsed/refractory DLBCL patients treated with axi-cel was analyzed.
• Radiomics features (size, shape, texture) were extracted from baseline PET/CT scans of dominant lesions.
• Principal component (PC) analysis was used to reduce feature dimensionality, and prognostic ability was assessed.

Main Results:

• Radiomic PCs, particularly those related to shape and size, were prognostic for OS on both CT and PET scans.
• Metabolic tumor volume (MTV) also demonstrated prognostic value.
• Radiomic PCs showed moderate to weak correlations with MTV.

Conclusions:

• Radiomics-based principal components derived from size and shape metrics can prognosticate treatment response to CAR-T therapy in DLBCL.
• This radiomic signature may aid in identifying patients likely to benefit from axi-cel therapy.
• Further validation is warranted to integrate radiomics into clinical decision-making for DLBCL treatment.