Deep learning-based image analysis of pretreatment FDG-PET/CT predicts CAR-T cell treatment outcome at month-12 for patients with Relapsed/Refractory large B-cell lymphomas
Analysis shows deep learning predicts treatment outcome in relapsed large B-cell lymphomas, indicating imaging may enhance prognostic evaluation using standard indices.
Key Points
DL-based prediction of CAR-T response status identified 77% of responders at Month-12, indicating substantial prognostic value.
Image analysis was performed on 66 evaluable patients with relapsed large B-cell lymphomas, utilizing advanced deep-learning methods.
Comparisons against serum LDH and International Prognostic Index highlighted varying sensitivity and specificity for predicting outcomes.
This predictive approach emphasizes the need for refinement and integration of clinical factors to optimize CAR-T therapy outcomes.