Predicting the risk of ibrutinib in combination with R-ICE in patients with relapsed or refractory DLBCL using...

Area of Science:

• Hematology
• Oncology
• Biostatistics

Background:

• Relapsed or refractory diffuse large B-cell lymphoma (DLBCL) presents treatment challenges due to varied patient outcomes.
• Predictive models are crucial for optimizing therapy in DLBCL.

Purpose of the Study:

• To assess the efficacy of the ibrutinib plus R-ICE regimen in DLBCL.
• To develop and validate explainable machine learning (ML) models for predicting treatment risks and outcomes in DLBCL.

Main Methods:

• Retrospective analysis of 28 DLBCL patients treated with ibrutinib plus R-ICE.
• Development and validation of ML models (CoxBoost+StepCox) using bootstrap methods.
• SMOTE-PSM for class imbalance; comparison with Cox proportional hazards model using decision curve, calibration, and ROC analyses.

Main Results:

• The CoxBoost+StepCox model demonstrated high prognostic performance with C-indices of 0.955 for overall survival (OS) and progression-free survival (PFS).
• Key predictors identified: elevated lactate dehydrogenase (LDH), initial treatment response, time to relapse >12 months, and CD5+ expression.
• CD5+ expression was most predictive for OS, while LDH was most predictive for PFS.

Conclusions:

• Explainable ML models offer high accuracy and clinical utility for predicting DLBCL treatment outcomes.
• These data-driven models show potential for guiding treatment decisions in relapsed/refractory DLBCL.