A Bayesian joint model for multivariate longitudinal and time-to-event data with application to ALL maintenance...

Area of Science:

• Pediatric Oncology
• Biostatistics
• Translational Cancer Research

Background:

• Acute Lymphocytic Leukemia (ALL) is the most common childhood cancer.
• Standard treatments involve drugs like 6-mercaptopurine (6MP) and methotrexate (MTx).
• Identifying biomarkers for relapse prediction and treatment effectiveness is crucial.

Purpose of the Study:

• To identify longitudinal biomarkers associated with time-to-relapse in pediatric ALL patients.
• To assess the effectiveness of 6MP and MTx treatments in relation to relapse.
• To develop and validate a statistical model for analyzing longitudinal biomarkers and time-to-relapse.

Main Methods:

• A Bayesian joint model was developed, integrating a linear mixed model for longitudinal biomarkers (white blood cell, neutrophil, platelet counts) and a semi-parametric proportional hazards model for time-to-relapse.
• The model analyzed data from 236 pediatric ALL patients treated at TTCRC, Kolkata.
• The model's ability to assess covariate effects and impute missing data was evaluated.

Main Results:

• Neutrophil and platelet counts were significantly associated with time-to-relapse in ALL patients.
• White blood cell count showed no significant association with time-to-relapse.
• Lower 6MP doses and higher MTx doses were linked to a reduced probability of relapse.
• Patients classified as "high-risk" at presentation exhibited the lowest relapse probability.

Conclusions:

• The developed Bayesian joint model effectively identifies biomarkers predicting relapse in pediatric ALL.
• Neutrophil and platelet counts are critical prognostic indicators for ALL relapse.
• Treatment strategies involving adjusted 6MP and MTx dosages may influence relapse rates, with potential implications for high-risk groups.