Dynamic Relapse Prediction by Peripheral Blood WT1mRNA after Allogeneic Hematopoietic Cell Transplantation for
Soichiro Nakako1, Hiroshi Okamura2, Isao Yokota3
1Department of Hematology, Osaka Metropolitan University Graduate School of Medicine, Osaka, Japan.
Transplantation and Cellular Therapy
|August 15, 2024
Summary
A new dynamic model accurately predicts relapse after allogeneic hematopoietic cell transplantation (allo-HCT) for acute myeloid leukemia (AML) and myelodysplastic syndrome (MDS) using Wilms' tumor 1 messenger RNA (WT1mRNA) levels, enabling timely interventions.
Area of Science:
- Hematology
- Oncology
- Transplantation Medicine
Background:
- Existing relapse prediction models for allogeneic hematopoietic cell transplantation (allo-HCT) lack dynamic updating capabilities.
- These static models are insufficient for guiding post-transplant treatment adjustments and preemptive interventions.
Purpose of the Study:
- To develop and evaluate a dynamic relapse prediction model for AML and MDS patients post-allo-HCT.
- To incorporate time-dependent Wilms' tumor 1 messenger RNA (WT1mRNA) levels for improved predictive accuracy.
Main Methods:
- Retrospective analysis of 238 allo-HCT cases for AML and MDS.
- Utilized the landmarking supermodel approach with pre- and post-transplant WT1mRNA levels as time-dependent covariates.
- Compared dynamic model performance against conventional models using time-dependent ROC curves.
Main Results:
- The dynamic model incorporating WT1mRNA kinetics demonstrated superior predictive performance (AUC=0.89) compared to static models (AUC=0.73) and non-kinetic dynamic models (AUC=0.87).
- Relapse probability increased significantly around 90 days prior to the event.
- A user-friendly web application was developed for real-time predictions.
Conclusions:
- Dynamic WT1mRNA monitoring offers a powerful tool for personalized relapse prediction after allo-HCT.
- This model supports timely clinical decision-making and intervention strategies for AML and MDS patients.
- The developed model enhances patient management through objective, real-time relapse forecasting.
