A modified single-stacked ensemble framework for prognostic analysis of leukemia relapse
Bone marrow transplant (BMT) is a life-saving treatment for acute leukemia, though relapse remains a risk due to donor compatibility, clinical variability, and healthcare resource limitations. This chapter introduces a modified single-stacked ensemble prognostic algorithm optimized for interpretabil...
Saved in:
| Main Author: | |
|---|---|
| Other Authors: | |
| Format: | Book Chapter |
| Language: | en |
| Published: |
IGI Global
2026
|
| Subjects: | |
| Online Access: | https://umpir.ump.edu.my/id/eprint/47627/1/Book%20Chapter%20%28Biomedical%20Data%20Science%29.pdf https://umpir.ump.edu.my/id/eprint/47627/ https://doi.org/10.4018/979-8-2600-0986-4.ch004 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | Bone marrow transplant (BMT) is a life-saving treatment for acute leukemia, though relapse remains a risk due to donor compatibility, clinical variability, and healthcare resource limitations. This chapter introduces a modified single-stacked ensemble prognostic algorithm optimized for interpretability, transparency, and reliability, especially for small and imbalanced datasets. While its predictive performance is average, the algorithm identifies key relapse-related covariates, including acute GVHD time, acute and chronic GVHD occurrence, platelet recovery time and status, and French-American-British (FAB) classification. By embedding structured statistical modeling within an AI framework, the approach integrates clinical insight with algorithmic reasoning. It promotes e-collaboration by enabling clinicians to engage with AI outputs, fostering trust and transparency. This study contributes to explainable AI in hematology, supports clinical decision-making, enhances patient care, and aligns with United Nations Sustainable Development Goal 3 (Good Health and Well-Being). |
|---|