All content on this site is intended for healthcare professionals only. By acknowledging this message and accessing the information on this website you are confirming that you are a Healthcare Professional.
The GvHD Hub uses cookies on this website. They help us give you the best online experience. By continuing to use our website without changing your cookie settings, you agree to our use of cookies in accordance with our updated Cookie Policy
Introducing
Now you can personalise
your GvHD Hub experience!
Bookmark content to read later
Select your specific areas of interest
View content recommended for you
Find out moreThe GvHD Hub website uses a third-party service provided by Google that dynamically translates web content. Translations are machine generated, so may not be an exact or complete translation, and the GvHD Hub cannot guarantee the accuracy of translated content. The GvHD Hub and its employees will not be liable for any direct, indirect, or consequential damages (even if foreseeable) resulting from use of the Google Translate feature. For further support with Google Translate, visit Google Translate Help.
The GvHD Hub is an independent medical education platform, sponsored by Medac and supported through grants from Sanofi and Therakos. The funders are allowed no direct influence on our content. The levels of sponsorship listed are reflective of the amount of funding given. View funders.
Machine learning-based analysis of the impact of donor characteristics on post-HCT outcomes
By Ella Dixon
Bookmark this article
|
In patients undergoing HCT, the optimal donor choice has historically been an HLA-matched sibling donor; however, this is only typically available for up to 30% of patients.1 Prioritization of a donor in the absence of a matched sibling is based on HLA-matching and non-HLA factors including donor age, CMV matching, sex matching, and donor parity.1
A study by the Center for International Blood & Marrow Transplant Research (CIBMTR), published in Blood Advances, employed a novel statistical approach using advanced machine learning (NFT BART) to analyze the impact of donor characteristics on post-HCT outcomes, with the aim of improving donor selection.1 A total of 11,818 patients, with a median age of 57 years, who received allo-HCT from HLA-matched (8/8) unrelated donors between 2016 and 2019 were included in the study. The NFT-BART model was trained using ~85% of the study population and validated using the remaining population.1 |
| Key learnings |
| In this observational study, use of donors aged 18–30 years offered optimal outcomes, with a significant reduction in OS observed with donors aged ≥31 years. Improved EFS was found with the use of donors aged 18 years compared with donors aged ≥32 years. |
| A clinically meaningful impact was defined as >1% difference in predicted outcome (OS or EFS) at 3 years; donor CMV, parity, HLA-DQB1 and HLA-DPB1 T cell epitope matching were not found to have a meaningful impact. |
| Male donors were associated with meaningful improvement in EFS compared with female donors; however, sex had no significant effect on OS, allowing clinicians flexibility in donor selection without compromising survival outcomes. |
| To improve outcomes following HCT, the first available donor aged 18–30 years should be used, with a male donor preferred if multiple donors are available. |
Abbreviations: CMV, cytomegalovirus; EFS, event-free survival; HCT, hematopoietic cell transplantation; HLA, human leukocyte antigen; NFT-BART, Nonparametric Failure Time Bayesian Additive Regression Trees; OS, overall survival.
- Spellman SR, Sparapani R, Maiers M, et al. Novel machine learning technique further clarifies unrelated donor selection to optimize transplantation outcomes. Blood Adv. 2024;online ahead of print. DOI: 1182/bloodadvances.2024013756
More about...
Your opinion matters
8 votes - 27 days left ...
Related articles
Newsletter
Subscribe to get the best content related to GvHD delivered to your inbox