Machine learning algorithms show promise for pathogenesis of BMF

A new machine learning algorithm could help some clinicians determine whether a patient has hereditary bone marrow failure (BMF) or immune/acquired bone marrow failure (BMF), according to a study published in . blood. Although the sensitivity of this tool was only about 89%, it may help non-professional patients efficiently perform confirmatory genetic testing.

A major difficulty in treating BMF is that therapeutic decision-making relies on distinguishing between syndrome etiologies. The same treatments for hereditary BMF are not recommended for acquired BMF. Furthermore, genetic testing is not always available, especially in resource-poor settings.

Due to the potential for misdiagnosis and incorrect treatment, new tools for determining the etiology of BMF will be essential. For this study, the researchers used existing patient data to train and test a machine learning algorithm aimed at distinguishing between hereditary and acquired BMF.

keep reading

Overall, data from 359 and 127 patients were used for the algorithm’s training and validation datasets, respectively. In the training and validation sets, 64.6% and 72.4% of patients acquired BMF, respectively, 35.3% and 27.5% inherited BMF, with a median age of 28 and 23 years, and 48% of patients and 46% were women.

The researchers created a model using 25 variables that are frequently recorded during a patient’s first clinical visit. The algorithm fairly classified the case into her one of her two groups. Cluster A consisted primarily of immune-mediated or aplastic anemia, and cluster B consisted of a small group of his BMF phenotypes, which are underestimated. However, the latter cluster was not included in subsequent analyses, due to its small size.

Analysis of the model assignments showed that the algorithm was accurate in predicting BMF etiology in 89% of cases. Specifically, the model accurately predicted hereditary BMF in 79% of cases and the likelihood of immunity in 92% of cases.

“This practical tool is also part of our ongoing research as we continue to accumulate models to increase the number of cases and further refine our predictions.” [inherited BMF syndrome] This is an underestimated case in the current cohort, especially pediatric cases,” the authors said in their report.

Disclosure: Some study authors have declared affiliations with biotechnology, pharmaceutical, or device companies. Please refer to the original reference for a complete list of disclosures.