Differential diagnosis of bone marrow failure syndromes guided by machine learning

Author(s):Show less -

Gutierrez-Rodrigues, Fernanda ; Munger, Eric ; Ma, Xiaoyang ; Groarke, Emma M. ; Tang, Youbao ; Patel, Bhavisha A. ; Catto, Luiz Fernando B. ; V. Cle, Diego ; Niewisch, Marena R. ; Alves-Paiva, Raquel M. ; Donaires, Flavia S. ; Pinto, Andre Luiz ; Borges, Gustavo ; Santana, Barbara A. ; McReynolds, Lisa J. ; Giri, Neelam ; Altintas, Burak ; Fan, Xing ; Shalhoub, Ruba ; Siwy, Christopher M. ; Diamond, Carrie ; Raffo, Diego Quinones ; Craft, Kathleen ; Kajigaya, Sachiko ; Summers, Ronald M. ; Liu, Paul ; Cunningham, Lea ; Hickstein, Dennis D. ; Dunbar, Cynthia E. ; Pasquini, Ricardo ; Oliveira, Michel Michels De ; Velloso, Elvira D. R. P. ; Alter, Blanche P. ; Savage, Sharon A. ; Bonfim, Carmem ; Wu, Colin O. ; Calado, Rodrigo T. ; Young, Neal S.

• Document type: Journal article
• Source: Blood; v. 141, n. 17, p. 14-pg., 2023-04-27.

Abstract

The choice to postpone treatment while awaiting genetic testing can result in significant delay in definitive therapies in patients with severe pancytopenia. Conversely, the misdiagnosis of inherited bone marrow failure (BMF) can expose patients to ineffectual and expensive therapies, toxic transplant conditioning regimens, and inappropriate use of an affected family member as a stem cell donor. To predict the likelihood of patients having acquired or inherited BMF, we developed a 2-step data-driven machine-learning model using 25 clinical and laboratory variables typically recorded at the initial clinical encounter. For model development, patients were labeled as having acquired or inherited BMF depending on their genomic data. Data sets were unbiasedly clustered, and an ensemble model was trained with cases from the largest cluster of a training cohort (n = 359) and validated with an independent cohort (n = 127). Cluster A, the largest group, was mostly immune or inherited aplastic anemia, whereas cluster B comprised underrepresented BMF phenotypes and was not included in the next step of data modeling because of a small sample size. The ensemble cluster A-specific model was accurate (89%) to predict BMF etiology, correctly predicting inherited and likely immune BMF in 79% and 92% of cases, respectively. Our model represents a practical guide for BMF diagnosis and highlights the importance of clinical and laboratory variables in the initial evaluation, particularly telomere length. Our tool can be potentially used by general hematologists and health care providers not specialized in BMF, and in under-resourced centers, to prioritize patients for genetic testing or for expeditious treatment. (AU)

• FAPESP's process: 16/12799-1 - Correlation between genotype, and telomere length, and phenotype in telomeropathies
• Grantee: Rodrigo do Tocantins Calado de Saloma Rodrigues
• Support Opportunities: Regular Research Grants
• FAPESP's process: 13/08135-2 - CTC - Center for Cell-Based Therapy
• Grantee: Dimas Tadeu Covas
• Support Opportunities: Research Grants - Research, Innovation and Dissemination Centers - RIDC
• FAPESP's process: 17/09428-4 - Clonal hematopoiesis in patients with telomeric dysfunction
• Grantee: Flávia Sacilotto Donaires Ramos
• Support Opportunities: Scholarships in Brazil - Post-Doctoral
• FAPESP's process: 14/26379-9 - Mutational analysis of GATA2 gene and immunological alterations in patientes with bone marrow failure
• Grantee: Gustavo Borges
• Support Opportunities: Scholarships in Brazil - Master
• FAPESP's process: 16/03620-8 - Evaluation of the mitochondrial function in human patient and mice with telomeropathies
• Grantee: André Luiz Pinto Santos
• Support Opportunities: Scholarships in Brazil - Master