Differential diagnosis of bone marrow failure syndromes guided by machine learning

Abstract

The choice to postpone treatment while awaiting genetic testing can result in significant delay in definitive therapies in severely pancytopenic patients. Conversely, inherited bone marrow failure (BMF) misdiagnosis 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 two-step data-driven machine-learning model using 25 clinical and laboratory variables typically recorded at the initial clinical encounter. For model's development, patients were labeled as having acquired or inherited BMF depending on their genomic data. Datasets were unbiasedly clustered and an ensemble model was trained with cases from the largest cluster of the 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 was composed of underrepresented BMF phenotypes, and not included in the next step of data modeling due to small sample size. The ensemble model Cluster A-specific 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.

Identification of Adult Patients With Classical Dyskeratosis Congenita or Cryptic Telomere Biology Disorder by Telomere Length Screening Using Age-modified Criteria

“…However, based on recently published data about the predictive value of TL screening for the identification of inherited BMFS, 2 , 46 we feel we can confirm this approach to be justified both in terms of the underlying biology of TBD but also regarding a rational cost-benefit ratio. 47 …”

“…However, based on recently published data about the predictive value of TL screening for the identification of inherited BMFS, 2,46 we feel we can confirm this approach to be justified both in terms of the underlying biology of TBD but also regarding a rational cost-benefit ratio. 47 Particular strength of our study resides in its prospective real-life, multicenter character, the screening by using the validated gold standard for TL determination, namely flow-FISH, and the implementation of the extended screening, which led to the identification of another 7 VUS in TBD-associated genes, although their clinical relevance needs to be further evaluated.…”

Progress in medical therapy in aplastic anemia: why it took so long?

“…For patients up to 40 years old, diepoxybutane (DEB) testing is advised to exclude Fanconi anemia. Additionally, measuring telomere length is useful for identifying those at increased risk of genetic defects in the telomerase gene complex [10]. There is growing evidence that unsuspected germline mutations not classically associated with inherited forms of marrow failure forms could underlie AA pathogenesis, almost exclusively in younger patients [9,11,12].…”

Hematopoietic Stem Cell Transplantation for Myelodysplastic Syndromes

“…20 Exciting progress with machine learning algorithms may be changing the ability to quickly determine the likelihood that presenting cytopenia is related to bone marrow failure syndromes. Gutierrez-Rodrigues et al 21 recently presented findings that a machine learning algorithm could accurately predict whether bone marrow failure syndromes were of immune (79%) or acquired (92%) etiologies in a validation cohort of 127 patients. Previously, Nazha et al 22 also proposed a personalized model to predict leukemia transformation and survival in MDS patients.…”