A predictive model for bone marrow disease in cytopenia based on noninvasive procedures

Area of Science:

• Hematology
• Oncology
• Genetics

Background:

• Bone marrow examination is crucial for diagnosing cytopenia.
• Unexplained cytopenia necessitates thorough investigation for hematologic malignancy.
• Next-generation sequencing (NGS) offers potential for non-invasive diagnostic insights.

Purpose of the Study:

• To evaluate NGS's utility in predicting bone marrow malignancy without invasive biopsy.
• To develop a predictive model for bone marrow disease using mutational data and clinical factors.
• To assess if NGS can obviate the need for bone marrow specimens in specific patient groups.

Main Methods:

• Analysis of mutations in 508 patients with cytopenia using targeted NGS.
• Division of patients into discovery (n=340) and validation (n=168) cohorts.
• Multivariable logistic regression to identify mutation-morphology associations.

Main Results:

• Mutations were found in 53% of patients; abnormal bone marrow morphology in 37%.
• Specific mutations (TET2, SF3B1, U2AF1, TP53, RUNX1) strongly correlated with abnormal morphology.
• A predictive model using NGS and clinical data achieved 100% sensitivity in identifying patients without abnormal morphology in the validation cohort.

Conclusions:

• NGS combined with clinical data effectively predicts abnormal bone marrow morphology in unexplained cytopenia.
• This integrated approach can guide decisions regarding bone marrow biopsy necessity.
• NGS holds promise for refining the diagnostic workup of hematologic disorders.