A lasso and random forest model using flow cytometry data identifies primary myelofibrosis

Area of Science:

• Hematology
• Oncology
• Immunophenotyping

Background:

• Classical Philadelphia-negative myeloproliferative neoplasms (MPNs), including essential thrombocythemia (ET), polycythemia vera (PV), and primary myelofibrosis (PMF), are challenging to differentiate using morphology and molecular markers alone.
• Accurate diagnosis and differentiation are crucial for guiding appropriate treatment strategies in MPN patients.

Purpose of the Study:

• To clarify the application of flow cytometry in the diagnosis and differential diagnosis of classical Philadelphia-negative MPNs.
• To identify specific flow cytometry markers and models for distinguishing between ET, PV, and PMF subtypes.

Main Methods:

• Retrospective analysis of immunophenotypes, clinical characteristics, and laboratory findings from 211 Ph-negative MPN patients (ET, PV, pre-PMF, overt PMF) and 47 controls.
• Utilized lasso and random forest models to identify key variables for PMF diagnosis.
• Employed classification and regression tree models to differentiate between ET and pre-PMF.

Main Results:

• PMF showed distinct differences from ET and PV in white blood cells, hemoglobin, peripheral blood blast cells, abnormal karyotype, and WT1 gene expression.
• PMF differed from controls in CD34+ cells, granulocyte and monocyte phenotypes, plasma cell percentage, and dendritic cells, with a significantly lower plasma cell percentage.
• A five-variable flow cytometry panel identified PMF with 90% sensitivity and specificity. A model using CD34+CD38- cells and platelet counts distinguished ET from pre-PMF with 94.3% and 83.9% accuracy, respectively.

Conclusions:

• Flow immunophenotyping is a valuable tool for diagnosing PMF and differentiating it from ET and PV.
• Specific flow cytometry markers and models can effectively distinguish between ET and pre-PMF, aiding in treatment decisions.
• This approach enhances diagnostic accuracy for classical Ph-negative MPNs, improving patient management.