Automatic lineage assignment of acute leukemias by flow cytometry

Area of Science:

• Hematology
• Immunophenotyping
• Computational Biology

Background:

• Accurate lineage assignment is crucial for diagnosing and treating acute leukemias.
• Current methods for leukemia classification can be labor-intensive and subjective.
• Flow cytometry provides multi-parametric data essential for cellular analysis.

Purpose of the Study:

• To develop and validate an automated method for lineage assignment of acute leukemias.
• To improve the efficiency and objectivity of leukemia classification using flow cytometry data.
• To distinguish between normal hematopoietic cells and various acute leukemia subtypes.

Main Methods:

• Utilized eight list mode data files from FACScan flow cytometry, measuring scatter and fluorescence parameters.
• Employed a nearest neighbor clustering algorithm for independent data file analysis.
• Developed a decision tree approach to identify normal cell populations and classify residual abnormal cells.

Main Results:

• Successfully clustered and associated cell populations across multiple flow cytometry data files.
• Effectively identified and excluded normal cell populations (monocytes, lymphocytes, etc.).
• Accurately classified residual populations into B-lineage ALL, T-lineage ALL, AML, AUL, and B-CLL categories.

Conclusions:

• The developed automated method provides accurate lineage assignment for acute leukemias.
• This approach enhances the diagnostic process by efficiently classifying leukemia subtypes.
• The method demonstrates effectiveness in case studies involving B-lymphoid, T-lymphoid, and Myeloid acute leukemias.