Combined Multiple Clusterings on Flow Cytometry Data to Automatically Identify Chronic Lymphocytic Leukemia

摘要

We described a combined multiple clustering approach to automatically identify chronic lymphocytic leukemia neoplastic population by flow cytometry immunophenotyping. Flow cytometry data from various specimens were preprocessed by data cross-linking and subset selection before undergoing subspace and consensus clustering. This approach was implemented as a Server-side application, with results comparable to those performed by manual gating on commercial software.