An Algorithm for the Efficient Flow Cytometric Diagnosis of Acute Leukemia
N Haycocks, L Lawrence, JW Cain, XF Zhao. University of Maryland Medical Center, Baltimore, MD
Background: Flow cytometry is an indispensible tool for the diagnosis of acute leukemias. Since many immunological markers are employed in the analysis, its application is both expensive and time consuming. Despite this, clinical treatment depends on the immunophenotype of leukemia cells, in particular, the differentiation of acute myeloid leukemia (AML) from acute lymphoblastic leukemia (ALL). In an era of ever increasing demand for testing capabilities with static or declining funding, effectively triaging specimens to select more tailored panels could reduce unnecessary reagent use without compromising diagnosis. Having observed different patterns between AML and ALL when plotted by CD45 intensity versus side scatter (SCC), we set out to establish an efficient algorithm for distinguishing AML from ALL with limited immunological markers.
Design: We retrospectively reviewed the flow cytometry of 104 consecutive new acute leukemias in the archives of our Flow Cytometry Laboratory, dating from April, 2008 to October, 2009. Distinct CD45/SCC patterns were identified for AML and ALL. Using three categories – AML, ALL and indeterminate – two hematopathologists blind to the diagnosis independently reviewed all the cases. The diagnoses rendered by this approach were compared to the original diagnosis for each case, and the results were tabulated for analysis.
Results: The 104 new leukemias consist of 79 AML and 25 ALL. Our criteria alone correctly identified 96 acute leukemias (92.3%). Four ALL were misinterpreted as AML, but 2 AML was considered ALL (5.8%). Two cases (2%) were considered indeterminate. The interobserver consistency between the two hematopathologists is 97.1%. Based on these findings, we developed a stepwise algorithm for the analysis of acute leukemia: New leukemia → CD45/SSC → 1) AML → detailed AML studies; 2) ALL → detailed ALL studies; 3) Indeterminate → detailed ALL & AML studies. Using this approach, retrospectively we would have saved 1152 immunological markers (both reagent and labor) for analysis of the 104 leukemias.
Conclusions: Based on the fact that AML and ALL usually recapitulate the differentiation of their normal counterparts, we were able to identify the distinct patterns of AML and ALL. Our algorithm proved to be efficient and reliable. An enormous amount of money and labor will be saved without compromising the diagnostic accuracy and timeliness. This approach is particularly practical for the small flow cytometry laboratory operating under uncertain economic conditions, and anticipating future trends in healthcare reform.
Monday, March 22, 2010 9:30 AM
Poster Session I Stowell-Orbison/Surgical Pathology/Autopsy Awards Poster Session # 183, Monday Morning