Objective: To assess the diagnostic accuracy of flow cytometric immunophenotyping in comparison with classic cytomorphology for diagnosing CNS localizations of hematologic malignancies, and to evaluate the implications of CSF pleocytosis and protein content in this context.
Methods: We reviewed the results of diagnostic evaluations of all CSF samples analyzed for localization of a hematologic malignancy between 2001 and 2004 at our center.
Results: A total of 1,054 samples from 219 patients were available for analysis. Sixty patients had a CSF localization diagnosed by positive flow cytometry, cytomorphology, or both. The first sample was positive by flow cytometry in 44 (73%) patients, by cytomorphology in 19 (32%). Four first samples were positive by cytomorphology but negative by flow cytometry. Patients with positive cytomorphology had more frequent clinical symptomatology (95% vs 58%) and CSF pleocytosis (84% vs 25%), and tended to a poorer progression-free survival than patients with positive flow cytometry only. OR for CNS localization in case of CSF pleocytosis was 10.1 (95% CI 4.9 to 20.8); OR for CNS localization in case of elevated protein content was 2.9 (95% CI 1.5 to 5.4). Nevertheless, 26 of 137 (19%) patients with normal cell count and protein concentration had a CNS localization.
Conclusions: The diagnostic value of flow cytometry is more than twice that of cytomorphology. However, cytomorphologic examination of the CSF has additional diagnostic and possibly prognostic value, and should still be performed in conjunction with flow cytometry.