Utility of Cerebrospinal Fluid in the Diagnosis of Non-Hodgkin Lymphoma
Meredith E Pittman, Ling Chen, John L Frater, Anjum Hassan, TuDung T Nguyen, Friederike Kreisel. Washington University in Saint Louis, Saint Louis, MO
Background: Flow cytometric analysis (FCA) of cerebrospinal fluid (CSF) increases the likelihood of detecting central nervous system involvement by lymphoma when compared to cytopathologic examination alone. It has become routine practice, however, to submit CSF for FCA even when clinical suspicion for lymphoma is low. The purpose of this study is to assess the value of FCA with cytology as a screening tool in the work-up of patients with neurologic symptoms.
Design: Between June 2001 and June 2011, 368 patients had 496 CSF samples submitted for FCA and cytologic review. Electronic medical records were reviewed for patient symptomatology, history of malignancy, brain imaging, FCA results, cytology results, brain biopsy, and clinical follow-up, including autopsy results.
Results: Patients with neurologic symptoms but no history of hematologic malignancy made up 62% of patients. Thirty percent of patients had a history of lymphoma, and the remaining 8% were immunosuppresed. Most patients (93%) had brain imaging prior to lumbar puncture, and 52% of those had significant neurologic findings by imaging. Overall, FCA was positive in 22 cases (4%), negative in 280 (57%), and indeterminate in 194 (39%). Of the 22 cases positive by FCA, 12 (55%) were also positive by cytologic examination, but 9 (45%) were negative. Confirmatory brain biopsies were performed on 64 patients (17%). With open biopsy as the gold standard, the sensitivity and specificity of combined FCA and cytology are 24% and 96%, respectively. If one assumes that all positive FCA are true positives, as is done when making treatment decisions, the sensitivity/specificity of combined FCA and cytology increase to 64% and 96%, respectively. Patients with a positive FCA or brain biopsy were more likely to have either a history of hematologic malignancy or findings on brain imaging (p<0.05), making the positive predictive value of FCA in this population >95% despite the low sensitivity.
Conclusions: Although flow cytometric analysis of CSF is very specific for CNS lymphoma, the sensitivity of this test is too low for routine CSF screening of patients with neurologic symptoms. In an era of rising health care costs that necessitates judicious use of resources, FCA is not an appropriate initial test if the clinical suspicion for CNS lymphoma is low, and its use should be reserved for ancillary testing in the appropriate clinical setting.
Monday, March 19, 2012 9:30 AM
Poster Session I Stowell-Orbison/Surgical Pathology/Autopsy Awards Poster Session # 51, Monday Morning