Z Laura Tabatabai, Manon Auger, Rhona J Souers, Lisa Teot, Diane D Davey1. 1. From the Department of Pathology, University of California, San Francisco (Dr Tabatabai); the Department of Pathology, McGill University Health Sciences Center, Montreal, Quebec, Canada (Dr Auger); the Department of Biostatistics, College of American Pathologists, Northfield, Illinois (Ms. Souers); the Department of Pathology, Boston Children's Hospital, Boston, Massachusetts (Dr Teot); and the Department of Clinical Sciences, University of Central Florida, Orlando (Dr Davey).
Abstract
CONTEXT: - Cerebrospinal fluid cytology is a critical diagnostic tool for the diagnosis of many conditions affecting the central nervous system. OBJECTIVE: - To assess the performance characteristics of cerebrospinal fluid cytology samples by evaluating participant interpretations within the College of American Pathologists Nongynecologic Cytopathology Education program. DESIGN: - Participant interpretations (N = 46 264) evaluated in the College of American Pathologists Nongynecologic Cytopathology Education Program were examined for concordance with the general category and with the reference diagnosis. Two nonlinear mixed models were used to analyze the concordance rates. RESULTS: - The overall concordance rates for the general category and reference diagnosis were 92.1% and 81.0%, respectively. In the malignant category, the concordance rates with the reference diagnosis were lowest for diagnoses of nonhematopoietic small blue round cell tumors (54.8%) and metastatic malignancy (77.5%); the concordance rate with the reference diagnosis was highest for leukemia/lymphoma (94.0%). In the benign category, the concordance rate was lowest for normal cerebrospinal fluid reference diagnoses (58.6%), followed by acute and chronic inflammation (64.6%), fungal infection (80.8%), and macrophages (85.3%). Significant differences in concordance were uncovered when performance was evaluated by participant type and stain technique. Leukemia/lymphoma was the most common diagnosis for misclassified nonhematopoietic small blue round cell tumor cases and negative or inflammatory cerebrospinal fluid cases. CONCLUSIONS: - This study illustrates the difficulties in achieving accurate diagnoses from cerebrospinal fluid specimens, particularly for nonhematopoietic small blue round cell tumors and normal and inflammatory cerebrospinal fluid specimens.
CONTEXT: - Cerebrospinal fluid cytology is a critical diagnostic tool for the diagnosis of many conditions affecting the central nervous system. OBJECTIVE: - To assess the performance characteristics of cerebrospinal fluid cytology samples by evaluating participant interpretations within the College of American Pathologists Nongynecologic Cytopathology Education program. DESIGN: - Participant interpretations (N = 46 264) evaluated in the College of American Pathologists Nongynecologic Cytopathology Education Program were examined for concordance with the general category and with the reference diagnosis. Two nonlinear mixed models were used to analyze the concordance rates. RESULTS: - The overall concordance rates for the general category and reference diagnosis were 92.1% and 81.0%, respectively. In the malignant category, the concordance rates with the reference diagnosis were lowest for diagnoses of nonhematopoietic small blue round cell tumors (54.8%) and metastatic malignancy (77.5%); the concordance rate with the reference diagnosis was highest for leukemia/lymphoma (94.0%). In the benign category, the concordance rate was lowest for normal cerebrospinal fluid reference diagnoses (58.6%), followed by acute and chronic inflammation (64.6%), fungal infection (80.8%), and macrophages (85.3%). Significant differences in concordance were uncovered when performance was evaluated by participant type and stain technique. Leukemia/lymphoma was the most common diagnosis for misclassified nonhematopoietic small blue round cell tumor cases and negative or inflammatory cerebrospinal fluid cases. CONCLUSIONS: - This study illustrates the difficulties in achieving accurate diagnoses from cerebrospinal fluid specimens, particularly for nonhematopoietic small blue round cell tumors and normal and inflammatory cerebrospinal fluid specimens.
Authors: Celina M Abreu; Rebecca T Veenhuis; Lucio Gama; Janice E Clements; Claudia R Avalos; Shelby Graham; Daymond R Parrilla; Edna A Ferreira; Suzanne E Queen; Erin N Shirk; Brandon T Bullock; Ming Li; Kelly A Metcalf Pate; Sarah E Beck; Lisa M Mangus; Joseph L Mankowski; Feilim Mac Gabhann; Shelby L O'Connor Journal: mBio Date: 2019-08-20 Impact factor: 7.867