Accuracy of bronchial brush and wash specimens prepared by the ThinPrep method in the diagnosis of pulmonary small cell carcinoma.

Background: ThinPrep bronchial brush and wash accuracy in the diagnosis of pulmonary small cell carcinoma (pSCCa) and measured as sensitivity, specificity, positive and negative predictive values (PPV and NPV) is incompletely studied or unknown.

Methods: Specimens collected over 5 years from 199 pSCCa and 938 negative (Neg) for pulmonary cancer individuals were selected by linking the laboratory file with the cancer registry. Results other than unsatisfactory were classified as true-positive and -negative, and false-positive and -negative tests so as to calculate accuracy estimates. Slides of all false-negative and -positive and randomly selected samples of true-positive and -negative tests were evaluated for 11 abnormal cell features typical of pSCCa in conventional preparations: distribution differences by disease status were tested for significance.

Results: There were 129 brush and 170 wash in the pSCCa group and 365 brush and 1153 wash in the Neg group. Of all specimens, 1.2% were unsatisfactory. Brush sensitivity, specificity, PPV, and NPV were 61.9%, 99.4%, 97.5%, and 88%, respectively. Wash frequencies were 53.3%, 98.8%, 86.5%, and 93.5%, respectively. Abnormal cell features occurred in 29.9% of the selected pSCCa and 4.7% of the Neg specimens, and distribution differences were significant for each feature (P < .001).

Conclusions: Unsatisfactory brush and wash specimens are infrequent in the diagnosis of pSCCa, and both have moderate sensitivity and high specificity, PPV, and NPV. pSCCa abnormal cell features resemble those seen in conventional preparations and can distinguish specimens with pSCCa from those negative for pulmonary cancer.