Percutaneous needle biopsies were performed on 683 patients with solitary pulmonary nodules during 1976-84. A cytological diagnosis of malignancy was made from the first biopsy in 473 patients (69%). A second biopsy was performed in 43 patients, a diagnosis of malignancy being made in a further 16 cases (37%). Histological material was available for comparison with cytological findings in 203 patients. Cytological examination was reliable in the diagnosis of malignancy with a high yield (75%) and low false positive rate (1.5%). Specific benign lesions were correctly diagnosed in 10 patients (1.5%). There was a false negative rate for the diagnosis of malignancy of 18% for the patients with a subsequent histological diagnosis. This compares with a false negative rate of 9% overall; the true rate probably lies between these figures. These results imply that a cytology report indicating no evidence of malignancy, but not diagnostic of a specific benign condition, does not reliably exclude a malignant lesion. In this series cytological typing was not accurate at predicting the cell type determined by histological examination (61% agreement) and was not able to discriminate between small cell and non-small cell lung cancer.

Cytological examination of material obtained by percutaneous thin needle lymph node aspiration is an estab- lished diagnostic technique.'The method for obtaining such material has been well described,2 but there are no recent reports of its use in patients with bronchial car- cinoma who have enlarged peripheral lymph nodes at pre- sentation.We report our results of the past seven years to show the value of the technique in the diagnosis and man- agement of bronchial carcinoma.

The accuracy of diagnosis in 656 patients with the four common histopathological types of primary lung cancer has been assessed by comparing the cell type diagnosis made on cytological and histological investigation with that determined by examination of the surgically resected or necroscopy specimen. The accuracy of diagnosis achieved by cytological examination of sputum and bronchial aspirate, and by bronchial biopsy histology was over 85%. The least accurate diagnostic procedure was percutaneous needle biopsy (62%). Squamous and small cell tumours were accurately diagnosed by all four investigations but errors were made in the diagnosis of large cell and adenocarcinomas. Nearly half the number of patients (43%) with large cell carcinoma were later reclassified as having squamous carcinoma and of the patients with adenocarcinoma 32% had been predicted to be squamous and 18% large cell carcinoma. We consider such quality control of pretreatment diagnosis mandatory in management of individual patients and before enrollment in clinical trials.