Three year trial of endoscopic cytology of the stomach and duodenum.

The results of endoscopic cytology during the period 1974 to 1976 are presented. These consist of a pilot study on 76 patients and a main trial on 329 patients, of whom 61 proved to have malignant disease as the final diagnosis. Methods of cell collection, preparation, and reporting are given. Special emphasis was placed on making a definitive diagosis, especially the distinction between severe atypia and malignancy, and 'suspicious' reports were not issued. The overall accuracy achieved for all tumour types was 91.8%, with an accuracy of 98% in the diagnosis of adenocarcinoma of the stomach. There were no false positive cytology reports. Endoscopic histology on the same patients gave an accuracy of 68.9% and did not give a positive result in any of the five cases with false negative cytology. Adequate sampling, a variety of cell collection methods, and great care in technique are necessary in order to obtain optimum sensitivity and specificity in cytological diagnosis.