The diagnostic accuracy of ultrasonically guided biopsy-gun biopsies was assessed in a group of 47 patients with suspected pancreatic carcinoma. A correct diagnosis with the aid of biopsy was obtained in 44 of the 47 patients (94%). In 3 patients with a carcinoma of the pancreas the correct diagnosis was not obtained with the first biopsy. In 2 of these 3 patients, a simultaneous biopsy of a liver metastasis revealed the presence of malignant tumour growth. No major complications occurred. Biopsy-gun biopsy of the pancreas is considered a useful, reliable and non-traumatic method for the diagnosis of pancreatic malignancy. Twenty-five patients with known neuroendocrine tumour disease were biopsied with 1.2 mm and 0.9 mm biopsy-gun needles to evaluate the respective diagnostic accuracy of the 2 needle sizes. The influence of treatment-related fibrosis on the histopathological diagnosis was also evaluated. The overall diagnostic accuracy with the 0.9 mm needle was 69% as compared to 92% with the 1.2 mm needle. This difference, however, seems more related to needle guiding difficulties with the 0.9 mm needle than to insufficient tissue yield. The increased amount of fibrous tissue due to interferon treatment did not seem to negatively influence the diagnostic accuracy. In order to assess the diagnostic accuracy rate for radiologists with different experience of biopsy procedures 175 cases of renal biopsy-gun biopsies were evaluated. No statistical significant difference was found between the different operators. Provided with detailed instruction even operators with limited experience produced biopsy results equal to those of operators with extensive practice in ultrasound-guided biopsies. The automated sampling performance of the biopsy-gun, with a consistent high diagnostic sampling rate (96%), was believed to be responsible for these results. The role of duplex Doppler ultrasound in monitoring interferon treatment-related changes in carcinoid metastases was evaluated. The patients were divided into 4 groups: untreated (n = 10), progressive disease (n = 17), stable disease (n = 20) and objective response (n = 18). No significant differences in Doppler values were found between the groups, and at present duplex Doppler ultrasound does not seem to play a role in the evaluation of tumour therapy in carcinoid patients.(ABSTRACT TRUNCATED AT 400 WORDS)