Background and study aims: Although endoscopic ultrasonography (EUS) is a well-described examination method, there have been few reports concerning its clinical impact. The aim of this study was to describe EUS as it is performed at a county hospital, with an emphasis on the indications and clinical outcome.
Patients and methods: Patients examined using EUS between December 1997 and November 2000 were recorded prospectively. Follow-up was conducted by examining each patient's medical records at least 3 months after the investigation. The EUS findings were compared with the patient's final diagnosis, and the decisions made by the referring department on the basis of each investigation were recorded.
Results: A total of 344 EUS procedures were performed. In the third year, the distribution of patients relative to the various referral diagnoses was: 78 with suspected benign pancreaticobiliary disease, 33 for staging of known upper gastrointestinal tract malignancy, 15 with suspected mediastinal disease or for staging of lung cancer, 13 with suspected submucosal lesions, and five with unclassified disease. Follow-up was possible in 340 patients (99 %). Compared to the final diagnosis in each patient, the sensitivity, specificity, and accuracy rates of EUS were 86 %, 90 %, and 88 %, respectively. The EUS findings made more invasive procedures unnecessary in 199 patients (58 %). EUS led to a switch to less invasive procedures in 61 patients (18 %), and it had no influence on the further management strategy in 80 patients (24 %).
Conclusions: EUS has a high level of accuracy and a substantial clinical impact when performed in an unselected population. The estimated numbers of investigations needed appear to justify setting up an EUS center at institutions with a catchment population of 350 000 inhabitants.