Laparoscopic ultrasonography: a training model.

Abstract

PURPOSE: The purpose of this study was to develop a surgical training program and to test the accuracy of laparoscopic ultrasound in detecting injected lesions in pig livers. METHODS: Pig livers were divided into eight segments and injected with Surgilube "malignant" and silicone "benign" lesions. All were examined by laparoscopic ultrasound followed by liver explantation to confirm results. First, a pilot study was conducted on six swine by injecting Surgilube lesions and performing laparoscopic ultrasound through 3 different ports (left upper quadrant (I), umbilicus (II), and right lower quadrant (III)) to determine per-segment accuracy and to optimize port placement. Second, blinded injection of Surgilube and silicone implants was done on 18 pigs with laparoscopic ultrasound conducted through the two most accurate ports from the pilot study. This model was then tested during a resident training workshop. RESULTS: In the pilot study, per-lesion and per-segment sensitivity was 96 percent, with no difference among the three ports used. Ports I and II were chosen for the blinded study for their convenience in performing laparoscopic colectomy. In the blinded study, per-segment sensitivity, specificity, and accuracy were 97 percent, 94 percent, and 96 percent and 99 percent, 94 percent, and 97 percent for ports I and II, respectively. At the conclusion of a pilot workshop, trainee per-segment sensitivity, specificity, and accuracy were 60 percent, 80 percent, and 70 percent, respectively. The major difficulty was differentiating benign from malignant lesions. CONCLUSIONS: A useful liver laparoscopic ultrasound training model for surgeons was developed with good preliminary results. It is anticipated that further training will enhance laparoscopic ultrasound accuracy rates before application of this modality in humans.