PURPOSE: This study was conducted to assess the ability of the endoluminal sonographic probes to image upper gastrointestinal lesions detected with conventional endoscopy and to reconstruct 3-dimensional (3D) images. METHODS: Sonographic images were obtained from 25 gastrointestinal lesions in 22 patients using a sonographic miniprobe manually drawn at a set speed across lesions detected during conventional endoscopy. 3D images were then reconstructed using an in-house software program. RESULTS: In 16 lesions (64%), the 3D sectional images were evaluated as good, and all 25 lesions (100%) could be visualized both transversely and longitudinally. In 12 lesions (48%), the 3D display of the endosonographic surface image corresponded to the endoscopic appearance. In all 11 cases of gastric cancer, the depth of tumor invasion measured using the 3D image was consistent with the histologic diagnosis (100% accuracy). When videotaped sonographic images were stored on a hard disk, it took an average of 3.5 minutes to obtain and display 3D images of the targeted area. CONCLUSIONS: 3D sonographic reconstructions of digestive lesions were useful in diagnosis, and we believe that further improvements should enhance the clinical usefulness of 3D endosonography.
PURPOSE: This study was conducted to assess the ability of the endoluminal sonographic probes to image upper gastrointestinal lesions detected with conventional endoscopy and to reconstruct 3-dimensional (3D) images. METHODS: Sonographic images were obtained from 25 gastrointestinal lesions in 22 patients using a sonographic miniprobe manually drawn at a set speed across lesions detected during conventional endoscopy. 3D images were then reconstructed using an in-house software program. RESULTS: In 16 lesions (64%), the 3D sectional images were evaluated as good, and all 25 lesions (100%) could be visualized both transversely and longitudinally. In 12 lesions (48%), the 3D display of the endosonographic surface image corresponded to the endoscopic appearance. In all 11 cases of gastric cancer, the depth of tumor invasion measured using the 3D image was consistent with the histologic diagnosis (100% accuracy). When videotaped sonographic images were stored on a hard disk, it took an average of 3.5 minutes to obtain and display 3D images of the targeted area. CONCLUSIONS: 3D sonographic reconstructions of digestive lesions were useful in diagnosis, and we believe that further improvements should enhance the clinical usefulness of 3D endosonography.