Enhanced Recognition Confidence of Millimeter‐Sized Intrahepatic Targets by Real‐Time Virtual Sonography

Abstract

Objectives Real‐time virtual sonography (RVS) is an artificial‐intelligence‐assisted ultrasonographic navigation system that displays synchronized preoperative computed tomography (CT) images corresponding to real‐time intraoperative ultrasonograms (IOUS). This study aimed to investigate whether RVS can enhance IOUS identification of small intrahepatic targets found in preoperative CT. Methods Patients with small intrahepatic targets detected by preoperative thin‐slice dynamic CT before liver resection were included. The targets included millimeter‐sized liver tumors or a third‐order or more distal portal branch and were marked on CT images using 3D simulation software. After laparotomy, the targets were searched using fundamental IOUS, and participating liver surgeons subjectively scored the target identifying confidence on a scale of 1–5 (5 points for detection with the highest confidence and one point for undetectable). Then, the search procedure was repeated using the RVS, and the scores were compared. Results Totally, 55 patients with 117 small targets were investigated. The median target size was 6.0 mm, and the median registration time was 3.6 seconds. The target identification confidence score significantly increased from 2.78 to 4.52 points after using RVS. Seventeen targets (14.5%) were undetectable in fundamental IOUS, and 14 of them were identified by RVS. The detectability of small liver tumors (2–5 points of identification confidence) by IOUS was 81.1 and 96.7% by RVS. Conclusion RVS enhanced surgeons' confidence in identifying millimeter‐sized intrahepatic targets found in preoperative CT.

Artificial Intelligence, Machine Learning, and Deep Learning in the Diagnosis and Management of Hepatocellular Carcinoma

“…The algorithm achieved an AUC of 80.2% and an overall classification accuracy of 74.6% [56]. Furthermore, Takamoto et al used real-time virtual sonography (RVS), an AI-assisted platform that merges preoperative CT images with real-time IOUS, to enhance IOUS with the identification of small intrahepatic lesions [57]. The median liver lesion size was 6.0 mm and RVS significantly improved surgeon confidence in lesion identification.…”