HPB : the official journal of the International Hepato Pancreato Biliary Association · 2025
Accuracy and efficiency of artificial Intelligence-Assisted three-dimensional liver volumetry in living donor evaluation based on real world prospective data.
СтатьяНазвание и аннотация: Перевод в очереди
Обработка статьи · этапы, ошибки и управление
Загрузка состояния…
Ошибки и ограничения
История · было → стало
Состояние обновляется каждые 5 секунд, пока блок открыт. Процент показывается только там, где известен объём работы. Пауза останавливает продолжение этапа; текущий запрос к источнику может завершиться.
Do Hayeon, Baik Jiyoung, Gwon Suk Min, Lee Eunjin, Ryu Youngju, Kim Bogeun, Lim Soyoung, Oh Namkee, Rhu Jinsoo, Choi Gyu-Seong, Kim Jongman
Аннотация
BACKGROUND: Accurate liver volumetry is crucial for safe living donor liver transplantation (LDLT). Manual segmentation is time-consuming and delays surgical planning. Artificial intelligence (AI) enables automated 3D volumetry, improving workflow efficiency and accuracy.
METHODS: This retrospective study analyzed 583 living liver donor candidates who underwent preoperative CT volumetry between January 2023 and February 2025 at Samsung Medical Center. Donors were grouped into conventional 2D manual segmentation and AI-based 3D volumetry. Volumetric data before and after AI refinement were compared with intraoperative graft weights. Turnaround time, agreement between AI and manual results, and prediction accuracy were assessed. Subgroup analyses excluded donors with segmental resections or >5 % preoperative weight loss.
RESULTS: Of 583 donors, 271 underwent surgery, and 239 met inclusion criteria. The AI group had shorter turnaround times (1.5 ± 1.5 vs. 4.2 ± 3.6 days; p < 0.001). AI and manually edited volumes correlated strongly (R2 = 0.992; mean difference 23.7 ± 39.5 mL). After exclusions, AI-based volumetry showed superior predictive performance (R2 = 0.818; median error 54.0 mL) versus conventional methods (R2 = 0.707; median error 65.0 mL).
CONCLUSIONS: AI-assisted 3D volumetry enables fast, accurate graft volume estimation, reducing processing time and manual effort in LDLT.
CLINICAL TRIAL REGISTRATION: Not applicable.
METHODS: This retrospective study analyzed 583 living liver donor candidates who underwent preoperative CT volumetry between January 2023 and February 2025 at Samsung Medical Center. Donors were grouped into conventional 2D manual segmentation and AI-based 3D volumetry. Volumetric data before and after AI refinement were compared with intraoperative graft weights. Turnaround time, agreement between AI and manual results, and prediction accuracy were assessed. Subgroup analyses excluded donors with segmental resections or >5 % preoperative weight loss.
RESULTS: Of 583 donors, 271 underwent surgery, and 239 met inclusion criteria. The AI group had shorter turnaround times (1.5 ± 1.5 vs. 4.2 ± 3.6 days; p < 0.001). AI and manually edited volumes correlated strongly (R2 = 0.992; mean difference 23.7 ± 39.5 mL). After exclusions, AI-based volumetry showed superior predictive performance (R2 = 0.818; median error 54.0 mL) versus conventional methods (R2 = 0.707; median error 65.0 mL).
CONCLUSIONS: AI-assisted 3D volumetry enables fast, accurate graft volume estimation, reducing processing time and manual effort in LDLT.
CLINICAL TRIAL REGISTRATION: Not applicable.