Development of a simulation model for suprapancreatic marginal lymph node dissection using CT and intraoperative videos

2023 Fiscal Year Final Research Report

• Project/Area Number: 19K18151
• Research Category: Grant-in-Aid for Early-Career Scientists
• Allocation Type: Multi-year Fund
• Review Section: Basic Section 55020:Digestive surgery-related
• Research Institution: Kobe University
• Principal Investigator: Yamazaki Yuta 神戸大学, 医学研究科, 医学研究員 (60817823)
• Project Period (FY): 2019-04-01 – 2024-03-31
• Keywords: 腹腔鏡下胃切除術 / ディープラーニング

Outline of Final Research Achievements

1) Attempted to create organ vessel models from CT images using a 3D printer. We created a dataset from surgical videos, trained AI to create an automatic surgical instrument recognition system during laparoscopic gastrectomy, and reported the results in the Journal of the American College of Surgeons. In addition, we compared the use of surgical instruments during subpyloric and suprapancreatic marginal lymph node dissection at the level of surgeon skill, and published the results in the Journal of Gastrointestinal Surgery. (4) Automatic recognition of the right gastric reticular vein, right gastric artery, left gastric artery, and left gastric vein from videos of laparoscopic gastrectomy.

Free Research Field

消化器外科

Academic Significance and Societal Importance of the Research Achievements

腹腔鏡下胃切除術のシミュレーションモデル開発に必要な、手術動画からの臓器・手術器具の抽出及びトラッキングを、人工知能を用いて行うことに成功した。とくに術中動画からの手術器具の抽出システムの実装、さらにそれを用いた臨床研究（術者レベルごとの差異の評価）の成果報告は、消化器外科領域では本研究を除いてほとんど類を見ず、画期的であったとともに、消化器外科領域におけるAIを用いた研究の草分けとなった。上記を踏まえ、本研究は学術的にも社会的にも有意義であった。