Field experimental research using an infant stool color detection algorithm for early diagnosis of biliary atresia

2022 Fiscal Year Final Research Report

• Project/Area Number: 18K08575
• Research Category: Grant-in-Aid for Scientific Research (C)
• Allocation Type: Multi-year Fund
• Section: 一般
• Review Section: Basic Section 55010:General surgery and pediatric surgery-related
• Research Institution: Kochi University
• Principal Investigator: Obatake Masayuki 高知大学, 医学部附属病院, 特任教授 (50336153)
• Co-Investigator(Kenkyū-buntansha): 星野 絵里 立命館大学, 総合科学技術研究機構, 准教授 (50598521)
• Project Period (FY): 2018-04-01 – 2023-03-31
• Keywords: 人工知能 / 便色 / 早期発見 / 胆道閉鎖症 / 胆汁うっ滞性肝疾患 / アプリケーション

Outline of Final Research Achievements

The number of cases (mean birth weight, mean weeks of gestation) at the four facilities participating in this study was 2463, including facility A: 767 cases (2931 g, 38 weeks 6 days), facility B: 289 cases (3013 g, 39 weeks 1 day), facility C: 596 cases (2881 g, 38 weeks 2 days), and facility D: 811 cases (3066 g, 39 weeks 4 days). This corresponded to 71.5% of all births during the study period. The results were 96.59% "no abnormality", 0.14% "necessary observation,", and 3.29% "be careful". Surgical treatment was performed in 2 of the 84 cases judged to be "necessary observation" and "be careful", and the sensitivity and specificity for detecting liver disease requiring surgical treatment were 100% and 96.7%, respectively. This study was considered useful as a method for early detection of biliary atresia and biliary congestive disease.

Free Research Field

小児外科

Academic Significance and Societal Importance of the Research Achievements

胆道閉鎖症および胆汁うっ滞性肝疾患の早期発見のため母子手帳に添付されている便色カラーカードは簡便で安価なスクリーニング法であり、海外での使用も報告されている。しかし肉眼で児の便色と比較するため主観的判断による見逃しや診断遅れなどが問題となっている。本研究のAIによる判定は人間の目で認識・判断する必要が無く、プログラムが微細な色調変化を的確に診断することでより早期の疾患の発見が可能となり、便色カラーカードに替わる方法として期待される。