Explainable machine learning for predicting postoperative length of stay after gastrectomy: a nationwide study using XGBoost and SHAP
We developed a machine learning model that predicts postoperative length of stay with an error range of 2-4 days using admission data. This proof-of-concept study demonstrates the feasibility of predicting length of stay from admission data, showing that explainable AI can replicate intuitive patterns in surgical oncology while simultaneously identifying unexpected insights from administrative data. These findings highlight the clinical potential of explainable AI for perioperative workflow optimization.
Authors
- Yoshimasa Akashi (ORCID: https://orcid.org/0000-0003-0589-3816)
- Tsunehiko Maruyama (ORCID: https://orcid.org/0000-0002-9701-7604)
- Hideo Suzuki (ORCID: https://orcid.org/0000-0003-2941-7726)
- Kazuto Ikezawa (ORCID: https://orcid.org/0000-0002-6281-8743)
- Tatsuya Oda (ORCID: https://orcid.org/0000-0001-6115-0158)
- Tomohiro Kurokawa
Institutions
- Fukushima Medical University (JP)
- University of Tsukuba (JP)
- Mito Saiseikai General Hospital (JP)
- Tsukuba Memorial Hospital (JP)
Publication Details
- Journal
- Frontiers in Medical Technology
- Published
- 2025-12-05
- DOI
- https://doi.org/10.3389/fmedt.2025.1732580
- Citations
- 1
- Primary Topic
- Artificial Intelligence in Healthcare and Education
- Type
- article
- Field-Weighted Citation Impact
- 0.65
Funders
- University of Tsukuba
- Japan Society for the Promotion of Science