Surgical Data Scientist · Colorectal Surgeon
clinical problems → structured data → deployable systems
I build evidence-backed surgical data systems — turning raw CT, intraoperative video, and postoperative records into reproducible analysis, open-source software, and deployable clinical tools.
高雄榮總大腸直腸外科主治醫師
專注於將臨床問題轉化為可計算、可重現、可部署的資料科學系統
| Clinical problem | Peer-reviewed evidence | Reproducible code | Deployed artifact |
|---|---|---|---|
| Mid-pelvic workspace from CT | Auto-ISD · IJCARS 2026 Scalable annotation-free CT pelvimetry · JIIM 2026 |
auto-isd-pelvimetry |
Interactive demo · ctpelvimetry |
| Robotic surgery learning curves | J Robotic Surg. 2026 | rissa-ML-learning-curve |
Reusable analysis workflow |
| Postoperative symptom monitoring | IRB-approved study | hemorrhoids-postop |
Live clinical-research PWA |
auto-isd-pelvimetry— two peer-reviewed validation studies (IJCARS · JIIM), an open-source pipeline, and an interactive democtpelvimetry— installable PyPI package with a Python API, CLI, automated QC, and tests
rissa-ML-learning-curve— reproducible ML analysis of robotic surgery learning curves; published in the Journal of Robotic Surgerycholec80-phase-recognition— surgical phase recognition pipeline comparing non-causal MS-TCN with causal TeCNO models
Stage_III_Colon_EDR— externally validated early-recurrence modeling workflow with manuscript-aligned notebooks; patient-level data excludedllm-extractor— local-first LLM pipeline for extracting structured fields from rectal cancer pathology reports
hemorrhoids-postop— postoperative symptom monitoring and AI-assisted patient education PWA, deployed in an IRB-approved studyMedFeedJournalTracker— automated literature monitoring with LLM filtering and daily LINE notifications
Focus: surgical AI, learning curves, and outcome modeling
- A Scalable, Annotation-Free Pipeline for Automated CT Pelvimetry: A Validation Study with Landmark Uncertainty Analysis and Clinical Correlation — Journal of Imaging Informatics in Medicine 2026
doi - Automated CT-based pelvimetry — Int J Comput Assist Radiol Surg. 2026
doi - Machine learning–based learning curve analysis — J Robotic Surg. 2026
doi - Video-based RA-CUSUM proficiency assessment — Int J Colorectal Dis. 2026
doi
Full list → ORCID
- Data Science: pandas, scikit-learn, lifelines, PyTorch
- Imaging: CT processing, TotalSegmentator, 3D Slicer
- Causal Inference: overlap weighting, RMST, survival modeling
- Engineering: Python packaging, pytest, GitHub Actions, Next.js, TypeScript, PostgreSQL
Building the infrastructure of Surgical Data Science:
- From clinical intuition → quantitative modeling
- From retrospective data → real-time systems
- From isolated studies → reproducible pipelines
- Now building: surgical video analytics — automated phase recognition & workflow decomposition (
cholec80-phase-recognition: MS-TCN vs causal TeCNO)