林協霆 M.D. · Hsieh-Ting Lin
血液腫瘤內科 (Hematology & Medical Oncology)
和信治癌中心醫院腫瘤內科部血液腫瘤科專研醫師。臨床筆記與閱讀記錄。
Hematology & Medical Oncology Fellow at Koo Foundation Sun Yat-Sen Cancer Center, Taipei. Clinical notes and reading log.
關於我
林協霆，和信治癌中心醫院血液腫瘤科專研醫師（2023–），國立成功大學醫學系畢業，內科與腫瘤內科雙專科。臨床涵蓋肺、乳、大腸、肝、攝護腺等實體瘤，與淋巴瘤、白血病、骨髓瘤、再生不良性貧血等血液腫瘤；熟悉 NCCN／ASCO／ESMO 跨指引差異與台灣健保條件。
研究專注於統合分析、網絡統合分析、生存分析、Bayesian target trial emulation、因果推論（DAG／傾向分數／TMLE），以 R + Quarto IMRAD pipeline 完成 BTK 抑制劑、PARP 抑制劑、ICI 時序等多個專案；同時建構 NGS tertiary 分析、TCGA／SCAN-B 再分析、AML 多體學機器學習等精準腫瘤學工具。
在「臨床 × AI」交界，釋出多個開源 MCP／Vercel Skills 套件（OpenEvidence、NCCN、EBMT、Micromedex 風格 DDI），核心設計為「parallel Haiku + anti-hallucination 引用驗證」；正在準備 NEJM AI、PLOS Medicine 等期刊投稿。長期目標是把可重現研究、引用驗證與臨床決策整合在一起。
關於本站
本站為腫瘤內科臨床筆記、衛教整理與閱讀記錄，主要服務兩種讀者：
- 病友與家屬：以「病人真正會問」的問題切入，整理治療選項、副作用、與決策時可帶去診間的問題清單。
- 同業與住院醫師：整理新藥臨床試驗、跨指引差異、以及臨床上未必能立刻找到實證的灰色地帶。
本站不取代主治醫師之診療判斷，亦不從事醫療業務招徠。實際治療決策請與您的醫療團隊面對面討論。
編輯方針 · Editorial Policy
- 依國際指引 — 臨床主題以 NCCN、ASCO、ESMO 最新版指引、以及 PubMed 索引論文為主要來源，並比對台灣腫瘤醫學會、國健署資料。
- 署名審稿、標註日期 — 每篇醫療相關文章皆由本人署名審稿，文末顯示審稿日期與最後更新日期；新版指引釋出時依年度重審。
- 符合醫療法 §87 — 本站屬醫療法第 87 條第 2 項所定之醫學新知與病人衛生教育，不從事醫療業務招徠，避免「保證、根治、最有效、永不復發」等違反法規之用語。
- 個案資料去識別化 — 若引用個案經驗，必經充分去識別化處理，並僅以教學、衛教為目的；不公開可辨識病人身分之內容。
技能與作品 · Skills
以下技能項目皆對應公開的 GitHub repos；點擊任一專案標籤可直接跳到該 repo 檢視程式碼、文件與最近更新。
- 臨床研究方法學 · Clinical research methodology
Meta-analysis、network meta-analysis、survival analysis（含 competing risks、immortal time、left truncation）、Bayesian target trial emulation、causal inference（DAG / propensity / TMLE）、real-world evidence（MMRM / CMH）。R + Quarto IMRAD pipeline，已用於 BTK 抑制劑、PARP 抑制劑、ICI 時序、咖啡因等多個 RCT 統合分析。
- 精準腫瘤學與多體學分析 · Precision oncology & multi-omics
NGS tertiary analysis（BAM → ESMO 2024 clinical report，含 OncoKB、AMP/ASCO/CAP、ESCAT 分級）、TCGA / METABRIC / SCAN-B 再分析、單細胞 RNA-seq 與 bulk transcriptomic 拓樸風險分層、機器學習於 AML 多體學藥物反應與乳癌放療反應預測。
- 臨床決策 AI 與 MCP 工具開發 · Clinical AI tooling & MCP servers
建構與維護多個 MCP / Vercel Skills 套件：OpenEvidence 雙向整合 + Crossref 驗證、NCCN PDF 轉結構化 skill、EBMT 8e、Micromedex 風格 DDI、護理實證升等教練、Anki 整合。多套件以「parallel Haiku、anti-hallucination 引用驗證」為設計核心。發表 NEJM AI Perspective 級論文討論 LLM 診斷失敗模式。
- 學術寫作與可重現研究 · Academic writing & reproducible research
Quarto / Pandoc / Typst / LaTeX 全流程；AMA、APA CSL；CrossRef DOI 驗證；BibTeX 多格式輸出；Docker 化 R 環境；醫療專用 cookiecutter 範本；已準備中的手稿涵蓋 NEJM AI、PLOS Medicine、American Psychologist、JCO 等期刊投稿風格。
- 教學與訓練資源 · Teaching & training resources
ABIM 血液內科自製 72 主題 + 216 題、台灣 N1–N4 護理實證升等教練、臨床統計與因果推論視覺化（含 TMLE / DAG / propensity score 動畫）、Network meta-analysis with R 中文教學；ASH 2026 Asia-Pacific 與 ASH B-cell lymphoma 系列 Reveal.js 簡報。
- 醫院流程自動化 · Hospital workflow automation
KFSYSCC IRB 文件批次生成（43 表單、11 類別、一鍵）、TMA 月度繼續教育、教育部學術倫理課程、醫學會行事曆 → Google Calendar、北榮泌尿部排班系統、e-portfolio 教學評核、新生兒奶量計算、病理面積六角格量測。
完整 210 個公開 repos（醫療 79 + IT 131）依時間排序見 /portfolio。學術發表以 Google Scholar 為準；歷年口頭報告與證照見 CV。
現職
- 2023.08 –
血液腫瘤科專研醫師（Fellow）
學歷
- 2013 – 2019
醫學系（Doctor of Medicine）
國立成功大學 National Cheng Kung University
專科證照
- 台灣內科醫學會 內科專科醫師
- 台灣癌症醫學會 腫瘤內科專科醫師
完整學經歷與發表
完整現職、經歷、發表、口頭報告與專科證照清單，請見 CV 頁；學術發表與引用以 Google Scholar 為準。
聯絡與更新
本站源碼與版本紀錄公開於 GitHub；內容修訂以 git commit 留下時間戳。學術合作、糾錯、衛教合作請開 issue 或透過 ORCID 聯絡。
Email： [email protected] （勘誤、衛教提問、學術合作；不提供個別診療建議）
讀者回饋管道：LINE 官方帳號 @927pjtfa ，接受文章勘誤建議、衛教問題提問、學術討論。 本管道不提供個別診療建議，亦不受理任何形式之醫療業務預約；具體病情、用藥、治療方案請與您的主治醫師面對面討論。
About me
Hsieh-Ting Lin, M.D., is a Hematology and Medical Oncology fellow at Koo Foundation Sun Yat-Sen Cancer Center, Taipei (since 2023). Trained at National Cheng Kung University and double board-certified in Internal Medicine and Medical Oncology, his clinical work spans solid tumors (lung, breast, colorectal, hepatocellular, prostate) and hematologic malignancies (lymphoma, leukemia, myeloma, aplastic anemia), with cross-guideline fluency in NCCN, ASCO, and ESMO alongside Taiwan's NHI reimbursement landscape.
His research focuses on meta-analysis, network meta-analysis, survival analysis, Bayesian target trial emulation, and causal inference (DAG, propensity scoring, TMLE), built on an R + Quarto IMRAD pipeline. Completed projects cover BTK inhibitors in Waldenström macroglobulinemia, PARP inhibitors in BRCA-mutated breast cancer, and immune checkpoint inhibitor timing in early-stage NSCLC. He also maintains NGS tertiary analysis pipelines, TCGA / SCAN-B reanalyses, and machine learning for AML multi-omics drug-response prediction.
At the interface of clinical medicine and AI, he is the author of multiple open-source MCP servers and Vercel Skills (OpenEvidence, NCCN, EBMT, Micromedex-style DDI) centred on parallel Haiku conversion and anti-hallucination citation verification; manuscripts are in preparation for NEJM AI and PLOS Medicine. The long-term goal: weave reproducible research, citation-verified AI, and frontline clinical decisions into a single workflow.
About this site
A clinical notebook, patient-education resource, and reading log in medical oncology — written for two audiences:
- Patients & families — starts from the questions patients actually ask, then organises treatment options, side effects, and questions to bring back to clinic.
- Peers & residents — summaries of new trials, cross-guideline differences, and the grey areas where evidence has not settled.
The site does not replace clinical judgement and does not solicit medical business. Always discuss specific treatment decisions with your own care team.
Editorial Policy
- Guideline-grounded — Clinical topics anchored on current NCCN, ASCO, and ESMO guidelines and PubMed-indexed literature; cross-checked against Taiwan Oncology Society and HPA sources.
- Signed clinical review with dated stamps — Every medical post is reviewed under my name; review and last-updated dates are shown at the foot of each article and refreshed when guidelines update.
- Compliant with Medical Care Act §87 — This site operates under Article 87.2 as medical news and patient education; absolute terms (e.g. "guarantee", "cure", "most effective", "no recurrence") prohibited by the law are avoided.
- De-identified case material — Any patient case used is fully de-identified — age range only, no unit or date — and used solely for teaching purposes.
Skills & portfolio
Every skill below maps to one or more public GitHub repos; click any project chip to jump to the source, docs, and latest commits.
- Clinical research methodology · 臨床研究方法學
Meta-analysis, network meta-analysis, survival analysis (competing risks, immortal-time bias, left truncation), Bayesian target trial emulation, causal inference (DAG / propensity scores / TMLE), and real-world evidence (MMRM / CMH). R + Quarto IMRAD pipeline applied to BTK inhibitors, PARP inhibitors, ICI timing, and other multi-RCT meta-analyses.
- Precision oncology & multi-omics · 精準腫瘤學與多體學分析
NGS tertiary analysis (BAM → ESMO 2024 clinical report with OncoKB, AMP/ASCO/CAP, ESCAT classification), TCGA / METABRIC / SCAN-B reanalysis, single-cell RNA-seq and bulk transcriptomic topological risk stratification, and machine learning for AML multi-omics drug response and breast cancer radiotherapy response prediction.
- Clinical AI tooling & MCP servers · 臨床決策 AI 與 MCP 工具開發
Author and maintainer of multiple MCP and Vercel Skills packages: OpenEvidence bi-directional integration with Crossref verification, NCCN PDF → structured AI skill, EBMT 8e guideline skill, Micromedex-style DDI assessment, Taiwan nursing-evidence promotion coach, and Anki integration. Design centred on parallel Haiku conversion and anti-hallucination citation enforcement. NEJM AI Perspective-style manuscript on LLM failure modes in diagnosis.
- Academic writing & reproducible research · 學術寫作與可重現研究
Full Quarto / Pandoc / Typst / LaTeX pipeline; AMA and APA CSL; CrossRef DOI verification; multi-format BibTeX output; Docker-ised R environments; medical cookiecutter templates. Manuscripts in preparation target NEJM AI, PLOS Medicine, American Psychologist, and JCO.
- Teaching & training resources · 教學與訓練資源
Self-authored ABIM Hematology bank (72 topics + 216 questions); Taiwan N1–N4 nursing evidence-based promotion coach; visualised clinical statistics and causal inference (TMLE / DAG / propensity score animations); Chinese-language Network Meta-analysis with R tutorial; Reveal.js decks for ASH 2026 Asia-Pacific and ASH B-cell lymphoma updates.
- Hospital workflow automation · 醫院流程自動化
KFSYSCC IRB document batch generator (43 forms, 11 categories, one command); Taiwan Medical Association monthly CME automation; Ministry of Education research-ethics course completion; medical-society calendar → Google Calendar sync; Taipei VGH urology scheduler; e-portfolio teaching-evaluation filler; neonatal milk calculator; pathology hex-grid area measurement.
Full 210 public repos (79 medical + 131 dev tooling), sorted by latest push, live at /portfolio. Publications and citations tracked on Google Scholar; full training history and oral presentations live at CV.
Current Position
- 2023.08 –
Hematology & Medical Oncology Fellow
Department of Hematology & Medical Oncology, Koo Foundation Sun Yat-Sen Cancer Center
Education
- 2013 – 2019
Doctor of Medicine (M.D.)
National Cheng Kung University, Tainan, Taiwan
Board Certifications
- Board-Certified Internist, Taiwan Society of Internal Medicine
- Board-Certified Medical Oncologist, Taiwan Oncology Society
Full CV & publications
Full clinical training, publications, oral presentations, and board certifications are listed on the CV page. Academic publications and citation count are tracked on Google Scholar.
Contact & updates
Source code and revision history are public on GitHub; every revision is timestamped via git. For corrections, academic collaboration, or patient-education collaboration, please open an issue or reach me through ORCID.
Email: [email protected] — corrections, education questions, and academic collaboration; not for individual medical advice.
Reader feedback channel: LINE official account @927pjtfa accepts corrections, education questions, and academic discussion. This channel does not provide individual medical advice and does not accept any form of appointment booking ; specific symptoms, medications, or treatment plans should be discussed face-to-face with your own physician.