6.3 🩺 內科專科考前版
對象:R2-R3 / Fellow,準備台灣內科專科考。本章 specialist 重點:USPSTF vs 國健署差異、篩檢爭議、polygenic risk scores、新興工具。
6.3.0.1 📌 一頁重點整理
- USPSTF 是美國權威,但台灣國健署有本土化建議(四癌篩檢、成人預防保健)
- 22E 強調 polygenic risk scores 興起:未來個人化篩檢核心
- Liquid biopsy / ctDNA:multi-cancer early detection 試驗中(Galleri test)
- AI in screening:mammography、retinal、skin lesion 已 FDA 核准多項
- USPSTF 2021 colon screening update:起始 45 yo(從 50 → 45,BIPOC 健康差距驅動)
- USPSTF 2024 mammography update:40-49 yo 從 C → B
- 📍 台灣:四癌篩檢(mammo、Pap、FIT、口腔)+ HBV/HCV + 成人預防保健 40+
6.3.0.2 📜 22E vs 21E 主要差異
| 項目 | 21E | 22E |
|---|---|---|
| Polygenic risk scores | 略提 | 正式整合 |
| Liquid biopsy | 無 | 新增 |
| AI in screening | 略提 | 大幅擴充 |
| Mammography 40-49 | C 級 | B 級(2024 USPSTF update) |
| Colon screening start age | 50 | 45(2021 update) |
| HCV screening | 1945-65 | 18-79 universal(2020 update) |
| Lung CT | 55-80 + 30 PY | 50-80 + 20 PY(2021 update) |
| HIV | 15-65 | 同 |
6.3.0.3 🧠 深度概念
6.3.0.3.1 Cost-Effectiveness Analysis
- ICER (Incremental Cost-Effectiveness Ratio) = ΔCost / ΔQALY
- < $50,000/QALY:highly cost-effective
- $50,000-150,000/QALY:generally cost-effective
$150,000/QALY:questionable
例: - Universal HCV screening 18-79:< $50K/QALY ✅ - Annual lung CT:~$60-150K/QALY(依 risk) - Annual mammography 40-49:$70-100K/QALY(爭議)
6.3.0.3.2 Lead Time Bias 解析
關鍵:mortality 是 robust endpoint - 假設疾病自然 history 不變 - 篩檢提前診斷 → survival 看似 +X 年 - 但 mortality date 沒變 → mortality rate 才能反映真實效益
如何避免: - RCT with mortality endpoint - 用 disease-specific mortality
6.3.0.3.3 Length Time Bias 解析
- 進展慢的 case 在 screening interval 中累積
- 進展快的 case 多在 symptom 階段才被診斷
- → 篩檢族群「artificially enriched」for indolent
- 例:mammo 偵測到的多是 ER+ luminal A(慢);triple-negative 較少 detected
6.3.0.3.4 Polygenic Risk Scores (PRS)
- 結合幾百個 SNPs
- 重要應用:
- CV disease:CARDIoGRAMplusC4D PRS → identify top 5% extreme risk
- Breast CA:integrated with BRCA testing → 個人化 risk
- AF:identify pre-clinical AF risk
- 限制:
- 多基於 European 人群 → ancestry bias
- 亞裔模型仍 emerging(Taiwan Biobank 進度中)
- 未充分整合 environmental factors
6.3.0.3.5 Liquid Biopsy / Multi-Cancer Early Detection
- 偵測 ctDNA(circulating tumor DNA)+ methylation patterns
- Galleri test (GRAIL):50+ cancer types screening
- PATHFINDER study:~1.4% positivity rate
- 限制:
- PPV 仍待提升
- False positive 引發 unnecessary workup
- 22E:「emerging, not standard」
6.3.0.3.6 AI in Screening 進展
| 領域 | FDA approved | 應用 |
|---|---|---|
| Mammography | iCAD ProFound、ScreenPoint | Computer-aided detection 整合 |
| DR (Diabetic retinopathy) | IDx-DR | Primary care direct dx |
| Skin lesion | DERM | Triage |
| Lung CT | Multiple | Nodule detection |
| ECG | AliveCor、AppleWatch | AF detection |
風險:bias、generalizability、deskilling、false reassurance
6.3.0.4 🩺 USPSTF Recommendation 詳解
6.3.0.4.1 Cancer Screening 詳細
6.3.0.4.1.1 Breast Cancer
- 40-49: B level(2024 update from C)
- 50-74: A level
- 75+: I(insufficient)
- 頻率:q2y(USPSTF)vs q1y(ACS)
- BRCA1/2 carrier:MRI + mammo annual from 25-30
- 男性:no screening(low risk)
6.3.0.4.1.2 Cervical Cancer
- 21-29 yo:Pap q3y
- 30-65 yo:Pap q3y OR HPV q5y OR co-test q5y
- < 21 / > 65 (with adequate prior screening):not recommended
6.3.0.4.1.3 Colorectal Cancer (2021 update)
- 起始 45 yo(vs ACS 從 45)— 2021 update from 50
- 結束:75 yo
- Tools (各有頻率):
- FIT yearly
- FIT-DNA (Cologuard) q1-3y
- Sigmoidoscopy q5y (or q10y + FIT q1y)
- Colonoscopy q10y
6.3.0.4.2 CV Risk Screening
- BP:≥ 18 yo periodic
- Lipid:40-75 yo periodic
- DM:overweight ≥ 35 yo q3y
- AAA US:M 65-75 ever smoker once
- ASCVD risk calculator:用於 statin / aspirin decision
6.3.0.4.3 Infectious Disease
- HCV:18-79 once(2020 update from 1945-65)
- HIV:15-65 at least once
- Chlamydia/GC:F < 25 sexually active
6.3.0.5 🌟 Clinical Pearls (8 條)
- 「Lung CT screening 後仍需戒菸」:篩檢不能當「continue smoking 的 license」
- Mammography 過度自信:DCIS 過度診斷有時會 overtreat
- PSA 是 individual decision:強烈推薦 → 過度治療;強烈反對 → 偶爾漏 high-grade
- Colon screening 工具選擇看 patient preference:依從性比 modality 重要
- HIV 一定 screen at least once:早期治療大幅改善 outcome
- BRCA carrier surveillance ≠ general population screening:MRI + mammo annual
- Polygenic risk scores 仍 emerging:謹慎解讀,不要 over-rely
- Stop screening 在 life expectancy < 5-10 yr:mortality benefit 來不及表現
6.3.0.6 🔍 特殊情境
6.3.0.6.1 1. BRCA1/2 carrier 管理
- Breast CA risk lifetime:60-70%
- Ovarian CA risk lifetime:BRCA1 40-60%、BRCA2 15-30%
- Surveillance:annual mammo + MRI from 25-30
- Prophylactic:bilateral mastectomy / oophorectomy(after childbearing)
- Chemoprevention:Tamoxifen for premenopausal、AI for postmenopausal
6.3.0.6.2 2. Lynch Syndrome (HNPCC)
- Mismatch repair gene mutations
- Colon CA risk lifetime:50-80%
- Endometrial CA risk:F 40-60%
- Surveillance:colonoscopy q1-2y from 20-25
- Aspirin chemoprevention(CAPP2 trial)
6.3.0.6.3 3. CF Carrier Screening
- 美國 ACOG 推薦 universal preconception
- Detect ~88% mutations in white population
- 較低 detection in non-white populations
6.3.0.6.4 4. Familial Hypercholesterolemia
- 1/250 prevalence
- Cascade screening:identify family members
- Statin from age 8-10
- LDL-C goal < 100 mg/dL
6.3.0.7 🔬 22E 提到的新進展
6.3.0.7.1 Multi-Cancer Early Detection Tests
- Galleri (GRAIL):detects 50+ cancers via methylation patterns
- PATHFINDER study:99% specificity
- Limitations:cost、PPV、access
- 22E:「promising but not yet standard」
6.3.0.7.2 AI Mammography
- FDA approved multiple AI tools
- COSMOS-3 / others:comparable / better than radiologist
- Workflow integration challenges
6.3.0.8 📍 台灣 Context 專區
6.3.0.8.1 國健署四癌篩檢
| 癌症 | 工具 | 對象 | 頻率 |
|---|---|---|---|
| Pap | Pap smear | F ≥ 30 | q3y |
| Mammography | Mammo | F 45-69(45+ 2 fold history → 40+) | q2y |
| 大腸癌 FIT | FIT | 50-74(2024 起 45-74) | q2y |
| 口腔癌 | 視診 | ≥ 30 yo + 抽菸 / 嚼檳 | q2y |
6.3.0.8.2 成人預防保健(2025 起每年)
對象: - 40-64:每 3 年 - ≥ 65:每年 - 35 yo + 高 risk
內容: - 身體理學檢查 - BP、BMI - Blood:fasting glucose、cholesterol、TG、HDL、creatinine - Urine routine - 健康行為諮詢
6.3.0.9 ⚠️ 老闆地雷區
- 未調整 screening to risk profile:BRCA carrier 用 general population schedule
- >75 仍積極 screen:浪費 + harm
- 依賴 PSA 而不 shared decision:過度治療
- Mammography 不解釋 overdiagnosis:病人錯誤期待
- 不問 family history:漏掉 hereditary syndromes
- Lung CT 後不戒菸介入:lose major effect
- AI 無條件 trust:bias / generalizability 問題
- Polygenic risk scores 在亞裔過度自信:model 多 European-based
6.3.0.10 🎓 內科專科考重點預測
6.3.0.10.1 高機率題型
- USPSTF 篩檢年齡 + 頻率配對
- Lead/Length time bias 概念辨識
- USPSTF A/B/D grade 應用
- 特定 mutation surveillance(BRCA、Lynch)
6.3.0.11 📖 延伸閱讀
- USPSTF:uspreventiveservicestaskforce.org
- 國健署四癌篩檢:hpa.gov.tw
- Khera AV et al. Genome-wide polygenic scores. Nat Genet 50:1219, 2018.
- NLST. Reduced lung-cancer mortality with low-dose CT. NEJM 365:395, 2011.
- Krilaviciute A et al. Comparison of cancer detection methods. Lancet Oncol 25:101, 2024.
6.3.0.12 📚 三階段教材索引
- 醫學生概念 → Ch 6 medstudent.md
- 國考衝刺 → Ch 6 board-prep.md
- Harrison 22E 原文 → Ch 6
⚠️ 本 md 為 claude-opus-4-7 撰寫(2026-05-07),未經盧醫師驗證。台灣篩檢政策請以最新公告為準。