Medical audit of screening mammography – performing and interpreting

Presentation on theme: "Medical audit of screening mammography – performing and interpreting"— Presentation transcript:

1 Medical audit of screening mammography – performing and interpreting
陳勁宇 醫師 Chin-Yu Chen, M.D., M.S. 奇美醫學中心 影像醫學部 Department of Radiology, Chi-Mei Medical Center

2 What is medical audit? Follow-up and outcome monitoring of a test, i.e., mammography. - 資料監測  Provide quantifiable evidence of the test (mammography) performance Definition and standard: ACR BI-RADS® 4th ed. (2003) section III. ACR: American College of Radiology BI-RADS: Breast Imaging Reporting and Data System

3 Why doing medical audit?
A concrete feedback for radiologists in mammography interpretation – 判讀成效之具體回饋  Government regulation – 乳房攝影品管的要求  MQSA (Mammography Quality Standard Act) in U.S.A. (1992) Mammography quality assurance program by DOH (Department of Health) in Taiwan

4 How to ? 明確區分乳房攝影之檢查屬性: screening or diagnostic
依照ACR BI-RADS® 之診斷分類 (assessment category) 完成報告 原始資料 (raw data) 之收集 衍生性資料 (derived data) 之計算 - 陳勁宇等, 使用Microsoft® Office 開發乳房攝之BI-RADS® 資料監測系統, 中華放射醫學雜誌, 2005; 30:

5 原始資料 (raw data)

6 衍生性資料 (derived data)

7 What’s the meaning? FP – false positive
PPV – positive predictive value Cancer detection rate Percentage of minimal cancers Recall rate

8 False Positives (FP) FP1 = initially screening category 0,4,5 - TP
數量最大, 包含原始篩檢結果異常之所有個案 FP2 = biopsied benign + non-biopsied cases = screening final category 4,5 – TP 數量較小, 包含所有篩檢之最終結果為 BR4 or 5者 FP3 = biopsied benign cases 數量最小, 只包含所有已知做 biopsy 的個案

9 Positive Predictive Value (PPV)
PPV1 = TP / (TP + FP1) 數值最小, 分母包含原始篩檢結果異常之所有個案 PPV2 = TP / (TP + FP2) 數值較大, 分母包含所有篩檢之最終結果為 4or5者 PPV3 = TP / (TP + FP3) 數值更大, 分母只包含所有已知做 biopsy 的個案 若所有最終結果為 BR4 or 5 者皆完成 biopsy, 則 PPV2 = PPV3

10 Data Interpretation Is it good for the Taiwanese population ?

11 台灣之乳癌發生率 ( )

12 美國之乳癌發生率 ( ) 白種人 非洲裔 亞洲裔 西班牙裔

13

14

15 Reference: Chen CY, et al. Adjusting mammography-audit recommendations in lower-incidence Taiwanese population. - JACR 5(9): , Sep

16 The Reasonable Initial Screening Results?

17 May I remind you… Data in    Information out
Garbage in    Garbage out

18 Thank you very much for your attention 
陳勁宇 醫師 Chin-Yu Chen, M.D., M.S. 奇美醫學中心 影像醫學部 Department of Radiology, Chi-Mei Medical Center