雲嘉南地區核醫冠狀動脈造影異常之風險因子分析

Risk Factors of Nuclear Coronary Angiography Anomaly in Yunlin, Chiayi and Tainan Areas

蔡志欣(Chih-Hsin Tsai) ； 莊紫翎(Tayy-Lin Chuang) ； 邱靜娥(Jing-Er Chiu) ； 王昱豐(Yuh-Feng Wang)

《南臺灣醫學雜誌》 6卷1期 (2010/06) Pp. 22-29

https://doi.org/10.6726/MJST.201006_6(1).0003

摘要

目的：本研究以雲嘉南地區接受核醫心肌灌注掃瞄的381位病患為例，建立冠狀動脈風險因子與造影異常與否的預測模式，藉以輔助醫師的臨床判斷，降低侵入性檢測的風險。方法：本研究以病人的特徵（性別、年齡、體重）、危險因子（胸悶胸痛、高血壓、高膽固醇、家族心臟病史、抽菸、糖尿病）、過去病史（PTCA、CABG、氣喘）及持續症狀（藥物過敏、胸悶胸痛時間、悶痛至部位、頭暈、冒冷汗/全身無力、心悸）共19個因子為獨立變數，以冠狀動脈造影異常與否為依變數，應用支撐向量機與邏輯斯迴歸的統計方法建立模式。結果：支撐向量機分析結果，以性別、年齡、抽菸等三項因子判斷造影異常與否，準確率達65%；邏輯斯迴歸分析結果以體重、抽菸、PTCA、心悸等四項因子，準確率達67%。結論：從本研究模型中了解風險因子與冠狀動脈造影間的關係，未來將可作為臨床醫師判斷冠狀動脈疾病的折衷參考準則。

關鍵字

冠狀動脈疾病 ； 心肌灌注掃瞄 ； 風險因子

並列摘要

Objective: In order to investigate indications for cardiac catheterization versus conservative examination in patients, we analyzed data of 381 patients who received myocardial perfusion scan in Yunlin, Chiayi and Tainan areas. Method: We want to establish the prediction model between risk factors of CAD and coronary angiography results. In this study, support vector machine (SVM) and logistic regression (LR) were used to predict the risk characteristics of CAD based on the accuracy.Results: The results showed that the accuracy was 65% with the risk factors of sex, age and smoking by using SVM method. The accuracy was 67% by LR method with the risk factors of weight, smoking, percutaneous transluminal coronary angioplasty (PTCA) and palpitations.Conclusions: It shows that the risk characteristics suggested in SVM and LR models can be used to predict the result of coronary angiography, and thus is a useful judgment resource for clinicians.

並列關鍵字

coronary artery disease ； myocardial perfusion scan ； risk factor