Predicting extent of coronary disease: fuzzy cluster analysis vs. Duke treadmill score

Journal of Clinical and Basic Cardiology 2000; 3 (1): 39-41

PDF    Summary    Figures   




Keywords: Duke treadmill score,  Fuzzy-Cluster-Analyse,  Stresstest,  Duke treadmill score,  fuzzy cluster analysis,  stress testing

We used fuzzy cluster analysis (FCA) to classify 166 positive stress tests as mildly, moderately, or severely abnormal. The method combines ST-segment change with 5 other stress test variables, and then computes a similarity measure to determine how closely each patient?s stress test resembles a prototypical mildly, moderately, or severely abnormal stress test. All of the patients had coronary angiography within one month of their stress tests. A Duke treadmill score was also calculated for each patient?s stress test: this score is derived from total exercise time, ST-segment change, and the presence or absence of angina. FCA showed better overall correlation with extent of coronary artery disease (CAD) (r = 0.74) than the treadmill score (r = 0.47). Tests classified as mild by FCA were more strongly associated with single vessel CAD or normal coronaries than treadmill scores of >= +5 (low risk range). For patients with triple vessel CAD, only 44 % had a treadmill score of -11 or less (high risk range), while 76 % were classified as severely abnormal by FCA (p < 0.01). For left main CAD, 86 % had severely abnormal tests by FCA compared with 77 % having a score of -11 or less (P = NS). For the combined group with high-grade CAD (left main or triple vessel), 79 % of their tests were classified as severe by FCA compared to 55 % of these tests having a treadmill score of -11 or less (p < 0.01). Thus in patients with positive stress tests, FCA better predicts the extent of CAD, especially the presence of triple vessel disease, than the treadmill score. J Clin Basic Cardiol 2000; 3: 39-41.