Bigi Riccardo, Gregori Dario, Cortigiani Lauro, Desideri Alessandro, Chiarotto Francesco A, Toffolo Gianna M
CNR Clinical Physiology Institute Niguarda Ca' Granda Hospital P.zza Ospedale Maggiore, Milan, Italy.
Int J Cardiol. 2005 Jun 8;101(3):481-7. doi: 10.1016/j.ijcard.2004.07.008.
To compare artificial neural networks (ANN) and robust Bayesian classifiers (RBC) in predicting outcome following acute myocardial infarction (AMI).
Clinical, exercise ECG and stress echo variables by 496 patients with AMI were used to predict the cumulative end-point of cardiac death, nonfatal reinfarction and unstable angina. Revascularized patients were censored. Short (200 days)-, medium (400 days)- and long (1000 days)-term observation intervals, including 50%, 75% and 90% of the events, respectively, were considered. At each interval, any patient was binary assigned to the "event" or "no event" class. A multilayer feedforward ANN, trained by a back propagation algorithm, was used. RBC, using the leave-one-out technique, were derived. The accuracy of both techniques was compared to the default accuracy (DA) obtained by assigning all subjects to the largest class.
14 death, 27 reinfarction and 29 unstable angina were observed during a mean follow-up of 24 [95% confidence interval (CI) 19 to 22] months. The accuracy of ANN and RBC and DA were 70%, 81% and 74% at short, 67%, 73% and 56% at medium and 64%, 68% and 62% at long-term follow-up.
(1) ANN do not improve the prognostic classification of patients with uncomplicated AMI as compared to RBC. (2) In particular, short-term prognostic accuracy seems insufficient.
比较人工神经网络（ANN）和稳健贝叶斯分类器（RBC）在预测急性心肌梗死（AMI）后结局方面的表现。
采用496例AMI患者的临床、运动心电图和负荷超声心动图变量来预测心源性死亡、非致死性再梗死和不稳定型心绞痛的累积终点。接受血运重建的患者被截尾。分别考虑短期（200天）、中期（400天）和长期（1000天）的观察期，这些观察期分别包含50%、75%和90%的事件。在每个观察期，将任何患者二分类为“事件”或“无事件”类别。使用了一种由反向传播算法训练的多层前馈人工神经网络。通过留一法技术得出稳健贝叶斯分类器。将这两种技术的准确性与通过将所有受试者分配到最大类别而获得的默认准确性（DA）进行比较。
在平均24[95%置信区间（CI）19至22]个月的随访期间，观察到14例死亡、27例再梗死和29例不稳定型心绞痛。在短期随访中，人工神经网络、稳健贝叶斯分类器和默认准确性的准确率分别为70%、81%和74%；在中期随访中分别为67%、73%和56%；在长期随访中分别为64%、68%和62%。
（1）与稳健贝叶斯分类器相比，人工神经网络并不能改善无并发症AMI患者的预后分类。（2）特别是，短期预后准确性似乎不足。
