Acute myocardial infarction (AMI) is a common and complex disease; pathogenesis of
AMI is not completely understood and genetic, clinical and phenotypic variables are involved in the
The aim of this paper was to assess: 1) the predictive capacity of Artificial Neural Networks (ANNs) in consistently distinguishing
the two different conditions (AMI or control). 2) the identification of the variables with the maximal amount
of relevant information for AMI.
Genetic variances in inflammatory genes, along with clinical and classical risk factors from 149 AMI patients and 72
healthy subjects were investigated. From the data base of this case/control study 36 genetic, clinical and phenotypic variables
were selected. The TWIST system, an evolutionary algorithm able to remove redundant and noisy information from
complex data sets, further selected 18 variables. Moreover, fitness, sensitivity, specificity, overall accuracy and areas under
the receiver-operating curves (AUC) of the 18 selected variables associated with AMI risk were investigated.
Our findings showed that ANNs are useful in distinguishing risk factors selectively associated with the disease. Finally,
the new mix of variables, including classical risk factors and genetic markers, generated a new risk cluster of variables
able to discriminate between AMI and control subjects with an overall accuracy of 90%.
This approach may be used to assess individual AMI risk in unaffected subjects with increased risk of the disease such as,
first relative with positive parental history of cardiovascular diseases.