Machine learning-based prediction of infarct size in patients with ST-segment elevation myocardial infarction: Misinterpretation

•Diagnosis and prediction of a clinical outcome are two different methodological issues. By evaluating the value of actual (CMR) and ML-IS, they evaluated diagnostic accuracy of ML based analysis for IS.•Performing a linear regression, any associations even statistically significant have nothing to do with prediction.•In reality, there are numerous variables at multiple levels, including genetic, individual, population and environmental that interact simultaneously to cause a health event.•Limitation in detecting interactions among numerous variables leads to misleading results especially in terms of prediction.