Optimizing myocardial hypothermia: I. Temperature probe design and clinical inferences