We studied 306 patients with chest pain (262 with coronary artery disease and 44 with no coronary artery disease) to determine which of 23 clinical, exercise, thallium, and angiographic variables best discriminate between patients with increased lung/heart ratios of thallium versus those with normal ratios. Normal lung/heart ratio values were defined using an additional 45 subjects with less than 1% probability of coronary artery disease. The number of diseased vessels was the best discriminator between patients with increased ratios versus those with normal ratios. Double product at peak exercise, number of segments with abnormal wall motion, patient gender, and duration of exercise were also significant discriminators. Using discriminant function analysis these variables could correctly identify 81% of cases with increased lung/heart ratios and 72% of cases with normal ratios. These results indicate that an increased lung/heart ratio of thallium reflects exercise-induced left ventricular dysfunction and affords a better understanding of why this thallium parameter is a powerful prognostic indicator in patients with chest pain.