Abstract
BACKGROUND: Septal penetration causes collimator-dependent differences in the heart-to-mediastinum (H/M) ratio in (123)I-metaiodobenzylguanidine (MIBG) cardiac imaging. We investigated generally applicable methods to correct such differences. METHODS AND RESULTS: Four hours after (123)I-MIBG injection, 40 patients underwent anterior chest imaging successively with medium-energy (ME) and various non-ME collimators. The H/M ratios obtained with the non-ME collimators before and after (123)I-dual-window penetration correction were compared with the ME-derived standard values to determine patient-based conversion equations for empiric and combined corrections, respectively. A (123)I point source was imaged with various collimators, and the central ratio, the ratio of count in a small central region of interest to count in a large one, was calculated. The method of predicting the conversion equations from the central ratios was determined. Correction using the patient-based conversion equations removed systematic underestimation of the H/M ratios obtained with the non-ME collimators, and combined correction depressed residual random errors to some degree. Point-source-based equations yielded results comparable to the patient-based equations. CONCLUSIONS: Empiric and combined corrections effectively reduce collimator-dependent differences in the H/M ratio. The conversion equations can be predicted from simple point-source imaging, which would allow to apply these corrections to data obtained with various collimators.