Standardization of metaiodobenzylguanidine heart to mediastinum ratio using a calibration phantom: effects of correction on normal databases and a multicentre study.

This study was performed to demonstrate that the results obtained with a calibration phantom could be used as a tool for standardizing measurement of heart to mediastinum (H/M) ratio in cardiac metaiodobenzylguanidine (MIBG) imaging. Images of the phantom containing (123)I-MIBG were acquired on the cameras in 10 hospitals (11 camera types) to determine the relationship between H/M ratios using different collimators: low-energy (LE) and medium-energy (ME)/low-medium-energy (LME) collimators. The effect of standardization on the ME-comparable H/M ratio was examined in two settings: a Japanese standard MIBG database (n = 62) and multicentre studies (n = 49). In a multicentre study, probable Alzheimer's disease (AD, n = 18) and probable dementia with Lewy bodies (DLB, n = 31) were studied and standardized by the calibration phantom method. Linear regression equations between LE and ME collimators were obtained for the phantom study in all institutions. When the H/M ratio with an LE collimator was corrected based upon the calibration phantom, the corrected values were comparable to those obtained using ME collimators. The standard database also exhibited a normal distribution after standardization as determined by skewness and goodness-of-fit test. A mixture of the populations by LE and ME collimators showed significant separation of AD and DLB groups (F ratio = 24.9 for the late H/M), but the corrected values resulted in higher F ratios for both early and late H/M (F ratio = 34.9 for the late H/M). Standardization of H/M ratios by the heart-chest calibration phantom method is feasible among different collimator types. This method could be practically used for multicentre comparison of H/M ratios.