RESUMO
BACKGROUND AND PURPOSE: Because it is a small subcortical structure, the precise measurement of the human LGN is still a technical challenge. In this article, we identify the LGN in vivo, measure its volume based on high-resolution MR imaging, and then relate its volume to subject age to evaluate the potential clinical application. MATERIALS AND METHODS: A semiautomatic LGN isolation method was developed on scans obtained with 1.5T MR imaging, which involves highlighting the surrounding landmarks, obtaining candidate LGN voxels with a region-growing algorithm, and isolating the LGN from the ventral diencephalon. The method was accessed with a test-retest reliability on the results from 55 healthy subjects at different ages. RESULTS: This method showed high test-retest within-subject reliability (ICC, 0.950 and 0.948 in left and right hemispheres, respectively) among 3 independent measurements in each subject. The unilateral volume was highly variable, ranging from 52 to 102 mm(3) in the left and 66 to 105 mm(3) in the right hemisphere, with significantly larger volumes on the right (86 mm(3)) than on the left (77 mm(3)). The combined bilateral volumes (controlled for ICV) significantly decreased in size with progressing age from 20 to 65 years (r = -0.512, P = .000). There was no sex difference in bilateral LGN volumes (male/female: 163.1 ± 18.2/162.2 ± 21.4 mm(3)). CONCLUSIONS: Using our new technique, we were able to reliably determine the human LGN volume in vivo, which was found to decline with age. The volumes obtained by our method corresponded well with previously reported postmortem values, so our method may be considered to be superior for investigating the pathology of LGN.