Open AccessPostmortem validation of M RI cortical volume measurements in M S
Author(s) -
Popescu Veronica,
Klaver Roel,
Versteeg Adriaan,
Voorn Pieter,
Twisk Jos W.R.,
Barkhof Frederik,
Geurts Jeroen J.G.,
Vrenken Hugo
Publication year - 2016
Publication title -
human brain mapping
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 2.005
H-Index - 191
eISSN - 1097-0193
pISSN - 1065-9471
DOI - 10.1002/hbm.23168
Subject(s) - atrophy , magnetic resonance imaging , brain size , grey matter , pathology , cortex (anatomy) , medicine , anatomy , nuclear medicine , white matter , biology , radiology , neuroscience
Grey matter (GM) atrophy is a prominent aspect of multiple sclerosis pathology and an important outcome in studies. GM atrophy measurement requires accurate GM segmentation. Several methods are used in vivo for measuring GM volumes in MS, but assessing their validity in vivo remains challenging. In this postmortem study, we evaluated the correlation between postmortem MRI cortical volume or thickness and the cortical thickness measured on histological sections. Sixteen MS brains were scanned in situ using 3DT1‐weighted MRI and these images were used to measure regional cortical volume using FSL‐SIENAX, FreeSurfer, and SPM, and regional cortical thickness using FreeSurfer. Subsequently, cortical thickness was measured histologically in 5 systematically sampled cortical areas. Linear regression analyses were used to evaluate the relation between MRI regional cortical volume or thickness and histological cortical thickness to determine which postprocessing technique was most valid. After correction for multiple comparisons, we observed a significant correlation with the histological cortical thickness for FSL‐SIENAX cortical volume with manual editing (std. β = 0.345, adjusted R 2 = 0.105, P = 0.005), and FreeSurfer cortical volume with manual editing (std. β = 0.379, adjusted R 2 = 0.129, P = 0.003). In addition, there was a significant correlation between FreeSurfer cortical thickness with manual editing and histological cortical thickness (std. β = 0.381, adjusted R 2 = 0.130, P = 0.003). The results support the use of FSL‐SIENAX and FreeSurfer in cases of severe MS pathology. Interestingly none of the methods were significant in automated mode, which supports the use of manual editing to improve the automated segmentation. Hum Brain Mapp 37:2223–2233, 2016 . © 2016 Wiley Periodicals, Inc.