Changes in estimated total intracranial volume with age across 6 different datasets
Poster No:
2308
Submission Type:
Abstract Submission
Authors:
Rodrigo de Luis Garcia1, Rafael Navarro-González1, Álvaro Planchuelo-Gómez1, Santiago Aja-Fernández1, Juan Calabia del Campo2
Institutions:
1Universidad de Valladolid, Valladolid, Valladolid, 2Hospital Clínico Universitario de Valladolid, Valladolid, Valladolid
First Author:
Co-Author(s):
Introduction:
The Estimated Total Intracranial Volume (eTIV) is a measure of the total volume inside the skull, encompassing the brain, cerebrospinal fluid (CSF) and other intracranial structures. eTIV can be estimated from neuroimaging in number of ways. Freesurfer, for instance, uses a registration process to a standard space and computes the eTIV from the determinant of the transformation matrix [1].
Apart from using eTIV as a normalization factor in order to account for individual variations in head size when studying brain structures, eTIV can be employed as a biomarker on its own. In schizophrenia, for instance, several studies have found decreases in eTIV in patients with respect to controls [2, 3, 4], with reductions ranging from 0.13% to 4%.
If eTIV is to be employed as a neuroimaging biomarker, a detailed understanding of its aging trajectory is needed. Several cross-sectional studies have analyzed eTIV changes as a function of age [5, 6, 7]. Some of them identified age-related eTIV decreases at least from the 4th decade of life ranging from 0.1%/year to 0.29%/year, while others reported no changes. Assuming that intracranial volume must remain essentially stable because of biological reasons, those changes, when found, were interpreted as a consequence of generational growth in some populations. Finally, a few other studies have employed longitudinal data to estimate aging-related trajectories of eTIV. In [8], a small but detectable nonlinear aging pattern was detected in eTIV (with an average rate of change of 0.03%/year at age 20 and -0.09%/year at age 55).
In this abstract we aim at describing age-related changes in eTIV, as measured using Fastsurfer, across several neuroimaging datasets. Secondarily, assuming that these changes are primarily due to generational growth, we model the implications of such changes on neuroimaging studies focusing on eTIV as a biomarker.
Apart from using eTIV as a normalization factor in order to account for individual variations in head size when studying brain structures, eTIV can be employed as a biomarker on its own. In schizophrenia, for instance, several studies have found decreases in eTIV in patients with respect to controls [2, 3, 4], with reductions ranging from 0.13% to 4%.
If eTIV is to be employed as a neuroimaging biomarker, a detailed understanding of its aging trajectory is needed. Several cross-sectional studies have analyzed eTIV changes as a function of age [5, 6, 7]. Some of them identified age-related eTIV decreases at least from the 4th decade of life ranging from 0.1%/year to 0.29%/year, while others reported no changes. Assuming that intracranial volume must remain essentially stable because of biological reasons, those changes, when found, were interpreted as a consequence of generational growth in some populations. Finally, a few other studies have employed longitudinal data to estimate aging-related trajectories of eTIV. In [8], a small but detectable nonlinear aging pattern was detected in eTIV (with an average rate of change of 0.03%/year at age 20 and -0.09%/year at age 55).
In this abstract we aim at describing age-related changes in eTIV, as measured using Fastsurfer, across several neuroimaging datasets. Secondarily, assuming that these changes are primarily due to generational growth, we model the implications of such changes on neuroimaging studies focusing on eTIV as a biomarker.
Methods:
3602 subjects from 6 different publicly available datasets were included (see Figure 1). From the T1w MRI images, Fastsurfer [9] was employed to extract morphometric features, and eTIV was selected among them. Fastsurfer employs Deep Learning to perform brain segmentation based on the Desikan-Killiany atlas.
Separate linear regression models were fitted for each dataset, considering age and eTIV independently for female and male subjects.
If eTIV varies strongly with age in a certain population, this will cause a dispersion in the eTIV values of the subjects of that cohort that in principle could threaten the ability of the study to find significant differences in eTIV, even if they do exist. In order to model this effect, we assumed an age range of 20-70 years. Next, we considered different scenarios regarding the difference in eTIV between diagnostic groups (2%, 2.5% and 3%). For each scenario, 10,000 simulations were run, where in each simulation 200 synthetic eTIV values were generated, 100 for each diagnostic group. Finally, we computed the median p-value obtained from performing t-tests on the eTIV values for both diagnostic groups.
Separate linear regression models were fitted for each dataset, considering age and eTIV independently for female and male subjects.
If eTIV varies strongly with age in a certain population, this will cause a dispersion in the eTIV values of the subjects of that cohort that in principle could threaten the ability of the study to find significant differences in eTIV, even if they do exist. In order to model this effect, we assumed an age range of 20-70 years. Next, we considered different scenarios regarding the difference in eTIV between diagnostic groups (2%, 2.5% and 3%). For each scenario, 10,000 simulations were run, where in each simulation 200 synthetic eTIV values were generated, 100 for each diagnostic group. Finally, we computed the median p-value obtained from performing t-tests on the eTIV values for both diagnostic groups.
Results:
Figure 1 represents the values of eTIV as a function of age for the different datasets, together with the fitted linear regression models. Most (but not all) datasets showed a significant effect of age on eTIV, with changes ranging from -0.02%/year to -0.22%/year. Changes were more pronounced for females than for males.
Figure 2 shows the evolution of the median p-value for our simulations as a function of the rate of change of eTIV vs age. Although in principle increases in the slope could hamper the ability to detect significant changes between diagnostic groups, for the values encountered in our results over the different datasets or reported in the literature (slope < 0.30%/year) this effect is almost negligible.
Figure 2 shows the evolution of the median p-value for our simulations as a function of the rate of change of eTIV vs age. Although in principle increases in the slope could hamper the ability to detect significant changes between diagnostic groups, for the values encountered in our results over the different datasets or reported in the literature (slope < 0.30%/year) this effect is almost negligible.
Conclusions:
Our results suggest that changes of eTIV with age in a particular population are not a relevant confounding factor in studies using eTIV as a biomarker.
Lifespan Development:
Aging 2
Modeling and Analysis Methods:
Other Methods
Neuroanatomy, Physiology, Metabolism and Neurotransmission:
Cortical Anatomy and Brain Mapping
Neuroanatomy Other
Novel Imaging Acquisition Methods:
Anatomical MRI 1
Keywords:
Aging
MRI
Other - Estimated Intracranial Volume
1|2Indicates the priority used for review
Provide references using author date format
[2] Baaré, W. F. et al. (2001), ‘Volumes of brain structures in twins discordant for schizophrenia’, Archives of General Psychiatry, 58(1), 33-40.
[3] Haijma, S. V. et al. (2013), ‘Brain volumes in schizophrenia: a meta-analysis in over 18 000 subjects’, Schizophrenia Bulletin, 39(5), 1129-1138.
[4] Van Erp, T. G. et al. (2016), ‘Subcortical brain volume abnormalities in 2028 individuals with schizophrenia and 2540 healthy controls via the ENIGMA consortium’, Molecular Psychiatry, 21(4), 547-553.
[5] DeCarli, C. et al. (2005), ‘Measures of brain morphology and infarction in the Framingham heart study: establishing what is normal’, Neurobiology of Aging, 26(4), 491-510.
[6] Fillmore, P. T. et al. (2015). ‘Age-specific MRI brain and head templates for healthy adults from 20 through 89 years of age’, Frontiers in Aging Neuroscience, 7, 44.
[7] Ricard, A. S. et al. (2010), ‘On two equations about brain volume, cranial capacity and age’, Surgical and Radiologic Anatomy, 32, 989-995.
[8] Caspi, Y. et al. (2020), ‘Changes in the intracranial volume from early adulthood to the sixth decade of life: A longitudinal study’, NeuroImage, 220, 116842.
[9] Henschel L. et al. (2020), ‘Fastsurfer - A fast and accurate deep learning based neuroimaging pipeline’, NeuroImage 219:117012.