Voxel-based Morphometry in Single Subjects Without a Scanner-specific Normal Database Using a Convolutional Neural Network

Overview

Journal Eur Radiol

Specialty Radiology

Date 2023 Nov 9

PMID 37943313

Authors

Julia Kruger

Roland Opfer

Lothar Spies

Dennis Hedderich

Ralph Buchert

Affiliations

Soon will be listed here.

Abstract

Objectives: Reliable detection of disease-specific atrophy in individual T1w-MRI by voxel-based morphometry (VBM) requires scanner-specific normal databases (NDB), which often are not available. The aim of this retrospective study was to design, train, and test a deep convolutional neural network (CNN) for single-subject VBM without the need for a NDB (CNN-VBM).

Materials And Methods: The training dataset comprised 8945 T1w scans from 65 different scanners. The gold standard VBM maps were obtained by conventional VBM with a scanner-specific NDB for each of the 65 scanners. CNN-VBM was tested in an independent dataset comprising healthy controls (n = 37) and subjects with Alzheimer's disease (AD, n = 51) or frontotemporal lobar degeneration (FTLD, n = 30). A scanner-specific NDB for the generation of the gold standard VBM maps was available also for the test set. The technical performance of CNN-VBM was characterized by the Dice coefficient of CNN-VBM maps relative to VBM maps from scanner-specific VBM. For clinical testing, VBM maps were categorized visually according to the clinical diagnoses in the test set by two independent readers, separately for both VBM methods.

Results: The VBM maps from CNN-VBM were similar to the scanner-specific VBM maps (median Dice coefficient 0.85, interquartile range [0.81, 0.90]). Overall accuracy of the visual categorization of the VBM maps for the detection of AD or FTLD was 89.8% for CNN-VBM and 89.0% for scanner-specific VBM.

Conclusion: CNN-VBM without NDB provides a similar performance in the detection of AD- and FTLD-specific atrophy as conventional VBM.

Clinical Relevance Statement: A deep convolutional neural network for voxel-based morphometry eliminates the need of scanner-specific normal databases without relevant performance loss and, therefore, could pave the way for the widespread clinical use of voxel-based morphometry to support the diagnosis of neurodegenerative diseases.

Key Points: • The need of normal databases is a barrier for widespread use of voxel-based brain morphometry. • A convolutional neural network achieved a similar performance for detection of atrophy than conventional voxel-based morphometry. • Convolutional neural networks can pave the way for widespread clinical use of voxel-based morphometry.

Citing Articles

Searching for the grail: may machine learning be a road to clinical use of brain MRI segmentation?.

Quarantelli M Eur Radiol. 2023; 34(6):3575-3577.

PMID: 37975922 DOI: 10.1007/s00330-023-10438-0.

References

Ashburner J, Friston K . Voxel-based morphometry--the methods. Neuroimage. 2000; 11(6 Pt 1):805-21. DOI: 10.1006/nimg.2000.0582. View

Ashburner J, Friston K . Why voxel-based morphometry should be used. Neuroimage. 2001; 14(6):1238-43. DOI: 10.1006/nimg.2001.0961. View

Goto M, Abe O, Hagiwara A, Fujita S, Kamagata K, Hori M . Advantages of Using Both Voxel- and Surface-based Morphometry in Cortical Morphology Analysis: A Review of Various Applications. Magn Reson Med Sci. 2022; 21(1):41-57. PMC: 9199978. DOI: 10.2463/mrms.rev.2021-0096. View

Caspers J, Heeger A, Turowski B, Rubbert C . Automated age- and sex-specific volumetric estimation of regional brain atrophy: workflow and feasibility. Eur Radiol. 2020; 31(2):1043-1048. PMC: 7813701. DOI: 10.1007/s00330-020-07196-8. View

Sato R, Kudo K, Udo N, Matsushima M, Yabe I, Yamaguchi A . A diagnostic index based on quantitative susceptibility mapping and voxel-based morphometry may improve early diagnosis of Alzheimer's disease. Eur Radiol. 2022; 32(7):4479-4488. DOI: 10.1007/s00330-022-08547-3. View

Koikkalainen J, Rhodius-Meester H, Frederiksen K, Bruun M, Hasselbalch S, Baroni M . Automatically computed rating scales from MRI for patients with cognitive disorders. Eur Radiol. 2019; 29(9):4937-4947. DOI: 10.1007/s00330-019-06067-1. View

Dash S, Stezin A, Takalkar T, George L, Kamble N, Netravathi M . Abnormalities of white and grey matter in early multiple system atrophy: comparison of parkinsonian and cerebellar variants. Eur Radiol. 2018; 29(2):716-724. DOI: 10.1007/s00330-018-5594-9. View

Geng B, Gao M, Wu J, Yang G, Liu C, Piao R . Smaller volume and altered functional connectivity of the amygdala in patients with lifelong premature ejaculation. Eur Radiol. 2021; 31(11):8429-8437. DOI: 10.1007/s00330-021-08002-9. View

Li S, Wang B, Li C, Feng Y, Li M, Wang T . Progressive gray matter hypertrophy with severity stages of insomnia disorder and its relevance for mood symptoms. Eur Radiol. 2021; 31(8):6312-6322. DOI: 10.1007/s00330-021-07701-7. View

10.

Chen Y, Chen H, Lu C, Chen M, Chou K, Tsai N . Reduced lateral occipital gray matter volume is associated with physical frailty and cognitive impairment in Parkinson's disease. Eur Radiol. 2018; 29(5):2659-2668. DOI: 10.1007/s00330-018-5855-7. View

11.

Yoshiura T, Hiwatashi A, Yamashita K, Ohyagi Y, Monji A, Takayama Y . Deterioration of abstract reasoning ability in mild cognitive impairment and Alzheimer's disease: correlation with regional grey matter volume loss revealed by diffeomorphic anatomical registration through exponentiated lie algebra analysis. Eur Radiol. 2010; 21(2):419-25. DOI: 10.1007/s00330-010-1939-8. View

12.

Hall A, Moore R, Lopez O, Kuller L, Becker J . Basal forebrain atrophy is a presymptomatic marker for Alzheimer's disease. Alzheimers Dement. 2008; 4(4):271-9. PMC: 2517158. DOI: 10.1016/j.jalz.2008.04.005. View

13.

Wu J, Zhang Q, Sun P, Zhang H, Gao M, Ma M . Gray matter microstructural alterations in manganese-exposed welders: a preliminary neuroimaging study. Eur Radiol. 2022; 32(12):8649-8658. DOI: 10.1007/s00330-022-08908-y. View

14.

Zikou A, Kosmidou M, Astrakas L, Tzarouchi L, Tsianos E, Argyropoulou M . Brain involvement in patients with inflammatory bowel disease: a voxel-based morphometry and diffusion tensor imaging study. Eur Radiol. 2014; 24(10):2499-506. DOI: 10.1007/s00330-014-3242-6. View

15.

Porto L, Preibisch C, Hattingen E, Bartels M, Lehrnbecher T, Dewitz R . Voxel-based morphometry and diffusion-tensor MR imaging of the brain in long-term survivors of childhood leukemia. Eur Radiol. 2008; 18(11):2691-700. DOI: 10.1007/s00330-008-1038-2. View

16.

Salmond C, de Haan M, Friston K, Gadian D, Vargha-Khadem F . Investigating individual differences in brain abnormalities in autism. Philos Trans R Soc Lond B Biol Sci. 2003; 358(1430):405-13. PMC: 1693120. DOI: 10.1098/rstb.2002.1210. View

17.

Colliot O, Bernasconi N, Khalili N, Antel S, Naessens V, Bernasconi A . Individual voxel-based analysis of gray matter in focal cortical dysplasia. Neuroimage. 2005; 29(1):162-71. DOI: 10.1016/j.neuroimage.2005.07.021. View

18.

Scarpazza C, Sartori G, De Simone M, Mechelli A . When the single matters more than the group: very high false positive rates in single case Voxel Based Morphometry. Neuroimage. 2013; 70:175-88. DOI: 10.1016/j.neuroimage.2012.12.045. View

19.

Suzuki F, Sato N, Ota M, Sugiyama A, Shigemoto Y, Morimoto E . Discriminating chorea-acanthocytosis from Huntington's disease with single-case voxel-based morphometry analysis. J Neurol Sci. 2019; 408:116545. DOI: 10.1016/j.jns.2019.116545. View

20.

Hedderich D, Dieckmeyer M, Andrisan T, Ortner M, Grundl L, Schon S . Normative brain volume reports may improve differential diagnosis of dementing neurodegenerative diseases in clinical practice. Eur Radiol. 2020; 30(5):2821-2829. DOI: 10.1007/s00330-019-06602-0. View