Quantifying lumbar paraspinal intramuscular fat: Accuracy and reliability of automated thresholding models

E O Wesselink; J M Elliott; A Pool-Goudzwaard; M W Coppieters; P P Pevenage; A Di Ieva; K A Weber Ii

doi:10.1016/j.xnsj.2024.100313

Quantifying lumbar paraspinal intramuscular fat: Accuracy and reliability of automated thresholding models

N Am Spine Soc J. 2024 Jan 24:17:100313. doi: 10.1016/j.xnsj.2024.100313. eCollection 2024 Mar.

Authors

E O Wesselink^{1

2}, J M Elliott³, A Pool-Goudzwaard^{1

4}, M W Coppieters^{1

5}, P P Pevenage⁶, A Di Ieva⁷, K A Weber Ii²

Affiliations

¹ Faculty of Behavioural and Movement Sciences, Amsterdam Movement Sciences - Program Musculoskeletal Health, Vrije Universiteit Amsterdam, Amsterdam, The Netherlands.
² Division of Pain Medicine, Department of Anesthesiology, Perioperative and Pain Medicine, Stanford University School of Medicine, Palo Alto, CA, United States.
³ The University of Sydney, Faculty of Medicine and Health and the Northern Sydney Local Health District, The Kolling Institute, Sydney, Australia.
⁴ SOMT University of Physiotherapy, Amersfoort, The Netherlands.
⁵ Menzies Health Institute Queensland, School of Health Sciences and Social Work, Griffith University, Brisbane and Gold Coast, Australia.
⁶ MRI Centrum, Amsterdam, The Netherlands.
⁷ Computational Neurosurgery (CNS) Lab, Macquarie Medical School, Faculty of Medicine, Health and Human Sciences, Macquarie University, Level 1, 75 Talavera Road, Sydney, NSW 2109, Australia.

Abstract

Background: The reported level of lumbar paraspinal intramuscular fat (IMF) in people with low back pain (LBP) varies considerably across studies using conventional T₁- and T₂-weighted magnetic resonance imaging (MRI) sequences. This may be due to the different thresholding models employed to quantify IMF. In this study we investigated the accuracy and reliability of established (two-component) and novel (three-component) thresholding models to measure lumbar paraspinal IMF from T₂-weighted MRI.

Methods: In this cross-sectional study, we included MRI scans from 30 people with LBP (50% female; mean (SD) age: 46.3 (15.0) years). Gaussian mixture modelling (GMM) and K-means clustering were used to quantify IMF bilaterally from the lumbar multifidus, erector spinae, and psoas major using two and three-component thresholding approaches (GMM_2C; K-means_2C; GMM_3C; and K-means_3C). Dixon fat-water MRI was used as the reference for IMF. Accuracy was measured using Bland-Altman analyses, and reliability was measured using ICC_3,1. The mean absolute error between thresholding models was compared using repeated-measures ANOVA and post-hoc paired sample t-tests (α = 0.05).

Results: We found poor reliability for K-means_2C (ICC_3,1 ≤ 0.38), moderate to good reliability for K-means_3C (ICC_3,1 ≥ 0.68), moderate reliability for GMM_2C (ICC_3,1 ≥ 0.63) and good reliability for GMM_3C (ICC_3,1 ≥ 0.77). The GMM (p < .001) and three-component models (p < .001) had smaller mean absolute errors than K-means and two-component models, respectively. None of the investigated models adequately quantified IMF for psoas major (ICC_3,1 ≤ 0.01).

Conclusions: The performance of automated thresholding models is strongly dependent on the choice of algorithms, number of components, and muscle assessed. Compared to Dixon MRI, the GMM performed better than K-means and three-component performed better than two-component models for quantifying lumbar multifidus and erector spinae IMF. None of the investigated models accurately quantified IMF for psoas major. Future research is needed to investigate the performance of thresholding models in a more heterogeneous clinical dataset and across different sites and vendors.

Keywords: Adiposity; Back muscles; Low back pain; Machine learning; Magnetic resonance imaging; Thresholding.

Abstract

Grants and funding