Manually defining regions of interest when quantifying paravertebral muscles fatty infiltration from axial magnetic resonance imaging: a proposed method for the lumbar spine with anatomical cross-reference

Abstract

Background: There is increasing interest in paravertebral muscle composition as a potential prognostic and diagnostic element in lumbar spine health. As a consequence, it is becoming popular to use magnetic resonance imaging (MRI) to examine muscle volume and fatty infiltration in lumbar paravertebral muscles to assess both age-related change and their clinical relevance in low back pain (LBP). A variety of imaging methods exist for both measuring key variables (fat, muscle) and for defining regions of interest, making pooled comparisons between studies difficult and rendering post-production analysis of MRIs confusing. We therefore propose and define a method as an option for use as a standardized MRI procedure for measuring lumbar paravertebral muscle composition, and to stimulate discussion towards establishing consensus for the analysis of skeletal muscle composition amongst clinician researchers.

Method: In this descriptive methodological study we explain our method by providing an examination of regional lumbar morphology, followed by a detailed description of the proposed technique. Identification of paravertebral muscles and vertebral anatomy includes axial E12 sheet-plastinates from cadaveric material, combined with a series of axial MRIs that encompass sequencing commonly used for investigations of muscle quality (fat-water DIXON, T1-, and T2-weighted) to illustrate regional morphology; these images are shown for L1 and L4 levels to highlight differences in regional morphology. The method for defining regions of interest (ROI) for multifidus (MF), and erector spinae (ES) is then described.

Results: Our method for defining ROIs for lumbar paravertebral muscles on axial MRIs is outlined and discussed in relation to existing literature. The method provides a foundation for standardising the quantification of muscle quality that particularly centres on examining fatty infiltration and composition. We provide recommendations relating to imaging parameters that should additionally inform a priori decisions when planning studies examining lumbar muscle tissues with MRI.

Conclusions: We intend this method to provide a platform towards developing and delivering meaningful comparisons between MRI data on lumbar paravertebral muscle quality.

Keywords: Erector spinae; Fat infiltration; Lumbar spine; Magnetic resonance imaging; Manual segmentation; Multifidus; Paravertebral muscles; Region of interest.

Fig. 1

Axial E12 plastinated sections (a, c) and schematic illustrations (b, d) at approximately L1 (a, b) and L4 (c, d) highlighting anatomical structures at these vertebral levels. b, d Dotted lines and shading, Green - psoas major muscle; Blue – quadratus lumborum muscle; Purple – erector spinae muscles; Red – spinotransverse muscles. b round white dotted regions (bilateral) denote 12th rib. d square dotted box surrounds enlarged inset; round dotted circle indicates morphological feature of interest (ILB fatty ‘tent’). Legend: A – aorta; ES – erector spinae muscles; ESA – erector spinae aponeurosis; ILB – iliocostalis – longissimus boundary and indentation; ISL – interspinous ligament; IT – intertransversarii muscle; IVC – inferior vena cava; K – kidney; L – liver; P – psoas major muscle; QL – quadratus lumborum muscle; SAF – superior articular facet; SP – spinous process; SPC – spinal canal; SPT – spinotransverse muscle group; ZJ – zygapophysial joint

Fig. 2

Axial MRIs at the L1/2 disc level of a 47 year old male depicting fat- (a) and water-separated (b) chemical shift, and T2- (c) and T1-weighted (d and e, same) images. e Dotted lines and shading depicting the regions of interest for: Green - psoas muscle; Blue – quadratus lumborum muscle; Purple – erector spinae group (longissimus and iliocostalis together); and Red – spinotransverse muscles (predominantly multifidus)

Fig. 3

Axial MRIs at the L4/5 disc level of a 47 year old male depicting fat- (a) and water-separated (b) chemical shift, and T2- (c) and T1-weighted (d and e; same image cropped anteriorly therein truncating psoas) images. e Dotted lines and shading depicting the regions of interest for: Green - psoas muscle (truncated at the ventral border due to limited visibility in image); Blue – quadratus lumborum muscle; Purple – erector spinae group (longissimus and iliocostalis together); and Red – spinotransverse muscles (predominantly multifidus)