Objective analysis of partial three-dimensional rotator cuff muscle volume and fat infiltration across ages and sex from clinical MRI scans

Abstract

Objective analysis of rotator cuff (RC) atrophy and fatty infiltration (FI) from clinical MRI is limited by qualitative measures and variation in scapular coverage. The goals of this study were to: develop/evaluate a method to quantify RC muscle size, atrophy, and FI from clinical MRIs (with typical lateral only coverage) and then quantify the effects of age and sex on RC muscle. To develop the method, 47 full scapula coverage CTs with matching clinical MRIs were used to: correct for variation in scan capture, and ensure impactful information of the RC is measured. Utilizing this methodology and automated artificial intelligence, 170 healthy clinical shoulder MRIs of varying age and sex were segmented, and each RC muscle's size, relative contribution, and FI as a function of scapula location were quantified. A two-way ANOVA was used to examine the effect of age and sex on RC musculature. The analysis revealed significant (p < 0.05): decreases in size of the supraspinatus, teres minor, and subscapularis with age; decreased supraspinatus and increased infraspinatus relative contribution with age; and increased FI in the infraspinatus with age and in females. This study demonstrated that clinically obtained MRIs can be utilized for automatic 3D analysis of the RC. This method is not susceptible to coverage variation or patient size. Application of methodology in a healthy population revealed differences in RC musculature across ages and FI level between sexes. This large database can be used to reference expected muscle characteristics as a function of scapula location and could eventually be used in conjunction with the proposed methodology for analysis in patient populations.

Figure 1

Overview of the paper’s methodology workflow and steps used for this paper. The process of analyzing and interpreting clinical rotator cuff MRI scans in this paper are summarized in four steps.

Figure 2

(A) Example segmentation of the scapula on a MRI (top) and matching CT (bottom) scan. (B) Registering the full CT (red) to the original partial MRI (blue) 3D rendering of the scapula to get final label maps. (C) Measuring the points used to relate the partial MRI scan’s scapula to the total scapula length from the CT scan. (Left) The scapula’s CSA as a function of its sagittal length for the MRI and CT scan. (Right) the scapula morphology at the point of peak scapula CSA. Sagittal distance of peak CSA (PD), CSA at point (PC), vertical (VB) and horizontal (HB) bounding length of the scapula at this point, and total scapula length.

Figure 3

Visual overview of the rotator cuff processing. (A) Segmenting the RoIs from the MRI scan. (B) 3D-volume rendering of the RoI segmentation. (C) Cumulative volume (ml) of the scapula (solid line), a muscle (dotted line), and intramuscular fat (dashed line) RoI as a function of distance along the scapula (mm) moving medially. (D) Predicting the total scapula length (mm) from morphological features of the lateral scapula. (E) The cumulative muscle volume normalized by the scapula volume and presented as a function of percent distant into scapula (%) for the muscle size, relative contribution, and fat infiltration volumetric score.

Figure 4

Demonstration of muscle morphology of the RC as a function of percent distance of scapula. (A) Anterior and posterior views of the coverage range as well as MRI slice examples (scapula labeled in white). (B) The number of scans from the control database (n = 170) that had image coverage at each percent of the scapula. Based on the 170 scans collected from the control database, approximately 90% of the scans captured at least 40% of the scapula, while less than 50% of the scans had 50% of the scapula captured.

Figure 5

Correlation results for values found from partial lateral scapula coverages (10%, 20%, 30%, and 40%) as compared to values found from total coverage for the supraspinatus (blue), infraspinatus (orange), teres minor (yellow), and subscapularis (green) with example 3D volume rendering from an analyzed patient. (A) Raw volume (ml) correlation for all muscles from volume found at total coverage compared to that found at 40% lateral scapula coverage. (B) Raw volume (ml) correlation for the supraspinatus found at total coverage compared to that found at 10% (black), 20% (grey), 30% (light blue), and 40% (blue) lateral scapula coverage. (C) Normalized volume correlation for all muscles found at total coverage compared to that found at 40% lateral scapula coverage. (D) Relative contribution (%) correlation for all muscles found at total coverage compared to that found at 40% lateral scapula coverage.

Figure 6

Relationship between partial scapula volume and rotator cuff muscle volume as a function of location along the scapula. (A) Correlation coefficient for the partial scapula volume related to the supraspinatus (blue), infraspinatus (orange), teres minor (yellow), and subscapularis (green) as a function of location along scapula (%) when analyzed as a group (top) and when split by sex (bottom) via males (solid) and females (dashed). A Pearson’s correlation was used if both datasets were normally distributed (using Shapiro wilks), otherwise Spearman’s correlation was used. All values had a p < 0.05. (B) Example correlation between the scapula and supraspinatus volume (ml) at 30%. Male participants are a circle symbol, while female participants are an x symbol. Patients between 15–29 years (dark blue), 30–39 years. (blue), 40–49 (light blue), 50–59 years. (yellow), 60–69 (orange), and 70–89 (red) are shown. While patients older than 40 where not used for analysis, they are shown here for visualization.

Figure 7

Analysis of the effects of age and sex on each rotator cuff muscle at 30% coverage for the metrics of (A) muscle size (normalized volume), (B) relative contribution (%), and (C) fat infiltration (%) for all six age subgroups for males (black) and females (light grey). Means are demonstrated by a horizontal line, and individual data points as dots. Shown is post-hoc analysis, in which * signifies a significance level of p < 0.05, ° is p < 0.01, and • is p < 0.001.