Intra- and inter-rater reliability in the cross-sectional area of feline lumbar epaxial musculature evaluated via abdominal CT scan

Abstract

Objectives: The aim of this study was to evaluate the intra- and inter-rater reliability of epaxial muscle cross-sectional area measurement on feline CT images and to determine the relationship between normalized epaxial muscle area (EMA) and subjective muscle condition score (MCS).

Methods: Feline transverse CT images including the junction of the 13th thoracic vertebrae/13th rib head were retrospectively reviewed. Right and left epaxial muscle circumference and vertebral body height were measured and an average normalized EMA (ratio of epaxial area:vertebral height) was calculated for each image. Measurements were performed by three individuals blinded to the clinical data and were repeated 1 month later. Intra- and inter-rater reliability of EMA was assessed with concordance correlation coefficient (CCC), and Bland-Altman analysis was performed to assess bias and limits of agreement (LoA) between and within observers at different time points. In cats for which MCS data were available, EMA was compared between differing MCSs via the Kruskal-Wallis test, with Bonferroni-corrected Wilcoxon rank-sum post-hoc analysis.

Results: In total, 101 CT scans met the inclusion criteria for reliability analysis, 29 of which had muscle condition information available for analysis. Intra-rater EMA CCC ranged from 0.84 to 0.99 with minimal bias (range -0.16 to 0.08) and narrow LoA. Inter-rater EMA CCC ranged from 0.87 to 0.94, bias was larger (range -0.46 to 0.66) and LoA were wider when assessed between observers. Median EMA was significantly lower in cats with severe muscle atrophy (2.76, range 1.28-3.96) than in all other MCS groups (P <0.0001 for all comparisons).

Conclusions and relevance: Measurement of EMA on CT showed strong intra-rater reliability, and median EMA measurements were significantly lower in cats with severe muscle wasting, as assessed on physical examination. Further studies correlating EMA to lean muscle mass in cats are needed to determine whether this method may be useful to quantify muscle mass in patients undergoing a CT scan.

Keywords: Muscle atrophy; cachexia; muscle mass; sarcopenia.

Figure 1

Bland–Altman plots showing the intra-rater reliability for overall epaxial muscle area measured approximately 1 month apart: (a) rater EMG; (b) rater JMQ; (c) rater LHR. The muscle condition score is indicated for outlying points, when available. n = 101 feline CT scans

Figure 2

Bland–Altman plots showing the inter-rater reliability for overall epaxial muscle area: (a) rater EMG vs JMQ; (b) rater EMG vs LHR; (c) rater JMQ vs LHR. Muscle condition score is indicated for outlying points, when available. n = 101 feline CT scans

Figure 3

Box and whisker plots of epaxial muscle area (EMA) in cats with known muscle condition scores (n = 29) for each rater at each time point