Ultrasound transducer shape has no effect on measurements of lumbar multifidus muscle size

Evidence is currently lacking for guidance on ultrasound transducer configuration (shape) when imaging muscle to measure its size. This study compared measurements made of lumbar multifidus on images obtained using curvilinear and linear transducers.

Fifteen asymptomatic males (aged 21-32 years) had their right lumbar multifidus imaged at L3. Two transverse images were taken with two transducers (5 MHz curvilinear and 6 MHz linear), and linear and cross-sectional area (CSA) measurements were made off-line. Reliability of image interpretation was shown using intra-class correlation coefficients (0.78-0.99). Muscle measurements were compared between transducers using Bland and Altman plots and paired t-tests. Relationships between CSA and linear measurements were examined using Pearson's Correlation Coefficients.

There were no significant differences (p > 0.05) in the measurements of the two transducers. Thickness and CSA measurements had small differences between transducers, with mean differences of 0.01 cm (SDdiff = 0.21 cm) and 0.03 cm(2) (SDdiff = 0.58 cm(2)) respectively. Width measures had a mean difference of 0.14 cm, with the linear transducer giving larger measures. Significant correlations (p < 0.001) were found between all linear measures and CSA, with both transducers (r = 0.78-0.89).

Measurements of multifidus at L3 were not influenced by the configuration of transducers of similar frequency. For the purposes of image interpretation, the curvilinear transducer produced better definition of the lateral muscle border, suggesting it as the preferable transducer for imaging lumbar multifidus.