The reliability of measuring quadriceps cross-sectional area with compound B ultrasound scanning

Many therapists use thigh circumference as a measure of quadriceps size to evaluate the effect of an injury or effectiveness of an intervention. However, this technique has been shown to be unreliable, thus more accurate methods of measuring muscle size are required. The purpose of this study was to investigate the validity and reliability of measuring quadriceps cross-sectional area (CSA) with static compound B ultrasound scanning at the level of the mid-thigh. Repeated measures of known linear distances and CSAs were made with a planimetry device. Both face and criterion validity were demonstrated indicating that the device is reliable when measuring CSAs. A single rater located quadriceps muscle borders and measured CSAs on a series of ultrasound scans. A mean coefficient of variation (%CV = 1.7) was observed, signifying high intra-rater reliability. Two raters performed six scans on each of 15 subjects. Initial observation of these results suggest that the measurement of quadriceps CSA with compound B ultrasound scanning is reliable, mean %CVs = 2.8 and 1.9 for the two raters and 2.4 when their results were combined. When raters performed scans on 18 subjects on two occasions on the same day intra-class correlation coefficients (ICC) were high, 0.991 and 0.980 for raters A and B, respectively. Paired Student's t-tests revealed a significant difference (p < 0.05) for Rater B between CSAs measured for initial and relocated scans. However, this difference was calculated as 1.4-2.9% and was not of clinical importance. A significant difference (p < 0.05) was observed between measurements of CSAs made by both raters on the same subject. A calibration factor was calculated for transforming the measurements of Rater B, the use of which gave comparable values. ICCs using a fixed model for inter-rater reliability were 0.963 and 0.974 for actual and transformed values, respectively. This study observed differences between measures obtained by different raters and recommends that the calculation and employment of a calibration factor may be useful when comparing sequential measures made by different raters.