Insufficient accuracy of the ultrasound-based determination of Achilles tendon cross-sectional area.

Abstract

The accurate assessment of the Achilles tendon cross-sectional area (CSA) is a crucial prerequisite to investigate tendon adaptation and to calculate the tendon Young׳s modulus. Besides magnetic resonance imaging (MRI), ultrasonography emerged as an alternative imaging technique. The purpose of the present study was to assess the objectivity and reliability of the ultrasound-based methodology and its validity with respect to the established MRI-based assessment. The Achilles tendon CSA from 17 healthy males was assessed by ultrasonography at the proximal, medial and distal position five times each on two separate days and three observers segmented the CSAs. For the validation, MRI-based CSA measurements were provided accordingly. The analysis of the ultrasound-based method revealed significant observer and day effects (p<0.05), despite high intra-class correlation coefficients for the three observers (>0.93) and both days (>0.89), respectively. The mean typical percentage error of both days was 7.1%. Comparing both methods, ultrasonography underestimated (19%) the CSA values obtained by MRI (p<0.05). Although the correlation coefficients of both methods were high in all three positions (>0.88), the respective absolute difference of in average 12mm² and limits of agreement of up to +28mm² and -2mm² (~37% of the mean CSA) evidence considerable inconsistency. Besides the dependency of the CSA determination from the observer, the ultrasound method is not sensitive enough to detect physiological changes in tendon morphology of the magnitude that is reported by intervention studies. Furthermore, the validity analysis revealed a lack of agreement between both methods. Therefore, the ultrasound-based methodology cannot be recommended for an accurate Achilles tendon CSA determination in vivo.