INTRA AND INTER-OBSERVER RELIABILITY AND REPRODUCIBILITY OF A SEMIAUTOMATIC IMAGE ANALYSIS METHOD OVER ACHILLES TENDON ULTRASONOGRAPHY

Abstract

 Objectives: quantitative ultrasonography is a non-invasive technique that is useful for obtaining information for therapeutic evaluation and control of tendinous injuries; but, like ultrasonography, it is an examiner-dependent technique for which necessary to know the method’s reliability before clinical application. The aim of this study was to analyze intra- and inter-examiner reliability of the Achilles tendon border outlining. Material and method: it was recruited 15 (6 female, 9 males between 25 to 35 years old) asymptomatic volunteers of San Antonio Catholic University staff (Murcia-Spain) along 2009. It was taken cross-sectional and bilateral ultra sonographies of the Achilles tendon at 20, 40 and 60 mm from the insertion point onto calcaneus. Therefore, 90 images were analyzed. We used an ultrasound scanner (Sonosite S180) fitted with a linear transducer (L-38; 5 MHz-10 MHz). The Achilles tendon was contoured manually by two blinded examiners at two different moments. A training program was carried out by examiner B to increase the inter-examiner agreement. The image analyses were carried out with ImageJ software (v1.35). The intraclass correlation coefficient (ICC) was calculated and complemented with Bland and Altman plots.
Results: intra-examiner reliability of examiner A was excellent (ICC > 0.90) in all three scans. For examiner B, the ICC = 0.799 (CI: 0.620-0.899) denoted only a moderate reproducibility. The confidence interval of the ICC to 40 and 60 mm slides previous to training ranged from 0.528 to 0.869 and from 0.157 to 0.717 respectively. Conclusions: the reproducibility of method changes with the region of analysis from one examiner to another. Reliable data of the Achilles tendon perimeter can be obtained if the examiners undergo a previous agreement period.