A Novel Sonographic Method of Measuring Patellar Tendon Length

Abstract

Obtaining accurate and readily repeatable measurements is a prerequisite for using measures of soft tissue structures both clinically and in the research setting. Few studies have evaluated the interrater reliability of ultrasound measurements of tendons. The objective of this study was to determine the accuracy and reliability of a new method of sonographic measurement of patellar tendon length using direct dissection as the gold standard. Four cadaveric knees were sonographically evaluated by two independent investigators. Two custom designed straps with nylon strapping and stainless steel wire were used to firmly mark position on the leg and create an acoustic shadow on the ultrasound image. Anatomic landmarks were the distal patellar pole and the bony ridge on the anterior proximal tibia. After sonographic evaluation, the knee was dissected to expose the patellar tendon, which was measured using digital calipers. Intraclass correlation coefficients (ICC) were used to determine reliability of measurements between observers, where ICC >0.75 was considered good and >0.9 was considered excellent. Validity was measured using a Bland-Altman plot, which measures bias between measurement methods as well as variability of scatter. Three sonographic measurements were made by each investigator on each tendon. The length of each of the four tendons based on the mean values of sonographic measurements was 53.8 mm, 53.4 mm, 49.4 mm and 46.8 mm. The length based on visual inspection of the dissected tissue was 54.6 mm, 52.8 mm, 49.8 mm and 46.9 mm. The calculated ICC between raters was 0.96. On the Bland-Altman plot, the bias, or mean difference between sonographic and visual measures, was 0.17 mm, with a standard deviation of 0.71. The 95% limit of agreement was −1.55 to 1.22 mm. Measurement of patellar tendon length with ultrasound using adjustable surface markers and calipers is highly accurate and has good interrater reliability.

Introduction

High resolution sonography is an imaging modality well suited to evaluation of superficial soft tissue structures, with advantages including high axial resolution, wide availability, relatively low cost, relatively short time to conduct the test, real-time image capture, lack of ionizing radiation and capacity to image tissues dynamically. As a typical ultrasound transducer will have an axial resolving power of between 0.04 and 0.2 mm (Grassi and Cervini, 1998, Maganaris, 2005), this high spatial resolution theoretically makes it ideally suited for the calculation of physical dimensions of soft tissue structures such as tendon length and width. However, the absence of standardized protocols and the operator dependence of scanning technique have limited the wide adoption of ultrasound for this use.

The physical measurements of the patellar tendon are important in both the clinical and research settings. The importance of patellar tendon width and cross-sectional area has been established as an important marker of patellar tendinopathy and ultrasound is well suited to measure this (Davies et al. 1991). The length of the patellar tendon is also important clinically: it is the basis for the diagnosis of both patella alta and patella infera (Lancourt and Cristini 1975) and is also considered an important factor in patellar instability (Neyret 2002). In its use as a graft for anterior cruciate ligament (ACL) reconstruction, the patellar tendon’s length may be an important determinant of graft suitability, depending on the surgical approach. In a research context, the patellar tendon length is critical for accurate measurement of tendon strain, which is defined as the relative longitudinal deformation of a tendon under an applied load.

Obtaining accurate and readily repeatable measurements is a prerequisite for using measures of soft tissue structures both clinically and in the research setting. Few studies have evaluated the accuracy, inter- and intrarater reliability of ultrasound measurements of tendons. One study evaluating the in vivo measurement of cross-sectional tendon size demonstrated significant variation between observers when assessing a number of tendon measurements (O’Connor et al. 2004). This lack of inter- and intrarater reliability may be due to a number of factors, including operator experience, non-standardized imaging protocols and transducer positioning relative to the tendon under study. Compared with cross-sectional measurements, sonographic calculation of tendon length presents a different type of challenge due to the limited field of view on most ultrasound transducers. While measuring the length of a tendon would be relatively straightforward if the entire tendon could be visualized in a single image, the length of many tendons exceeds the length of the transducer. Therefore, visualization of the entire length of the tendon is often impossible with standard transducers.

Two methods have been proposed to overcome this difficulty and allow measurement of tendon length but neither has been validated with comparison to surgical or cadaveric measurements of the tendon. The first method utilizes extended field of view (EFOV) imaging, which enables generation of images of tendon length when the tendon is longer than the transducer (Pang and Ying 2006). This method relies on specialized software that combines overlapping images to produce a panoramic image, utilizing probe motion estimates; the image is created as the probe is swept manually across a length of the surface skin. Although this method has been evaluated in sonographic imaging phantoms (Weng et al., 1997, Fornage et al., 2000) and dissected animal tissue (Noorkoiv et al. 2010), to our knowledge, no studies of accuracy and reliability have been performed on human tendons. In imaging phantoms, the EFOV technique has been reported to have relative errors of 4% to 5%, which may be unacceptably high for some clinical and research applications. The second method of sonographically measuring tendon length utilizes surface markers, which correspond to the proximal and distal ends of the tendon and which can then be measured manually. The surface markers are usually strands of wire or needles, either manually stabilized or glued to the skin. While these methods have been used to measure tendon displacement in a number of biomechanical studies (Arya and Kulig 2010), the validity of the measurements has been questioned based on substantial movements of the skin markers, with some investigators finding that these methods may introduce measurement error of as much as 35% (Maganaris 2005). In addition to problems with accuracy, gluing markers to the skin may not be practical in many clinical settings where such measurements may be desired.

Our present study investigated a novel method of calculating tendon length using easily adjustable surface tissue markers and calipers. Cadaveric dissection was used as the measurement gold standard. Our primary hypothesis was that measured tendon lengths, as obtained with this technique, have a high level of accuracy compared with cadaveric dissection. Our secondary hypothesis was that these measurements have a high level of interrater reliability.