Assessments of trochlear depth for patients with medial patellar luxation have traditionally been estimated from skyline radiographs or visually during surgery. Ultrasound offers a non-invasive evaluation method which could avoid the need for arthrotomy. We compared anatomic, radiographic and ultrasonographic measurements of trochlear depth using the red fox hind limb as a canine surrogate, dividing the trochlea into five regions from the origin of the caudal cruciate ligament to the proximal aspect of the trochlea.

We found reasonable agreement between anatomic and ultrasonographic measurements as assessed by Bland-Altman difference charting and limits of agreement (approximately ±0.7mm). Differences may be accounted for by ultrasound probe positioning errors, which should be taken into consideration during clinical investigations. Skyline views were difficult to standardise to a distinct position in the trochlea but appeared to centre around the middle of the trochlea: the proximal section of the trochlea has been suggested to be most critical in medial patellar luxation, and thus the usefulness of this view is questionable.

Trochlear depth in this study population was approximately 30% of patellar depth over the five regions examined. Measurement reliability for the ultrasonographic examinations was markedly less that for the skyline views (repeatability coefficient of 0.5mm vs 0.25mm), but still adequate for clinical use. A likely reason is the relative amount of detail recorded in the respective DICOM files which restricts ultrasonographic landmark resolution. Further work on quantifying interobserver variation and repeat measurement variation is required to ensure confidence in this technique.