PURPOSE: Choosing the optimal radiographic methods to diagnose the cartilage height and degree of knee osteoarthritis is crucial to determine suitability for unicompartmental knee replacement. This study aims to evaluate and compare articular cartilage thickness measured using the Rosenberg view and coronal stress radiography. Intra- and interrater agreement and test-retest reliability of each method were determined. The hypothesis of the study was that the Rosenberg view and coronal stress radiographs provide similar assessments of articular cartilage height in the medial and lateral knee compartments of osteoarthritic knees.
METHODS: A prospective diagnostic study, including 73 patients was performed. Inclusion criteria were enrollment for either a medial unicompartmental or a total knee replacement. Radiographs were taken as the Rosenberg view, and coronal stress radiography using the Telos stress device. Repeated measurements were performed. Experienced knee surgeons performed measurements of cartilage height at a standardized location of joint space width (JSW), and a rater-perceived location of minimal joint space width (mJSW), thus allowing for reliability and agreement analysesusing weighted kappa. Coronal stress measurements were ultimately compared to the Rosenberg view using Spearman's rank correlation.
RESULTS: A total of 12,264 measurements were performed. The radiographic methods proved substantial reliability. Intra- and interrater agreement showed substantial to almost perfect agreement. A very strong correlation was observed in the medial knee compartment (r = 0.91; CI = 0.84-0.95; p < 0.001), with a mean difference of 0.1 mm and limits of agreement of - 1.5 to 1.7 mm, when comparing JSW between the Rosenberg view and varus stress. Only a strong correlation was observed medially when using mJSW, and when using this measurementmore incidences of bone-on-bone were observed than when measuring with JSW. A Strong correlation was observed in the lateral knee compartment (r = 0.83; CI = 0.71-0.89; p < 0.001), with a mean difference of 0.62 mm and limits of agreement of - 1.5 to 2.7 mm, when comparing JSW between the Rosenberg view and valgus stress.
CONCLUSION: The Rosenberg view is similar to 20° coronal valgus-varus stress radiography for determining articular cartilage thickness. Both techniques can be used in a clinical setting. Therefore, extra radiographs, equipment and expertise could be saved, when solely utilizing the Rosenberg view which is simple to perform.
LEVEL OF EVIDENCE: III.