Test-retest reliability of tibiofemoral joint space width measurements made using a low-dose standing CT scanner.

Objective: To determine the test-retest reliability of knee joint space width (JSW) measurements made using standing CT (SCT) imaging.
Subjects and Methods: This prospective two-visit study included 50 knees from 30 subjects (66% female; mean ± SD age 58.2 ± 11.3 years; BMI 29.1 ± 5.6 kg/m ² ; 38% KL grade 0-1). Tibiofemoral geometry was obtained from bilateral, approximately 20° fixed-flexed SCT images acquired at visits 2 weeks apart. For each compartment, the total joint area was defined as the area with a JSW <10 mm. The summary measurements of interest were the percentage of the total joint area with a JSW less than 0.5-mm thresholds between 2.0 and 5.0 mm in each tibiofemoral compartment. Test-retest reliability of the summary JSW measurements was assessed by intraclass correlation coefficients (ICC 2,1) for the percentage area engaged at each threshold of JSW and root-mean-square errors (RMSE) were calculated to assess reproducibility.
Results: The ICCs were excellent for each threshold assessed, ranging from 0.95 to 0.97 for the lateral and 0.90 to 0.97 for the medial compartment. RMSE ranged from 1.1 to 7.2% for the lateral and from 3.1 to 9.1% for the medial compartment, with better reproducibility at smaller JSW thresholds.
Conclusion: The knee joint positioning protocol used demonstrated high day-to-day reliability for SCT 3D tibiofemoral JSW summary measurements repeated 2 weeks apart. Low-dose SCT provides a great deal of information about the joint while maintaining high reliability, making it a suitable alternative to plain radiographs for evaluating JSW in people with knee OA.
Competing Interests: The corresponding author is named as an inventor on a patent application for the positioning frame used in this study, but has neither intellectual property rights nor financial rights to that patent and has not been employed or received honoraria, stock options, or other sources of financial support in relation to this work. The other authors report no conflicts of interest.