High-Fidelity Measurement-Based Detection Values and Roles of Three-Dimensional Joint Space Width in Evaluation of Knee Osteoarthritis.

objective3D joint space width (JSW) mapping quantifies structural narrowing in knee osteoarthritis (OA), but the effect of surface mesh resolution on measurement reliability and precision remains unclear. This study assessed how mesh resolution influences 3D JSW measurements, whether effects differ by metric and OA severity, and which resolutions best preserve OA group-level discrimination and subject-level agreement.

methodsKnees with medial knee OA (KL grades 0-4) were analyzed. For each knee, 3D JSW maps were generated on a 0.1 mm reference mesh and seven coarser meshes (0.3, 0.5, 0.7, 1.0, 2.0, 3.0, and 4.0 mm). Minimum and median JSW (minJSW, medJSW) were calculated within a standardized medial region and compared to the reference. Systematic error, agreement (Bland-Altman), and subject-level interchangeability were evaluated across resolutions for the overall cohort and early versus advanced OA subgroups. Absolute-error thresholds were set at 0.2 mm (sensitivity) and 0.5 mm (clinical significance). Mesh-by-severity interactions were analyzed using mixed-effects models, and group-level discrimination was assessed at each resolution.

resultsSeventy-two knees from 54 patients (mean age 65 ± 10.4 years; 48.6% advanced OA) were included. The 0.1 mm reference mesh had excellent geometric fidelity (RMS deviation 0.0011 ± 0.0010 mm) and high vertex density (median 118,300 vertices), which dropped to 14,848 at 0.3 mm, 5033 at 0.5 mm, and 72 at 4.0 mm. From 0.1 to 4.0 mm, medJSW increased from 4.07 ± 0.77 mm to 5.14 ± 0.96 mm and minJSW from 1.98 ± 0.81 to 2.34 ± 0.81 mm, converging at 0.3 mm. Subject-level agreement for medJSW declined beyond 0.3 mm: at 0.5 mm and coarser meshes, ≥ 44.4% of knees exceeded the 0.2 mm error threshold; at 0.7 mm and coarser meshes, ≥ 12.5% exceeded the 0.5 mm threshold. In contrast, minJSW remained within the 0.5 mm clinical threshold through 1.0 mm. Early OA showed a larger absolute medJSW inflation (p = 0.023), whereas advanced OA had greater relative minJSW error (p < 0.05). Group-level OA discrimination remained preserved at all resolutions (p < 0.001), while subject-level agreement deteriorated at coarser meshes.

conclusionsMesh resolution significantly influences 3D JSW measurements, with effects modulated by selected metric, disease stage, and level of analysis. Using a rigorous 0.1-mm reference mesh, specific thresholds were established: a 0.3-mm mesh provided the most consistent performance across all conditions; a 0.5-mm mesh may be acceptable for medJSW under a less stringent error tolerance; and up to 1.0 mm sufficed only for identifying the narrowest point (minJSW). At coarser resolutions, baseline JSW overestimation can be large enough to alter perceived structural severity of medial compartment narrowing, carrying risks for clinical misinterpretation and treatment decisions.