Malrotated lateral radiographs do not allow for proper assessment of patellar height using the Caton‐Deschamps Index

TL;DR

Malrotated knee X-rays lead to inaccurate patellar height measurements using the Caton-Deschamps Index, affecting clinical decisions.

Contribution

This study identifies a clinically relevant threshold for malrotation in lateral knee radiographs affecting patellar height assessment.

Findings

Malrotated radiographs significantly alter Caton-Deschamps Index values compared to properly aligned images.

34.9% of patients showed a clinically important difference in patellar height measurements due to malrotation.

An APD cut-off of 6.3 mm corresponds to a 0.1 change in the Caton-Deschamps Index.

Abstract

To compare patellar height measurement (PHM) using the Caton‐Deschamps Index (CDI) on malrotated versus properly positioned lateral knee radiographs within the same patients. It was hypothesised that malrotation would significantly alter CDI values. A secondary aim was to establish an anterior–posterior distance (APD) cut‐off for malrotation corresponding to a minimally clinically important difference (MCID) of 0.1 in ΔCDI.

This retrospective analysis included patients with lateral knee radiographs between January 2020 and March 2023 at a single academic institution. Patients were included if at least two radiographs were available: one with malrotation (APD ≥ 1 mm) and one properly aligned (APD < 1 mm). Radiographs with tilt, prior osseous surgery, or fractures were excluded. Patellar height was measured using the CDI. Inter and intrarater reliability were assessed via intraclass correlation coefficients (ICC). Paired t‐tests compared CDI values. Cases with a ΔCDI exceeding the MCID of ≥0.1 were identified. The APD cut‐off corresponding to a ΔCDI of 0.1 was determined by regression of ΔCDI on APD.

A total of 126 lateral knee radiographs from 63 patients (57% female) were analysed. Inter and intrarater reliability was excellent (ICC > 0.8 for all comparisons). The mean CDI was 0.96 ± 0.06 on properly aligned and 1.03 ± 0.08 on malrotated radiographs, with a significant mean difference of 0.07 ± 0.05 (p < 0.001). In 34.9% of patients, CDI differed by ≥0.1 between imaging conditions. Linear regression of ΔCDI on APD showed a slope of 0.016 per mm (R² = 0.57). The APD cut‐off corresponding to a ΔCDI of 0.1 was 6.3 mm (95% CI, 5.7–6.9 mm).

Malrotation significantly alters PHM using the CDI on lateral knee radiographs. A clinically relevant difference of ≥0.1 in ΔCDI occurred in over one‐third of patients. Malrotated radiographs should be used with caution, and repeat imaging should be considered when CDI values approach clinical decision thresholds.

Level III, diagnostic study.

Figures