Development of a Reliable Grading System for the Radiographic Analysis of the Subtalar Joint

Abstract

Category: Hindfoot Introduction/Purpose: Subtalar arthritis is frequently diagnosed in the clinical setting by correlating patients’ presenting signs and symptoms with weightbearing radiographs. Due to the naturally undulating surface of the subtalar joint and frequent concomitant hindfoot deformity, radiographic assessment of the severity of subtalar arthrosis can be difficult. The Kellgren-Lawrence (KL) scale for arthritis was not designed for such a unique joint and has not been validated for generalization at the subtalar joint. Previous studies have shown poor interobserver and intraobserver reliability of this grading system in patients with post-traumatic subtalar arthritis and following total ankle replacement. The authors hypothesized this scale would not be reliable for classifying the severity of subtalar arthrosis and sought to establish a novel scoring system with high interobserver reliability and clinical relevance. Methods: A collection of 40 lateral weightbearing ankle radiographs were reviewed by nine independent reviewers, including foot and ankle orthopaedic surgeons and certified musculoskeletal (MSK) radiologists for the severity of subtalar arthrosis using the KL scale. These radiographs were selected from patients with a variety of complaints of pain in the ankle and foot and not necessarily for subtalar pathology. Patients were excluded if they did not have advanced imaging, had a coalition, or prior hindfoot fusion. Interobserver reliability was assessed for KL scores using a marginal kappa statistic. An MSK radiologist graded available advanced imaging on all 40 radiographs and the advanced imaging scores were correlated to the radiographic scores. A novel scoring system, Grades 0 to 2 was created. In a separate sitting, three foot and ankle orthopedic surgeons and one MSK radiologist evaluated the radiographs using the new scale. Interobserver reliability of this new scale was then calculated. Results: There was overall fair reliability amongst reviewers with the traditional KL score, kappa = 0.26. There was moderate agreement amongst reviewers with patients deemed to have zero subtalar arthrosis (k = 0.50). There was fair interobserver reliability with Grade 4 scores (k= 0.28). Grouping Grades 3 and 4 together provided a slight increase in scoring reliability (k= 0.46). There remained only fair agreement when grades 1-2 (k = 0.31) and grades 1-3 (k =0.37) were combined. Radiographic scores did have a moderate correlation with advanced imaging (r=0.56). The nine reviewers only recorded the same score as the advanced imaging score 37.8% of the time, and were within one grade 82.0% of the time. The simpler proposed grading system had improved interobserver reliability (kappa =0.68). Conclusion: The KL scale is unreliable for assessment of the severity of subtalar arthrosis on lateral weightbearing radiographs and carries no clinical correlation. In our study, there was a correlation between radiographic severity of disease and advanced imaging. Our proposed novel scoring system has a substantially improved reliability when compared to the KL scale, is reproducible, and carries potential clinical implications and relevance. This system can be used for clinical reasons and also reliably used for tracking progression of clinically significant radiographic disease of the subtalar joint.