Abstract
BACKGROUND: Talus injuries are clinically significant due to the talus bone's essential function in foot biomechanics and the complexity of its healing process. Although plain radiographs are more accessible and involve lower radiation exposure than CT scans, they often lack sufficient detail for accurate diagnosis. Given their frequent use in initial assessments, this study aims to evaluate the diagnostic accuracy of plain radiographs in classifying talus injuries within a patient group with confirmed talar fractures. METHODS: Sample size determination aimed to adequately represent rare fracture subtypes while resembling real-world prevalence, indicating a minimum of 140 cases. Radiographs of 159 patients with talus fractures were randomly selected from a database. Two authors interpreted CT scans to establish an accurate gold standard diagnosis. Three diagnosticians assessed the radiographs independently, searching for 10 talus injuries and fracture characteristics. Sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), and their confidence intervals were calculated for each injury type and overall injuries. Individual and pooled diagnosticians' responses were used to compute diagnostic metrics. RESULTS: Diagnosticians achieved only moderate interrater agreement, with a Kappa coefficient of 0.445 (p < 0.001). The overall sensitivity of plain radiographs for detecting talus injuries was 60% (95% CI: 59.10%-61.87%). Sensitivity for identifying fracture characteristics, such as comminution and displacement was lower, at around 53%. Fractures of the head, body, and neck exhibited variable sensitivities, with talus neck fractures demonstrating the highest sensitivity at 76.73% followed by head (52.78%) and body (51.92%). The lateral process, posterior process, and medial tubercle of talus displayed sensitivity rates ranging from 51.52 to 55.56%, while osteochondral fractures showed the lowest sensitivity at 33.33%. Sensitivities for tibiotalar (82.22%) and talonavicular (75%) dislocations were satisfactory compared to subtalar dislocations (57.36%). Hawkins type 2 yielded the lowest sensitivity at 29.76% (95% CI: 25.66%-33.86%). CONCLUSION: While plain radiographs are accessible for detecting certain talus injuries, they fall short in precisely classifying injury locations and identifying details, such as displacement and comminution. CT scans are recommended when there is high clinical suspicion, as they offer greater accuracy for complex and subtle injuries. LEVEL OF EVIDENCE: Level II.