Observer Reliability of CT Tibiofibular Line for Syndesmosis Reduction Assessment

CATEGORY: Ankle INTRODUCTION/PURPOSE: When necessary, surgical treatment of ankle syndesmosis disruption typically yields excellent functional outcomes. That is, when the syndesmosis is correctly reduced. However, accurate reduction of the ankle syndesmosis is difficult to achieve reproducibly. A clinically important step would be to translate the precision of a computed tomography (CT)-based technique for syndesmosis reduction accuracy assessment into a validated clinical approach. Reb et al. previously demonstrated the clinical feasibility of translating the CT tibiofibular line (CT-TFL) to a clinical technique. However, this study raised questions regarding several potentially significant confounding effects. These included the effects of fibula morphology. The present study evaluated the validity and reliability of the CT-TFL technique, particularly with regard to variations in fibula morphology. METHODS: An IRB approved, retrospective cross sectional study of consecutive foot and ankle CT scans obtained for non-ankle complaints and without radiographic evidence of ankle injury or prior surgery. 3 trained observers repeatedly performed measurements on the 52 patients meeting criteria. Anterolateral fibula shape was categorized in the axial plane 10 mm superior to the plafond. The CT-TFL was drawn contacting the longest section of straight anterolateral cortical surface of fibula with this length recorded. The CT-TFL distance measurement was made connecting the CT-TFL to the closest point of tibial surface (Figure 1). Means, standard deviations, and 95% confidence intervals were calculated for continuous variables and frequencies were calculated for categorical variables. Intraobserver and interobserver consistency were assessed using intraclass correlation for continuous variables and Fleiss’ kappa for categorical variables. The optimal range of straight anterolateral fibula cortical length was determined by plotting CT-TFL contact length versus observer consistency of CT-TFL distances. RESULTS: Mean fibula cortical contact length ranged from 3.74 mm (95% CI 3.24 to 4.24 mm) to 11.45 mm (95% CI 11.21 to 11.69mm). Mean TFL distance ranged from -2.83 mm (95% CI -3.65 to -2.00) to 5.03 mm (95% CI 4.48 to 5.57 mm). Intraobserver and interobserver consistencies were excellent (minimum ICC, 0.87) for these measurements. For fibula shape categorization, intraobserver consistency ranged from substantial to almost perfect (Fleiss’ Kappa range, 0.73 to 0.97). However, interobserver consistency was moderate (Fleiss’ Kappa, 0.55). 6 mm to 10 mm of fibula contact length corresponded to a minimum of excellent observer consistency for CT-TFL distance (ICC 0.80 to 0.98). CONCLUSION: The study refines and expands upon the initial description of the CT-TFL technique. Among uninjured syndesmoses, a broader range of CT-TFL values were observed than previously reported, including negative values of which some were relatively large. Instead of fibula shape, anterolateral fibula cortical contact length was more useful in discerning which subjects were best suited for this technique. Excellent observer consistency occurred when 6 mm to 10 mm of fibula contact length was present. These results support the validity of the TFL technique and further characterize its role in the clinical setting.