Serial comparison of a predefined ultrasonographic cortical bridging appearance and conventional radiography in tibial fractures after intramedullary nailing: A prospective comparative cohort study.

Abstract

Conventional radiography is the standard method for monitoring tibial fracture healing, but radiographic mineralization may be inconspicuous when earlier sonographic changes are detectable. This study compared serial ultrasonographic and conventional radiographic detection of a predefined cortical bridging finding using an identical four-cortex threshold and assessed interobserver reliability of the Radiographic Union Score for Tibial fractures (RUST). This prospective comparative cohort study included 269 patients with tibial fractures treated with intramedullary nailing. Ultrasonography was performed at approximately 4, 6, 8, 12, 16, and 24 weeks; radiographic comparisons were available from 6 weeks onward. The medial, lateral, anterior, and posterior cortical surfaces were assessed. An examination was classified as positive if ultrasound showed continuous hyperechoic cortical bridging in at least three of four cortices, or if radiographs showed bridging callus in at least three of four cortices. Paired findings at the same follow-up visits were compared using the exact McNemar test. RUST was independently scored by two observers and assessed using ICC (2,1), weighted kappa, and Bland-Altman analysis. Ultrasound-positive examinations increased from 36 (13.4%) at 4 weeks to 90 (33.5%) at 6 weeks, 125 (46.5%) at 8 weeks, 197 (73.2%) at 12 weeks, 200 (74.3%) at 16 weeks, and 224 (83.3%) at 24 weeks. Radiographically positive examinations were 17 (6.3%) at 6 weeks, 125 (46.5%) at 8 weeks, 198 (73.6%) at 12 weeks, 198 (73.6%) at 16 weeks, and 216 (80.3%) at 24 weeks. At 6 weeks, 74 patients were ultrasound-positive/radiograph-negative and one was ultrasound-negative/radiograph-positive (exact McNemar P < 0.001). At 8 weeks, the discordant counts were 8 and 8 (P = 1.000); at 12 weeks, 5 and 6 (P = 1.000); at 16 weeks, 7 and 5 (P = 0.774); and at 24 weeks, 10 and 2 (P = 0.039). Study-defined union at 24 weeks was documented in 234 of 269 patients (87.0%). RUST scores showed excellent interobserver reliability (ICC[2,1] = 0.979; 95% CI, 0.969-0.986). At 6 weeks, the predefined sonographic cortical bridging appearance was observed more frequently than radiographic bridging, with the two modalities converging during later follow-up. Because the sonographic appearance was not histologically validated or correlated with CT at the same cortical surface, it should be regarded as a descriptive imaging finding rather than proof of osseous union or mechanical stability. Ultrasonography may play a complementary role in early assessment, while conventional radiography remains important for structural assessment and confirmation of established union.