A Modified Keshishyan Index Demonstrates Three-dimensional Assessment of Asymmetry in Pediatric Pelvic Fractures.

Abstract

Background: Orthopedic surgeons use a combination of plain radiographs and computed tomography (CT) to diagnose and manage pelvic ring injuries. Historically, pelvic ring asymmetry has been calculated using the Keshishyan index (KI) measured in 2 dimensions. Three-dimensional (3D) CT modeling may assess deformity more comprehensively, aiding in treatment decisions. This study aimed to evaluate the accuracy and reliability of a modified KI (MKI) derived from 3D models of pediatric patients with pelvic fractures.

Materials and methods: Two separate raters identified landmarks 3 separate times on 29 generated models. Traditional KI (TKI) was calculated using the contralateral distance between the most anterior portion of the sacroiliac joint (SAAS) and the deepest point of the acetabulum. The MKI used the contralateral distance between the SAAS and pubic tubercle. KI was calculated on correlating radiographs (RxKI). Bland-Altman analyses and intra-/interclass correlation coefficients (ICC) were used to evaluate agreement between measurement methods and reliability.

Results: RxKI and TKI measurements exhibited similar agreement with a mean bias (MB) of -0.0205. RxKI and MKI had a MB of 0.0012, and TKI and MKI had a MB of 0.0210. TKI reliability ranged from good to excellent (Intra = 0.923, 95% confidence interval [CI], 0.892-0.947; Inter = 0.897, 95% CI, 0.852-0.928), and MKI reliability was excellent (Intra = 0.938, 95% CI, 0.911-0.959; Inter = 0.912, 95% CI, 0.876-0.937).

Conclusion: All methods exhibited excellent reliability, with MKI having the highest ICCs. MKI provided similar, reproducible values compared with RxKI, suggesting that clinicians can compare pelvic asymmetry values between CT and radiograph images.