Institutional members access full text with Ovid®

Share this article on:

Three-dimensional Magnetic Resonance Imaging of Physeal Injury: Reliability and Clinical Utility

Lurie, Brett MBBS*; Koff, Matthew F. PhD*; Shah, Parina MS*; Feldmann, Eric James MD*; Amacker, Nadja MD*; Downey-Zayas, Timothy BA; Green, Daniel MD; Potter, Hollis G. MD*

Journal of Pediatric Orthopaedics: April/May 2014 - Volume 34 - Issue 3 - p 239–245
doi: 10.1097/BPO.0000000000000104

Background: Injuries to the physis are common in children with a subset resulting in an osseous bar and potential growth disturbance. Magnetic resonance imaging allows for detailed assessment of the physis with the ability to generate 3-dimensional physeal models from volumetric data. The purpose of this study was to assess the interrater reliability of physeal bar area measurements generated using a validated semiautomated segmentation technique and to highlight the clinical utility of quantitative 3-dimensional (3D) physeal mapping in pediatric orthopaedic practice.

Methods: The Radiology Information System/Picture Archiving Communication System (PACS) at our institution was searched to find consecutive patients who were imaged for the purpose of assessing a physeal bar or growth disturbance between December 2006 and October 2011. Physeal segmentation was retrospectively performed by 2 independent operators using semiautomated software to generate physeal maps and bar area measurements from 3-dimensional spoiled gradient recalled echo sequences. Inter-reliability was statistically analyzed. Subsequent surgical management for each patient was recorded from the patient notes and surgical records.

Results: We analyzed 24 patients (12M/12F) with a mean age of 11.4 years (range, 5-year to 15-year olds) and 25 physeal bars. Of the physeal bars: 9 (36%) were located in the distal tibia; 8 (32%) in the proximal tibia; 5 (20%) in the distal femur; 1 (4%) in the proximal femur; 1 (4%) in the proximal humerus; and 1 (4%) in the distal radius. The independent operator measurements of physeal bar area were highly correlated with a Pearson correlation coefficient (r) of 0.96 and an intraclass correlation coefficient for average measures of 0.99 (95% confidence interval, 0.97-0.99). Four patients underwent resection of the identified physeal bars, 9 patients were treated with epiphysiodesis, and 1 patient underwent bilateral tibial osteotomies.

Conclusions: Semiautomated segmentation of the physis is a reproducible technique for generating physeal maps and accurately measuring physeal bars, providing quantitative and anatomic information that may inform surgical management and prognosis in patients with physeal injury.

Level of Evidence: Level IV.

Departments of *Radiology and Imaging—MRI

Orthopedics, Hospital for Special Surgery, New York, NY

None of the authors received financial support for this study.

H.S.S. has an institutional research agreement in place with General Electric Healthcare. The other authors have no conflict of interest to declare.

Reprints: Hollis G. Potter, MD, Department of Radiology and Imaging—MRI, Hospital for Special Surgery, 535 East 70th Street, West Basement—MRI, New York, NY 10021. E-mail:

© 2014 by Lippincott Williams & Wilkins