200 - A 3D Knowledge-based Grading System for Heterotopic Ossification Follo...

#200 A 3D Knowledge-based Grading System for Heterotopic Ossification Following Cervical Disc Arthroplasty


Poster Presented by: M.F. Gornet


M.F. Gornet (1)
J.A. Hipp (2)
A.G. Copay (3)
F.W. Schranck (3)

(1) Orthopedic Center of St. Louis, St Louis, MO, United States
(2) Medical Metrics, Inc., Houston, TX, United States
(3) SPIRITT Research, St Louis, MO, United States


Introduction: Heterotopic ossification (HO) at levels implanted with a cervical disc arthroplasty (CDA) has been reported in many publications. The clinical consequences of HO are poorly understood, but it is unlikely that HO contributes to the motion-preserving goal of CDA. Our understanding of the clinical consequences of HO may be compromised by the difficulty of documenting the true extent of HO from radiographs in clinical studies. The goal of the research described herein was to develop a more reliable grading system for HO based on a detailed understanding of the typical patterns of HO and the similarities and differences from osteophytes that form in the absence of surgical interventions.

Methods: Volume compositing reconstructions were created from thin-slice CT exams used in a prior study of trauma patients and from CT exams of CDA patients where HO had been detected from radiographs. The posterior elements were virtually removed to allow for visualization of the posterior aspects of the disc spaces. Animations were created showing the osseous anatomy being rotated 360 deg about the cranial-caudal axis. These animations were critically assessed to identify common patterns seen in the osteophytes observed in trauma CT exams and HO in CDA patients.

Results: Osteophytes observed in the unoperated spine can be classified into five patterns starting with early and modest endplate rim changes, progressing through beak or lip-like, overhanging, or interlocking patterns, through to solid bone bridging between vertebrae. Isolated islands of ossification were also observed that likely fuse as the islands and osteophytes increase in size through the interlocking pattern. These patterns can be symmetric across the mid-sagittal plane, but are often asymmetric. Similar patterns form along the anterior borders of the endplates and along the uncovertebral joints. The osteophytes that form along the posterior borders of the endplates are less variable. Heterotopic ossification that forms at CDA levels can be classified using the same five patterns. Comparison of the radiographs to the 3D reconstructions from thin-slice CT exams can be used to train observers to more reliably document the pattern and extent of HO using only radiographs.

Conclusion: The 3D reconstructions showing osteophytes that form in the absence of operative intervention informed the interpretation of heterotopic ossification observed in 3D reconstructions of CDA patients, and that knowledge informed the interpretation of radiographs of the CDA patients. This educational process will be used to train the observers that will determine the presence and extent of HO following CDA in clinical studies. Although there are multiple patterns in the way osteophytes and HO form, this granularity may not be needed in a large clinical study where reliable documentation of the presence and extent of HO may be sufficient. It is expected that observers educated in osteophyte/HO pattern recognition can more reliably document the presence and extent of HO following CDA. This may improve our ability to determine the true clinical consequences of HO as well as to identify interventions to minimize HO.