P83. Preoperative Hounsfield units at the planned upper instrumented vertebrae (UIV) may predict proximal junctional kyphosis (PJK) in adult spinal deformity

• Published in: The spine journal 2020-09, Vol.20 (9), p.S186-S187
• Main Authors: Yao, Yu-Cheng, Elysee, Jonathan, McCarthy, Michael H., Louie, Philip, Lafage, Renaud, Weissmann, Karen, Alshabab, Basel Sheikh, Lafage, Virginie, Schwab, Frank J., Kim, Han Jo
• Format: Article
• Language: English
• ISSN: 1529-9430; 1878-1632
• DOI: 10.1016/j.spinee.2020.05.481

Proximal Junctional Kyphosis (PJK) following surgery of ASD is common and poor bone quality is noted to be one of the risk factors. Bone Densitometry (DEXA) of the spine does not accurately assess bone quality, especially in those who have severe degenerative conditions, deformity or prior fusions. HUs on standard CT can be used as an alternative in evaluating localized bone quality and correlation between local HU and the incidence of PJK are not clear. The study aimed to investigate the association between the local bone marrow density represented in Hounsfield Units (HU) of vertebral body of the planned UIV and UIV+1 on preoperative CT scan and the correlation with PJK in patients after surgery for ASD. Retrospective case control study. Sixty-three patients were included from a single institution database from 2013-18. Inclusion criteria are patients: age≥18 years, diagnosis of spinal deformity, preoperative CT scan present, and a minimum of 1-year follow-up. Mean HU of UIV and UIV+1, incidence of PJK, and change of PJK angle. Patient demographic and radiographic parameters were recorded. Local vertebral HU of UIV and UIV+1 was measured on preoperative CT scan. Patients were separated into three groups, no PJK, non-bony PJK, and bony PJK groups, which were distinguished by cause of PJK. The risk factors between the three groups and the correlation of mean HU and progression of PJK angle were also analyzed. The incidence of PJK was 36.5%(n=23) and there were two revisions for PJK (3.2%). There was no difference between groups in the preop, postop, and corrections of the spinopelvic and sagittal radiographic parameters. The ANOVA showed that the mean HU was significantly different within the three groups (p=0.025). The posthoc analysis for the three groups showed the mean HU in the bony PJK group (HU:109.0) was significantly lower compared to the no PJK group (HU:168.7, p=0.038). The mean HU in the non-bony PJK group (HU:141.7) showed no difference compared to the other two groups. Multivariate analysis demonstrated a significant negative correlation between mean HU and PJK angle (r=0.475, p