Friday, September 28, 2018 4:05 PM–5:05 PM abstracts: tumor and deformity: 248. Can postoperative cSVA, C2 slope and T1 slope be predicted accurately in cervical deformity surgery?

Abstract

BACKGROUND CONTEXT Cervical malalignment is associated with severe disability. In cervical deformity surgery, failure to correct cSVA and C2 slope (C2S) is associated with poor clinical outcomes. Current surgical planning and intraoperative measurements are limited to cervical lordosis (CL) correction. We aim to develop a predictive model for postop cSVA, C2S and T1S using more than CL change. PURPOSE To propose reliable formulas that may be utilized to plan postoperative correction of C2S, T1S, and cSVA. STUDY DESIGN/SETTING Retrospective review of a prospectively collected cervical deformity database. PATIENT SAMPLE A total of 153 patients with cervical deformity. OUTCOME MEASURES Postoperative sagittal alignment parameters, health related outcome measures, and surgical parameters. METHODS A prospective database of operative CD patients was analyzed. Inclusion criteria were cervical kyphosis>10°, cervical scoliosis>10°, cSVA>4cm or CBVA>25°. The patients were randomly filtered to include 66.7% of the cohort for model development. Predictive models were developed to predict postop T1S, cSVA, and C2S using linear regression. The new predictive equations were validated in the remaining 33.3% of the cohort. RESULTS A total of 153 patients with CD met inclusion criteria. T1S changed significantly (32.4°–35.2°, p=.05) from baseline to 3M follow-up. The mean DJKA change was −6.59. A total of 101 patients were included in model development. To predict postop T1S, CL change and preT1S explained 62.4% of the variability of data (R2=0.624). By including DJKA, R2 improved to 0.724. When predicting postop cSVA, CL change and pre-opcSVA accounted for 57.2% of variability (R2=0.572). With change in DJKA, the R2 improved to 0.661. The model was optimized with the change in T1S (R2=0.777). Preop C2S and CL change lead to poor predictability in postop C2S (R2=0.348). Using change in DJKA, the R2 improved to 0.550. By including DJKA and T1S change, the model was optimized (R2=0.926).The formulas obtained from the predictive model were then simplified by utilizing the mean DJKA of the cohort used to develop the predictive formulas and by utilizing the calculated T1S change obtained from the first equation to yield the following equations: (1) PostOperative T1S=−0.207 + (0.95*Pre-Op T1S) + (1.15*CL Change) (2) PostOperative cSVA =5.803 + (Pre-Op cSVA) + (0.146*CL Change) – (0.028*Pre-Op T1S) (3) PostOperative C2S = 1.223 + (1.38*Pre-Op C2S) – (0.302*CL Change) – (0.053*Pre-OP T1S). These simplified equations were applied to the remaining 52 patients to validate their clinical application. Predicted postop alignments correlated to postop T1S, cSVA, and C2S (R=0.712, R=0.736, and R=0.584 respectively, p CONCLUSIONS Restoring sagittal alignment is critical to obtaining good postoperative outcomes following correction of CD. Current surgical planning and intraoperative measurements are currently limited to cervical lordosis (CL) correction. However, by predicting T1 Slope (T1S) change from baseline to 3 months postop and adding a correction factor for the change in distal junctional kyphosis angle (DJKA), cSVA and C2S can be predicted more accurately. The formulas proposed for the purposes of this study are simple, effective, and can be effectively utilized in a clinical setting to plan successful postoperative correction.