Development and Validation of a Multivariable Model to Predict Surgical Outcome in Patients with Chiari malformations.

Abstract

Surgical outcome predictive models for Chiari malformations (CM) which are applicable to all age groups and simple enough to use on outpatient basis are lacking. The aim of this study was to develop and validate a preoperative index for predicting long-term outcomes in Chiari 1 (CM1) and Chiari 0 (CM0) malformations. It was a single-institution, ambidirectional, cohort study from 2014 to 2019, having patients between 5 to 70 years. Outcome was assessed using Chicago Chiari outcome score (CCOS) over 2 years follow-up. Preoperative clinical and radiological factors were analyzed using Chi-square test and Mann Whitney U test, in relation to CCOS and those attaining P value ≤0.05, were used to develop model - Chiari Outcome Predictive Index (COPI). COPI was internally validated using 10-fold cross-validation and c-statistic for discrimination. A total of 88 patients (66 in development and 22 in validation cohort) were included in the study. Outcome was negatively associated with presence of motor, sensory or cranial nerve symptoms, poor functional status, basilar invagination, and tonsillar descent. It was positively associated with shorter duration of presenting symptom (<9 months) and syrinx diameter <6 mm. COPI predicted CCOS with 91.1% accuracy (10-fold cross-validation). It had excellent discrimination for improved outcome (c = 0.968 in development and 0.976 in validation cohort), at threshold index of -1. COPI is simple tool that can be administered in outpatient setting. It can facilitate evidence-based preoperative counseling of patients, to help them develop reasonable expectations regarding surgical outcomes.