Cerebral collateral validation
Description
Data derived from acute ischemic stroke patients undergoing endovascular thrombectomy and a trained neurologist performed collateral grading based on the ASITN scale on the DSA images acquired. Perfusion MRIs were evaluated with the Rapid software (https://www.rapidai.com/) version 5.2.2, (c) iSchemaView, Inc. 2011-2019. Data were tabulated and statistical analyses performed.
Files
Steps to reproduce
Primary outcome and analysis: ASITN/SIR collateral grade (ordinal, 0–4 scale) can be modeled as the primary outcome using a proportional-odds (ordinal logistic) regression, since this respects the ordered, non-continuous nature of the scale without assuming equal spacing between grades. HIR will be the primary predictor. Effect estimates will be reported as odds ratios with 95% Wald confidence intervals. Secondary analyses: • Continuous approximation: Linear regression of ASITN on HIR, unadjusted and adjusted for age and sex, reported for comparability with prior literature that treats ASITN as continuous. Residual diagnostics (normality, homoscedasticity) will be reported. • Association strength: Pearson and Spearman correlation coefficients between HIR and ASITN, to characterize linear and monotonic association independent of a regression model. • Dichotomized outcome: Logistic regression of "poor collaterals" (binary, defined as ASITN ≤ [cutoff]) on HIR, with odds ratios, 95% CIs, and a likelihood-ratio test against the null model. This mirrors how collateral status is often dichotomized clinically (e.g., for reperfusion therapy decisions). Covariates: Age and sex can be included as adjustment variables based on prior literature.
Institutions
- The University of Texas SystemTexas, Austin