Combined clinical and imaging information as an early stroke outcome measure

Abstract

Background and purpose: Imaging information has been proposed as a potential surrogate outcome in stroke clinical trials. The purpose of this study was to determine whether an early outcome measure combining clinical and imaging information is better than either alone in predicting 3-month outcome in acute ischemic stroke patients.

Methods: Clinical information (National Institutes of Health Stroke Scale) and imaging information (CT infarct volume), measured at 1 week from 201 patients from the Randomized Trial of Tirilazad Mesylate in Acute Stroke (RANTTAS) study, were used in a multivariable logistic regression analysis to predict excellent and devastating 3-month outcome. The combined models were compared with the infarct volume models and the clinical models. Discrimination, calibration, and change in global model chi-square were assessed.

Results: The combined models and models using clinical information alone had areas under the receiver operating characteristic curves that did not differ significantly (probability value = 0.092 to 0.4), ranging from 0.83 to 0.95. The imaging alone models performed less well (P<0.005) and had areas under the receiver operating characteristic curves that ranged from 0.70 to 0.80.

Conclusions: The National Institutes of Health Stroke Scale at 1 week is highly predictive of 3-month outcome in ischemic stroke patients. The addition of 1-week infarct volume does not improve the accuracy of the predictive model.

Figure 1

Area under the receiver operating characteristic (ROC) curves for each of the models. Differences in model chi-squares were tested. For all 6 outcome measures, the combined model was not significantly better than the clinical alone model. For all 6 outcome measures, the combined model was significantly better (P<0.005) than the imaging alone model. NIH indicates National Institutes of Health stroke scale score; BI, Barthel Index score; GOS, Glasgow Outcome Scale score.

Figure 2

Calibration curves. Top, Calibration curve for excellent outcome as determined by Barthel Index model. The bias corrected line is almost superimposed on the ideal line, suggesting excellent calibration. Data density is demonstrated by the small hatch lines at the top of the graph. Bottom, Calibration curve for devastating outcome as determined by National Institutes of Health stroke scale model. The bias corrected line deviates from the ideal line, suggesting not as good calibration. Data density is demonstrated by the small hatch lines at the top of the graph.

Figure 3

Partial predictive power. Top, Role of NIHSS score measured at 1 week in predicting probability of outcome for the combined excellent outcome model as determined by Barthel Index. The narrow band of predicted probabilities suggests that the NIHSS is dominating the prediction. Bottom, Role of infarct volume measured at 1 week in predicting probability of outcome for the combined excellent outcome model as determined by the Barthel Index. The broad range of predicted probabilities suggests that the other variable (NIHSS) is controlling the prediction of outcome. NIHSS indicates National Institutes of Health Stroke Scale; BI, Barthel Index.