Sensitivity of the Montreal Cognitive Assessment in screening for cognitive impairment in patients with newly diagnosed high-grade glioma.

Abstract

Cognitive impairment is frequent in patients with high-grade glioma and requires cognitive follow-up. Cognitive screening tools such as the Montreal Cognitive Assessment (MoCA) have been used to assess cognition in these patients. Here we assessed the sensitivity of the MoCA in screening for cognitive impairment in a cohort of 156 patients with newly-diagnosed high-grade glioma, after surgery and before radiochemotherapy. We assessed cognitive performance with the MoCA and a neuropsychological battery. Cognitive scores were analyzed in terms of a previously validated framework designed to control false positives and data for 1003 control participants from the GRECOGVASC study. After comparison of performance on the tests, we used stepwise logistic regression to produce a cognitive summary score from the neuropsychological battery. Then we analyzed sensitivity and specificity of the MoCA with receiver operator characteristic (ROC) curve analysis. Both raw and adjusted MoCA scores showed only moderate sensitivity. The area under the ROC curve was 0.759 (95% CI 0.703-0.815) for the raw score and 0.788 (95% CI 0.734-0.842) for the adjusted score. Optimal discrimination was obtained with a raw score ≤ 25 (sensitivity: 0.526; specificity: 0.832; positive predictive value: 0.2; negative predictive value: 0.96) and an adjusted score - 0.603 (sensitivity: 0.716; specificity: 0.768; positive predictive value: 0.24; negative predictive value: 0.96). The moderate sensitivity of MoCA indicates that it is not a suitable screening tool for detecting cognitive impairment in patients with newly-diagnosed high-grade glioma.