A risk prediction model of post-stroke cognitive impairment based on magnetic resonance spectroscopy imaging

Objective: To explore the clinical value of a risk prediction model of post-stroke cognitive impairment (PSCI) based on proton magnetic resonance spectroscopy (1H-MRS).Methods:A retrospective analysis was conducted on 376 stroke patients hospitalized between March 2016 and March 2019. Their relevant clinical baseline data were collected at admission. After the patients' condition was stabilized, 1H-MRS was performed to detect the related indices of the bilateral prefrontal lobe, thalamus, basal ganglia, hippocampus, precuneus, and angular gyrus. Within 12 months of the onset of stroke, cognitive impairment tests were performed monthly. Based on score results, stroke patients were divided into two groups: PSCI and post-stroke non-PSCI (N-PSCI). Thirty-four characteristic parameters of baseline and imaging data were extracted from the PSCI and N-PSCI groups. The least absolute shrinkage and selection operator (LASSO) regression was used for optimal feature selection, and a nomogram prediction model was established. The predictive ability of the model was validated by a calibration plot and the area under the curve (AUC) of the receiver operating characteristic curve.Results: Six risk factors were identified from clinical baseline data and MRS indices based on screening by LASSO dimensionality reduction. The consistency test of the correction curve showed that the prediction probability of the PSCI nomogram had good correlation with actual diagnosis. The AUCs of internal and external validation were 0.8935 and 0.8523, respectively.Discussion: A PSCI risk prediction model based on MRS serves to assist clinicians in estimating the risk of cognitive impairment after stroke.