Assessment of Parkinson's disease severity based on T2-weighted magnetic resonance imaging radiomic modeling
Objective:To explore the value of assessing the severity of patients with Parkinson's disease(PD)based on the T2WI radiology model.Materials and Methods:A total of 201 patients with a clinical diagnosis of PD were retrospectively collected,according to Hoehn-Yahr(H-Y)stage.They were divided into the early group(n=113)and middle-late group(n=88),into a training set of 140 cases and a test set of 61 cases using a 7∶3 ratio at the same time.We established by extracting and screening radiomics features from substantia nigra(SN),red nucleus(RN),caudate nucleus(CN),putamen(PUT),globus pallidus(GP),and used a logistic regression(LR)classifier to build the corresponding models separately.Clinical and imaging data were sequentially incorporated into univariate and multivariate logistic regression analysis to identify independent risk factors about PD severity.Receiver operating characteristic(ROC)curves were used to assess the efficacy of each model,and calibration curves were used to assess calibration accuracy.Results:19 optimal features on T2WI were selected from SN,and the AUC(0.817,0.733)in the training and test set were higher than those of the RN model(0.758,0.704),the CN model(0.712,0.643),the PUT model(0.713,0.708),and the GP model(0.705,0.708).The white matter hyperintensities burden and the duration of PD were independent risk factors for diagnosing the severity of PD patients,and the AUC of the combined model with the SN model was 0.865(training set)and 0.836(test set),but there is no significant difference in diagnostic performance between the two models.The calibration curve indicates good consistency between the diagnostic results of the six models and the actual outcomes.Conclusions:SN signatures on T2WI achieved better performance in assessing the severity of PD patients.The assessed efficacy of the combined model established by the combined WMH burden and duration of PD was further improved and provide imaging guidance for timely clinical intervention and treatment.
Parkinson's diseasemagnetic resonance imagingT2-weighted magnetic resonance imagingradiomicswhite matter hyperintensities
国家科技期刊平台
NSTL
万方数据
