RESUMO
Objective:To explore the impact of different segmentation methods on differential diagnostic efficiency of 18F-FDG PET/MR radiomics to distinguish Parkinson′s disease (PD) from multiple system atrophy (MSA). Methods:From December 2017 to June 2019, 90 patients (60 with PD and 30 with MSA; 37 males, 53 females; age (55.8±9.5) years) who underwent 18F-FDG PET/MR in Union Hospital, Tongji Medical College, Huazhong University of Science and Technology were retrospectively collected. Patients were randomized to training set and validation set in a ratio of 7∶3. The bilateral putamina and caudate nuclei, as the ROIs, were segmented by automatic segmentation of brain regions based on anatomical automatic labeling (AAL) template and manual segmentation using ITK-SNAP software. A total of 1 172 radiomics features were extracted from T 1 weighted imaging (WI) and 18F-FDG PET images. The minimal redundancy maximal relevance (mRMR) and least absolute shrinkage and selection operator (LASSO) algorithm were used for features selection and radiomics signatures (Radscore) construction, with 10-fold cross-validation for preventing overfitting. The diagnostic performance of the models was assessed by ROC curve analysis, and the differences between models were calculated by Delong test. Results:There were 63 cases in training set (42 PD, 21 MSA) and 27 cases in validation set (18 PD, 9 MSA). The Radscore values were significantly different between the PD group and the MSA group in all training set and validation set of radiomics models ( 18F-FDG_Radscore and T 1WI_Radscore) based on automatic or manual segmentation methods ( z values: from -5.15 to -2.83, all P<0.05). ROC curve analysis showed that AUCs of 18F-FDG_Radscore and T 1WI_Radscore based on automatic segmentation in training and validation sets were 0.848, 0.840 and 0.892, 0.877, while AUCs were 0.900, 0.883 and 0.895, 0.870 based on manual segmentation. There were no significant differences in training and validation sets between Radiomics models based on different segmentation methods ( z values: 0.04-0.77, all P>0.05). Conclusions:The 18F-FDG PET/MR radiomics models based on different segmentation methods achieve promising diagnostic efficacy for distinguishing PD from MSA. The radiomics analysis based on automatic segmentation shows greater potential and practical value in the differential diagnosis of PD and MSA in view of the advantages including time-saving, labor-saving, and high repeatability.
RESUMO
Objective:To explore the feasibility of statistical parametric mapping (SPM) aided semi-quantitative analysis in 11C-Pittsburgh compound B (PIB) β-amyloid (Aβ) PET imaging acquired by hybrid PET/MR, and evaluate its possibility in assisting the diagnosis or differential diagnosis for cognitive impairment. Methods:From January 2018 to September 2019, 13 Alzheimer′s disease (AD) patients (4 males, 9 females; age (59.2±5.8) years) and 10 vascular cognitive disorders (VCD) patients (9 males, 1 female; age (59.5±11.5) years) who underwent 11C-PIB PET/MR in PET center of Union Hospital, Tongji Medical College, Huazhong University of Science and Technology were retrospectively analyzed. The standardized uptake value ratio (SUVR) of eight key brain regions (cerebral white matter, striatum, thalamus, posterior cingulate gyrus, frontal cortex, posterior parietal cortex, lateral temporal cortex and occipital cortex) to cerebellum cortex were obtained by manual delineation and SPM-aided semi-automatic segmentation with the help of synchronous three-dimensional T 1 weighted imaging (3D T 1WI). Pearson correlation analysis was carried out on the SUVR obtained by the two methods. Independent-sample t test and paired t test were used to analyze the data. Results:There was no significant difference between AD group and VCD group in age and Mini-Mental State Examination (MMSE) score (19.7±4.7 vs 21.7±3.8; t values: 0.095 and 1.098, both P>0.05). Except thalamus( r=0.179, P=0.413), there were good correlations between SUVR obtained by segmentation and delineation in the other 7 key regions ( r values: 0.678-0.893, all P<0.05). The SUVR of 8 key regions obtained by the two methods in AD group was significantly higher than that in VCD group (1.519-2.055 vs 1.105-1.618; t values: 2.799-11.582, all P<0.01). The SUVR of striatum (1.942±0.205), posterior cingulate gyrus (1.915±0.249), frontal lobe (1.983±0.264), parietal lobe (2.008±0.296) and temporal cortex (1.931±0.254) in AD group was significantly higher than that of cerebral white matter (1.746±0.192; t values: 3.793-6.992, all P<0.01). But in VCD group, there was no region with the SUVR higher than that of cerebral white matter. Conclusions:Hybrid PET/MR can acquire the PET and MRI images synchronously, which can realize the accurate brain segmentation and obtain the semi-quantitative data of key brain regions aided by SPM. The method can analyze the characteristics and differences of amyloid imaging in AD and VCD, which is expected to provide an accurate imaging analysis method for the diagnosis and differential diagnosis of cognitive disorders.