Clinical and imaging data of 11 patients with dysembryoplastic neuroepithelial tumors (DNET) and 15 patients with low-grade glioma (LGG) admitted in Northern Jiangsu People's Hospital were analyzed retrospectively.Routine MRI scan,diffusion weighted imaging (DWI) and enhanced scan were performed.The workstation automatically generated apparent diffusion coefficient (ADC) maps and then to obtain ADC values of the tumor parenchymal area and the contralateral normal reference area.Relative tumor/reference ADC values (rADC) were also calculated.The ADC values of parenchymal regions of tumor and contralateral normal reference areas and the rADC between DNET and LGG were compared.There was significant difference in age distribution between the two groups [(16.6± 13.1) vs.(43.0± 19.2) years,t=3.938,P＜0.01].Six out of 11 DNET cases and none of 15 LGG cases were cuneiform or fan-shaped (P＜0.01);5/11 DNET and 0/15 LGG showed circular high signal in fluid attenuated inversion recovery-T2 weighted imaging (T2FLAIR) sequence (P＜0.01),while there no significant differences in intracapsular segmentation,peritumor edema and mass effect,enhancement,and skull compression between two groups (all P＞0.05).The ADC values of tumor parenchymal regions in both groups were significantly higher than those in contralateral reference regions (both P＜0.01),the rADC of DNET was significantly higher than that of LGG (P＜0.01).It is difficult to identify DNET and LGG by conventional image morphology,however the rADC value of DNET in DWI is significantly higher than that of LGG,and can provide important reference for differential diagnosis between them.

Objective:To evaluate the effect of deep learning based on DWI and fluid attenuated inversion recovery (FLAIR) to construct a prediction model of the onset time in acute stroke.Methods:A total of 324 cases of acute stroke with clear onset time, from January 2017 to May 2020 in Nanjing First Hospital, were retrospectively enrolled and analyzed. The patients were divided into a training set of 226 patients and a test set of 98 patients according to the complete randomization method using a 7∶3 ratio, and the patients were divided into ≤ 4.5 h and >4.5 h according to symptom onset time in each group. The acute infarction areas on DWI and the corresponding high signal area on FLAIR were manually outlined by physician. Using the InceptionV3 model as the basic model for image features extraction, the deep learning prediction model based on single sequence (DWI, FLAIR) and multi sequences (DWI+FLAIR) were established and verified. Then the area under curve (AUC), accuracy of human readings, single sequence model and multi sequence model in predicting the acute stroke onset time from imaging were compared.Results:DWI-FLAIR mismatch was found in 94 cases (94/207) of patients with symptom onset time from imaging ≤ 4.5 h, while in 28 cases (28/117) of patients with symptom onset time from imaging >4.5 h. ROC analysis showed that the AUC of DWI-FLAIR mismatch in predicting acute stroke onset time from imaging was 0.607, and the accuracy was 60.2%. The prediction model of deep learning based on single sequence showed that the AUC of FLAIR was 0.761 and the accuracy was 71.4%; the AUC of DWI was 0.836 and the accuracy was 81.6%. The AUC of predicting stroke onset time based on the multi-sequence (DWI+FLAIR) deep learning model was 0.852, which was significantly better than that of manual identification ( Z = 0.617, P = 0.002), FLAIR sequence deep learning model ( Z = 2.133, P = 0.006) and DWI sequence deep learning model ( Z = 1.846, P = 0.012). Conclusion:The deep learning model based on DWI and FLAIR is superior to human readings in predicting acute stroke onset time from imaging, which could provide guidance for intravenous thrombolytic therapy for acute stroke patients with unknown onset time.

Objective:To evaluate image quality and diagnostic performance of silent MR angiography (MRA) and discuss the feasibility of silent MRA in diagnosing intracranial aneurysms.Methods:Twenty seven patients suspected with cerebrovascular disorders and 30 intracranial aneurysms in Northern Jiangsu People's Hospital, were enrolled prospectively in this study from December 2015 to December 2018. Silent and time of flight (TOF) MRA were performed on the same day prior to CTA examination. The corresponding MRA images were independently and blindly evaluated by two experienced neuroradiologists in the aspects of signal homogeneity, lesion conspicuity, venous signal/artifact and diagnostic confidence (4-point scale). The aneurysms were divided into tiny (≤ 3 mm) and non-tinyaneurysm groups(> 3 mm) according to the measured diameters of aneurysms. The differences in image quality ratings between silent MRA and TOF MRA were analyzed using Wilcoxon signed rank tests. Intra-class correlation coefficients (ICC) were used to test the consistency of measurements between MRAs (silent MRA, TOF MRA) and CTA.Results:CTA revealed 32 intracranial aneurysms. For silent MRA and TOF MRA, the scores of signal homogeneity were 3.38±0.49 and 3.00±0.62, andthe scores of venous signal/artifact were 3.77±0.42 and 2.65±0.48.Significant differences were found between the two MRAs in these aspects ( Z=-2.21, P=0.02; Z=-5.69, P=0.01). The scores of lesion conspicuity were 3.19±0.56 and 3.15±0.46, and the scores of diagnostic confidence were 3.27±0.44 and 3.12±0.51.There were no significant differences found in these aspects ( P>0.05).The ICC coefficient was excellentfor silent MRA (0.94, 95%CI 0.82- 0.98)and was good for TOF MRA (0.72, 95%CI 0.30-0.91) in tiny aneurysm group. The ICC coefficient was excellent (silent MRA, 0.98, 95%CI 0.95-0.99; TOF MRA, 0.95, 95%CI 0.87-0.98) for both MRA in non-tiny aneurysm group. Conclusions:Compared with TOF MRA, silent MRA could achieve higher image quality and higher diagnostic confidence, and higher consistency with CTA. Silent MRA can be a promising non-contrast-enhanced alternative MRA technique in clinical setting.