No abstract

BackgroundAutoimmune encephalitis (AE) is a noninfectious emergency with severe clinical attacks. It is difficult for the earlier diagnosis of acute AE due to the lack of antibody detection resources.PurposeTo construct a deep learning (DL) algorithm using multi‐sequence magnetic resonance imaging (MRI) for the identification of acute AE.Study TypeRetrospective.PopulationOne hundred and sixty AE patients (90 women; median age 36), 177 herpes simplex virus encephalitis (HSVE) (89 women; median age 39), and 184 healthy controls (HC) (95 women; median age 39) were included. Fifty‐two patients from another site were enrolled for external validation.Field Strength/Sequence3.0 T; fast spin‐echo (T1WI, T2WI, fluid attenuated inversion recovery imaging) and spin‐echo echo‐planar diffusion weighted imaging.AssessmentFive DL models based on individual or combined four MRI sequences to classify the datasets as AE, HSVE, or HC. Reader experiment was further carried out by radiologists.Statistical TestsThe discriminative performance of different models was assessed using the area under the receiver operating characteristic curve (AUC). The optimal threshold cut‐off was identified when sensitivity and specificity were maximized (sensitivity + specificity − 1) in the validation set. Classification performance using confusion matrices was reported to evaluate the diagnostic value of the models and the radiologists' assessments before being assessed by the paired t‐test (P < 0.05 was considered significant).ResultsIn the internal test set, the fusion model achieved the significantly greatest diagnostic performance than single‐sequence DL models with AUCs of 0.828, 0.884, and 0.899 for AE, HSVE, and HC, respectively. The model demonstrated a consistently high performance in the external validation set with AUCs of 0.831 (AE), 0.882 (HSVE), and 0.892 (HC). The fusion model also demonstrated significantly higher performance than all radiologists in identifying AE (accuracy between the fuse model vs. average radiologist: 83% vs. 72%).Data ConclusionThe proposed DL algorithm derived from multi‐sequence MRI provided desirable identification and classification of acute AE.Level of Evidence3Technical EfficacyStage 2

ObjectiveTo develop a fusion model combining clinical variables, deep learning (DL), and radiomics features to predict the functional outcomes early in patients with adult anti-N-methyl-D-aspartate receptor (NMDAR) encephalitis in Southwest China.MethodsFrom January 2012, a two-center study of anti-NMDAR encephalitis was initiated to collect clinical and MRI data from acute patients in Southwest China. Two experienced neurologists independently assessed the patients’ prognosis at 24 moths based on the modified Rankin Scale (mRS) (good outcome defined as mRS 0–2; bad outcome defined as mRS 3-6). Risk factors influencing the prognosis of patients with acute anti-NMDAR encephalitis were investigated using clinical data. Five DL and radiomics models trained with four single or combined four MRI sequences (T1-weighted imaging, T2-weighted imaging, fluid-attenuated inversion recovery imaging and diffusion weighted imaging) and a clinical model were developed to predict the prognosis of anti-NMDAR encephalitis. A fusion model combing a clinical model and two machine learning-based models was built. The performances of the fusion model, clinical model, DL-based models and radiomics-based models were compared using the area under the receiver operating characteristic curve (AUC) and accuracy and then assessed by paired t-tests (P &lt; 0.05 was considered significant).ResultsThe fusion model achieved the significantly greatest predictive performance in the internal test dataset with an AUC of 0.963 [95% CI: (0.874-0.999)], and also significantly exhibited an equally good performance in the external validation dataset, with an AUC of 0.927 [95% CI: (0.688-0.975)]. The radiomics_combined model (AUC: 0.889; accuracy: 0.857) provided significantly superior predictive performance than the DL_combined (AUC: 0.845; accuracy: 0.857) and clinical models (AUC: 0.840; accuracy: 0.905), whereas the clinical model showed significantly higher accuracy. Compared with all single-sequence models, the DL_combined model and the radiomics_combined model had significantly greater AUCs and accuracies.ConclusionsThe fusion model combining clinical variables and machine learning-based models may have early predictive value for poor outcomes associated with anti-NMDAR encephalitis.

ObjectiveThis study aimed to investigate iron deposition and thickness and signal changes in optic radiation (OR) by enhanced T2*-weighted angiography imaging (ESWAN) in patients with relapsing-remitting multiple sclerosis (RRMS) with unilateral and bilateral lesions or no lesions.MethodsFifty-one RRMS patients (42 patients with a disease duration [DD] ≥ 2 years [group Mor], nine patients with a DD < 2 years [group Les]) and 51 healthy controls (group Con) underwent conventional magnetic resonance imaging (MRI) and ESWAN at 3.0 T. The mean phase value (MPV) of the OR was measured on the phase image, and thickness and signal changes of the OR were observed on the magnitude image.ResultsThe average MPVs for the OR were 1,981.55 ± 7.75 in group Mor, 1,998.45 ± 2.01 in group Les, and 2,000.48 ± 5.53 in group Con. In group Mor, 28 patients with bilateral OR lesions showed bilateral OR thinning with a heterogeneous signal, and 14 patients with unilateral OR lesions showed ipsilateral OR thinning with a heterogeneous signal. In the remaining nine patients without OR lesions and in group Con, the bilateral OR had a normal appearance. In the patients, a negative correlation was found between DD and OR thickness and a positive correlation was found between MPV and OR thickness.ConclusionsWe confirmed iron deposition in the OR in the RRMS patients, and the OR thickness was lower in the patients than in the controls.Key Points • Enhanced T 2 * -weighted magnetic resonance angiography (ESWAN) provides new insights into multiple sclerosis (MS). • Focal destruction of the optic radiation (OR) is detectable by ESWAN. • Iron deposition in OR can be measured on ESWAN phase image in MS patients. • OR thickness was lower in the patients than in the controls. • Iron deposition and thickness changes of the OR are associated with disease duration. Electronic supplementary materialThe online version of this article (10.1007/s00330-018-5461-8) contains supplementary material, which is available to authorized users.

ObjectivesTo compare parameters of diffusion tensor imaging (DTI) and diffusion kurtosis imaging (DKI) to evaluate which can better describe the microstructural changes of anti-N-methyl-D-aspartate receptor (NMDAR) encephalitis patients and to characterize the non-Gaussian diffusion patterns of the whole brain and their correlation with neuropsychological impairments in these patients.Materials and methodsDTI and DKI parameters were measured in 57 patients with anti-NMDAR encephalitis and 42 healthy controls. Voxel-based analysis was used to evaluate group differences between white matter and gray matter separately. The modified Rankin Scale (mRS) was used to evaluate the severity of the neurofunctional recovery of patients, the Montreal Cognitive Assessment (MoCA) was used to assess global cognitive performance, and the Hamilton Depression Scale (HAMD) and fatigue severity scale (FSS) were used to evaluate depressive and fatigue states.ResultsPatients with anti-NMDAR encephalitis showed significantly decreased radial kurtosis (RK) in the right extranucleus in white matter (P &lt; 0.001) and notably decreased kurtosis fractional anisotropy (KFA) in the right precuneus, the right superior parietal gyrus (SPG), the left precuneus, left middle occipital gyrus, and left superior occipital gyrus in gray matter (P &lt; 0.001). Gray matter regions with decreased KFA overlapped with those with decreased RK in the left middle temporal gyrus, superior temporal gyrus (STG), supramarginal gyrus (SMG), postcentral gyrus (POCG), inferior parietal but supramarginal gyrus, angular gyrus (IPL) and angular gyrus (ANG) (P &lt; 0.001). The KFA and RK in the left ANG, IPL and POCG correlated positively with MoCA scores. KFA and RK in the left ANG, IPL, POCG and SMG correlated negatively with mRS scores. KFA in the left precuneus and right SPG as well as RK in the left STG correlated negatively with mRS scores. No significant correlation between KFA and RK in the abnormal brain regions and HAMD and FSS scores was found.ConclusionThe microstructural changes in gray matter were much more extensive than those in white matter in patients with anti-NMDAR encephalitis. The brain damage reflected by DKI parameters, which have higher sensitivity than parameters of DTI, correlated with cognitive impairment and the severity of the neurofunctional recovery.