Although various advanced MRI technologies have rapidly developed in recent years, these new technologies rely on advanced post-processing software and relatively complicated post-processing technologies. Therefore, conventional MRI plays a vital role in diagnosing CNS disease. At present, research on anti-NMDA receptor encephalitis mainly focuses on the clinical aspect, while imaging studies on this disease are few, and awareness about this disease is insufficient[16]. Therefore, we analyzed the conventional MRI and DWI characteristics of anti-NMDA receptor encephalitis lesions diagnosed in our hospital in recent years, which is conducive to the correct diagnosis of this disease and improvements in understanding this disease. As a quantitative indicator of DWI, ADC can be used to determine the phase of the lesion.
First, we found only 22/66 patients with positive intracerebral lesions were observed on conventional MRI. The incidence rate of positive lesions was 33.33%, consistent with 30–50%[17]. In addition, there were more females than males in this group, with an average age of 28.82 years, which was in line with previous research reports[17]. In the past, this disease was reported to mainly affect the medial temporal lobe and the hippocampus based on positron emission computed tomography (PET-CT)[18]. Similarly, Harald Hegen et al. observed abnormal MRI signals in the bilateral insular lobes and the bilateral hippocampal gyrus in one patient with subacute anti-NMDA receptor encephalitis, which was hypermetabolic on corresponding PET-CT[19]. However, in this study, the lesions involved nearly all cerebral regions but were mainly involved in the cortical and subcortical regions. Still they were consistent with the decreased cortical metabolism in anti-NMDA receptor encephalitis[18]. Furthermore, one study by the Xuanwu Hospital Capital Medical University showed that 29 out of 62 (46.8%) patients with anti-NMDA receptor encephalitis had abnormal conventional MRI. Five lesions extensively involved the cerebral cortex[17]. Their incidence of positive lesions and involved cerebral areas were similar to our results. This study revealed anti-NMDA receptor encephalitis involves the cortex because the subcortical white matter is mainly composed of the myelin sheath and oligodendrocytes. The latter highly expressed NMDA receptors, which allows antibodies able to attack these receptors[20].
Next, we found that the ADC values of the 19 lesions with isointensiy on T1WI in the 17 patients with relative short DDs were lower than those of the contralateral normal brain tissues. The sulci vessels were thickened and the adjacent meninges were thickened in 7 cortical lesions. It has been suggested that immune-mediated microvascular diseases, lymphocytic vascular diseases, and antigen-antibody reactions lead to hyperemia within the lesion and perivascular inflammation, resulting in intracellular water-sodium retention and cytotoxic edema. However, because the cytomembrane is still intact, the diffusion of water molecules is limited, so it presents as hyperintensity on DWI. This is similar to an acute cerebral infarction (ACI)[21], which is hyperintense on DWI and hypointense on the corresponding ADC images. However, compared with patients with anti-NMDA receptor encephalitis, patients with ACI are older and have a more rapid onset. The typical imaging manifestations of ACI are consistent with the cerebrovascular distribution, whereas anti-NMDA receptor encephalitis is not. Therefore, imaging should be closely combined with a clinical diagnosis to distinguish anti-NMDA receptor encephalitis from ACI. Furthermore, this study also found that lesions were more clearly observed on DWI than on T2WI. In other words, the lesions are not clear or cannot be observed on conventional T2WI, but they are hyperintensity on DWI and hypointensity on ADC maps. This also suggests that the pathophysiological process of this disease in the early stage is cytotoxic edema, but the degree is too mild to be shown on conventional T2WI. Wang et al. found that FLAIR combined with DWI is more sensitive in detecting abnormalities than conventional MRI sequences[22]. Such abnormalities on DWI resemble that of demyelinating disease such as acute disseminated encephalomyelitis[15], multiple sclerosis[23]. Therefore, DWI can identify more occult lesions and can predict the occurrence of new lesions to a certain extent, further indicating the advantages of this technique in the diagnosis of anti-NMDA receptor encephalitis.
Furthermore, the results of this study also showed that the ADC values of the 22 hyperintense lesions were higher than those of the contralateral normal cerebral tissues on the ADC images, and these patients also had relatively long DDs. The possible reason for this is that the vasculitis gradually subsides, the cytomembrane is damaged, and the diffusion rate of the water molecules increases, leading to an increase in the ADC value with the progression of the disease. At this stage, the cytotoxic edema gradually evolves into vasogenic edema, so gliosis, demyelination, neuronal necrosis, and other pathological changes play a dominant role. Tsubouchi et al. also showed that the ADC value of chronic encephalitis was lower than that of acute encephalitis, which is consistent with the results of this study[24–26]. Both are inflammatory diseases, which supports the reliability of the results in another sense.A biopsy study from Camdessanche et al.[25–27] reported that the main pathological changes were gliosis, microglial hyperplasia, and IgG deposition in the right frontal lobe of one patient with anti-NMDA receptor encephalitis more than one month after onset. However, no demyelination or neuronal necrosis was found in that study, possibly due to the short DD and no progression of this pathological change. Additionally, the average ADC value of all 29 lesions was still higher than that of the contralateral normal cerebral tissue in this group, possibly because hyperintense lesions on the ADC images accounted for the majority. Therefore, the lesions should not be analyzed in general, which cannot truly reflect the pathological changes of the lesions. Lesions with different ADC signals should be analyzed separately, which is more conducive to judging the different pathological stages of the lesions.
In this group, most of the lesions were hyperintense on both T2WI but presented isointensity or hypointensity on T1WIs. Additionally, the ADC values of the isointense lesions were lower than those of the hypointense lesions. This quantitative value of the ADC could suggest that the pathological changes are decided by the stage of the disease. The isointense lesions on T1WI had reduced ADC values and were hypointense on the corresponding ADC images in this study, indicating that the disease is at an early stage. The possible pathological changes are inflammatory cell infiltration and cytotoxic edema. Meanwhile, the hypointensi lesions on T1WI had increased ADC values and were hyperintense on the corresponding ADC images, indicating that the disease is at a late stage, possibly accompanied by tissue destruction. Therefore, the diffusion of water molecules is accelerated. Thus, the inverse relationship between the ADC values and the signal intensity on T1WI further suggests that ADC is an effective quantitative parameter, which could provide a very important value to identify the pathological changes of the lesions with a different signal on T1WI in anti-NMDAR encephalitis.
In addition, correlation analysis showed that the ADC values of the lesions were positively correlated with the DDs. That is, the longer the DD, the higher the ADC value, which further indicates that the ADC value is different at variable stages of the disease. To some extent, it is a supplement to the lesions with a different signal on T1WI and on ADC images. We also found that there was no correlation between ADC values and mRS scores. The possible reason for this is that the mRS score is relatively simple and is determined subjectively to some extent. However, Probasco et al.[25] found that the metabolism of the lateral lobe and medial occipital lobe in patients with anti-NMDAR encephalitis with mRS scores of 4–5 was lower than that in patients with mRS scores of 0–3. The reasons for this difference may be related to the different number of patients and the different stages of the disease, so more patients need to be collected for analysis.
Limitations
Although we believe that our finds in this study are substantial and have a good pathophysiological correlation, they are not without limitations. First, the cross-sectional design and relatively small number of sbjects studied may not provide an accurate reflection of the characteristic of the lesions in patients with anti-NMDA receptor encephalitis. Second, owing to the nature of DWI and the associated “blooming” effects, our measurements of ADC values are inherently relatively rather than absolute. Third, ADC values may be influenced by lesions; however, the assessments of the correlations between the ADC values and the DD and the mRS scores have been a topic of previous investigation and are beyond the scope of our study.