MR Perfusion in the Workup of NMDA Receptor Encephalitis
Abby Scurfield1, Kristin Galetta3, Bryan Lanzman2, Rachelle Dugue1
1Neurology, 2Neuroradiology, Stanford, 3Neurology, Stanford University
Objective:
Assess the utility of MR perfusion in the diagnosis of NMDA receptor autoimmune encephalitis.  
Background:
Autoimmune encephalitis (AE) is frequently misdiagnosed, contributing to treatment delays and poor outcomes. Standard brain MRI is abnormal in approximately 50% of patients, often even less in NMDA encephalitis. In contrast to MRI, positron emission tomography (PET) scans are more sensitive, with findings in up to 99% of cases. PET, however, is not widely available, costly, and exposes patients to radiation. Similar to PET, MR perfusion imaging characterizes regional blood flow, perhaps representing an accessible alternative to PET to aid in early diagnosis of AE. 
Design/Methods:
We conducted a retrospective analysis of confirmed NMDAR+ encephalitis cases from Stanford Research Repository (STARR) database. Rstudio was used for descriptive and inferential analyses. Firth’s penalized logistic regression was used to examine potential association between MR perfusion abnormalities and APEII scores, seizures, ICU admission or age.   
Results:
Of 17 NMDA cases, 16 cases had MR perfusion studies, 50% of which were abnormal. Amongst those, 50% had normal FLAIR sequences. PET scans were performed in 11 cases, 4 of which were positive. Amongst these positive cases, 3/4 had MRI perfusion abnormalities, 2 out of 3 of which matched the abnormal region on PET. There were no significant associations between the presence of a perfusion abnormality and APEII score, age, ICU admission, or seizures (p>0.05).  
Conclusions:
50% of NMDA patients had MR perfusion changes. Notably, half of these cases had negative MRI sequences otherwise. Only a small proportion of PET studies were positive, the majority of which had MR perfusion abnormalities. These findings support use of MR perfusion in AE workup, especially when conventional MRI is unrevealing. Future analyses will assess the utility of MR perfusion as a tool to discriminate AE cases through a blinded-imaging re-analysis of cases and age matched controls.   
Generative AI Usage
No, did not use generative AI in the drafting or editing in this abstract.
Disclaimer: Abstracts were not reviewed by Neurology® and do not reflect the views of Neurology® editors or staff.