Authors :
Presenting Author: Kia Gilani, MD – Icahn School of Medicine at Mount Sinai, New York
Timothy Shepherd, MD – NYU
Leah Blank, MD, MPH – Icahn School of Medicine at Mount Sinai, New York
Claude Steriade, MD – Department of Neurology, NYU Langone Medical Center, New York City
Rationale:
PET abnormalities are common in autoimmune encephalitis (AE), and PET hypermetabolism has been proposed as a marker of ongoing brain autoimmunity. We aimed to examine PET findings in relation to immunotherapy (IT) response in AE.
Methods:
Demographics, presenting features, IT, seizure burden were identified in a retrospective cohort of 32 patients with AE (neural autoantibody positive or meeting diagnostic criteria for probable seronegative AE) with FDG-PET scans. For each initial PET scan, seizure frequency, IT escalation (within 3 months or initial IT treatment if >3 months) and response, and (if available) simultaneous MRI findings were compared between those with and without hypermetabolism, using Fisher’s Exact and Mann-Whitney U tests. Results:
32 patients (mean age at diagnosis 42.1, SD 21.0, etiology: GAD65 (9), LGI1 (11), seronegative (7), Hu (2), Ma2 (2), GABA-B (1)) had 43 PET scans in total (7 individuals had 2 scans and 4 had three PET scans). Of 43 PET scans (7 preceded AE diagnosis), hypermetabolism was present in 19, hypometabolism in 39, one PET was normal. Hypermetabolism was more likely present if PET was closer in time to initiation of IT (11.2 vs. 49.8 months, P=0.043). Hypermetabolism did not correlate with daily seizure frequency at the time of PET (p=0.26). Of the 21 patients with IT initiation following or escalation near initial PET, 18 improved (12 with hypermetabolism, 9 with hypometabolism only). All patients with IT escalation surrounding observation of hypermetabolism on FDG PET with known outcomes improved. The three patients without improvement had PET hypometabolism only. MRI anatomic correlates (atrophy, edema, or normal findings) to PET abnormalities appeared similar between hyper- and hypo-metabolic regions on initial PET (p=0.83).
Increase (n=3) or decrease (n=3) in hypometabolism, increase (n=3) or decrease (n=2 ) in hypermetabolism, and no change (n=3) were seen in the 11 follow up PET scans. MRIs showed no change (6), less edema (1), and increased atrophy of mesial temporal structures (3). All 3 follow up PET scans with interval reduction in hypometabolism had IT escalation at the time of the prior scan and 2 had improved clinically.
Conclusions:
FDG PET abnormalities are common in AE. Hypermetabolism is found closest in time to initiation of IT and occurs independently of seizure frequency, suggesting that it reflects active brain autoimmunity and inflammation rather than seizure activity. MRI correlates of hypometabolism and hypermetabolism are similar, suggesting added value of FDG-PET to MRI imaging. In this cohort, patients with hypermetabolism on FDG PET uniformly benefited from IT escalation. A subset of patients with only hypometabolism benefited from IT escalation, some of whom had reduced hypometabolism on follow-up PET, suggesting that hypometabolism may also correlate with active brain autoimmunity. The absence of hypermetabolism should therefore not preclude IT escalation if otherwise indicated. Funding: None