Fisher, Ronald E. MD, PhD*; Patel, Niraj R. MD†; Lai, Eugene C. MD, PhD‡; Schulz, Paul E. MD§
Encephalitis predominantly affecting the limbic areas was first described in 1960 in a series of 3 patients, 2 of whom were suspected to have lung carcinoma.1 Subsequent reports of the disease classified limbic encephalitis as an autoimmune, paraneoplastic syndrome characterized by the subacute onset of memory impairment, mood abnormalities, and seizures. More recently, it has become clear that about half of the cases of autoimmune limbic encephalitis (ALE) occur in the absence of cancer.2–4 Evidence is building that both forms of this disorder are more common than previously believed and are widely underdiagnosed,2,5 partly due to lack of consistent, reliable imaging patterns on MRI or PET. When correctly diagnosed, many patients show a partial-to-complete response to intravenous methylprednisolone and other immunotherapies.6–8
Many patients with autoimmune encephalitis have antibodies in their cerebrospinal fluid (CSF) directed against neuronal proteins, several of which have been identified, including antibodies to glutamate receptors (NMDA and AMPA type),2,4,9 GABA receptors,3 glutamic acid decarboxylase,10 and the neurally secreted LGI1 protein, an antigen recently demonstrated to be responsible for the encephalitis formerly thought to result from antibodies against a voltage-gated potassium channel.6,8,11,12 Many more are likely to exist.13 This implies the presence of multiple distinct subtypes of ALE.
2-[18F] fluoro-2-deoxy-D-glucose (FDG) brain PET imaging produces a map of regional cerebral glucose metabolism, which closely reflects regional metabolic brain activity.14 In neurology, clinical applications of PET have been most successful in patients with dementia, in large part due to the presence of well-defined, disease-specific FDG scan patterns. Although striking, diffuse cortical 18F FDG patterns of Alzheimer disease (AD) and dementia with Lewy bodies (DLB) overlap with each other, differing mainly in terms of occipital cortex metabolism, they are quite distinct from other causes of dementia, including frontotemporal dementia, multi-infarct dementia, Creutzfeldt-Jakob disease, and Huntington chorea.15,16
18F FDG PET brain metabolism patterns in ALE have been described in several case reports, clinical reports, and small series of patients, but results have been variable and a disease-specific pattern has not yet emerged.17–26
However, PET imaging was not the focus of most of these reports and, with a few exceptions, images were not presented or discussed in great detail. MRI abnormalities have been reported in ALE, most commonly hyperintensity of the medial temporal lobe on T2-weighted and FLAIR images.3,6,11 However, many patients have normal or nonspecific findings on MRI.4,19 These publications reveal the need and potential value of PET imaging but have not yet provided specific, reliable scan patterns that can be used by readers of clinical brain PET studies to diagnose this disorder. We attempt to provide such guidelines in this report.
MATERIALS AND METHODS
Between December 2007 and December 2010, 9 patients, ages 21 to 80, presented with subacute onset of cognitive decline. Some patients also had prominent affective or psychotic symptoms. Five patients progressed over days to weeks to an unresponsive or nearly unresponsive state. Three patients had seizures early in the course of their disease.
All patients showed inflammatory abnormalities in their CSF, most consistently an elevated IgG index. Some also had an elevated IgG synthesis rate or mild CSF lymphocytosis (Table 1). After extensive evaluation by neurologists with particular expertise in dementia (E.L. and P.S.), which included MRI and 18F FDG PET imaging in all subjects, ALE was diagnosed. PET scanning was ordered in all cases because of diagnostic uncertainty. The final diagnosis, made subsequent to the PET scan, was based on (1) the clinical features, (2) elevated CSF parameters for inflammation, and (3) the clinical response, namely, improvement with immunosuppression.
The presence of an underlying malignancy was excluded by whole-body PET/CT and/or whole-body CT imaging, clinical follow-up, and in some cases tumor markers. Other etiologies of encephalitis were excluded by CSF markers and specific polymerase chain reaction tests, cultures, or stains.
All patients responded at least partially to intravenous corticosteroid treatment over several days to weeks. All patients were started on other immunosuppressive agents to allow weaning of their corticosteroids.
Five patients underwent 18F FDG PET/CT imaging between 2 and 9 days after initiating therapy with intravenous corticosteroids, and 3 patients had not yet received steroid therapy. The time between onset of symptoms and PET/CT scanning varied from 4 days to 2 months, as diagnosis of this uncommon entity was often elusive early in the course of the disease. PET/CT scanning of the brain was performed in 3D mode approximately 1 hour after intravenous injection of 10 mCi 18F FDG. All imaging was performed on a GE Discovery ST scanner, with a PET acquisition time of 8 minutes. Two patients also received a follow-up 18F FDG PET scan. All PET scans were ordered by clinicians for clinical and diagnostic purposes.
FDG PET Scan Pattern 1 (Mixed Hyper-Hypometabolic Pattern)
Five patients, ages 21 to 47, presented with subacute cognitive decline, but differed clinically from the second group in that bizarre behavior, psychosis, and/or nonsensical speech were prominent features in 3 patients, 3 also had dysautonomia, and 3 had seizures. Their PET scans demonstrated a unique pattern, with the following features (Figs. 1–3):
* Strikingly reduced metabolism in the occipital cortex, including primary visual cortex, and mildly reduced metabolism in the primary sensorimotor strips
* Increased metabolism throughout much of the temporal lobes, especially laterally (marked and bilateral in 3 patients, mild in 1, and mild and predominantly unilateral in 1)
* Increased metabolism in the orbitofrontal cortex bilaterally
* Borderline increased metabolism in the cerebellum diffusely and bilaterally
* Borderline reduced metabolism in the parietal cortex
* Inconsistent findings: mildly reduced thalamic metabolism in 2 patients; increased metabolism of the hypothalamus in 2 patients
The final patient in our study (patient 5), who presented after we had identified the metabolic pattern just described, nicely illustrates the clinical value of specific PET scan patterns in this disorder. This 47-year-old woman presented with 6 weeks of progressive short-term memory loss, anxiety, and depression. Onset was shortly after her mother was diagnosed with cancer, and her symptoms were initially attributed to this emotional trauma. A psychiatrist diagnosed Mood Disorder, NOS. A consulting neurologist ordered a PET scan, but autoimmune encephalitis, at this point, was not being seriously considered as the etiology. The PET scan showed the typical findings of our 4 previous ALE patients (Fig. 3). Following the striking PET scan result, high-dose intravenous corticosteroid therapy was initiated, and the patient demonstrated dramatic clinical improvement, returning nearly to her normal baseline in 4 weeks.
Commercial testing for anti-NMDA receptor antibodies has become available only recently. Two patients in the mixed hyper-hypometabolic group tested positive for anti-NMDA receptor antibodies, 1 had negative findings, and the 2 others were not tested due to unavailability of the test at the time of their hospitalization. Testing was not available for any of the patients in the older group.
FDG PET Scan Pattern 2 (“Neurodegenerative Pattern”)
Four patients presenting with subacute cognitive decline, ages 52–80, showed a very different PET scan pattern. These patients were older than the first group, had less striking symptoms, and did not have seizures. Of the 4, 3 displayed significant confusion, 1 was combative, 1 incontinent, and 1 depressed. None of these patients had any history of dementia prior to the onset of their illness, and all were functioning normally a few weeks prior to onset of the illness. Onset of symptoms was gradual over days to weeks. Two patients had 18F FDG PET scans indistinguishable from advanced AD, including moderately to markedly reduced metabolism in the cortex diffusely, particularly in the posterior temporal lobes, with sparing of the visual cortex and primary sensorimotor strips (Fig. 4). Uptake in the subcortical structures was normal. One patient additionally demonstrated reduced uptake in the visual cortex and thus was more typical for advanced DLB (Fig. 5). The fourth patient had an AD-like pattern but milder. She recovered more rapidly in response to intravenous corticosteroid therapy than the others, returning to her normal cognitive baseline in approximately 5 days (3 days after her PET scan).
It is worth noting that corticosteroids are known to reduce brain glucose metabolism,27 although this has not been studied extensively. They do not cause cortical hypermetabolism, but to exclude the possibility that they were causing the neurodegenerative type PET patterns in our patients, we examined the brain images from 5 patients who had received 18F FDG PET scans for cancer staging or restaging, who had been receiving corticosteroid therapy for at least 3 days prior to the scan, and who had not received chemotherapy in the previous 3 months. Two had never received chemotherapy. They were well matched in age to the ALE patients (ages, 47–79). In addition to visual inspection, we analyzed the scans semiquantitatively by measuring the maximum standard uptake value in the basal ganglia and in the temporoparietal cortex using all axial slices that included the basal ganglia. We found that there was no specific pattern to the 18F FDG hypometabolism—it affected all brain structures equally. Specifically, the sensorimotor strips, visual cortex, and subcortical structures were reduced comparably to the rest of the cortex. Thus, corticosteroid therapy did not reproduce the scan patterns of our ALE patients.
Two patients had a follow-up PET scan. One patient with the first 18F FDG pattern had a repeat scan 1 month after the first, when his clinical state had deteriorated into a coma. Surprisingly, the PET scan showed significant improvement (Fig. 2B). This was followed by clinical improvement that began approximately 1 week later and gradually progressed to a full recovery to normal baseline at 1 year. A patient with the neurodegenerative 18F FDG pattern had a repeat scan 10 months later, during a mild recurrence (Fig. 5B), which showed partial recovery of his scan findings. This patient had a fluctuating, but generally improving, clinical course. Note that the clear improvement in his scan confirms that his original PET abnormalities were not due to advanced neurodegenerative disease.
Four of the patients in the younger group (mixed hyper/hypometabolism PET pattern) recovered essentially to their normal baseline after 1 to 12 months of immunosuppression. The fifth gradually recovered nearly to baseline, but has continued to have mild, periodic relapses for the past 3 years. Of the 4 patients with the neurodegenerative type PET pattern, 1 recovered very quickly following immunosuppression (5 days); this patient had the mildest of the PET scan abnormalities. Two of the other patients recovered partially following immunosuppression, but continue to have occasional relapses, which respond well to corticosteroids. The fourth patient in this group had additional medical conditions and died of sepsis, although some improvement in cognition had already occurred.
It should be noted that none of our patients have yet been diagnosed with a tumor on follow-up, although the follow-up period is relatively brief at this point—less than 18 months for the majority of patients.
The diagnosis of ALE is often delayed or missed.2,5 This can have enormous consequences, as the illness can be fatal if untreated but often responds to immunosuppression.5,28 Therefore, imaging potentially plays an important role in the diagnostic evaluation.2,10
The most commonly used imaging study in ALE patients is MRI, but previous publications have found mixed results. Some reports indicate that hyperintensities in the medial temporal lobes on T2 and FLAIR images are often observed in these patients.3,6,10,11 However, many patients have normal or nonspecific MRI scans.19 In most of our patients, the MRI was unremarkable, and it did not yield the diagnosis in any. This is similar to one recent, large study, which found that 89% of initial MRI scan findings were negative.28
Previous reports have found that 18F FDG PET scans are usually abnormal in ALE, but the findings have been widely variable. Our goal is to summarize and organize these results and combine them with our new data to generate clinically useful interpretive criteria for ALE.
The predominant abnormalities in several case reports, a series of 6 patients, and a large clinical study that briefly mentions PET results on 12 patients, have been located in the medial temporal lobes, mainly focal hypermetabolism,10,20–24,29 occasionally hypometabolism.10,24 The common feature of all of the earlier reports is that PET abnormalities were confined mainly to the temporal lobes. It may be noted that, while PET imaging is mentioned in the previous studies, none of them focuses on PET imaging or describes the PET results in great detail. A recent report of 2 young children showed a very different, asymmetric hypometabolic pattern, which perhaps is restricted to children.18
More recently, a case report of anti-NMDA-R encephalitis has appeared describing a pattern similar to our mixed hyper-hypometabolic pattern.17 This pattern has not been reported for any other disorder. Another recent case report shows limited PET images, but the occipital hypometabolism of our mixed pattern can clearly be observed.30 A third recent case report includes just one axial slice, but occipital hypometabolism is clearly evident, as pointed out by the authors.26 A fourth recent case report claims occipital hypometabolism, not included in the limited images in the article.31 Finally, one of the patients in one report is briefly described as having increased metabolism in the right temporal lobe and decreased in the occipital lobes, which likely resembles our scan pattern, though no images of this patient are shown.24 In summary, 5 previously reported patients have had scan findings similar to our mixed hyper-hypometabolic pattern.
Previous studies have not noted any resemblance between the PET scan appearance of autoimmune encephalitis patients and those with severe neurodegenerative disease. Yet, this was the predominant scan finding in our older patients. The only explanation we can offer is that our subgroup of patients with this pattern was older than in most of the previous studies, and may represent a different subtype of the disease. Most of the previous PET studies focused on young patients.
Our 2 different scan patterns may represent 2 etiologically different subtypes of ALE, but commercial antibody testing has become available only recently and for a limited number of antibodies; we do not have such results for most of our patients. Two patients in our mixed hyper-hypometabolic group tested positive for anti-NMDA receptor antibodies. Because no antibody results are available for any of the patients in the older group, it is not possible to say whether the 2 scan patterns we found represent immunologically different subtypes of disease.
Previous publications have been divided on the explanation of the hypermetabolic abnormalities sometimes observed in ALE.20–24 We believe these are most likely due to either inflammation, supported by the finding of increased CSF markers of inflammation in all subjects, or to locally increased neuronal activity caused by direct alteration of synaptic and neuronal properties by the antibodies, for which there is experimental evidence.9,32,33 We think seizure activity is unlikely to be the cause, as most seizures, even if they happen to occur during the FDG uptake period, are insufficient to induce a detectible hypermetabolic scan focus,34 with the few reported exceptions being cases of status epilepticus35,36 or rapidly recurring seizures.37 Whatever the cellular mechanism, the temporal and orbitofrontal neuronal hyperactivity could underlie some of the prominent psychomotor features of this subgroup of patients.
We are interested in the possibility that 18F FDG PET may prove useful in predicting response to therapy in ALE patients. Our limited data raise the possibility that a mildly abnormal PET scan (patient 9) or improving scan (patient 3) presages a short disease course and/or excellent response to steroid therapy, although more cases obviously are needed for confirmation. Similar, though less dramatic, prognostic value has been reported.23,24
In all of our cases, PET scans were ordered before a clinical diagnosis was established, and a variety of potential diagnoses were still being entertained. Our mixed pattern is particularly useful, as it strongly supports a diagnosis of ALE. Our neurodegenerative pattern is obviously less specific, but still useful if clinical information is provided to the PET reader: in the setting of subacute onset dementia and/or prominent psychiatric features or seizures, then ALE can reasonably be raised as a possible diagnosis, and the scan not mistaken for neurodegenerative disease. Both of the PET scan patterns discussed in this report can help exclude psychiatric illness, epilepsy, or drug effects, among other things, as the etiology.
A third PET scan pattern occasionally reported in the literature mainly in patients with paraneoplastic autoimmune encephalitis, namely, focal hypermetabolism in the medial temporal lobes, was not observed in our patients, none of whom had evidence of an associated neoplasm. However, based on the literature, this might reasonably be considered a third type of 18F FDG scan pattern in autoimmune encephalitis.
We conclude that at least 2 very distinct 18F FDG PET patterns can be found in ALE. Our younger patients had a mixed hyper-hypometabolic 18F FDG pattern that is similar to findings reported recently for 5 other patients.17,24,26,30,31 This pattern is easily recognized by the striking disparity between temporal hypermetabolism and occipital hypometabolism. This pattern may be highly suggestive of ALE, as it has not been reported in the scientific literature for any other disorder. Our older patients had a pattern that was indistinguishable from neurodegenerative disease, a finding not previously reported. Awareness of both scan patterns can potentially contribute to the elusive diagnosis of this treatable disorder.
The authors thank the Lou DeGeorge family for its invaluable contributions.
1. Brierley JB, Corsellis CA, Hierons R, et al.. Subacute encephalitis of later adult life mainly affecting the limbic areas. Brain. 1960; 83: 357–370.
2. Wandinger KP, Saschenbrecker S, Stoecker W, et al.. Anti-NMDA-receptor encephalitis: a severe, multistage, treatable disorder presenting with psychosis. J Neuroimmunol. 2011; 231: 86–91.
3. Lancaster E, Lai M, Peng X, et al.. Antibodies to the GABA(B) receptor in limbic encephalitis with seizures: case series and characterisation of the antigen. Lancet Neurol. 2010; 9: 67–76.
4. Dalmau J, Gleichman AJ, Hughes EG, et al.. Anti-NMDA-receptor encephalitis: case series and analysis of the effects of antibodies. Lancet Neurol. 2008; 7: 1091–1098.
5. Dalmau J, Lancaster E, Martinez-Hernandez E, et al.. Clinical experience and laboratory investigations in patients with anti-NMDAR encephalitis. Lancet Neurol. 2011; 10: 63–74.
6. Wong SH, Saunders MD, Larner AJ, et al.. An effective immunotherapy regimen for VGKC antibody-positive limbic encephalitis. J Neurol Neurosurg Psychiatry. 2010; 81: 1167–1169.
7. Mori M, Kuwabara S, Yoshiyama M, et al.. Successful immune treatment for non-paraneoplastic limbic encephalitis. J Neurol Sci. 2002; 201: 85–88.
8. Vincent A, Buckley C, Schott JM, et al.. Potassium channel antibody-associated encephalopathy: a potentially immunotherapy-responsive form of limbic encephalitis. Brain. 2004; 127: 701–712.
9. Lai M, Hughes EG, Peng X, et al.. AMPA receptor antibodies in limbic encephalitis alter synaptic receptor location. Ann Neurol. 2009; 65: 424–434.
10. Malter MP, Helmstaedter C, Urbach H, et al.. Antibodies to glutamic acid decarboxylase define a form of limbic encephalitis. Ann Neurol. 2010; 67: 470–478.
11. Schott JM, Harkness K, Barnes J, et al.. Amnesia, cerebral atrophy, and autoimmunity. Lancet. 2003; 361: 1266.
12. Lai M, Huijbers MG, Lancaster E, et al.. Investigation of LGI1 as the antigen in limbic encephalitis previously attributed to potassium channels: a case series. Lancet Neurol. 2010; 9: 776–785.
13. Bataller L, Kleopa KA, Wu GF, et al.. Autoimmune limbic encephalitis in 39 patients: immunophenotypes and outcomes. J Neurol Neurosurg Psychiatry. 2007; 78: 381–385.
14. Sokoloff L. Modeling metabolic processes in the brain in vivo. Ann Neurol. 1984; 15 (suppl): S1–S11.
15. Silverman DH. Brain 18F-FDG PET in the diagnosis of neurodegenerative dementias: comparison with perfusion SPECT and with clinical evaluations lacking nuclear imaging. J Nucl Med. 2004; 45: 594–607.
16. Henkel K, Zerr I, Hertel A, et al.. Positron emission tomography with [(18)F]FDG in the diagnosis of Creutzfeldt-Jakob disease (CJD). J Neurol. 2002; 249: 699–705.
17. Mohr BC, Minoshima S. 18F fluorodeoxyglucose PET/CT findings in a case of anti-NMDA receptor encephalitis. Clin Nucl Med. 2010; 35: 461–463.
18. Pillai SC, Gill D, Webster R, et al.. Cortical hypometabolism demonstrated by PET in relapsing NMDA receptor encephalitis. Pediatr Neurol. 2010; 43: 217–220.
19. Basu S, Alavi A. Role of FDG-PET in the clinical management of paraneoplastic neurological syndrome: detection of the underlying malignancy and the brain PET-MRI correlates. Mol Imaging Biol. 2008; 10: 131–137.
20. Provenzale JM, Barboriak DP, Coleman RE. Limbic encephalitis: comparison of FDG PET and MR imaging findings. Am J Roentgenol. 1998; 170: 1659–1660.
21. Fakhoury T, Abou-Khalil B, Kessler RM. Limbic encephalitis and hyperactive foci on PET scan. Seizure. 1999; 8: 427–431.
22. Kassubek J, Juengling FD, Nitzsche EU, et al.. Limbic encephalitis investigated by 18FDG-PET and 3D MRI. J Neuroimaging. 2001; 11: 55–59.
23. Fauser S, Talazko J, Wagner K, et al.. FDG-PET and MRI in potassium channel antibody-associated non-paraneoplastic limbic encephalitis: correlation with clinical course and neuropsychology. Acta Neurol Scand. 2005; 111: 338–343.
24. Ances BM, Vitaliani R, Taylor RA, et al.. Treatment-responsive limbic encephalitis identified by neuropil antibodies: MRI and PET correlates. Brain. 2005; 128: 1764–1777.
25. Gast H, Schindler K, Z’Graggen WJ, et al.. Improvement of non-paraneoplastic voltage-gated potassium channel antibody-associated limbic encephalitis without immunosuppressive therapy. Epilepsy Behav. 2010; 17: 555–557.
26. Maeder-Ingvar M, Prior JO, Irani SR, et al.. FDG-PET hyperactivity in basal ganglia correlating with clinical course in anti-NDMA-R antibodies encephalitis. J Neurol Neurosurg Psychiatry. 2011; 82: 235–236.
27. Fulham MJ, Brunetti A, Aloj L, et al.. Decreased cerebral glucose metabolism in patients with brain tumors: an effect of corticosteroids. J Neurosurg. 1995; 83: 657–664.
28. Irani SR, Bera K, Waters P, et al.. N-methyl-D-aspartate antibody encephalitis: temporal progression of clinical and paraclinical observations in a predominantly non-paraneoplastic disorder of both sexes. Brain. 2010; 133: 1655–1667.
29. Dadparvar S, Anderson GS, Bhargava P, et al.. Paraneoplastic encephalitis associated with cystic teratoma is detected by fluorodeoxyglucose positron emission tomography with negative magnetic resonance image findings. Clin Nucl Med. 2003; 28: 893–896.
30. Iizuka T, Sakai F, Ide T, et al.. Anti-NMDA receptor encephalitis in Japan: long-term outcome without tumor removal. Neurology. 2008; 70: 504–511.
31. Maqbool M, Oleske DA, Huq AH, et al.. Novel FDG-PET findings in anti-NMDA receptor encephalitis: a case based report. J Child Neurol. 2011; 26: 1325–1328.
32. Hughes EG, Peng X, Gleichman AJ, et al.. Cellular and synaptic mechanisms of anti-NMDA receptor encephalitis. J Neurosci. 2010; 30: 5866–5875.
33. Moscato EH, Jain A, Peng X, et al.. Mechanisms underlying autoimmune synaptic encephalitis leading to disorders of memory, behavior and cognition: insights from molecular, cellular and synaptic studies. Eur J Neurosci. 2010; 32: 298–309.
34. Barrington SF, Koutroumanidis M, Agathonikou A, et al.. Clinical value of “ictal” FDG-positron emission tomography and the routine use of simultaneous scalp EEG studies in patients with intractable partial epilepsies. Epilepsia. 1998; 39: 753–766.
35. Dong C, Sriram S, Delbeke D, et al.. Aphasic or amnesic status epilepticus detected on PET but not EEG. Epilepsia. 2009; 50: 251–255.
36. Handforth A, Cheng JT, Mandelkern MA, et al.. Markedly increased mesiotemporal lobe metabolism in a case with PLEDs: further evidence that PLEDs are a manifestation of partial status epilepticus. Epilepsia. 1994; 35: 876–881.
37. Engel J Jr, Kuhl DE, Phelps ME. Patterns of human local cerebral glucose metabolism during epileptic seizures. Science. 1982; 218: 64–66.
© 2012 Lippincott Williams & Wilkins, Inc.