- Email: [email protected]
Neuroimaging in status epilepticus secondary to paraneoplastic autoimmune encephalitis
Clinical Radiology xxx (2014) e1ee9
Contents lists available at ScienceDirect
Clinical Radiology journal homepage: www.clinicalradiologyonline.net
Neuroimaging in status epilepticus secondary to paraneoplastic autoimmune encephalitis s c, S. Sarria-Estrada a, *, M. Toledo b, C. Lorenzo-Bosquet c, G. Cuberas-Borro Rovira a C. Auger a, S. Siurana a, A. n University Hospital, Barcelona, Spain Magnetic Resonance Unit, Radiology Department, Vall d’Hebro n University Hospital, Barcelona, Spain Epilepsy Unit, Neurology Department, Vall d’Hebro c n University Hospital, Barcelona, Spain Nuclear Medicine Department, Vall d’Hebro a
b
art icl e i nformat ion Article history: Received 18 August 2013 Received in revised form 16 March 2014 Accepted 18 March 2014
AIM: To describe the characteristic magnetic resonance imaging (MRI) findings of paraneoplastic autoimmune encephalitis in patients with new-onset status epilepticus. MATERIALS AND METHODS: The neuroimaging and clinical data of five patients with paraneoplastic autoimmune encephalitis debuting as status epilepticus were retrospectively reviewed. All patients met the criteria for definite paraneoplastic syndrome and all underwent brain MRI during the status epilepticus episode or immediately after recovery. RESULTS: All patients showed hyperintense lesions on T2-weighted imaging (WI) involving the limbic structures, specifically the hippocampus. Three of them showed additional extralimbic areas of signal abnormalities. The areas of T2 hyperintensity were related to the electroclinical onset of the seizures. In three patients, various techniques were used to study cerebral perfusion, such as arterial spin labelling MRI, single photon-emission computed tomography (SPECT) and 2-[18F]-fluoro-2-deoxy-D-glucose (FDG)-positron-emission tomography (PET). Arterial spin labelling showed hyperperfusion overlapping the inflammatory lesions, whereas PET and SPECT disclosed increased perfusion and increased metabolism. The subtraction SPECT co-registered to MRI (SISCOM) demonstrated hypermetabolism outside the areas of encephalitis. After clinical recovery, follow-up MRI revealed the development of atrophy in the initially affected hippocampus. Two patients who had recurrent paraneoplastic autoimmune encephalitis manifesting as status epilepticus showed new T2 lesions involving different structures. CONCLUSION: The presence of limbic and extra-limbic T2 signal abnormalities in new-onset status epilepticus should suggest the diagnosis of a paraneoplastic syndrome, especially when status epilepticus is refractory to treatment. The lesions are consistently seen as hyperintense on T2WI. Ó 2014 The Royal College of Radiologists. Published by Elsevier Ltd. All rights reserved.
Introduction
n Univer* Guarantor and correspondent: S. Sarria-Estrada, Vall d’Hebro n, 119-129, 08035 Barcelona, Spain. sity Hospital, Passeig de la Vall d’Hebro Tel.: þ34 932746731. E-mail address: [email protected] (S. Sarria-Estrada).
Paraneoplastic autoimmune encephalitis (PE) is considered one of the classic paraneoplastic neurological syndromes (PNS), which are defined as neurological symptoms that are not associated with local or metastatic activity of a malignancy and sometimes supported by the presence of
0009-9260/$ e see front matter Ó 2014 The Royal College of Radiologists. Published by Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.crad.2014.03.012
Please cite this article in press as: Sarria-Estrada S, et al., Neuroimaging in status epilepticus secondary to paraneoplastic autoimmune encephalitis, Clinical Radiology (2014), http://dx.doi.org/10.1016/j.crad.2014.03.012
e2
S. Sarria-Estrada et al. / Clinical Radiology xxx (2014) e1ee9
specific onconeural antibodies.1 PE has been termed “limbic encephalitis”, as it is often confined to the limbic system, especially the hippocampus and temporal cortex. However, some reports have described involvement of extra-limbic structures, such as the basal ganglia, mesencephalon, or neocortex, constituting the so-called “extra-limbic encephalitis”.2 Despite relevant scientific advances, the pathophysiology of PE remains elusive. The diagnosis can be difficult because of the wide spectrum of clinical presentations of this condition. At onset, PE commonly appears as an insidious subacute cognitive-behavioural disorder, developing over a few days or up to 12 weeks.3 Seizures and epileptiform discharges have been described in nearly 29% of patients with PE.4 However, convulsive status epilepticus (SE) is less commonly associated with PE.2,5,6 Epilepsia partialis continua has been reported in 7% of patients with extra-limbic anti-Hu encephalitis.4 Other types of SE are rare in PE, and cases of non-convulsive SE are often misdiagnosed as confusional states.4,5 PNS can be classified as “definite” or “possible” by combinations of a set of criteria1 and both diagnoses require exclusion of other causes to explain the symptoms in a patient with malignant disease. The presence of wellcharacterized onconeural antibodies, even in the absence of a known tumour, can be used to classify the associated disorder as definite PNS.1,3 Magnetic resonance imaging (MRI) combined with 2(FDG)-positron-emission [18F]-fluoro-2-deoxy-D-glucose tomography (PET) and single photon-emission computed tomography (SPECT), are fundamental for detecting neurological complications of systemic cancer such as brain metastasis or carcinomatosis.7 Although MRI findings are not considered diagnostic for PE and normal findings are not uncommon, the patients in the present study with SE secondary to PE showed abnormal MRI. Hypersignal on T2weighted imaging (WI) involving various brain regions and indicating inflammation have been described and can support the clinical diagnosis, particularly in the absence of antibodies. Nonetheless, the rate at which these imaging abnormalities occur and their relationship with the clinical features of the condition are uncertain.2 The aim of the present study was to describe the neuroimaging findings in patients with SE secondary to PE.
Materials and methods The neuroimaging and clinical data of five consecutive n University Hospital, patients admitted to Vall d’Hebro Barcelona, Spain, with a diagnosis of SE secondary to PE between August 2005 and March 2011 were retrospectively reviewed. The study was approved by the hospital ethics committee. All patients had seizures captured using video-EEG (electroencephalography) monitoring, and all underwent contrast-enhanced brain CT at hospital admittance, brain MRI, and determination of anti-neural antibodies (anti-Hu,
anti-Ma2, anti-Yo, anti-CV2, anti-Ri, and anti-amphiphysin) in the cerebrospinal fluid (CSF). Subsequent MRI was carried out within 6 months of SE using a 1.5 T magnet in three patients and a 3 T system in one patient. Whole-brain study with 5 mm section thickness was obtained by using (1) transverse T2WI; (2) transverse and coronal fluid-attenuated inversion recovery (FLAIR); (3) diffusion-weighted imaging (DWI) with b ¼ 500 and 1000 s/mm2; and (4) unenhanced and contrastenhanced (0.1 mmol/kg body weight gadolinium-based contrast agent) transverse T1WI. A pulsed arterial spin labelling (ASL) sequence was also obtained during the 3 T imaging and was used to perform cerebral blood flow (CBF) mapping. All patients underwent nuclear medicine examination to detect the underlying malignancy or to stage disseminated disease. Three patients were also studied using SPECT or PET to characterize the nature of the brain lesions. SPECT was obtained with 740 MBq of Tc-99m ethyl cysteinate dimer, using a double-headed gamma camera with highresolution and low-energy parallel collimators. The FDGPET examinations were performed with a threedimensional mode in a single imaging session after injection of 2.5 MBq/kg of the radiotracer.
Results Five patients with SE with definite criteria for PNS were reviewed. Four patients were men, and the mean age was 59 years. The existence of underlying cancer was identified with body PET/CT in three of the five patients at the onset of PE, and the two remaining patients were diagnosed during follow-up. Four lung cancers and one colon cancer were identified (Table 1). At the time of the diagnosis, four patients were seen to have non-disseminated disease. Anti-Hu antibody was detected in the two patients with small cell lung carcinoma, which was discovered in one patient 4 years before the diagnosis of the primary tumour. In all cases, CSF analysis showed an inflammatory profile with high protein concentrations and low lymphocyte count. In four patients, the clinical onset manifested as nonspecific neuropsychiatric symptoms, such as mood lability, confusional state, and short-term memory impairment. The symptoms were insidious and had a waxing and waning course lasting from a few days to a month. Behavioural disorders accompanied different types of epileptic seizures in all cases. SE was confirmed by clinical findings in two patients and by video-EEG in three patients with non-convulsive SE. Two patients had focal clonic seizures (cases 1 and 3). Three had autonomic seizures with diaphoresis, tachycardia, and piloerection (cases 2 and 3, and second episode case 4). Ictal speech arrest occurred in four patients (cases 2e5). One patient had convulsive SE that required sedation (case 4). Multiple anti-epileptic drugs were needed to decrease the seizures. A significant improvement in epileptic activity was only achieved when patients received treatment with gammaglobulin,
Please cite this article in press as: Sarria-Estrada S, et al., Neuroimaging in status epilepticus secondary to paraneoplastic autoimmune encephalitis, Clinical Radiology (2014), http://dx.doi.org/10.1016/j.crad.2014.03.012
S. Sarria-Estrada et al. / Clinical Radiology xxx (2014) e1ee9
e3
Table 1 Patients’ demographics, clinical and systemic disease characteristics. Patient Tumour (age/sex)
Clinical onset and symptoms
Anti-neuronal Body PET-CT antibodies
Follow-up
1 (55/\)
Small-cell lung carcinoma
Anti-Hu
Initial: negative malignancy
Controlled with cyclosporine serial bolus Cancer diagnosed 4 years after PE onset Died a few months later
2 (77/_)
Mixed small-cell lung carcinoma and adenocarcinoma
Progressive memory, language, and writing impairment over a few months Occasional auditive illusions SE: simple partial seizures Duration: continuous Subacute confusional state, behaviour disorder of 2 days duration; auditive illusions; ictal speech arrest SE: complex partial seizure Duration: 1 week
Negative
Hypermetabolic node in the Short-term memory deficit upper left lung lobe. Negative Recurrence with abdominal lymph nodes and metastasis seizures (simple partial seizures) and behavioural disorder Died 1 year later due to brain metastasis Hypermetabolic lesion in the Good response to initial right lung lobe, invading the treatment with gammaglobulins pleura, bilateral lymph nodes, and chemotherapy and mediastinal metastases Recurrence of PE 4 months later, along with tumour enlargement Died 6 months after due to complex partial SE Not done
Short-term memory impairment, Anti-Hu fatigue, and weight loss during 1 month, followed by 2 days of confusion. Ictal speech arrest SE: simple partial seizures Duration: 3 weeks 30 2nd SE Asthenia and fatigue, acute confusional state SE: complex partial seizures/ secondary generalization Duration: 8 weeks 4 (60/_) Colorectal adeno- 1st SE Headache, visual illusions, and Negative carcinoma language impairment for a few days Ictal speech arrest SE: generalized toniceclonic seizure Duration: 24 h 40 (62/_) 2nd SE Intense fatigue and acute confusional state SE: complex partial seizure Duration: 2 weeks Negative 5 (57/_) Lung adenocarcinoma Progressive apathy, anorexia, (chronic mantel weight loss, aphasia, and cell lymphoma) hypersomnolence SE: complex partial seizure Duration: 1 week 3 (49/_)
Small-cell lung carcinoma
1st SE
Initial not done One-year follow-up PET-CT hypermetabolic infiltrative colon lesion. Negative lymph nodes or metastasis
Patient remained seizure-free, but had a chronic short-term memory deficit. After 5 years follow-up was in complete remission
Not done
Once the tumour was resected, the patient improved and remained seizure-free Died 2 years later due to systemic complications of background lymphoma
PET-CT, combined positron-emission tomography and computed tomography; SE, status epilepticus.
corticosteroids, and chemotherapy. At short-term followup, seizures were partially controlled with anti-epileptic medication in all patients. The seizure-freedom was achieved when patients received gammaglobulins and chemotherapy. At long-term follow-up, four patients remained seizure-free. The patient with epilepsia partialis continua had occasional simple partial hemifacial clonic seizures. Three patients had seizures or status epilepticus relapse associated with the progression of the malignancy. The memory deficit remained in all cases (Table 1). Four patients died, one as a result of SE, two secondary to late-onset brain metastasis, and one due to systemic complications of cancer (Table 1). The initial brain CT showed non-contrast enhanced hypodense lesions in three patients. In one patient, focal contrast-enhancing lesions were seen in the left insula and right hippocampus. The CT lesions correlated with areas of hypersignal on T2WI (Fig 1). All MRI examinations were performed during the SE episode or within 1 h after cessation of the electroclinical
features. T2-weighted sequences depicted a variable degree of hypersignal associated with increased volume affecting the hippocampus and amygdala unilaterally or bilaterally (Figs 2e4). Furthermore, three patients showed scattered extra-limbic areas of hypersignal affecting the cortical and subcortical regions of the parietal and frontal lobes (Fig 5). DWI showed hypersignal in the affected structures, associated with a corresponding mild to moderate signal increase in ADC (apparent diffusion coefficient), indicating that the signal changes corresponded to a T2 shine-through effect, probably reflecting extracellular oedema. These signal changes were well correlated to the electroclinical findings. In the non-convulsive SE patient in whom an ASL sequence was performed during the ictal phase, focal areas of increased relative CBF were demonstrated, matching the T2 signal abnormalities, which resolved once the patient recovered from SE (case 3; Figs 1 and 2). Moderate and heterogeneous enhancement in the hippocampus was identified on contrast-enhanced T1WI in
Please cite this article in press as: Sarria-Estrada S, et al., Neuroimaging in status epilepticus secondary to paraneoplastic autoimmune encephalitis, Clinical Radiology (2014), http://dx.doi.org/10.1016/j.crad.2014.03.012
e4
S. Sarria-Estrada et al. / Clinical Radiology xxx (2014) e1ee9
Figure 1 A 49-year-old man with anti-Hu antibody encephalitis (case 3, Table 2). Non-contrast and contrast-medium-enhanced axial brain CT (a,b) showed a small lesion in the left insula with mild contrast medium uptake (white arrow), which was hyperintense on FLAIR (c). CBF map obtained with ASL showed hyperperfusion overlapping the area of abnormality seen on both CT and MRI (d). DWI demonstrated mild hyperintensity of the lesion, likely corresponding to T2 shine-through effect (e), as it was also hyperintense on the ADC (f). Ictal (g) and interictal (h) SPECT, and SISCOM (i), showed propagated epileptic activity distant from the encephalitic areas in the left temporal and frontal lobes.
two patients, who had multiple limbic and extra-limbic lesions on T2WI (cases 3 and 4). Three of the four patients in whom at least one follow-up MRI examination within 6 months was obtained after SE, showed volume loss and hypersignal on T2WI in the previously affected hippocampi (Figs 2 and 4). No signs of cortical atrophy or laminar cortical necrosis were seen in the areas that had been affected on the initial MRI study obtained during SE (Table 2). SPECT and PET performed in the ictal phase in two patients revealed areas of increased metabolism in the
same location as the abnormal T2WI. The SISCOM in one patient showed increased perfusion that exceeded the hyperperfusion areas obtained with the ASL (Fig 1). Follow-up PET and SPECT studies in two patients demonstrated temporal lobe hypometabolism or hypoperfusion where follow-up MRI had shown hippocampal atrophy (Table 2). Recurrent PE manifesting as SE occurred in three cases at 6, 8, and 12 months after the first episode. MRI in these patients showed new T2 lesions with different topography as compared to the baseline study (Figs 4 and 5).
Please cite this article in press as: Sarria-Estrada S, et al., Neuroimaging in status epilepticus secondary to paraneoplastic autoimmune encephalitis, Clinical Radiology (2014), http://dx.doi.org/10.1016/j.crad.2014.03.012
S. Sarria-Estrada et al. / Clinical Radiology xxx (2014) e1ee9
e5
Figure 2 A 49-year-old man with anti-Hu antibody encephalitis associated with small-cell lung carcinoma (case 3, Table 2). During the first clinical episode of encephalitis (a, b, and c) MRI showed hyperintense foci in the right hippocampus and left superior frontal gyrus (a, white arrow) on coronal FLAIR. The lesion in the right hippocampus showed contrast medium uptake (b, black arrow), whereas hyperperfusion in both mesial temporal regions were seen on ASL (c). At MRI 6 months later, a recurrence showed hypersignal foci on FLAIR involving the left hippocampus (d, arrowhead) that demonstrated contrast medium uptake (e, curved arrow) and hyperperfusion on ASL (f). Note the development of right hippocampal atrophy as a consequence of the first encephalitic episode (d).
Discussion The neuroimaging findings in five patients with SE secondary to PE were reported. The clinical onset of PE was mainly characterized by neuropsychiatric symptoms and several types of epileptic seizures, in keeping with the clinical data from previous reports.4,5 The main imaging features were increased signal on T2WI involving limbic and extra-limbic areas, associated with increased perfusion and hypermetabolism, demonstrated by ASL, SPECT, and PET. PE is a rare entity that commonly manifests with neuropsychiatric symptoms occurring before, or at the same time as, the diagnosis of a malignant disease.1,3 The most common features are short-term memory impairment and mood disorders, developing over several days to a few months, as was seen in the present patients before the seizures occurred. In the context of PE, epilepsia partialis continua seems to be the most characteristic type of SE.2,4e6 Various ictal symptoms, such as speech arrest and autonomic signs, were observed more often than continuous focal motor activity. The fact that most of the present patients had subtle, nonspecific ictal symptoms suggests that the incidence of epileptic seizures in the course of PE may be underestimated, as has been proposed by Kaplan et al.8 In accordance with the literature,2,3 small cell lung cancer were found to be the malignant disease most commonly
associated with PE, although other types of neoplasms have also been reported, such as adenocarcinomas,9 which affected two of the present patients. As would be expected, SE was refractory to anti-epileptic drugs, and the outcome of epilepsy depended on the prognosis of the malignancy.3 Anti-neural antibodies are useful markers of an underlying cancer, although some patients with these antibodies never develop cancer after a follow-up of several years.1,2 PNS can occur without the presence of onconeural antibodies, and antibodies can occur without a neurological syndrome. Anti-Hu antibodies are detected in nearly 50% of patients with PE and small cell lung carcinoma.3 Likewise, anti-Hu antibodies have been detected in nearly half of PE cases with epileptic seizures.4 Anti-Hu associated with small cell lung carcinoma was seen in two of the present patients. Nonetheless, the three remaining patients did not have distinctive onconeural antibodies, and also experienced SE. The presence of inflammatory CSF findings supported the diagnosis in 100% of patients, a higher percentage than would be expected for PNS. This may be because SE, in itself, can induce similar CSF characteristics.1,10 Hypersignal in the limbic system on T2WI, which has been described in up to 70% of PE cases, was seen in all of the present patients with SE and PE. The less commonly reported extra-limbic involvement was observed only in the anti-Hu-positive patients.2,4,11 The hyperintense signal on T2WI lesions, involving both limbic and neocortical
Please cite this article in press as: Sarria-Estrada S, et al., Neuroimaging in status epilepticus secondary to paraneoplastic autoimmune encephalitis, Clinical Radiology (2014), http://dx.doi.org/10.1016/j.crad.2014.03.012
e6
S. Sarria-Estrada et al. / Clinical Radiology xxx (2014) e1ee9
Figure 3 A 57-year-old man with PE secondary to lung adenocarcinoma (case 5, Table 2). Unenhanced CT showed hypodensity in both hippocampi (a, arrows). Axial and coronal FLAIR (b and c) and coronal contrast-enhanced T1WI (d) demonstrated swelling and hypersignal of both hippocampi, but no contrast medium uptake (d).
structures, was not likely a consequence of prolonged epileptic activity. The non-restriction on DWI supports the concept that the T2WI abnormalities are related to extracellular oedema, which is not attributable to seizure, in which the characteristic phenomenon is intracellular oedema. These findings are not exclusively observed in this syndrome and other disorders must be always considered. Systemic metabolic disorders, posterior reversible encephalopathy, or non-paraneoplastic autoimmune encephalitis, are some disease entities that may share common neuroimaging findings with the PE, although they usually occur in a different clinical setting.12,13 The contrast enhancement seen at CT and MRI in two patients, reflected increased bloodebrain barrier (BBB) permeability as a consequence of inflammatory activity or prolonged SE, a feature that has been only sporadically described.2,3,10 At follow-up MRI, no residual lesions were identified in relation to the neocortical abnormalities, as would be
expected for prolonged focal SE or encephalitis. However, the mesial temporal structures showed hypometabolism, atrophy, and sclerosis after PE recovery, as a result of an inflammatory process involving the limbic system.14,15 In contrast to the results obtained in patients with chronic epilepsy, who commonly show normal or “cold” metabolic areas surrounding the epileptogenic source in SPECT/PET, interictal hyperperfusion or hypermetabolism was observed using SPECT/PET and hyperperfusion using ASL, matching the areas of increased T2 signal and reflecting the baseline inflammation.16 In the ictal period, the volume of the inflammatory lesions was largely exceeded by the hyperperfusion areas as seen at ASL or SPECT; this can be explained by spreading of the epileptic activity from the epileptogenic region to adjacent cerebral areas.17,18 In conclusion, in those patients who satisfy the criteria for PNS, the presence of single or scattered limbic and extralimbic T2WI hyperintense signal abnormalities at MRI with non-restricted diffusion and sometimes showing contrast
Please cite this article in press as: Sarria-Estrada S, et al., Neuroimaging in status epilepticus secondary to paraneoplastic autoimmune encephalitis, Clinical Radiology (2014), http://dx.doi.org/10.1016/j.crad.2014.03.012
S. Sarria-Estrada et al. / Clinical Radiology xxx (2014) e1ee9
e7
Figure 4 MRI after recurrence 12 months later (case 4, Table 2) showed enlargement and hypersignal of the right amygdala on axial and coronal FLAIR (a, c, arrows), with irregular contrast-medium uptake on contrast-enhanced T1WI (b). A follow-up coronal FLAIR obtained 1 year later demonstrated right hippocampal atrophy (d, dashed arrow).
Figure 5 A 60-year-old man with a first episode of PE associated with colorectal adenocarcinoma (case 4, Table 2). Axial FLAIR showed scattered areas of hypersignal affecting cortical and subcortical regions of the right brain hemisphere. Please cite this article in press as: Sarria-Estrada S, et al., Neuroimaging in status epilepticus secondary to paraneoplastic autoimmune encephalitis, Clinical Radiology (2014), http://dx.doi.org/10.1016/j.crad.2014.03.012
e8
S. Sarria-Estrada et al. / Clinical Radiology xxx (2014) e1ee9
Table 2 Neuroimaging features. Patient
CT cerebral
Brain MRI
Follow-up MRI
Cerebral FDG-PET/SPECT
1
Non-contrast-enhanced low-density area in the temporal lobes and parietal cortex
Recovered MRI lesions
FDG-PET and SPECT showed increased metabolism in the left superior temporal gyrus
2
Normal
Not done
Not done
3
Subtle contrast-enhanced low-density area in the left insular region
1st SE
Right hippocampal atrophy
Ictal SPECT showed increased perfusion in the same areas of increased perfusion on ASL Interictal SPECT showed hyperperfusion areas also Follow-up SPECT was normal
Subtle contrast enhancement in the right hippocampus
2nd SE
Diffuse hypodensity of right hemisphere
1st SE
Right hippocampal atrophy
Follow-up FDG-PET showed hypometabolism in the right temporal lobe
Hypodensity in the right hippocampus
2nd SE
Increased FLAIR/T2WI signal in both mesial temporal lobes and left superior temporal gyrus and adjacent parietal cortex DWI: negative Contrast enhancement: negative Increased FLAIR/T2WI signal of the hippocampus DWI: negative Contrast enhancement: negative Increased FLAIR/T2WI signal in both amygdalae and hippocampi, and the anterior temporal cortex, as well as in subcortical areas of both parietal lobes and left frontal lobe DWI: negative Contrast enhancement: Right hippocampus ASL: increased perfusion within the affected areas Increased FLAIR/T2WI signal in both hippocampi DWI: negative Contrast enhancement: moderate in right hippocampus ASL: focal hyperperfusion limited to the right hippocampus Increased FLAIR/T2WI signal in the right temporoparietal lobe DWI: negative Contrast enhancement: Negative. Increased FLAIR/T2WI signal in the right hippocampus DWI: negative Contrast enhancement: right hippocampus Increased FLAIR/T2WI signal in both hippocampi DWI: negative Contrast enhancement: negative
Bilateral hippocampal atrophy
Not done
4
5
Bilateral hippocampal hypodensity
CT, computed tomography; MRI, magnetic resonance imaging; FDG-PET, 2-[18F]-fluoro-2-deoxy-d-glucose positron-emission tomography; SPECT, single photon-emission computed tomography; FLAIR, fluid-attenuated inversion recovery; T2WI, T2-weighted imaging; SE, status epilepticus; DWI, diffusionweighted imaging; ASL, arterial spin labelling.
enhancement is likely to reflect extracellular oedema in patients with new-onset refractory SE secondary to PE. During the SE, hypermetabolism and hyperperfusion of the affected areas as observed by SPECT/PET or ASL can exceed the extent of the lesion on MRI because of propagation of the epileptic activity.
References 1. Graus F, Delattre JY, Antoine JC, et al. Recommended diagnostic criteria for paraneoplastic neurological syndromes. J Neurol Neurosurg Psychiatry 2004;75:1135e40. 2. Demaerel P, Van Dessel W, Van Paesschen W, et al. Autoimmunemediated encephalitis. Neuroradiology 2011;53:837e51. 3. Gultekin SH, Rosenfeld MR, Voltz R, et al. Paraneoplastic limbic encephalitis: neurological symptoms, immunological findings and tumour association in 50 patients. Brain 2000;123:1481e94. 4. Rudzinski LA, Pittock SJ, McKeon A, et al. Extratemporal EEG and MRI findings in ANNA-1 (anti-Hu) encephalitis. Epilepsy Res 2011;95:255e62.
5. Shavit YB, Graus F, Probst A, et al. Epilepsia partialis continua: a new manifestation of anti-Hu-associated paraneoplastic encephalomyelitis. Ann Neurol 1999;45:255e8. 6. Drislane FW. Nonconvulsive status epilepticus in patients with cancer. Clin Neurol Neurosurg 1994;96:314e8. 7. Titulaer MJ, Soffietti R, Dalmau J, et al. Screening for tumours in paraneoplastic syndromes: report of an EFNS task force. Eur J Neurol 2011;18. 19ee3. 8. Kaplan PW, Rossetti AO, Kaplan EH, et al. Proposition: limbic encephalitis may represent limbic status epilepticus. A review of clinical and EEG characteristics. Epilepsy Behav 2012;24:1e6. 9. Tsukamoto T, Mochizuki R, Mochizuki H, et al. Paraneoplastic cerebellar degeneration and limbic encephalitis in a patient with adenocarcinoma of the colon. J Neurol Neurosurg Psychiatr 1993;56:713e6. 10. Correale J, Rabinowicz AL, Heck CN, et al. Status epilepticus increases CSF levels of neuron-specific enolase and alters the bloodebrain barrier. Neurology 1998;50:1388e91. 11. Dalmau J, Rosenfeld MR. Paraneoplastic syndromes of the CNS. Lancet Neurol 2008;7:327e40. €rster A, Griebe M, Gass A, et al. Diffusion-weighted imaging for the 12. Fo differential diagnosis of disorders affecting the hippocampus. Cerebrovasc Dis 2012;33:104e15.
Please cite this article in press as: Sarria-Estrada S, et al., Neuroimaging in status epilepticus secondary to paraneoplastic autoimmune encephalitis, Clinical Radiology (2014), http://dx.doi.org/10.1016/j.crad.2014.03.012
S. Sarria-Estrada et al. / Clinical Radiology xxx (2014) e1ee9 € ster A, et al. Features of acute DWI ab13. Chatzonstantinou A, Gass A, Fo normalities related to status epilepticus. Epilepsy Res 2011;97:45e51. 14. Kaiboriboon K, Hogan RE. Hippocampal shape analysis in status epilepticus associated with acute encephalitis. AJNR Am J Neuroradiol 2002;23:1003e6. 15. Herrmann EK, Hahn K, Kratzer C, et al. Status epilepticus as a risk factor for postencephalitic parenchyma loss evaluated by ventricle brain ratio measurement on MR imaging. AJNR Am J Neuroradiol 2006;27:1245e51.
e9
16. Pendse N, Wissmeyer M, Altrichter S, et al. Interictal arterial spinlabeling MRI perfusion in intractable epilepsy. J Neuroradiol 2010;37:60e3. 17. Toledo M, Munuera J, Sueiras M, et al. MRI findings in aphasic status epilepticus. Epilepsia 2008;49:1465e9. 18. Toledo M, Munuera J, Salas-Puig X, et al. Localisation value of ictal arterial spin-labeled sequences in partial seizures. Epileptic Disord 2011;13:336e9.
Please cite this article in press as: Sarria-Estrada S, et al., Neuroimaging in status epilepticus secondary to paraneoplastic autoimmune encephalitis, Clinical Radiology (2014), http://dx.doi.org/10.1016/j.crad.2014.03.012
Contents lists available at ScienceDirect
Clinical Radiology journal homepage: www.clinicalradiologyonline.net
Neuroimaging in status epilepticus secondary to paraneoplastic autoimmune encephalitis s c, S. Sarria-Estrada a, *, M. Toledo b, C. Lorenzo-Bosquet c, G. Cuberas-Borro Rovira a C. Auger a, S. Siurana a, A. n University Hospital, Barcelona, Spain Magnetic Resonance Unit, Radiology Department, Vall d’Hebro n University Hospital, Barcelona, Spain Epilepsy Unit, Neurology Department, Vall d’Hebro c n University Hospital, Barcelona, Spain Nuclear Medicine Department, Vall d’Hebro a
b
art icl e i nformat ion Article history: Received 18 August 2013 Received in revised form 16 March 2014 Accepted 18 March 2014
AIM: To describe the characteristic magnetic resonance imaging (MRI) findings of paraneoplastic autoimmune encephalitis in patients with new-onset status epilepticus. MATERIALS AND METHODS: The neuroimaging and clinical data of five patients with paraneoplastic autoimmune encephalitis debuting as status epilepticus were retrospectively reviewed. All patients met the criteria for definite paraneoplastic syndrome and all underwent brain MRI during the status epilepticus episode or immediately after recovery. RESULTS: All patients showed hyperintense lesions on T2-weighted imaging (WI) involving the limbic structures, specifically the hippocampus. Three of them showed additional extralimbic areas of signal abnormalities. The areas of T2 hyperintensity were related to the electroclinical onset of the seizures. In three patients, various techniques were used to study cerebral perfusion, such as arterial spin labelling MRI, single photon-emission computed tomography (SPECT) and 2-[18F]-fluoro-2-deoxy-D-glucose (FDG)-positron-emission tomography (PET). Arterial spin labelling showed hyperperfusion overlapping the inflammatory lesions, whereas PET and SPECT disclosed increased perfusion and increased metabolism. The subtraction SPECT co-registered to MRI (SISCOM) demonstrated hypermetabolism outside the areas of encephalitis. After clinical recovery, follow-up MRI revealed the development of atrophy in the initially affected hippocampus. Two patients who had recurrent paraneoplastic autoimmune encephalitis manifesting as status epilepticus showed new T2 lesions involving different structures. CONCLUSION: The presence of limbic and extra-limbic T2 signal abnormalities in new-onset status epilepticus should suggest the diagnosis of a paraneoplastic syndrome, especially when status epilepticus is refractory to treatment. The lesions are consistently seen as hyperintense on T2WI. Ó 2014 The Royal College of Radiologists. Published by Elsevier Ltd. All rights reserved.
Introduction
n Univer* Guarantor and correspondent: S. Sarria-Estrada, Vall d’Hebro n, 119-129, 08035 Barcelona, Spain. sity Hospital, Passeig de la Vall d’Hebro Tel.: þ34 932746731. E-mail address: [email protected] (S. Sarria-Estrada).
Paraneoplastic autoimmune encephalitis (PE) is considered one of the classic paraneoplastic neurological syndromes (PNS), which are defined as neurological symptoms that are not associated with local or metastatic activity of a malignancy and sometimes supported by the presence of
0009-9260/$ e see front matter Ó 2014 The Royal College of Radiologists. Published by Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.crad.2014.03.012
Please cite this article in press as: Sarria-Estrada S, et al., Neuroimaging in status epilepticus secondary to paraneoplastic autoimmune encephalitis, Clinical Radiology (2014), http://dx.doi.org/10.1016/j.crad.2014.03.012
e2
S. Sarria-Estrada et al. / Clinical Radiology xxx (2014) e1ee9
specific onconeural antibodies.1 PE has been termed “limbic encephalitis”, as it is often confined to the limbic system, especially the hippocampus and temporal cortex. However, some reports have described involvement of extra-limbic structures, such as the basal ganglia, mesencephalon, or neocortex, constituting the so-called “extra-limbic encephalitis”.2 Despite relevant scientific advances, the pathophysiology of PE remains elusive. The diagnosis can be difficult because of the wide spectrum of clinical presentations of this condition. At onset, PE commonly appears as an insidious subacute cognitive-behavioural disorder, developing over a few days or up to 12 weeks.3 Seizures and epileptiform discharges have been described in nearly 29% of patients with PE.4 However, convulsive status epilepticus (SE) is less commonly associated with PE.2,5,6 Epilepsia partialis continua has been reported in 7% of patients with extra-limbic anti-Hu encephalitis.4 Other types of SE are rare in PE, and cases of non-convulsive SE are often misdiagnosed as confusional states.4,5 PNS can be classified as “definite” or “possible” by combinations of a set of criteria1 and both diagnoses require exclusion of other causes to explain the symptoms in a patient with malignant disease. The presence of wellcharacterized onconeural antibodies, even in the absence of a known tumour, can be used to classify the associated disorder as definite PNS.1,3 Magnetic resonance imaging (MRI) combined with 2(FDG)-positron-emission [18F]-fluoro-2-deoxy-D-glucose tomography (PET) and single photon-emission computed tomography (SPECT), are fundamental for detecting neurological complications of systemic cancer such as brain metastasis or carcinomatosis.7 Although MRI findings are not considered diagnostic for PE and normal findings are not uncommon, the patients in the present study with SE secondary to PE showed abnormal MRI. Hypersignal on T2weighted imaging (WI) involving various brain regions and indicating inflammation have been described and can support the clinical diagnosis, particularly in the absence of antibodies. Nonetheless, the rate at which these imaging abnormalities occur and their relationship with the clinical features of the condition are uncertain.2 The aim of the present study was to describe the neuroimaging findings in patients with SE secondary to PE.
Materials and methods The neuroimaging and clinical data of five consecutive n University Hospital, patients admitted to Vall d’Hebro Barcelona, Spain, with a diagnosis of SE secondary to PE between August 2005 and March 2011 were retrospectively reviewed. The study was approved by the hospital ethics committee. All patients had seizures captured using video-EEG (electroencephalography) monitoring, and all underwent contrast-enhanced brain CT at hospital admittance, brain MRI, and determination of anti-neural antibodies (anti-Hu,
anti-Ma2, anti-Yo, anti-CV2, anti-Ri, and anti-amphiphysin) in the cerebrospinal fluid (CSF). Subsequent MRI was carried out within 6 months of SE using a 1.5 T magnet in three patients and a 3 T system in one patient. Whole-brain study with 5 mm section thickness was obtained by using (1) transverse T2WI; (2) transverse and coronal fluid-attenuated inversion recovery (FLAIR); (3) diffusion-weighted imaging (DWI) with b ¼ 500 and 1000 s/mm2; and (4) unenhanced and contrastenhanced (0.1 mmol/kg body weight gadolinium-based contrast agent) transverse T1WI. A pulsed arterial spin labelling (ASL) sequence was also obtained during the 3 T imaging and was used to perform cerebral blood flow (CBF) mapping. All patients underwent nuclear medicine examination to detect the underlying malignancy or to stage disseminated disease. Three patients were also studied using SPECT or PET to characterize the nature of the brain lesions. SPECT was obtained with 740 MBq of Tc-99m ethyl cysteinate dimer, using a double-headed gamma camera with highresolution and low-energy parallel collimators. The FDGPET examinations were performed with a threedimensional mode in a single imaging session after injection of 2.5 MBq/kg of the radiotracer.
Results Five patients with SE with definite criteria for PNS were reviewed. Four patients were men, and the mean age was 59 years. The existence of underlying cancer was identified with body PET/CT in three of the five patients at the onset of PE, and the two remaining patients were diagnosed during follow-up. Four lung cancers and one colon cancer were identified (Table 1). At the time of the diagnosis, four patients were seen to have non-disseminated disease. Anti-Hu antibody was detected in the two patients with small cell lung carcinoma, which was discovered in one patient 4 years before the diagnosis of the primary tumour. In all cases, CSF analysis showed an inflammatory profile with high protein concentrations and low lymphocyte count. In four patients, the clinical onset manifested as nonspecific neuropsychiatric symptoms, such as mood lability, confusional state, and short-term memory impairment. The symptoms were insidious and had a waxing and waning course lasting from a few days to a month. Behavioural disorders accompanied different types of epileptic seizures in all cases. SE was confirmed by clinical findings in two patients and by video-EEG in three patients with non-convulsive SE. Two patients had focal clonic seizures (cases 1 and 3). Three had autonomic seizures with diaphoresis, tachycardia, and piloerection (cases 2 and 3, and second episode case 4). Ictal speech arrest occurred in four patients (cases 2e5). One patient had convulsive SE that required sedation (case 4). Multiple anti-epileptic drugs were needed to decrease the seizures. A significant improvement in epileptic activity was only achieved when patients received treatment with gammaglobulin,
Please cite this article in press as: Sarria-Estrada S, et al., Neuroimaging in status epilepticus secondary to paraneoplastic autoimmune encephalitis, Clinical Radiology (2014), http://dx.doi.org/10.1016/j.crad.2014.03.012
S. Sarria-Estrada et al. / Clinical Radiology xxx (2014) e1ee9
e3
Table 1 Patients’ demographics, clinical and systemic disease characteristics. Patient Tumour (age/sex)
Clinical onset and symptoms
Anti-neuronal Body PET-CT antibodies
Follow-up
1 (55/\)
Small-cell lung carcinoma
Anti-Hu
Initial: negative malignancy
Controlled with cyclosporine serial bolus Cancer diagnosed 4 years after PE onset Died a few months later
2 (77/_)
Mixed small-cell lung carcinoma and adenocarcinoma
Progressive memory, language, and writing impairment over a few months Occasional auditive illusions SE: simple partial seizures Duration: continuous Subacute confusional state, behaviour disorder of 2 days duration; auditive illusions; ictal speech arrest SE: complex partial seizure Duration: 1 week
Negative
Hypermetabolic node in the Short-term memory deficit upper left lung lobe. Negative Recurrence with abdominal lymph nodes and metastasis seizures (simple partial seizures) and behavioural disorder Died 1 year later due to brain metastasis Hypermetabolic lesion in the Good response to initial right lung lobe, invading the treatment with gammaglobulins pleura, bilateral lymph nodes, and chemotherapy and mediastinal metastases Recurrence of PE 4 months later, along with tumour enlargement Died 6 months after due to complex partial SE Not done
Short-term memory impairment, Anti-Hu fatigue, and weight loss during 1 month, followed by 2 days of confusion. Ictal speech arrest SE: simple partial seizures Duration: 3 weeks 30 2nd SE Asthenia and fatigue, acute confusional state SE: complex partial seizures/ secondary generalization Duration: 8 weeks 4 (60/_) Colorectal adeno- 1st SE Headache, visual illusions, and Negative carcinoma language impairment for a few days Ictal speech arrest SE: generalized toniceclonic seizure Duration: 24 h 40 (62/_) 2nd SE Intense fatigue and acute confusional state SE: complex partial seizure Duration: 2 weeks Negative 5 (57/_) Lung adenocarcinoma Progressive apathy, anorexia, (chronic mantel weight loss, aphasia, and cell lymphoma) hypersomnolence SE: complex partial seizure Duration: 1 week 3 (49/_)
Small-cell lung carcinoma
1st SE
Initial not done One-year follow-up PET-CT hypermetabolic infiltrative colon lesion. Negative lymph nodes or metastasis
Patient remained seizure-free, but had a chronic short-term memory deficit. After 5 years follow-up was in complete remission
Not done
Once the tumour was resected, the patient improved and remained seizure-free Died 2 years later due to systemic complications of background lymphoma
PET-CT, combined positron-emission tomography and computed tomography; SE, status epilepticus.
corticosteroids, and chemotherapy. At short-term followup, seizures were partially controlled with anti-epileptic medication in all patients. The seizure-freedom was achieved when patients received gammaglobulins and chemotherapy. At long-term follow-up, four patients remained seizure-free. The patient with epilepsia partialis continua had occasional simple partial hemifacial clonic seizures. Three patients had seizures or status epilepticus relapse associated with the progression of the malignancy. The memory deficit remained in all cases (Table 1). Four patients died, one as a result of SE, two secondary to late-onset brain metastasis, and one due to systemic complications of cancer (Table 1). The initial brain CT showed non-contrast enhanced hypodense lesions in three patients. In one patient, focal contrast-enhancing lesions were seen in the left insula and right hippocampus. The CT lesions correlated with areas of hypersignal on T2WI (Fig 1). All MRI examinations were performed during the SE episode or within 1 h after cessation of the electroclinical
features. T2-weighted sequences depicted a variable degree of hypersignal associated with increased volume affecting the hippocampus and amygdala unilaterally or bilaterally (Figs 2e4). Furthermore, three patients showed scattered extra-limbic areas of hypersignal affecting the cortical and subcortical regions of the parietal and frontal lobes (Fig 5). DWI showed hypersignal in the affected structures, associated with a corresponding mild to moderate signal increase in ADC (apparent diffusion coefficient), indicating that the signal changes corresponded to a T2 shine-through effect, probably reflecting extracellular oedema. These signal changes were well correlated to the electroclinical findings. In the non-convulsive SE patient in whom an ASL sequence was performed during the ictal phase, focal areas of increased relative CBF were demonstrated, matching the T2 signal abnormalities, which resolved once the patient recovered from SE (case 3; Figs 1 and 2). Moderate and heterogeneous enhancement in the hippocampus was identified on contrast-enhanced T1WI in
Please cite this article in press as: Sarria-Estrada S, et al., Neuroimaging in status epilepticus secondary to paraneoplastic autoimmune encephalitis, Clinical Radiology (2014), http://dx.doi.org/10.1016/j.crad.2014.03.012
e4
S. Sarria-Estrada et al. / Clinical Radiology xxx (2014) e1ee9
Figure 1 A 49-year-old man with anti-Hu antibody encephalitis (case 3, Table 2). Non-contrast and contrast-medium-enhanced axial brain CT (a,b) showed a small lesion in the left insula with mild contrast medium uptake (white arrow), which was hyperintense on FLAIR (c). CBF map obtained with ASL showed hyperperfusion overlapping the area of abnormality seen on both CT and MRI (d). DWI demonstrated mild hyperintensity of the lesion, likely corresponding to T2 shine-through effect (e), as it was also hyperintense on the ADC (f). Ictal (g) and interictal (h) SPECT, and SISCOM (i), showed propagated epileptic activity distant from the encephalitic areas in the left temporal and frontal lobes.
two patients, who had multiple limbic and extra-limbic lesions on T2WI (cases 3 and 4). Three of the four patients in whom at least one follow-up MRI examination within 6 months was obtained after SE, showed volume loss and hypersignal on T2WI in the previously affected hippocampi (Figs 2 and 4). No signs of cortical atrophy or laminar cortical necrosis were seen in the areas that had been affected on the initial MRI study obtained during SE (Table 2). SPECT and PET performed in the ictal phase in two patients revealed areas of increased metabolism in the
same location as the abnormal T2WI. The SISCOM in one patient showed increased perfusion that exceeded the hyperperfusion areas obtained with the ASL (Fig 1). Follow-up PET and SPECT studies in two patients demonstrated temporal lobe hypometabolism or hypoperfusion where follow-up MRI had shown hippocampal atrophy (Table 2). Recurrent PE manifesting as SE occurred in three cases at 6, 8, and 12 months after the first episode. MRI in these patients showed new T2 lesions with different topography as compared to the baseline study (Figs 4 and 5).
Please cite this article in press as: Sarria-Estrada S, et al., Neuroimaging in status epilepticus secondary to paraneoplastic autoimmune encephalitis, Clinical Radiology (2014), http://dx.doi.org/10.1016/j.crad.2014.03.012
S. Sarria-Estrada et al. / Clinical Radiology xxx (2014) e1ee9
e5
Figure 2 A 49-year-old man with anti-Hu antibody encephalitis associated with small-cell lung carcinoma (case 3, Table 2). During the first clinical episode of encephalitis (a, b, and c) MRI showed hyperintense foci in the right hippocampus and left superior frontal gyrus (a, white arrow) on coronal FLAIR. The lesion in the right hippocampus showed contrast medium uptake (b, black arrow), whereas hyperperfusion in both mesial temporal regions were seen on ASL (c). At MRI 6 months later, a recurrence showed hypersignal foci on FLAIR involving the left hippocampus (d, arrowhead) that demonstrated contrast medium uptake (e, curved arrow) and hyperperfusion on ASL (f). Note the development of right hippocampal atrophy as a consequence of the first encephalitic episode (d).
Discussion The neuroimaging findings in five patients with SE secondary to PE were reported. The clinical onset of PE was mainly characterized by neuropsychiatric symptoms and several types of epileptic seizures, in keeping with the clinical data from previous reports.4,5 The main imaging features were increased signal on T2WI involving limbic and extra-limbic areas, associated with increased perfusion and hypermetabolism, demonstrated by ASL, SPECT, and PET. PE is a rare entity that commonly manifests with neuropsychiatric symptoms occurring before, or at the same time as, the diagnosis of a malignant disease.1,3 The most common features are short-term memory impairment and mood disorders, developing over several days to a few months, as was seen in the present patients before the seizures occurred. In the context of PE, epilepsia partialis continua seems to be the most characteristic type of SE.2,4e6 Various ictal symptoms, such as speech arrest and autonomic signs, were observed more often than continuous focal motor activity. The fact that most of the present patients had subtle, nonspecific ictal symptoms suggests that the incidence of epileptic seizures in the course of PE may be underestimated, as has been proposed by Kaplan et al.8 In accordance with the literature,2,3 small cell lung cancer were found to be the malignant disease most commonly
associated with PE, although other types of neoplasms have also been reported, such as adenocarcinomas,9 which affected two of the present patients. As would be expected, SE was refractory to anti-epileptic drugs, and the outcome of epilepsy depended on the prognosis of the malignancy.3 Anti-neural antibodies are useful markers of an underlying cancer, although some patients with these antibodies never develop cancer after a follow-up of several years.1,2 PNS can occur without the presence of onconeural antibodies, and antibodies can occur without a neurological syndrome. Anti-Hu antibodies are detected in nearly 50% of patients with PE and small cell lung carcinoma.3 Likewise, anti-Hu antibodies have been detected in nearly half of PE cases with epileptic seizures.4 Anti-Hu associated with small cell lung carcinoma was seen in two of the present patients. Nonetheless, the three remaining patients did not have distinctive onconeural antibodies, and also experienced SE. The presence of inflammatory CSF findings supported the diagnosis in 100% of patients, a higher percentage than would be expected for PNS. This may be because SE, in itself, can induce similar CSF characteristics.1,10 Hypersignal in the limbic system on T2WI, which has been described in up to 70% of PE cases, was seen in all of the present patients with SE and PE. The less commonly reported extra-limbic involvement was observed only in the anti-Hu-positive patients.2,4,11 The hyperintense signal on T2WI lesions, involving both limbic and neocortical
Please cite this article in press as: Sarria-Estrada S, et al., Neuroimaging in status epilepticus secondary to paraneoplastic autoimmune encephalitis, Clinical Radiology (2014), http://dx.doi.org/10.1016/j.crad.2014.03.012
e6
S. Sarria-Estrada et al. / Clinical Radiology xxx (2014) e1ee9
Figure 3 A 57-year-old man with PE secondary to lung adenocarcinoma (case 5, Table 2). Unenhanced CT showed hypodensity in both hippocampi (a, arrows). Axial and coronal FLAIR (b and c) and coronal contrast-enhanced T1WI (d) demonstrated swelling and hypersignal of both hippocampi, but no contrast medium uptake (d).
structures, was not likely a consequence of prolonged epileptic activity. The non-restriction on DWI supports the concept that the T2WI abnormalities are related to extracellular oedema, which is not attributable to seizure, in which the characteristic phenomenon is intracellular oedema. These findings are not exclusively observed in this syndrome and other disorders must be always considered. Systemic metabolic disorders, posterior reversible encephalopathy, or non-paraneoplastic autoimmune encephalitis, are some disease entities that may share common neuroimaging findings with the PE, although they usually occur in a different clinical setting.12,13 The contrast enhancement seen at CT and MRI in two patients, reflected increased bloodebrain barrier (BBB) permeability as a consequence of inflammatory activity or prolonged SE, a feature that has been only sporadically described.2,3,10 At follow-up MRI, no residual lesions were identified in relation to the neocortical abnormalities, as would be
expected for prolonged focal SE or encephalitis. However, the mesial temporal structures showed hypometabolism, atrophy, and sclerosis after PE recovery, as a result of an inflammatory process involving the limbic system.14,15 In contrast to the results obtained in patients with chronic epilepsy, who commonly show normal or “cold” metabolic areas surrounding the epileptogenic source in SPECT/PET, interictal hyperperfusion or hypermetabolism was observed using SPECT/PET and hyperperfusion using ASL, matching the areas of increased T2 signal and reflecting the baseline inflammation.16 In the ictal period, the volume of the inflammatory lesions was largely exceeded by the hyperperfusion areas as seen at ASL or SPECT; this can be explained by spreading of the epileptic activity from the epileptogenic region to adjacent cerebral areas.17,18 In conclusion, in those patients who satisfy the criteria for PNS, the presence of single or scattered limbic and extralimbic T2WI hyperintense signal abnormalities at MRI with non-restricted diffusion and sometimes showing contrast
Please cite this article in press as: Sarria-Estrada S, et al., Neuroimaging in status epilepticus secondary to paraneoplastic autoimmune encephalitis, Clinical Radiology (2014), http://dx.doi.org/10.1016/j.crad.2014.03.012
S. Sarria-Estrada et al. / Clinical Radiology xxx (2014) e1ee9
e7
Figure 4 MRI after recurrence 12 months later (case 4, Table 2) showed enlargement and hypersignal of the right amygdala on axial and coronal FLAIR (a, c, arrows), with irregular contrast-medium uptake on contrast-enhanced T1WI (b). A follow-up coronal FLAIR obtained 1 year later demonstrated right hippocampal atrophy (d, dashed arrow).
Figure 5 A 60-year-old man with a first episode of PE associated with colorectal adenocarcinoma (case 4, Table 2). Axial FLAIR showed scattered areas of hypersignal affecting cortical and subcortical regions of the right brain hemisphere. Please cite this article in press as: Sarria-Estrada S, et al., Neuroimaging in status epilepticus secondary to paraneoplastic autoimmune encephalitis, Clinical Radiology (2014), http://dx.doi.org/10.1016/j.crad.2014.03.012
e8
S. Sarria-Estrada et al. / Clinical Radiology xxx (2014) e1ee9
Table 2 Neuroimaging features. Patient
CT cerebral
Brain MRI
Follow-up MRI
Cerebral FDG-PET/SPECT
1
Non-contrast-enhanced low-density area in the temporal lobes and parietal cortex
Recovered MRI lesions
FDG-PET and SPECT showed increased metabolism in the left superior temporal gyrus
2
Normal
Not done
Not done
3
Subtle contrast-enhanced low-density area in the left insular region
1st SE
Right hippocampal atrophy
Ictal SPECT showed increased perfusion in the same areas of increased perfusion on ASL Interictal SPECT showed hyperperfusion areas also Follow-up SPECT was normal
Subtle contrast enhancement in the right hippocampus
2nd SE
Diffuse hypodensity of right hemisphere
1st SE
Right hippocampal atrophy
Follow-up FDG-PET showed hypometabolism in the right temporal lobe
Hypodensity in the right hippocampus
2nd SE
Increased FLAIR/T2WI signal in both mesial temporal lobes and left superior temporal gyrus and adjacent parietal cortex DWI: negative Contrast enhancement: negative Increased FLAIR/T2WI signal of the hippocampus DWI: negative Contrast enhancement: negative Increased FLAIR/T2WI signal in both amygdalae and hippocampi, and the anterior temporal cortex, as well as in subcortical areas of both parietal lobes and left frontal lobe DWI: negative Contrast enhancement: Right hippocampus ASL: increased perfusion within the affected areas Increased FLAIR/T2WI signal in both hippocampi DWI: negative Contrast enhancement: moderate in right hippocampus ASL: focal hyperperfusion limited to the right hippocampus Increased FLAIR/T2WI signal in the right temporoparietal lobe DWI: negative Contrast enhancement: Negative. Increased FLAIR/T2WI signal in the right hippocampus DWI: negative Contrast enhancement: right hippocampus Increased FLAIR/T2WI signal in both hippocampi DWI: negative Contrast enhancement: negative
Bilateral hippocampal atrophy
Not done
4
5
Bilateral hippocampal hypodensity
CT, computed tomography; MRI, magnetic resonance imaging; FDG-PET, 2-[18F]-fluoro-2-deoxy-d-glucose positron-emission tomography; SPECT, single photon-emission computed tomography; FLAIR, fluid-attenuated inversion recovery; T2WI, T2-weighted imaging; SE, status epilepticus; DWI, diffusionweighted imaging; ASL, arterial spin labelling.
enhancement is likely to reflect extracellular oedema in patients with new-onset refractory SE secondary to PE. During the SE, hypermetabolism and hyperperfusion of the affected areas as observed by SPECT/PET or ASL can exceed the extent of the lesion on MRI because of propagation of the epileptic activity.
References 1. Graus F, Delattre JY, Antoine JC, et al. Recommended diagnostic criteria for paraneoplastic neurological syndromes. J Neurol Neurosurg Psychiatry 2004;75:1135e40. 2. Demaerel P, Van Dessel W, Van Paesschen W, et al. Autoimmunemediated encephalitis. Neuroradiology 2011;53:837e51. 3. Gultekin SH, Rosenfeld MR, Voltz R, et al. Paraneoplastic limbic encephalitis: neurological symptoms, immunological findings and tumour association in 50 patients. Brain 2000;123:1481e94. 4. Rudzinski LA, Pittock SJ, McKeon A, et al. Extratemporal EEG and MRI findings in ANNA-1 (anti-Hu) encephalitis. Epilepsy Res 2011;95:255e62.
5. Shavit YB, Graus F, Probst A, et al. Epilepsia partialis continua: a new manifestation of anti-Hu-associated paraneoplastic encephalomyelitis. Ann Neurol 1999;45:255e8. 6. Drislane FW. Nonconvulsive status epilepticus in patients with cancer. Clin Neurol Neurosurg 1994;96:314e8. 7. Titulaer MJ, Soffietti R, Dalmau J, et al. Screening for tumours in paraneoplastic syndromes: report of an EFNS task force. Eur J Neurol 2011;18. 19ee3. 8. Kaplan PW, Rossetti AO, Kaplan EH, et al. Proposition: limbic encephalitis may represent limbic status epilepticus. A review of clinical and EEG characteristics. Epilepsy Behav 2012;24:1e6. 9. Tsukamoto T, Mochizuki R, Mochizuki H, et al. Paraneoplastic cerebellar degeneration and limbic encephalitis in a patient with adenocarcinoma of the colon. J Neurol Neurosurg Psychiatr 1993;56:713e6. 10. Correale J, Rabinowicz AL, Heck CN, et al. Status epilepticus increases CSF levels of neuron-specific enolase and alters the bloodebrain barrier. Neurology 1998;50:1388e91. 11. Dalmau J, Rosenfeld MR. Paraneoplastic syndromes of the CNS. Lancet Neurol 2008;7:327e40. €rster A, Griebe M, Gass A, et al. Diffusion-weighted imaging for the 12. Fo differential diagnosis of disorders affecting the hippocampus. Cerebrovasc Dis 2012;33:104e15.
Please cite this article in press as: Sarria-Estrada S, et al., Neuroimaging in status epilepticus secondary to paraneoplastic autoimmune encephalitis, Clinical Radiology (2014), http://dx.doi.org/10.1016/j.crad.2014.03.012
S. Sarria-Estrada et al. / Clinical Radiology xxx (2014) e1ee9 € ster A, et al. Features of acute DWI ab13. Chatzonstantinou A, Gass A, Fo normalities related to status epilepticus. Epilepsy Res 2011;97:45e51. 14. Kaiboriboon K, Hogan RE. Hippocampal shape analysis in status epilepticus associated with acute encephalitis. AJNR Am J Neuroradiol 2002;23:1003e6. 15. Herrmann EK, Hahn K, Kratzer C, et al. Status epilepticus as a risk factor for postencephalitic parenchyma loss evaluated by ventricle brain ratio measurement on MR imaging. AJNR Am J Neuroradiol 2006;27:1245e51.
e9
16. Pendse N, Wissmeyer M, Altrichter S, et al. Interictal arterial spinlabeling MRI perfusion in intractable epilepsy. J Neuroradiol 2010;37:60e3. 17. Toledo M, Munuera J, Sueiras M, et al. MRI findings in aphasic status epilepticus. Epilepsia 2008;49:1465e9. 18. Toledo M, Munuera J, Salas-Puig X, et al. Localisation value of ictal arterial spin-labeled sequences in partial seizures. Epileptic Disord 2011;13:336e9.
Please cite this article in press as: Sarria-Estrada S, et al., Neuroimaging in status epilepticus secondary to paraneoplastic autoimmune encephalitis, Clinical Radiology (2014), http://dx.doi.org/10.1016/j.crad.2014.03.012