- Email: [email protected]
Cancer frequency detected by positron emission tomography-computed tomography in limbic encephalitis
Epilepsy & Behavior 89 (2018) 105–111
Contents lists available at ScienceDirect
Epilepsy & Behavior journal homepage: www.elsevier.com/locate/yebeh
Cancer frequency detected by positron emission tomography-computed tomography in limbic encephalitis Niels Hansen a,⁎,1, Guido Widman a,1, Svenja Stuff a, Albert J. Becker b, Juri-Alexander Witt a, Hojjat Ahmadzadehfar c, Christoph Helmstaedter a, Christian E. Elger a a b c
Department of Epileptology, University of Bonn, Sigmund-Freud-Straße 25, 53127 Bonn, Germany Department of Neuropathology, University of Bonn, Sigmund-Freud-Straße 25, 53127 Bonn, Germany Department of Nuclear Medicine, University of Bonn, Sigmund-Freud-Straße 25, 53127 Bonn, Germany
a r t i c l e
i n f o
Article history: Received 13 September 2018 Revised 27 September 2018 Accepted 27 September 2018 Available online xxxx Keywords: Temporal lobe epilepsy Limbic encephalitis Autoimmunity Cancer Neuronal autoantibodies
a b s t r a c t Objective: Paraneoplastic limbic encephalitis (LE) occurs frequently with considerable variability according to literature reports. We thus determined the cancer frequency in mixed LE subtypes sharing the diagnosis of temporal lobe epilepsy (TLE). Methods: All patients underwent magnetic resonance imaging (MRI) of the brain, electroencephalography (EEG) recordings, neuropsychological testing, immunohistochemistry, and clinical examination together with whole body 2-fluor-2-desoxy-D-glucose (FDG)-positron emission tomography (PET)/computed tomography (CT) to detect cancer in this observatory study. Results: Ninety-three patients (median: 52 years) with TLE due to autoimmune LE were investigated. Cancer was detected in the FDG-PET/CTs of 3 out of 93 (3.2%) patients with LE. Cancer was diagnosed upon, 5 years earlier and 5 years after FDG-PET/CT in 7 of 93 (7.5%) of all patients with LE. The cancer frequency in those patients was significantly lower than that reported in the largest series of patients with LE associated with and without different antibodies (7.5% vs. 23.5%, Bootstrap test, p b 0.05), but was indistinguishable from the estimated age-dependent cancer frequency in the German regional North-Rhine-Westfalian population without LE in 2014 (Chi-square test: p = 0.2). Conclusions: Our findings reveal that the cancer frequency in patients with TLE with LE detected by FDG-PET/CT is low and not different from the age-dependent natural cancer occurrence in a regional population. © 2018 Elsevier Inc. All rights reserved.
1. Introduction Cancer is one of the major causes of human mortality worldwide [1]. Its early diagnosis is therefore key to implementing effective treatment strategies. Neurological symptoms are a potential precursor or consequence of tumor-induced autoimmunity such as soAbbreviations: AMPAR, α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid receptor; Bq, Bequerel; CASPR2, contactin-associated protein-like 2; CSF, cerebrospinal fluid; CT, computed tomography; EEG, electroencephalography; FDG, 2-fluor-2-desoxyD-glucose; FLAIR, fluid attenuated inversion recovery images; GABAB, gammaaminobutyric acid B; GAD65, 67, glutamic acid decarboxylase antibodies with 65-kDa and 67-kDa isoforms; Hu, antinuclear neuronal antibody 1 (ANNA1); IgG, immunoglobulin G; LE, limbic encephalitis; LGI1, Leucine-rich glioma-inactivated 1; MRI, magnetic resonance imaging; NMDAR, N-methyl-D-aspartate receptor; PET, positron emission tomography; PB, peripheral blood; PNS, paraneoplastic neurological syndrome; TLE, temporal lobe epilepsy; Tr, delta notch-like epidermal growth factor related receptor (DNER); VGKC, voltage gated potassium channels. ⁎ Corresponding author. E-mail address: [email protected] (N. Hansen). 1 These authors contributed equally to this work.
https://doi.org/10.1016/j.yebeh.2018.09.043 1525-5050/© 2018 Elsevier Inc. All rights reserved.
called paraneoplastic neurological syndrome (PNS) [2]. One of these classic PNSs is limbic encephalitis (LE) [2], a disease entity characterized by temporal lobe seizures, memory deficits, or psychiatric abnormalities combined with signs of inflammation in mesiotemporal brain structures [3]. Tumor-triggered autoimmunity in LE is encountered often in patients with “well-characterized” paraneoplastic antibodies targeting intracellular antigens such as Hu- and Ma2/Ta-antibodies with cancer frequencies ranging from 94 to 100% in patient cohorts [4,5], but it is also seen in LE associated with antibodies against membrane surface antigens such as N-methyl-D-aspartate receptor (NMDAR)-antibodies with a lower tumor frequency (24–59%) in case series with autoimmune encephalitis [6,7]. The amount of NMDAR associated patients with autoimmune encephalitis with malignant tumors was even lower, amounting to about 6% [7]. Minor cancer frequencies are detected in patients with LE with voltage gated potassium channels (VGKC)-antibodies and their subgroups associated with LGI 1- and CASPR2-antibodies reaching frequencies of about 5% [8]. These observations suggest that the cancer frequency varies among LE subtypes. Data
106
N. Hansen et al. / Epilepsy & Behavior 89 (2018) 105–111
from cohorts with LE presenting various subtypes and containing seronegative patients with LE were pooled to increase the power because of an enlarged cohort. This procedure enabled us to reveal a moderate cancer frequency of 39% (Table 1A) [2,4,9,10]. Most of those study patients were recruited from neurology departments following referrals from clinics with patients with cancer. Hence, we wondered if the cancer frequency in patients with temporal lobe epilepsy (TLE) with mixed LE subtypes would be
distinct from that reported in other studies with mixed LE subtypes (39% pooled data from mixed studies and 23.5% from PNS Euronetwork data [2], (Table 1A). We therefore administered whole body 2-fluor-2-desoxy-D -glucose (FDG)-positron emission tomography (PET)/computed tomography (CT) in our patients to search for malignancies because of its high diagnostic accuracy, specificity, and sensitivity in localizing cancers in patients with a presumed paraneoplastic syndrome [11,12].
Table 1 Cancer frequency in LE from mixed studies (A) and characteristics of patients with LE (B). A: Cancer frequency of patients with LE from studies with different LE subtypes associated without and with neuronal antibodies Cancer LE/patients with LE (%)
Neuronal autoantibody subtypes
Cancer
Median age years (range)
References
23/98 (24)
Amphiphysin-, CV2/CRMP5Hu-, Ma2/Ta-, Ri, Tr-, VGKC-, Yo-, no abs
–
Giometto et al., 2010
50/79 (63)
Hu-, Ma2/Ta-, no abs
55 (11–75)
Gultekin et al., 2000
6/23 (26)
Hu-, Ma2/Ta-, VGKC abs
Breast, Colorectal, Eso/Gastro, HL, Kidney, Merkel, Neuroblastoma, NHL, NSCLC, Ovary, Prostate, SCLC, Testicular, Thymoma Breast, Colon, HL, Ovarian, SCLC, Teratoma, Testicular, Thymoma Rectal, SCLC, Testicular, Uterine Lyomyosarcoma Renal, SCLC, Thymoma
41 (23–66)
Soeder et al., 2009
4/16 (25) Hu-, NMDAR-, no abs 37 (15–69) Jagtap et al., 2014 Sum n = 83/216 (39%), median age: 41 years Abbreviations: abs = antibodies, AMPA = Amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid, CV2/CRMP5 = collapsing response mediator protein 5, Eso/Gastro = esophageal and gastrointestinal, GAD65 = glutamic acid decarboxylase 65-kDa isoform, HL = Hodgkin's lymphoma, mSCC = metastatic spindle cell carcinoma, NHL = non-Hodgkin's lymphoma, NMDAR = N-methyl-D-aspartate receptor, NSCLC = nonsmall cell lung cancer, SCLC = small cell lung ca-ncer, VGCC = voltage gated calcium channel, VGKC = voltage gated potassium channel, − = no data available. a. In the first column the first number refers to the number of patients with LE that had cancer in the referenced study, the second number after the diagonal slash refers to the number of patients with LE with and without cancer in the referenced study. The number in italics after the second number in the first column reflects the percentage of patients with LE that had cancer among all patients with LE with and without cancer in the respective study. b. After the bold word “sum n” the first appearing number means the sum of all patients with LE with cancer from the referenced studies and the second number after the diagonal slash means the sum of all patients with LE with and without cancer from the referenced studies. Thereafter, the numbers in brackets mean the percentage of the sum of all patients with LE and cancer from the sum of all patients with LE from the referenced studies. B: Clinical and laboratory characteristics of patients with LE Parameter at visit
Patients with LE with cancer
Patients with LE without cancer
χ2 U*
p-Level
Antibody+ patients with LE
Antibodypatients with LE
χ2 U*
p-Level
Clinical parameter Mean age (y) Gender: women Definitive autoimmune (%) Possible autoimmune (%) Mean No. AED Duration of epilepsy (y) Seizure frequency (m) Simple partial seizure (%) Complex partial seizure (%) Generalized seizure (%) Subjective memory disturbances (%) Subjective affective changes (%)
51.8 ± 16 3/7 (43) 0/7 (0) 7/7 (100) 1.66 ± 0.81 10.3 ± 6.7 5.5 ± 7.1 4/7 (57) 5/7 (71) 5/7 (71) 3/7 (43) 2/7 (29)
51.4 ± 16 35/86 (41) 20/86 (27) 66/86 (77) 1.40 ± 0.71 9.6 ± 9.1 16.5 ± 43.7 26/86 (30) 63/86 (74) 40/86 (47) 50/86 (58) 45/86 (52)
294* 0.01 2.07 2.07 195* 213* 146* 2.1 0.02 1.4 0.61 2.26
0.92 0.91 0.15 0.15 0.31 0.51 0.97 0.15 0.87 0.23 0.43 0.13
47.0 ± 15.5 17/40 (43) 7/40 (18) 33/40 (83) 1.5 ± 0.64 7.6 ± 4.2 11.4 ± 24.3 16/40 (40) 28/40 (70) 23/40 (58) 25/40 (63) 24/40 (60)
54.8 ± 15.5 22/53 (42) 13/53 (25) 40/53 (76) 1.21 ± 0.74 11.2 ± 11.1 19.1 ± 51.7 14/53 (26) 40/53 (75) 22/53 (42) 28/53 (53) 23/53 (43)
734.5* 0.01 0.67 0.66 798* 825* 669* 1.9 0.34 2.33 0.86 2.5
b0.05 0.92 0.41 0.41 0.08 0.13 0.49 0.16 0.55 0.12 0.35 0.11
4/6 (67) 0/6 (0)
46/86 (54) 25/86 (29)
0.39 2.39
0.53 0.14
19/40 (48) 9/40 (23)
31/52 (60) 16/52 (31)
1.1 0.68
0.29 0.41
3/7 (43) 4/7 (57)
68/86 (79) 26/86 (30)
4.7 2.14
b0.05 0.14
29/40 (73) 13/40 (33)
44/53 (83) 17/53 (32)
1.5 0.002
0.22 0.96
2/7 (29) 0/4 (0)
8/86 (9) 5/47 (11)
2.5 0.47
0.11 0.49
3/40 (8) 4/29 (14)
7/53 (13) 1/41 (2.4)
0.77 3.3
0.11 0.07
0/2 (0) 1/4 (25) 0/4 (50) 0/3 (0)
16/43 (37) 20/53 (38) 19/46 (41) 6/39 (15)
1.15 0.25 0.11 0.53
0.08 0.61 0.73 0.46
3/16 (19) 8/21 (38) 9/22 (41) 4/17 (24)
14/29 (48) 13/36 (36) 12/28 (43) 2/25 (8)
3.06 0.02 0.02 1.99
0.28 0.88 0.88 0.15
Laboratory parameter Mesiotemporal MRI+ Unilateral (%) Bilateral (%) Interictal temporal EEG Slowing (%) Epileptic activity (%) Ictal temporal EEG Epileptic activity (%) CSF Pleocytosis (%) Neuropsychological testing Impaired verbal memory (%) Impaired figural memory (%) Severely impaired verbal memory (%) Severely impaired figural memory (%)
Abbreviations: AED = antiepileptic drugs, FLAIR = fluid attenuated inversion recovery, m = month, mesiotemporal MRI+ = typical changes in T2/FLAIR weighted images in mesiotemporal MRI, NPT = neuropsychological testing, y = year; * = Mann Whitney U-test, U value instead of χ2 value is provided; for all other parameter a Chi-square test is used and χ2 is presented. The first number means the number of affected patients in the respective parameter separated by a diagonal slash. The second number means the number of (a) patients with LE and cancer in the first column or (b) number of patients with LE without cancer in the second column or (c) antibody+ patients with LE in the fifth column or (d) antibody- patients with LE in the sixth column with available data. The percentage means the number of affected patients in the given parameter of all patients with LE in the respective patient groups (columns 2, 3, 5 and 6).
N. Hansen et al. / Epilepsy & Behavior 89 (2018) 105–111
107
Fig. 1. Schematic study protocol (A) as well as cancer and neuronal autoantibody frequency of patients with LE (B). B: In (a) the cancer frequency in % of all patients with LE is depicted for different groups (antibody-negative vs. antibody-positive patients). (b) illustrates the frequency of the specific neuronal antibodies in % of all antibody-positive patients. Abbreviations: Ca+ = cancer positive, Ca- = cancer negative, CASPR2 = anti-contactin-associated protein-like 2, GAD65 = glutamic acid decarboxylase antibodies with the 65-kDa isoform, LGI1 = Leucine-rich glioma-inactivated 1, NMDA = N-methyl-D-aspartate receptor, VKGC = voltage gated potassium channel, y = year / years.
2. Material and methods 2.1. Patients and study design The patients in this observational, retrospective study were recruited between 2003 and 2015 at our Department of Epileptology. Within that time period, after applying the criteria for definitive and possible autoimmune LE [3], we ultimately enrolled 93 patients [women n = 39; median: 52 (14–81) years] with TLE due to LE from 200 patients with suspected LE, all of whom underwent an FDG-PET/ CT for cancer detection during their clinical workup (Fig. 1A). None of the investigated patients was diagnosed with a tumor at the time of FDG-PET/CT. However, a cancer was diagnosed in three patients 1 to 3 years before FDG-PET/CT. The FDG-PET/CT was part of routine clinical procedure for detecting cancer in patients with suspected LE. Definitive LE [3] was diagnosed if the following criteria (i–iv) were met: (i) a subacute memory deficit or clinical features of temporal lobe seizures or psychiatric symptoms, (ii) bilateral temporomesial signal abnormalities in magnetic resonance imaging (MRI) (for details see section Brain MRI), (iii) electroencephalographic evidence of slowing or epileptic
activity in the temporal lobe or a pleocytosis (N5/μl) in the cerebrospinal fluid (CSF), and (iv) other alternative causes had been carefully excluded (for detailed exclusion criteria see [3]). Possible LE [3] was assumed if criteria (i–iii) were fulfilled such as (i) memory or psychiatric disturbances, (ii) temporal lobe seizures or typical temporomesial signal abnormalities in MRI, or pleocytosis in CSF together with (iii) no fulfilled exclusion criterion. In addition, antibodies were determined in all patients with LE in the CSF and peripheral blood (PB) (section Immunopathology). For diagnosing purposes, patients underwent an electroencephalography (EEG) and neuropsychological testing. Each patient's EEG was recorded according to the international 10–20 EEG convention including additional T1/T2 electrodes [10–10: FT9 (T1), FT10 (T2)] to better identify the patients with TLE (System Plus evolution, Micromed S.p.A., Treviso, Italy). Neuropsychological investigations were done in conjunction with their clinical examination (section Neuropsychology). Whole body FDG-PET/CT scan was carried out in all patients at the Department of Nuclear Medicine at the University of Bonn to search for an underlying cancer according to the guidelines for paraneoplastic syndromes [12]. All patients provided informed consent prior to the clinical investigations. The study was approved by our
108
N. Hansen et al. / Epilepsy & Behavior 89 (2018) 105–111
local University of Bonn ethics committee and concurred with the Declaration of Helsinki. 2.2. Brain MRI Brain neuroimaging was done in all patients using a 3-T MRI (Magnetom Trio, Siemens, Germany at the Life and Brain Institute and/or Philips Medical Systems, at the Department of Neuroradiology, Bonn, Germany) to identify MRI signal alterations typical of LE. Individual patients' MRIs were inspected to see if typical LE criteria were fulfilled such as swelling and increased signal intensity of mesiotemporal brain structures in fluid attenuated inversion recovery (FLAIR) images and/or T2-weighted images. 2.3. Neuropsychology These patients underwent neuropsychological testing to detect any memory loss, to diagnose LE, and to phenotype them. To test verbal memory, we applied the “Verbaler Lern- und Merkfähigkeitstest” [13], and to assess figural memory capacity the revised version of the “Diagnosticum für Cerebralschädigung” was utilized [14]. To assess any deviation of figural and verbal memory performance from the mean, we used a calculated memory sum score of individual figural and verbal memory parameters (mean value of standardized memory parameters = 100 ± 10). We rated figural or verbal memory function as impaired if the memory score was below one standard deviation of the standardized mean value (b90). A severely impaired memory function was defined as a figural or verbal memory score lying below more than the twofold standard deviation of the standardized mean value (b80). 2.4. Immunopathology We used indirect immunohistochemistry to identify IgG neuronal autoantibodies in the PB or CSF. We searched for the following antibodies against cell membrane proteins: anti-α-amino-3-hydroxy-5methyl-4-isoxazolepropionic acid receptor (AMPAR) [containing the subunits AMPA1 and AMPA2], anti-aquaporin 4, anti-CASPR2, antiDPXX-6, anti-gamma-aminobutyric acid B (GABAB), anti-glycin and anti-LGI1, as well as NMDAR antibodies. We also determined paraneoplastic antibodies such as anti-amphiphysin, anti-Hu, anti-CV2 (CRMP5), anti-Ma-2/Ta (PNMA2), anti-Ri (ANNA-2), anti-Recoverin, anti-SOX1, anti-Titin, anti-Tr, anti-Yo (PCA-1), anti-Zic 4 and other intracellular antibodies such as anti-GAD65 and -67. In addition, a recent study has demonstrated that the detection of VGKC antibodies in patients is a questionable marker of autoimmunity [15], therefore it was not used as an additional criterion when applying Graus' criteria to our patients. 2.5. Cancer detection via positron emission tomography/computed tomography fusion imaging Whole body PET/CT fusion imaging (PET/CT, as the most accurate cancer-detection method in patients with paraneoplastic syndromes [12], was performed in the Department of Nuclear Medicine, University of Bonn to search for cancer. All imaging studies were carried out with a dual-modality PET/CT system (Biograph2; Siemens Medical Solutions Inc., Hoffman Estates, Illinois, USA). Positron emission tomography imaging started 80 ± 20 min after an intravenous injection of 370 ± 40 MBq FDG through an anterior cubital vein. Blood glucose measured before FDG injection was b 150 mg/dl in all cases. Whole body PET was acquired from the base of the skull to the mid femora at 5 min per bed position with the arms extended upwards. Diagnostic CT imaging was performed within 1 min before PET imaging with the patient in precisely the same position. One liter of an iodinated oral contrast agent was applied within 1 h before CT imaging for better delineation of
intestinal structures. An intravenous infusion of contrast medium (120 ml of Ultravist-300) was given at a dose rate of 2.5 ml/s for 50 s. Immediately after the infusion of contrast media, the whole-body CT data were acquired. A limited breath-hold technique was used for CT and shallow breathing for PET imaging to avoid motion-induced artifacts in the diaphragm area.
2.6. Statistics Statistics were performed by SigmaStat (Version 11, Systat Software GmbH, Erkrath, Germany) and the statistical program R (Version 3.4.3, R Foundation for Statistical Computing, Vienna, Austria). Normal data distribution was assessed via the Shapiro–Wilk test. Nonnormally distributed data of clinical and laboratory parameters were compared between antibody-positive and -negative patients as well as between patients with LE with and without cancer by the Mann–Whitney U test. Frequencies of clinical and laboratory parameters were analyzed between the above-mentioned groups by the Chi-square test. We compared the observed age-dependent frequency of cancer in our patients with LE with the mean estimated age-dependent frequency of all cancers in the North-Rhine-Westfalian population adapted to our patient cohort with LE via the Chi-square test [age-dependent data (5-year steps) of all patients with cancers mixed across genders are used from the cancer register of North Rhine-Westfalia [16] during 2014 and from a statistical report on the population of North Rhine-Westfalia in 2014 [17]. Pearson's Chi-squared test and Fisher's exact test were utilized to compare the relationship between neuronal autoantibodies and cancer occurrence in patients with LE. In addition, we used the bootstrap method to compare the cancer frequency we detected with that reported in the literature. A p-level of b 0.5 was considered as significant. 2.7. Data statement Data are available. 3. Results 3.1. Clinical and laboratory characteristics of patients with LE We identified 20 patients [median age: 53 (25–75) years] with definitive LE and 73 patients [median age: 51 (14–81) years] with possible autoimmune LE. We compared distinct groups of patients with LE with (n = 7) and without cancer (n = 86), as well as those with antibodies [n = 40: n = 10 (n = 10 PB, n = 2 CSF) anti-GAD65-, n = 10 (n = 10 PB) anti-VGKC-, n = 8 (n = 8 PB, n = 2 CSF) anti-NMDAR-, n = 5 (n = 5 PB, n = 1 CSF) anti-Ma2/Ta-, n = 3 (n = 3 PB, n = 2 CSF) CASPR2-, n = 3 (n = 3 PB), anti-LGI1-, n = 2 (n = 1 PB, n = 2 CSF) anti-Yo-, n = 2 (n = 2 PB) anti-Hu-, n = 1 (n = 1 PB and CSF) antiSOX1-, n = 1 (n = 1 PB) anti-CV2/CRMP5-, n = 1 (n = 1 PB) antiZic4-, n = 1 (n = 1 PB) anti-Glycin-antibodies (Fig. 1B)] and without (n = 53) antibodies to evaluate differences in clinical features and instrumental results for the early recognition of patients with LE who might suffer from malignancies. Neuronal antibodies were detected in 7 of 20 (35.7%) with definitive LE (n = 3 VGKC antibodies PB, n = 1 CASPR2 antibodies PB and CSF, n = 1 Ma/Ta2 antibodies PB, n = 1 LGI1 antibodies PB, n = 1 GAD65 and VGKC antibodies PB). However, we observed no differences in clinical and laboratory parameters between groups (Table 1B) with the exception that antibody-positive patients with LE were younger than antibody-negative patients with LE (Mann–Whitney U test: U = 734.5, p b 0.05; Table 1B). Moreover, temporal slowing in interictal EEG was observed more often in patients with LE without cancer than in those with (Chi-square test: χ2 = 4.7, p b 0.05, Table 1B).
N. Hansen et al. / Epilepsy & Behavior 89 (2018) 105–111 Table 2 Type of cancer (A) and estimated vs. observed cancer frequency in patients with LE (B). A: Cancer in patients with LE at visit ± 5 years No. of patient
Cancer
Autoantibody
Cancer detection
1 Acinar cell Ma2/Ta -abs 5 years prior to PET 2 Testicular Ma2/Ta -abs 3 years prior to PET 3 Prostate VGKC - abs 1 year prior to PET 4 Small cell lung Hu - abs upon PET 5 Ovarial Yo - abs upon PET 6 Bronchial No abs upon PET 7 Prostate Zic - 4 abs 1 year after PET Abbreviations: Abs = antibodies, No. = numbers, PET = positron emission tomography, VGKC = voltage gated potassium channels. B: Estimated vs. observed cancer frequency in all and antibody-positive patients with LE a) All patients with LE Age group
0–4 5–9 10–14 15–19 20–24 25–29 30–34 35–39 40–44 45–49 50–54 55–59 60–64 65–69 70–74 75–79 80–84 85+
Estimated
Observed
No. of patients
No. of patients
0 0 0.00012 0.00024 0.00076 0.00647 0.00675 0.00733 0.01693 0.05449 0.06123 0.07986 0.16268 0.20703 0.29297 0.06285 0.03440 0
0 0 0 1 0 0 0 1 0 0 1 0 3 1 0 0 0 0
b) Antibody-positive patients with LE Age group
0–4 5–9 10–14 15–19 20–24 25–29 30–34 35–39 40–44 45–49 50–54 55–59 60–64 65–69 70–74 75–79 80–84 85+
Estimated
Observed
No. of patients
No. of patients
0 0 0 0.00024 0.00076 0.00431 0.00113 0.00367 0.01129 0.03179 0.03402 0.02995 0.02957 0.08281 0.02663 0.06285 0 0
0 0 0 1 0 0 0 1 0 0 1 0 3 0 0 0 0 0
3.2. Cancer frequency in patients with LE A malignancy was detected via FDG-PET/CT in 3 of 93 (3.2%) patients with LE. No cancer was detected in patients with definitive LE associated with neuronal antibodies. In addition, four cancers were detected ±5 years before and after FDG-PET/CT examination, thus a cancer was diagnosed in 7 of 93 (7.5%) patients altogether [1 antibody-negative in 53 (1.89%) antibody-negative patients with LE and 6 antibody-positive in 40 (15.0%) antibody-positive patients with LE revealed a cancer, Fig. 1Ba]. We diagnosed acinar cell, ovarial, prostate, small cell lung, and testicular cancer in 5 of 12 (42%) patients with paraneoplastic antibodies
109
(anti-CV2/CRMP5, anti-Hu, anti-Ma-2/Ta, anti-SOX1, anti-Yo, antiZic4; Table 2A; Fig. 1Bb) prior to, upon, and after the FDG-PET/CT. Furthermore, prostate cancer was detected in one patient with VGKC-antibodies before his FDG-PET/CT (Table 2A). The FDG-PET/CT revealed bronchial cancer in another patient presenting with no antibodies (Table 2A). Antibody-positive patients with LE suffered significantly more from cancer diagnoses than antibody-negative patients with LE (Chi-Square test: χ2 = 5.6, p b 0.05, Fisher's exact test: p b 0.05; Fig. 1Ba). A positive antibody result implied a 9-fold higher risk for cancer. To estimate the difference in our paraneoplastic fraction of patients with LE with that of published data [2], we generated random samples from our sample of patients with LE. Not a single sample yielded a paraneoplastic LE fraction of 23.5%, demonstrating that the cancer frequency of about 23.5% in patients with LE from the PNS trial data [2] is significantly higher than the cancer frequency (7.5%) that we observed in all our patients with LE (Bootstrap test: p b 0.05; Fig. 2Aa). Moreover, the age-dependent cancer frequency observed in all and antibody-positive patients with LE did not differ from the estimated cancer frequency in Germany's North Rhine-Westfalia regional population during 2014 (Chi-square, all patients with LE: χ2 = 36, p = 0.2, Table 2Ba; antibody-positive patients with LE: χ2 = 36, p N 0.09, Table 2Bb; age distribution of all in Fig. 2Ba and antibody-positive patients with LE in Fig. 2Bb), indicating that our cohort's various LE subtypes do not have a higher risk of developing cancer than does the regional population. However, the cancer frequency (15.0%) of the antibody-positive patients with LE did not significantly differ from the 23.5% cancer frequency reported from the literature [2] (Bootstrap test: p = 0.14; Fig. 2Ab). 4. Discussion Our main findings provide evidence of a cancer frequency in our patients with TLE with LE detected by FDG-PET/CT that does not differ from the age-dependent cancer distribution pattern in the regional population in Germany's North Rhine-Westfalia in 2014. Furthermore, the cancer frequency of our patients with LE is significantly less (7.5%) than that of patients with LE with similarly diverse antibodies (23.5%) derived from a large cohort of patients with PNS [2] and that of pooled data from studies with heterogeneous LE subtypes [83/216 patients (39%, Table 1A)]. The low cancer rate in our patients with LE can be attributed to the following: (1) the composition of LE subtypes in our cohort with a large group of antibody-negative patients with LE and NMDAR-, VGKC-, and GAD65-antibodies-positive patients with LE, and (2) low referral of patients with cancer to our tertiary epilepsy center. Our selective cohort of patients with TLE might be a major reason for the few malignancies that we detected compared to the heterogeneous cohort of patients with PNS [2]. A further determining factor (3) for the low cancer rate in our patients with LE is the FDG-PET/CT's diagnostic drawback in identifying specific tumor types because of its inadequate sensitivity for detecting those tumors. Such tumors comprise, for example, lesser pelvis tumors containing certain variants of testicular tumors such as nonseminoma [18], subtypes of ovarian cancer (e.g., small or necrotic ones) [19] or cervical carcinoma [20]. If such a tumor is suspected clinically, additional investigations such as ultrasound and sometimes MRI in conjunction with urological and gynecological examinations are commendable in patients with PNS and thus patients with LE [12]. On the other hand, FDG-PET/CT is not recommended for diagnosing teratomas in PNS due to its poor detection sensitivity for these tumors [12]. This situation predominantly concerns patients with LE with the NMDAR antibodies often associated with ovarian teratoma [7]. The failure of FDG-PET/CT to detect teratomas might have led to the low rate of malignancies in our [0 of 8 (0%)] patients with NMDAR antibodies compared to reported rates (6%) [7]. Thus, further methods such as ultrasound, MRI, or CT thorax are advisable for cancer detection in patients with LE with NMDAR antibodies [12].
110
N. Hansen et al. / Epilepsy & Behavior 89 (2018) 105–111
Fig. 2. Fraction of paraneoplastic patients with LE from our and published data (A) and age distribution of patients with LE (B) A: Our sample revealed 7.5% who were paraneoplastic patients with LE. By generating samples from our patient cohort via the bootstrap method, we demonstrate that the reported frequency of 23.5% derived from the paraneoplastic neurological syndromes (PNS) Euronetwork clinical trial (Giometto et al., 2010) differs significantly from our fraction of paraneoplastic patients with LE (Bootstrap test: p b 0.05, Fig. 2Aa). However, the fraction of paraneoplastic antibody-positive patients with LE (15.0%) from our data did not differ from the reported cancer frequency of 23.5% (Bootstrap test, p = 0.14; Fig. 2Ab). The red lines indicate the 95% confidence interval of our calculated samples. The blue line indicates our fraction of paraneoplastic patients with LE (7.53%) in (a) and our fraction of paraneoplastic antibody-positive patients with LE (15.0%) in (b). The green lines in (a) and (b) denote the cancer frequency value from the literature (Giometto et al., 2010). B: The age distribution of all (Fig. 2Ba) and of antibody-positive (Fig. 2Bb) patients with LE is shown by age steps of 5 years. Abbreviation: no = number. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
Higher reported cancer frequencies are often associated with paraneoplastic LE subtypes such as those with Hu-antibodies [4], Zic4, or Yo-antibodies [2], concurring with our results. Surprisingly, we detected a moderate cancer frequency of 40% in patients with LE with Ma2/Ta antibodies (compared to a higher cancer frequency of 83% in patients with LE with Ma2/Ta antibodies reported by [5]), which supports the unexpectedly low cancer frequency in a few of our patients with TLE with typical paraneoplastic antibodies. However, we should keep in mind that another factor (4) is potentially responsible for the moderate cancer detection rate (42%) in our 12 of 93 (13%) patients with LE with paraneoplastic antibodies, namely the fact that the lack of cancer evidence in FDG-PET/CT does not rule out a paraneoplastic origin of LE, as the tumor-driven immunity might have abolished the tumor or reduced its growth or metastatic spread in these patients. Because of the small number of our patients with TLE with LE and the fact that FDG-PET/CT is not performed to search for cancer every 6 months because of PNS guidelines [12], we cannot assume clinical implications for diagnostic approaches. As not all patients underwent repeated investigations to detect cancer, the low number of cancers detected in our patients with LE might be due to occult cancers promptly eliminated by tumor-immunity control mechanisms.
frequencies being overestimated as a result of FDG-PET/CT methodology, as not all patients presumed to have LE are likely to have undergone FDG-PET/CT. It is more likely that our data reflect those patients who were already presenting suspicious clinical symptoms who were then referred for FDG-PET/CT to detect a malignancy. 4.2. Conclusions Taken together, our results reveal a relatively low cancer frequency detected via FDG-PET/CT in patients with TLE with LE compared to reports in the literature. However, our data do not reveal a clearcut phenotype of patients with LE that should be screened for cancer beyond the existing autoantibody profile. Acknowledgments We thank Denis Breuer for doing parts of the statistical analysis. Funding This study was supported by the “Verein zur Förderung der Epilepsieforschung, Bonn”.
4.1. Study limitations Declaration of interests Our strict criterion of including only patients with TLE with LE and completed FDG-PET/CT examination may have led to the cancer
None.
N. Hansen et al. / Epilepsy & Behavior 89 (2018) 105–111
References [1] Naghavi M, Abajobir AA, Abbafati C, Abbas KM, Abd-Allah F, Abera SF, et al. Global, regional, and national age-sex specific mortality for 264 causes of death, 1980– 2016: a systematic analysis for the Global Burden of Disease Study 2016. Lancet 2017;390:1151–210. [2] Giometto B, Grisold W, Vitaliani R, Graus F, Honnorat J, Bertolini G, et al. PNS Euronetwork. Paraneoplastic neurologic syndrome in the PNS Euronetwork database: a European study from 20 centers. Arch Neurol 2010;67:330–5. [3] Graus F, Titulaer MJ, Balu R, Benseler S, Bien CG, Cellucci T, et al. A clinical approach to diagnosis of autoimmune encephalitis. Lancet Neurol 2016;15:391–404. [4] Gultekin SH, Rosenfeld MR, Voltz R, Eichen J, Posner JB, Dalmau J. Paraneoplastic limbic encephalitis: neurological symptoms, immunological findings and tumour association in 50 patients. Brain 2000;123:1481–94. [5] Dalmau J, Graus F, Villarejo A, Posner JB, Blumenthal D, Thiessen B, et al. Clinical analysis of anti-Ma2-associated encephalitis. Brain 2004;127:1831–44. [6] Dalmau J, Gleichman AJ, Hughes EG, Rossi JE, Peng X, Lai M, et al. Anti-NMDA-receptor encephalitis: case series and analysis of the effects of antibodies. Lancet Neurol 2008;7:1091–8. [7] Bost C, Chanson E, Picard G, Meyronet D, Ducray F, Rogemond V, et al. Malignant tumors in autoimmune encephalitis with anti-NMDA receptor antibodies. J Neurol 2018. https://doi.org/10.1007/s00415-018-8970-0. [8] Dalmau J, Graus F. Antibody-mediated encephalitis. N Engl J Med 2018;378:840–51. [9] Soeder BM, Gleissner U, Urbach H, Clusmann H, Elger CE, Vincent A, et al. Causes, presentation and outcome of lesional adult onset mediotemporal lobe epilepsy. J Neurol Neurosurg Psychiatry 2009;80:894–9. [10] Jagtap SA, Das GK, Kambale HJ, Radhakrishnan A, Nair MD. Limbic encephalitis: clinical spectrum and long-term outcome from a developing country perspective. Ann Indian Acad Neurol 2014;17:161–5. [11] Sheikhbahaei S, Marcus CV, Fragomeni RS, Rowe SP, Javadi MS, Solnes LB. Wholebody 18F-FDG PET and 18F-FDG PET/CT in patients with suspected paraneoplastic
[12]
[13] [14]
[15]
[16]
[17]
[18]
[19] [20]
111
syndrome: a systematic review and meta-analysis of diagnostic accuracy. J Nucl Med 2017;58:1031–6. Titulaer MJ, Soffietti R, Dalmau J, Gilhus NE, Giometto B, Graus F, et al. Screening for tumours in paraneoplastic syndromes: report of an EFNS Task Force. Eur J Neurol 2011;18;19–e3. Helmstaedter C, Lendt M, Lux S, editors. VLMT: Verbaler Lern-und Merkfähigkeitstest. Testhandbuch. Göttingen: Hofgrefe; 2000. Helmstaedter C, Pohl C, Elger CE. Eine modifizierte Version des Diagnostikums für Cerebralschäden (DCS) zur Diagnostik räumlich-visueller Gedächtnisdefizite bei Patienten mit Temporallappenepilepsie. In: Scheffner D, editor. Epilepsie 90. Reinbek: Einhorn Presse; 1991. p. 272–9. Yeo T, Chen Z, Yong KP, Wong PYW, Chai JYH, Tan K. Distinction between antiVGKC-complex seropositive patients with and without anti-LGI1/CASPR2 antibodies. J Neurol Sci 2018;391:64–71. Landeskrebsregister NRW Germany, Bösartige Neubildungen gesamt (C00-C96), status of data base: 31. October 2016. http://www.krebsregister-nrw.de, Accessed date: 28 February 2018. Information und Technik Nordrhein-Westfalen, Geschäftsbereich Statistik. Statistische Berichte: Bevölkerung in Nordrhein-Westfalen 2014 nach Alter und Geschlecht. Fortschreibung des Bevölkerungsstandes auf Basis des Zensus vom 09.05.2011, Bestell-Nr. A133 2014 00, (Kennziffer A I – j /14). http://www.it.nrw. de; 2016, Accessed date: 28 February 2018. Ambrosini V, Zucchini G, Nicolini S, Berselli A, Nanni C, Allegri V, et al. 18F-FDG PET/ CT impact on testicular tumours clinical management. Eur J Nucl Med Mol Imaging 2014;41:668–73. Son H, Khan SM, Rahaman J, Cameron KL, Prasad-Hayes M, Chuang L, et al. Role of FDG PET/CT in staging of recurrent ovarian cancer. Radiographics 2011;31:569–83. Driscoll DO, Halpenny D, Johnston C, Sheehy N, Keogan M. 18F-FDG-PET/CT is of limited value in primary staging of early stage cervical cancer. Abdom Imaging 2015;40: 127–33.
Contents lists available at ScienceDirect
Epilepsy & Behavior journal homepage: www.elsevier.com/locate/yebeh
Cancer frequency detected by positron emission tomography-computed tomography in limbic encephalitis Niels Hansen a,⁎,1, Guido Widman a,1, Svenja Stuff a, Albert J. Becker b, Juri-Alexander Witt a, Hojjat Ahmadzadehfar c, Christoph Helmstaedter a, Christian E. Elger a a b c
Department of Epileptology, University of Bonn, Sigmund-Freud-Straße 25, 53127 Bonn, Germany Department of Neuropathology, University of Bonn, Sigmund-Freud-Straße 25, 53127 Bonn, Germany Department of Nuclear Medicine, University of Bonn, Sigmund-Freud-Straße 25, 53127 Bonn, Germany
a r t i c l e
i n f o
Article history: Received 13 September 2018 Revised 27 September 2018 Accepted 27 September 2018 Available online xxxx Keywords: Temporal lobe epilepsy Limbic encephalitis Autoimmunity Cancer Neuronal autoantibodies
a b s t r a c t Objective: Paraneoplastic limbic encephalitis (LE) occurs frequently with considerable variability according to literature reports. We thus determined the cancer frequency in mixed LE subtypes sharing the diagnosis of temporal lobe epilepsy (TLE). Methods: All patients underwent magnetic resonance imaging (MRI) of the brain, electroencephalography (EEG) recordings, neuropsychological testing, immunohistochemistry, and clinical examination together with whole body 2-fluor-2-desoxy-D-glucose (FDG)-positron emission tomography (PET)/computed tomography (CT) to detect cancer in this observatory study. Results: Ninety-three patients (median: 52 years) with TLE due to autoimmune LE were investigated. Cancer was detected in the FDG-PET/CTs of 3 out of 93 (3.2%) patients with LE. Cancer was diagnosed upon, 5 years earlier and 5 years after FDG-PET/CT in 7 of 93 (7.5%) of all patients with LE. The cancer frequency in those patients was significantly lower than that reported in the largest series of patients with LE associated with and without different antibodies (7.5% vs. 23.5%, Bootstrap test, p b 0.05), but was indistinguishable from the estimated age-dependent cancer frequency in the German regional North-Rhine-Westfalian population without LE in 2014 (Chi-square test: p = 0.2). Conclusions: Our findings reveal that the cancer frequency in patients with TLE with LE detected by FDG-PET/CT is low and not different from the age-dependent natural cancer occurrence in a regional population. © 2018 Elsevier Inc. All rights reserved.
1. Introduction Cancer is one of the major causes of human mortality worldwide [1]. Its early diagnosis is therefore key to implementing effective treatment strategies. Neurological symptoms are a potential precursor or consequence of tumor-induced autoimmunity such as soAbbreviations: AMPAR, α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid receptor; Bq, Bequerel; CASPR2, contactin-associated protein-like 2; CSF, cerebrospinal fluid; CT, computed tomography; EEG, electroencephalography; FDG, 2-fluor-2-desoxyD-glucose; FLAIR, fluid attenuated inversion recovery images; GABAB, gammaaminobutyric acid B; GAD65, 67, glutamic acid decarboxylase antibodies with 65-kDa and 67-kDa isoforms; Hu, antinuclear neuronal antibody 1 (ANNA1); IgG, immunoglobulin G; LE, limbic encephalitis; LGI1, Leucine-rich glioma-inactivated 1; MRI, magnetic resonance imaging; NMDAR, N-methyl-D-aspartate receptor; PET, positron emission tomography; PB, peripheral blood; PNS, paraneoplastic neurological syndrome; TLE, temporal lobe epilepsy; Tr, delta notch-like epidermal growth factor related receptor (DNER); VGKC, voltage gated potassium channels. ⁎ Corresponding author. E-mail address: [email protected] (N. Hansen). 1 These authors contributed equally to this work.
https://doi.org/10.1016/j.yebeh.2018.09.043 1525-5050/© 2018 Elsevier Inc. All rights reserved.
called paraneoplastic neurological syndrome (PNS) [2]. One of these classic PNSs is limbic encephalitis (LE) [2], a disease entity characterized by temporal lobe seizures, memory deficits, or psychiatric abnormalities combined with signs of inflammation in mesiotemporal brain structures [3]. Tumor-triggered autoimmunity in LE is encountered often in patients with “well-characterized” paraneoplastic antibodies targeting intracellular antigens such as Hu- and Ma2/Ta-antibodies with cancer frequencies ranging from 94 to 100% in patient cohorts [4,5], but it is also seen in LE associated with antibodies against membrane surface antigens such as N-methyl-D-aspartate receptor (NMDAR)-antibodies with a lower tumor frequency (24–59%) in case series with autoimmune encephalitis [6,7]. The amount of NMDAR associated patients with autoimmune encephalitis with malignant tumors was even lower, amounting to about 6% [7]. Minor cancer frequencies are detected in patients with LE with voltage gated potassium channels (VGKC)-antibodies and their subgroups associated with LGI 1- and CASPR2-antibodies reaching frequencies of about 5% [8]. These observations suggest that the cancer frequency varies among LE subtypes. Data
106
N. Hansen et al. / Epilepsy & Behavior 89 (2018) 105–111
from cohorts with LE presenting various subtypes and containing seronegative patients with LE were pooled to increase the power because of an enlarged cohort. This procedure enabled us to reveal a moderate cancer frequency of 39% (Table 1A) [2,4,9,10]. Most of those study patients were recruited from neurology departments following referrals from clinics with patients with cancer. Hence, we wondered if the cancer frequency in patients with temporal lobe epilepsy (TLE) with mixed LE subtypes would be
distinct from that reported in other studies with mixed LE subtypes (39% pooled data from mixed studies and 23.5% from PNS Euronetwork data [2], (Table 1A). We therefore administered whole body 2-fluor-2-desoxy-D -glucose (FDG)-positron emission tomography (PET)/computed tomography (CT) in our patients to search for malignancies because of its high diagnostic accuracy, specificity, and sensitivity in localizing cancers in patients with a presumed paraneoplastic syndrome [11,12].
Table 1 Cancer frequency in LE from mixed studies (A) and characteristics of patients with LE (B). A: Cancer frequency of patients with LE from studies with different LE subtypes associated without and with neuronal antibodies Cancer LE/patients with LE (%)
Neuronal autoantibody subtypes
Cancer
Median age years (range)
References
23/98 (24)
Amphiphysin-, CV2/CRMP5Hu-, Ma2/Ta-, Ri, Tr-, VGKC-, Yo-, no abs
–
Giometto et al., 2010
50/79 (63)
Hu-, Ma2/Ta-, no abs
55 (11–75)
Gultekin et al., 2000
6/23 (26)
Hu-, Ma2/Ta-, VGKC abs
Breast, Colorectal, Eso/Gastro, HL, Kidney, Merkel, Neuroblastoma, NHL, NSCLC, Ovary, Prostate, SCLC, Testicular, Thymoma Breast, Colon, HL, Ovarian, SCLC, Teratoma, Testicular, Thymoma Rectal, SCLC, Testicular, Uterine Lyomyosarcoma Renal, SCLC, Thymoma
41 (23–66)
Soeder et al., 2009
4/16 (25) Hu-, NMDAR-, no abs 37 (15–69) Jagtap et al., 2014 Sum n = 83/216 (39%), median age: 41 years Abbreviations: abs = antibodies, AMPA = Amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid, CV2/CRMP5 = collapsing response mediator protein 5, Eso/Gastro = esophageal and gastrointestinal, GAD65 = glutamic acid decarboxylase 65-kDa isoform, HL = Hodgkin's lymphoma, mSCC = metastatic spindle cell carcinoma, NHL = non-Hodgkin's lymphoma, NMDAR = N-methyl-D-aspartate receptor, NSCLC = nonsmall cell lung cancer, SCLC = small cell lung ca-ncer, VGCC = voltage gated calcium channel, VGKC = voltage gated potassium channel, − = no data available. a. In the first column the first number refers to the number of patients with LE that had cancer in the referenced study, the second number after the diagonal slash refers to the number of patients with LE with and without cancer in the referenced study. The number in italics after the second number in the first column reflects the percentage of patients with LE that had cancer among all patients with LE with and without cancer in the respective study. b. After the bold word “sum n” the first appearing number means the sum of all patients with LE with cancer from the referenced studies and the second number after the diagonal slash means the sum of all patients with LE with and without cancer from the referenced studies. Thereafter, the numbers in brackets mean the percentage of the sum of all patients with LE and cancer from the sum of all patients with LE from the referenced studies. B: Clinical and laboratory characteristics of patients with LE Parameter at visit
Patients with LE with cancer
Patients with LE without cancer
χ2 U*
p-Level
Antibody+ patients with LE
Antibodypatients with LE
χ2 U*
p-Level
Clinical parameter Mean age (y) Gender: women Definitive autoimmune (%) Possible autoimmune (%) Mean No. AED Duration of epilepsy (y) Seizure frequency (m) Simple partial seizure (%) Complex partial seizure (%) Generalized seizure (%) Subjective memory disturbances (%) Subjective affective changes (%)
51.8 ± 16 3/7 (43) 0/7 (0) 7/7 (100) 1.66 ± 0.81 10.3 ± 6.7 5.5 ± 7.1 4/7 (57) 5/7 (71) 5/7 (71) 3/7 (43) 2/7 (29)
51.4 ± 16 35/86 (41) 20/86 (27) 66/86 (77) 1.40 ± 0.71 9.6 ± 9.1 16.5 ± 43.7 26/86 (30) 63/86 (74) 40/86 (47) 50/86 (58) 45/86 (52)
294* 0.01 2.07 2.07 195* 213* 146* 2.1 0.02 1.4 0.61 2.26
0.92 0.91 0.15 0.15 0.31 0.51 0.97 0.15 0.87 0.23 0.43 0.13
47.0 ± 15.5 17/40 (43) 7/40 (18) 33/40 (83) 1.5 ± 0.64 7.6 ± 4.2 11.4 ± 24.3 16/40 (40) 28/40 (70) 23/40 (58) 25/40 (63) 24/40 (60)
54.8 ± 15.5 22/53 (42) 13/53 (25) 40/53 (76) 1.21 ± 0.74 11.2 ± 11.1 19.1 ± 51.7 14/53 (26) 40/53 (75) 22/53 (42) 28/53 (53) 23/53 (43)
734.5* 0.01 0.67 0.66 798* 825* 669* 1.9 0.34 2.33 0.86 2.5
b0.05 0.92 0.41 0.41 0.08 0.13 0.49 0.16 0.55 0.12 0.35 0.11
4/6 (67) 0/6 (0)
46/86 (54) 25/86 (29)
0.39 2.39
0.53 0.14
19/40 (48) 9/40 (23)
31/52 (60) 16/52 (31)
1.1 0.68
0.29 0.41
3/7 (43) 4/7 (57)
68/86 (79) 26/86 (30)
4.7 2.14
b0.05 0.14
29/40 (73) 13/40 (33)
44/53 (83) 17/53 (32)
1.5 0.002
0.22 0.96
2/7 (29) 0/4 (0)
8/86 (9) 5/47 (11)
2.5 0.47
0.11 0.49
3/40 (8) 4/29 (14)
7/53 (13) 1/41 (2.4)
0.77 3.3
0.11 0.07
0/2 (0) 1/4 (25) 0/4 (50) 0/3 (0)
16/43 (37) 20/53 (38) 19/46 (41) 6/39 (15)
1.15 0.25 0.11 0.53
0.08 0.61 0.73 0.46
3/16 (19) 8/21 (38) 9/22 (41) 4/17 (24)
14/29 (48) 13/36 (36) 12/28 (43) 2/25 (8)
3.06 0.02 0.02 1.99
0.28 0.88 0.88 0.15
Laboratory parameter Mesiotemporal MRI+ Unilateral (%) Bilateral (%) Interictal temporal EEG Slowing (%) Epileptic activity (%) Ictal temporal EEG Epileptic activity (%) CSF Pleocytosis (%) Neuropsychological testing Impaired verbal memory (%) Impaired figural memory (%) Severely impaired verbal memory (%) Severely impaired figural memory (%)
Abbreviations: AED = antiepileptic drugs, FLAIR = fluid attenuated inversion recovery, m = month, mesiotemporal MRI+ = typical changes in T2/FLAIR weighted images in mesiotemporal MRI, NPT = neuropsychological testing, y = year; * = Mann Whitney U-test, U value instead of χ2 value is provided; for all other parameter a Chi-square test is used and χ2 is presented. The first number means the number of affected patients in the respective parameter separated by a diagonal slash. The second number means the number of (a) patients with LE and cancer in the first column or (b) number of patients with LE without cancer in the second column or (c) antibody+ patients with LE in the fifth column or (d) antibody- patients with LE in the sixth column with available data. The percentage means the number of affected patients in the given parameter of all patients with LE in the respective patient groups (columns 2, 3, 5 and 6).
N. Hansen et al. / Epilepsy & Behavior 89 (2018) 105–111
107
Fig. 1. Schematic study protocol (A) as well as cancer and neuronal autoantibody frequency of patients with LE (B). B: In (a) the cancer frequency in % of all patients with LE is depicted for different groups (antibody-negative vs. antibody-positive patients). (b) illustrates the frequency of the specific neuronal antibodies in % of all antibody-positive patients. Abbreviations: Ca+ = cancer positive, Ca- = cancer negative, CASPR2 = anti-contactin-associated protein-like 2, GAD65 = glutamic acid decarboxylase antibodies with the 65-kDa isoform, LGI1 = Leucine-rich glioma-inactivated 1, NMDA = N-methyl-D-aspartate receptor, VKGC = voltage gated potassium channel, y = year / years.
2. Material and methods 2.1. Patients and study design The patients in this observational, retrospective study were recruited between 2003 and 2015 at our Department of Epileptology. Within that time period, after applying the criteria for definitive and possible autoimmune LE [3], we ultimately enrolled 93 patients [women n = 39; median: 52 (14–81) years] with TLE due to LE from 200 patients with suspected LE, all of whom underwent an FDG-PET/ CT for cancer detection during their clinical workup (Fig. 1A). None of the investigated patients was diagnosed with a tumor at the time of FDG-PET/CT. However, a cancer was diagnosed in three patients 1 to 3 years before FDG-PET/CT. The FDG-PET/CT was part of routine clinical procedure for detecting cancer in patients with suspected LE. Definitive LE [3] was diagnosed if the following criteria (i–iv) were met: (i) a subacute memory deficit or clinical features of temporal lobe seizures or psychiatric symptoms, (ii) bilateral temporomesial signal abnormalities in magnetic resonance imaging (MRI) (for details see section Brain MRI), (iii) electroencephalographic evidence of slowing or epileptic
activity in the temporal lobe or a pleocytosis (N5/μl) in the cerebrospinal fluid (CSF), and (iv) other alternative causes had been carefully excluded (for detailed exclusion criteria see [3]). Possible LE [3] was assumed if criteria (i–iii) were fulfilled such as (i) memory or psychiatric disturbances, (ii) temporal lobe seizures or typical temporomesial signal abnormalities in MRI, or pleocytosis in CSF together with (iii) no fulfilled exclusion criterion. In addition, antibodies were determined in all patients with LE in the CSF and peripheral blood (PB) (section Immunopathology). For diagnosing purposes, patients underwent an electroencephalography (EEG) and neuropsychological testing. Each patient's EEG was recorded according to the international 10–20 EEG convention including additional T1/T2 electrodes [10–10: FT9 (T1), FT10 (T2)] to better identify the patients with TLE (System Plus evolution, Micromed S.p.A., Treviso, Italy). Neuropsychological investigations were done in conjunction with their clinical examination (section Neuropsychology). Whole body FDG-PET/CT scan was carried out in all patients at the Department of Nuclear Medicine at the University of Bonn to search for an underlying cancer according to the guidelines for paraneoplastic syndromes [12]. All patients provided informed consent prior to the clinical investigations. The study was approved by our
108
N. Hansen et al. / Epilepsy & Behavior 89 (2018) 105–111
local University of Bonn ethics committee and concurred with the Declaration of Helsinki. 2.2. Brain MRI Brain neuroimaging was done in all patients using a 3-T MRI (Magnetom Trio, Siemens, Germany at the Life and Brain Institute and/or Philips Medical Systems, at the Department of Neuroradiology, Bonn, Germany) to identify MRI signal alterations typical of LE. Individual patients' MRIs were inspected to see if typical LE criteria were fulfilled such as swelling and increased signal intensity of mesiotemporal brain structures in fluid attenuated inversion recovery (FLAIR) images and/or T2-weighted images. 2.3. Neuropsychology These patients underwent neuropsychological testing to detect any memory loss, to diagnose LE, and to phenotype them. To test verbal memory, we applied the “Verbaler Lern- und Merkfähigkeitstest” [13], and to assess figural memory capacity the revised version of the “Diagnosticum für Cerebralschädigung” was utilized [14]. To assess any deviation of figural and verbal memory performance from the mean, we used a calculated memory sum score of individual figural and verbal memory parameters (mean value of standardized memory parameters = 100 ± 10). We rated figural or verbal memory function as impaired if the memory score was below one standard deviation of the standardized mean value (b90). A severely impaired memory function was defined as a figural or verbal memory score lying below more than the twofold standard deviation of the standardized mean value (b80). 2.4. Immunopathology We used indirect immunohistochemistry to identify IgG neuronal autoantibodies in the PB or CSF. We searched for the following antibodies against cell membrane proteins: anti-α-amino-3-hydroxy-5methyl-4-isoxazolepropionic acid receptor (AMPAR) [containing the subunits AMPA1 and AMPA2], anti-aquaporin 4, anti-CASPR2, antiDPXX-6, anti-gamma-aminobutyric acid B (GABAB), anti-glycin and anti-LGI1, as well as NMDAR antibodies. We also determined paraneoplastic antibodies such as anti-amphiphysin, anti-Hu, anti-CV2 (CRMP5), anti-Ma-2/Ta (PNMA2), anti-Ri (ANNA-2), anti-Recoverin, anti-SOX1, anti-Titin, anti-Tr, anti-Yo (PCA-1), anti-Zic 4 and other intracellular antibodies such as anti-GAD65 and -67. In addition, a recent study has demonstrated that the detection of VGKC antibodies in patients is a questionable marker of autoimmunity [15], therefore it was not used as an additional criterion when applying Graus' criteria to our patients. 2.5. Cancer detection via positron emission tomography/computed tomography fusion imaging Whole body PET/CT fusion imaging (PET/CT, as the most accurate cancer-detection method in patients with paraneoplastic syndromes [12], was performed in the Department of Nuclear Medicine, University of Bonn to search for cancer. All imaging studies were carried out with a dual-modality PET/CT system (Biograph2; Siemens Medical Solutions Inc., Hoffman Estates, Illinois, USA). Positron emission tomography imaging started 80 ± 20 min after an intravenous injection of 370 ± 40 MBq FDG through an anterior cubital vein. Blood glucose measured before FDG injection was b 150 mg/dl in all cases. Whole body PET was acquired from the base of the skull to the mid femora at 5 min per bed position with the arms extended upwards. Diagnostic CT imaging was performed within 1 min before PET imaging with the patient in precisely the same position. One liter of an iodinated oral contrast agent was applied within 1 h before CT imaging for better delineation of
intestinal structures. An intravenous infusion of contrast medium (120 ml of Ultravist-300) was given at a dose rate of 2.5 ml/s for 50 s. Immediately after the infusion of contrast media, the whole-body CT data were acquired. A limited breath-hold technique was used for CT and shallow breathing for PET imaging to avoid motion-induced artifacts in the diaphragm area.
2.6. Statistics Statistics were performed by SigmaStat (Version 11, Systat Software GmbH, Erkrath, Germany) and the statistical program R (Version 3.4.3, R Foundation for Statistical Computing, Vienna, Austria). Normal data distribution was assessed via the Shapiro–Wilk test. Nonnormally distributed data of clinical and laboratory parameters were compared between antibody-positive and -negative patients as well as between patients with LE with and without cancer by the Mann–Whitney U test. Frequencies of clinical and laboratory parameters were analyzed between the above-mentioned groups by the Chi-square test. We compared the observed age-dependent frequency of cancer in our patients with LE with the mean estimated age-dependent frequency of all cancers in the North-Rhine-Westfalian population adapted to our patient cohort with LE via the Chi-square test [age-dependent data (5-year steps) of all patients with cancers mixed across genders are used from the cancer register of North Rhine-Westfalia [16] during 2014 and from a statistical report on the population of North Rhine-Westfalia in 2014 [17]. Pearson's Chi-squared test and Fisher's exact test were utilized to compare the relationship between neuronal autoantibodies and cancer occurrence in patients with LE. In addition, we used the bootstrap method to compare the cancer frequency we detected with that reported in the literature. A p-level of b 0.5 was considered as significant. 2.7. Data statement Data are available. 3. Results 3.1. Clinical and laboratory characteristics of patients with LE We identified 20 patients [median age: 53 (25–75) years] with definitive LE and 73 patients [median age: 51 (14–81) years] with possible autoimmune LE. We compared distinct groups of patients with LE with (n = 7) and without cancer (n = 86), as well as those with antibodies [n = 40: n = 10 (n = 10 PB, n = 2 CSF) anti-GAD65-, n = 10 (n = 10 PB) anti-VGKC-, n = 8 (n = 8 PB, n = 2 CSF) anti-NMDAR-, n = 5 (n = 5 PB, n = 1 CSF) anti-Ma2/Ta-, n = 3 (n = 3 PB, n = 2 CSF) CASPR2-, n = 3 (n = 3 PB), anti-LGI1-, n = 2 (n = 1 PB, n = 2 CSF) anti-Yo-, n = 2 (n = 2 PB) anti-Hu-, n = 1 (n = 1 PB and CSF) antiSOX1-, n = 1 (n = 1 PB) anti-CV2/CRMP5-, n = 1 (n = 1 PB) antiZic4-, n = 1 (n = 1 PB) anti-Glycin-antibodies (Fig. 1B)] and without (n = 53) antibodies to evaluate differences in clinical features and instrumental results for the early recognition of patients with LE who might suffer from malignancies. Neuronal antibodies were detected in 7 of 20 (35.7%) with definitive LE (n = 3 VGKC antibodies PB, n = 1 CASPR2 antibodies PB and CSF, n = 1 Ma/Ta2 antibodies PB, n = 1 LGI1 antibodies PB, n = 1 GAD65 and VGKC antibodies PB). However, we observed no differences in clinical and laboratory parameters between groups (Table 1B) with the exception that antibody-positive patients with LE were younger than antibody-negative patients with LE (Mann–Whitney U test: U = 734.5, p b 0.05; Table 1B). Moreover, temporal slowing in interictal EEG was observed more often in patients with LE without cancer than in those with (Chi-square test: χ2 = 4.7, p b 0.05, Table 1B).
N. Hansen et al. / Epilepsy & Behavior 89 (2018) 105–111 Table 2 Type of cancer (A) and estimated vs. observed cancer frequency in patients with LE (B). A: Cancer in patients with LE at visit ± 5 years No. of patient
Cancer
Autoantibody
Cancer detection
1 Acinar cell Ma2/Ta -abs 5 years prior to PET 2 Testicular Ma2/Ta -abs 3 years prior to PET 3 Prostate VGKC - abs 1 year prior to PET 4 Small cell lung Hu - abs upon PET 5 Ovarial Yo - abs upon PET 6 Bronchial No abs upon PET 7 Prostate Zic - 4 abs 1 year after PET Abbreviations: Abs = antibodies, No. = numbers, PET = positron emission tomography, VGKC = voltage gated potassium channels. B: Estimated vs. observed cancer frequency in all and antibody-positive patients with LE a) All patients with LE Age group
0–4 5–9 10–14 15–19 20–24 25–29 30–34 35–39 40–44 45–49 50–54 55–59 60–64 65–69 70–74 75–79 80–84 85+
Estimated
Observed
No. of patients
No. of patients
0 0 0.00012 0.00024 0.00076 0.00647 0.00675 0.00733 0.01693 0.05449 0.06123 0.07986 0.16268 0.20703 0.29297 0.06285 0.03440 0
0 0 0 1 0 0 0 1 0 0 1 0 3 1 0 0 0 0
b) Antibody-positive patients with LE Age group
0–4 5–9 10–14 15–19 20–24 25–29 30–34 35–39 40–44 45–49 50–54 55–59 60–64 65–69 70–74 75–79 80–84 85+
Estimated
Observed
No. of patients
No. of patients
0 0 0 0.00024 0.00076 0.00431 0.00113 0.00367 0.01129 0.03179 0.03402 0.02995 0.02957 0.08281 0.02663 0.06285 0 0
0 0 0 1 0 0 0 1 0 0 1 0 3 0 0 0 0 0
3.2. Cancer frequency in patients with LE A malignancy was detected via FDG-PET/CT in 3 of 93 (3.2%) patients with LE. No cancer was detected in patients with definitive LE associated with neuronal antibodies. In addition, four cancers were detected ±5 years before and after FDG-PET/CT examination, thus a cancer was diagnosed in 7 of 93 (7.5%) patients altogether [1 antibody-negative in 53 (1.89%) antibody-negative patients with LE and 6 antibody-positive in 40 (15.0%) antibody-positive patients with LE revealed a cancer, Fig. 1Ba]. We diagnosed acinar cell, ovarial, prostate, small cell lung, and testicular cancer in 5 of 12 (42%) patients with paraneoplastic antibodies
109
(anti-CV2/CRMP5, anti-Hu, anti-Ma-2/Ta, anti-SOX1, anti-Yo, antiZic4; Table 2A; Fig. 1Bb) prior to, upon, and after the FDG-PET/CT. Furthermore, prostate cancer was detected in one patient with VGKC-antibodies before his FDG-PET/CT (Table 2A). The FDG-PET/CT revealed bronchial cancer in another patient presenting with no antibodies (Table 2A). Antibody-positive patients with LE suffered significantly more from cancer diagnoses than antibody-negative patients with LE (Chi-Square test: χ2 = 5.6, p b 0.05, Fisher's exact test: p b 0.05; Fig. 1Ba). A positive antibody result implied a 9-fold higher risk for cancer. To estimate the difference in our paraneoplastic fraction of patients with LE with that of published data [2], we generated random samples from our sample of patients with LE. Not a single sample yielded a paraneoplastic LE fraction of 23.5%, demonstrating that the cancer frequency of about 23.5% in patients with LE from the PNS trial data [2] is significantly higher than the cancer frequency (7.5%) that we observed in all our patients with LE (Bootstrap test: p b 0.05; Fig. 2Aa). Moreover, the age-dependent cancer frequency observed in all and antibody-positive patients with LE did not differ from the estimated cancer frequency in Germany's North Rhine-Westfalia regional population during 2014 (Chi-square, all patients with LE: χ2 = 36, p = 0.2, Table 2Ba; antibody-positive patients with LE: χ2 = 36, p N 0.09, Table 2Bb; age distribution of all in Fig. 2Ba and antibody-positive patients with LE in Fig. 2Bb), indicating that our cohort's various LE subtypes do not have a higher risk of developing cancer than does the regional population. However, the cancer frequency (15.0%) of the antibody-positive patients with LE did not significantly differ from the 23.5% cancer frequency reported from the literature [2] (Bootstrap test: p = 0.14; Fig. 2Ab). 4. Discussion Our main findings provide evidence of a cancer frequency in our patients with TLE with LE detected by FDG-PET/CT that does not differ from the age-dependent cancer distribution pattern in the regional population in Germany's North Rhine-Westfalia in 2014. Furthermore, the cancer frequency of our patients with LE is significantly less (7.5%) than that of patients with LE with similarly diverse antibodies (23.5%) derived from a large cohort of patients with PNS [2] and that of pooled data from studies with heterogeneous LE subtypes [83/216 patients (39%, Table 1A)]. The low cancer rate in our patients with LE can be attributed to the following: (1) the composition of LE subtypes in our cohort with a large group of antibody-negative patients with LE and NMDAR-, VGKC-, and GAD65-antibodies-positive patients with LE, and (2) low referral of patients with cancer to our tertiary epilepsy center. Our selective cohort of patients with TLE might be a major reason for the few malignancies that we detected compared to the heterogeneous cohort of patients with PNS [2]. A further determining factor (3) for the low cancer rate in our patients with LE is the FDG-PET/CT's diagnostic drawback in identifying specific tumor types because of its inadequate sensitivity for detecting those tumors. Such tumors comprise, for example, lesser pelvis tumors containing certain variants of testicular tumors such as nonseminoma [18], subtypes of ovarian cancer (e.g., small or necrotic ones) [19] or cervical carcinoma [20]. If such a tumor is suspected clinically, additional investigations such as ultrasound and sometimes MRI in conjunction with urological and gynecological examinations are commendable in patients with PNS and thus patients with LE [12]. On the other hand, FDG-PET/CT is not recommended for diagnosing teratomas in PNS due to its poor detection sensitivity for these tumors [12]. This situation predominantly concerns patients with LE with the NMDAR antibodies often associated with ovarian teratoma [7]. The failure of FDG-PET/CT to detect teratomas might have led to the low rate of malignancies in our [0 of 8 (0%)] patients with NMDAR antibodies compared to reported rates (6%) [7]. Thus, further methods such as ultrasound, MRI, or CT thorax are advisable for cancer detection in patients with LE with NMDAR antibodies [12].
110
N. Hansen et al. / Epilepsy & Behavior 89 (2018) 105–111
Fig. 2. Fraction of paraneoplastic patients with LE from our and published data (A) and age distribution of patients with LE (B) A: Our sample revealed 7.5% who were paraneoplastic patients with LE. By generating samples from our patient cohort via the bootstrap method, we demonstrate that the reported frequency of 23.5% derived from the paraneoplastic neurological syndromes (PNS) Euronetwork clinical trial (Giometto et al., 2010) differs significantly from our fraction of paraneoplastic patients with LE (Bootstrap test: p b 0.05, Fig. 2Aa). However, the fraction of paraneoplastic antibody-positive patients with LE (15.0%) from our data did not differ from the reported cancer frequency of 23.5% (Bootstrap test, p = 0.14; Fig. 2Ab). The red lines indicate the 95% confidence interval of our calculated samples. The blue line indicates our fraction of paraneoplastic patients with LE (7.53%) in (a) and our fraction of paraneoplastic antibody-positive patients with LE (15.0%) in (b). The green lines in (a) and (b) denote the cancer frequency value from the literature (Giometto et al., 2010). B: The age distribution of all (Fig. 2Ba) and of antibody-positive (Fig. 2Bb) patients with LE is shown by age steps of 5 years. Abbreviation: no = number. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
Higher reported cancer frequencies are often associated with paraneoplastic LE subtypes such as those with Hu-antibodies [4], Zic4, or Yo-antibodies [2], concurring with our results. Surprisingly, we detected a moderate cancer frequency of 40% in patients with LE with Ma2/Ta antibodies (compared to a higher cancer frequency of 83% in patients with LE with Ma2/Ta antibodies reported by [5]), which supports the unexpectedly low cancer frequency in a few of our patients with TLE with typical paraneoplastic antibodies. However, we should keep in mind that another factor (4) is potentially responsible for the moderate cancer detection rate (42%) in our 12 of 93 (13%) patients with LE with paraneoplastic antibodies, namely the fact that the lack of cancer evidence in FDG-PET/CT does not rule out a paraneoplastic origin of LE, as the tumor-driven immunity might have abolished the tumor or reduced its growth or metastatic spread in these patients. Because of the small number of our patients with TLE with LE and the fact that FDG-PET/CT is not performed to search for cancer every 6 months because of PNS guidelines [12], we cannot assume clinical implications for diagnostic approaches. As not all patients underwent repeated investigations to detect cancer, the low number of cancers detected in our patients with LE might be due to occult cancers promptly eliminated by tumor-immunity control mechanisms.
frequencies being overestimated as a result of FDG-PET/CT methodology, as not all patients presumed to have LE are likely to have undergone FDG-PET/CT. It is more likely that our data reflect those patients who were already presenting suspicious clinical symptoms who were then referred for FDG-PET/CT to detect a malignancy. 4.2. Conclusions Taken together, our results reveal a relatively low cancer frequency detected via FDG-PET/CT in patients with TLE with LE compared to reports in the literature. However, our data do not reveal a clearcut phenotype of patients with LE that should be screened for cancer beyond the existing autoantibody profile. Acknowledgments We thank Denis Breuer for doing parts of the statistical analysis. Funding This study was supported by the “Verein zur Förderung der Epilepsieforschung, Bonn”.
4.1. Study limitations Declaration of interests Our strict criterion of including only patients with TLE with LE and completed FDG-PET/CT examination may have led to the cancer
None.
N. Hansen et al. / Epilepsy & Behavior 89 (2018) 105–111
References [1] Naghavi M, Abajobir AA, Abbafati C, Abbas KM, Abd-Allah F, Abera SF, et al. Global, regional, and national age-sex specific mortality for 264 causes of death, 1980– 2016: a systematic analysis for the Global Burden of Disease Study 2016. Lancet 2017;390:1151–210. [2] Giometto B, Grisold W, Vitaliani R, Graus F, Honnorat J, Bertolini G, et al. PNS Euronetwork. Paraneoplastic neurologic syndrome in the PNS Euronetwork database: a European study from 20 centers. Arch Neurol 2010;67:330–5. [3] Graus F, Titulaer MJ, Balu R, Benseler S, Bien CG, Cellucci T, et al. A clinical approach to diagnosis of autoimmune encephalitis. Lancet Neurol 2016;15:391–404. [4] Gultekin SH, Rosenfeld MR, Voltz R, Eichen J, Posner JB, Dalmau J. Paraneoplastic limbic encephalitis: neurological symptoms, immunological findings and tumour association in 50 patients. Brain 2000;123:1481–94. [5] Dalmau J, Graus F, Villarejo A, Posner JB, Blumenthal D, Thiessen B, et al. Clinical analysis of anti-Ma2-associated encephalitis. Brain 2004;127:1831–44. [6] Dalmau J, Gleichman AJ, Hughes EG, Rossi JE, Peng X, Lai M, et al. Anti-NMDA-receptor encephalitis: case series and analysis of the effects of antibodies. Lancet Neurol 2008;7:1091–8. [7] Bost C, Chanson E, Picard G, Meyronet D, Ducray F, Rogemond V, et al. Malignant tumors in autoimmune encephalitis with anti-NMDA receptor antibodies. J Neurol 2018. https://doi.org/10.1007/s00415-018-8970-0. [8] Dalmau J, Graus F. Antibody-mediated encephalitis. N Engl J Med 2018;378:840–51. [9] Soeder BM, Gleissner U, Urbach H, Clusmann H, Elger CE, Vincent A, et al. Causes, presentation and outcome of lesional adult onset mediotemporal lobe epilepsy. J Neurol Neurosurg Psychiatry 2009;80:894–9. [10] Jagtap SA, Das GK, Kambale HJ, Radhakrishnan A, Nair MD. Limbic encephalitis: clinical spectrum and long-term outcome from a developing country perspective. Ann Indian Acad Neurol 2014;17:161–5. [11] Sheikhbahaei S, Marcus CV, Fragomeni RS, Rowe SP, Javadi MS, Solnes LB. Wholebody 18F-FDG PET and 18F-FDG PET/CT in patients with suspected paraneoplastic
[12]
[13] [14]
[15]
[16]
[17]
[18]
[19] [20]
111
syndrome: a systematic review and meta-analysis of diagnostic accuracy. J Nucl Med 2017;58:1031–6. Titulaer MJ, Soffietti R, Dalmau J, Gilhus NE, Giometto B, Graus F, et al. Screening for tumours in paraneoplastic syndromes: report of an EFNS Task Force. Eur J Neurol 2011;18;19–e3. Helmstaedter C, Lendt M, Lux S, editors. VLMT: Verbaler Lern-und Merkfähigkeitstest. Testhandbuch. Göttingen: Hofgrefe; 2000. Helmstaedter C, Pohl C, Elger CE. Eine modifizierte Version des Diagnostikums für Cerebralschäden (DCS) zur Diagnostik räumlich-visueller Gedächtnisdefizite bei Patienten mit Temporallappenepilepsie. In: Scheffner D, editor. Epilepsie 90. Reinbek: Einhorn Presse; 1991. p. 272–9. Yeo T, Chen Z, Yong KP, Wong PYW, Chai JYH, Tan K. Distinction between antiVGKC-complex seropositive patients with and without anti-LGI1/CASPR2 antibodies. J Neurol Sci 2018;391:64–71. Landeskrebsregister NRW Germany, Bösartige Neubildungen gesamt (C00-C96), status of data base: 31. October 2016. http://www.krebsregister-nrw.de, Accessed date: 28 February 2018. Information und Technik Nordrhein-Westfalen, Geschäftsbereich Statistik. Statistische Berichte: Bevölkerung in Nordrhein-Westfalen 2014 nach Alter und Geschlecht. Fortschreibung des Bevölkerungsstandes auf Basis des Zensus vom 09.05.2011, Bestell-Nr. A133 2014 00, (Kennziffer A I – j /14). http://www.it.nrw. de; 2016, Accessed date: 28 February 2018. Ambrosini V, Zucchini G, Nicolini S, Berselli A, Nanni C, Allegri V, et al. 18F-FDG PET/ CT impact on testicular tumours clinical management. Eur J Nucl Med Mol Imaging 2014;41:668–73. Son H, Khan SM, Rahaman J, Cameron KL, Prasad-Hayes M, Chuang L, et al. Role of FDG PET/CT in staging of recurrent ovarian cancer. Radiographics 2011;31:569–83. Driscoll DO, Halpenny D, Johnston C, Sheehy N, Keogan M. 18F-FDG-PET/CT is of limited value in primary staging of early stage cervical cancer. Abdom Imaging 2015;40: 127–33.