Abstract
Objective: To establish whether improved tumor survival in patients with Lambert-Eaton myasthenic syndrome (LEMS) and small-cell lung cancer (SCLC) was due to known prognostic risk factors or an effect of LEMS independently, perhaps as a result of circulating factors.
Methods: We undertook a prospective observational cohort study of patients with LEMS attending Nottingham University Hospitals, UK, or via the British Neurological Surveillance Unit. In parallel, patients with a new diagnosis of biopsy-proven SCLC were enrolled, examined for neurologic illness, and followed up until death or study end.
Results: Between May 2005 and November 2014, we recruited 31 patients with LEMS and SCLC and 279 patients with SCLC without neurologic illness. Allowing for known SCLC survival prognostic factors of disease extent, age, sex, performance status, and sodium values, multivariate Cox regression analysis showed that the presence of LEMS with SCLC conferred a significant survival advantage independently of the other prognostic variables (hazard ratio 1.756, 95% confidence interval 1.137–2.709, p = 0.011).
Conclusions: Improved SCLC tumor survival seen in patients with LEMS and SCLC may not be due solely to lead time bias, given that survival advantage remains after allowing for other prognostic factors and that the same degree of survival advantage is not seen in patients with paraneoplastic neurologic syndromes other than LEMS presenting before SCLC diagnosis.
GLOSSARY
- LEMS=
- Lambert-Eaton myasthenic syndrome;
- PNS=
- paraneoplastic neurologic syndromes;
- SCLC=
- small-cell lung cancer;
- SOX2=
- SRY-Box 2;
- TNM=
- tumor, node, metastasis;
- UK-BNSU=
- British Neurological Surveillance Unit;
- VGCC=
- voltage-gated calcium channel
For some time, it has been observed that some patients with paraneoplastic neurologic syndromes (PNS), often with associated small-cell lung cancer (SCLC), exhibit prolonged survival compared to cancer patients without neurologic illness.1,2 This has been noticed particularly in patients with Lambert-Eaton myasthenic syndrome (LEMS) and SCLC (LEMS-SCLC), in whom several retrospective and partly prospective studies have demonstrated improved survival.3,–,5 In contrast, survival advantage is seen less commonly in patients with PNS in whom the underlying antibodies are directed against intracellular antigens.6,–,9 The reason may be that these patients, particularly with anti-Hu antibodies, often die of complications from profound neurologic disease rather than SCLC. However, in LEMS-SCLC, specific autoantibodies against cell-surface voltage-gated calcium channels (VGCCs) seem to have a functionally pathogenic role.10,–,12
What remains unresolved is whether immune surveillance in PNS improves tumor prognosis. In most patients with LEMS-SCLC, the neurologic symptoms precede tumor diagnosis, offering clinicians a lead-time bias with early tumor diagnosis. However, this lead-time effect may not fully explain the survival advantage. First, in LEMS-SCLC, the associated SCLC is almost always diagnosed within 3 months, yet the survival advantage is often greater.13 Second, prolonged survival is also seen in patients with LEMS-SCLC whose tumor is extensive at initial diagnosis. Third, patients with LEMS who present concomitantly with lung cancer symptoms may also show improved survival.14
In this prospective study of patients with SCLC, we aimed to establish whether the presence of LEMS with detectable pathogenic VGCC antibodies confers a survival advantage independently of tumor detection at an earlier stage.
METHODS
Patients.
Patients were recruited either as part of an ongoing prospective study of the incidence of PNS in patients with newly diagnosed SCLC in the Nottingham Trent region of the United Kingdom15 or from national notification to one of us (P.M.) via the British Neurological Surveillance Unit (UK-BNSU). Recruitment occurred between May 2005 and October 2012, with follow-up complete on all cases. For survival comparisons, control data were obtained prospectively from patients with SCLC without PNS from the Nottingham Trent region over the same time period.15
Patients were diagnosed with LEMS on the basis of characteristic clinical features (proximal leg weakness, attenuated tendon reflexes, autonomic dysfunction) and either typical neurophysiologic findings (low resting compound muscle action potential amplitude in hand muscles, with incremental responses of >100% after maximal voluntary contraction) or positive P/Q-type VGCC antibodies.16,–,18 Tumor surveillance with PET/CT imaging was performed as per published guidelines.13,17,19
Staging of SCLC in all study patients was classified as limited or extensive as originally described by the Veterans Administration Lung Study Group and by the more recent tumor, node, metastasis (TNM) staging (stage I to IV) from the International Association for the Study of Lung Cancer.20,–,22 Patients with stage IV TNM classification were considered to have extensive disease.23 Complete response to oncological treatment was defined as no remaining evidence of tumor on follow-up imaging.
Standard protocol approvals, registrations, and patient consents.
Written informed consent was obtained from all patients (Nottingham Research Ethics Committee 04/Q2404/100; Oxfordshire Research Ethics Committee A 07/Q160X/28).
Antibody analysis.
Serum samples were taken either at the time of initial diagnosis of LEMS or when SCLC was first diagnosed before treatment, depending on which occurred first. Autoantibodies to P/Q-type VGCCs were measured in all patients by immunoprecipitation of VGCCs extracted from rabbit cerebellum and labelled with 125I-ω-conotoxin MVIIC as previously described.24 Antibodies titers >3 SD above normal and disease controls (50 pmol/L) were considered positive.
Antibodies to SRY-Box 2 (SOX2) were measured with a semiautomated ELISA, with positive signals defined as >3.5 SD above the mean of normal healthy controls (n = 279 controls matched for age, sex, smoking history).25
Statistics.
The number of patients recruited to the study was based pragmatically on the incidence of LEMS-SCLC in previously reported SCLC cohorts26 and the projected number of referrals of patients with LEMS-SCLC nationally via the UK-BNSU over a 10-year period to a figure calculated to exceed that shown to demonstrate survival advantage in patients with LEMS-SCLC in historical retrospective studies.4 Group statistics were compared with the Fisher exact test for frequency distribution and Mann-Whitney U test for mean numerical values. Survival from date of SCLC diagnosis was calculated with log-rank tests. Differences in survival according to LEMS status were evaluated with a Cox proportional hazards model incorporating known clinical predictors of survival (age, sex, sodium, Karnofsky score, and extent of disease)23 as covariates.
RESULTS
Three hundred thirteen consecutive, unselected patients with biopsy-proven SCLC were enrolled from Nottingham, UK. Of these 313 patients, 13 (4.2%) had LEMS-SCLC and 279 had SCLC with no evidence of an associated PNS on neurologic examination at diagnosis and on follow-up. A further 21 patients (6.7%) had SCLC and a PNS other than LEMS (Gozzard et al.15 and tables e-1 and e-2 at Neurology.org). In addition, a further 18 patients with LEMS-SCLC were recruited via the UK-BNSU up to November 2014. Of the SCLC cohort (n = 313), almost all deaths (92%) were attributable directly to SCLC on death certification. Seven patients with LEMS-SCLC (22.5%) were still alive at the end of the study period.
Almost all patients with LEMS-SCLC (29 of 31, 93.5%) were diagnosed with the neurologic syndrome before the SCLC (median time to SCLC diagnosis 4.6 months; 25 of 29 [86%] diagnosed with SCLC within 12 months of LEMS diagnosis). Only 10 patients with LEMS (34%) developed symptoms suggestive of an underlying chest malignancy (cough, chest pain, hemoptysis, breathlessness, weight loss) by the time of LEMS diagnosis. Two patients with LEMS-SCLC also had cerebellar ataxia, and one patient with LEMS-SCLC also had Hu-antibody–associated sensory neuronopathy.
Patients with LEMS-SCLC were younger than patients with SCLC and more likely to be female, but neither difference was significant (table 1). SCLC tumors were more often classified as limited in patients with LEMS (19 of 31, 61%) (stage 1A 3 patients, stage 1B 3 patients, stage II or III 13 patients) compared with patients with SCLC alone (96 of 279, 34%) (stage 1A 2 patients, stage 1B 10 patients, stage II or III 84 patients) (p = 0.005). There were similar rates of SCLC resection surgery in patients with LEMS-SCLC (1 of 31, 3%) and patients with SCLC without neurologic illness (6 of 279, 2%). All patients received oncologic treatment in the same unit, and initial chemotherapeutic regimens were similar in patients with LEMS-SCLC (carboplatin/cisplatin and etoposide median 5 [mean 4.2] cycles; thoracic radiotherapy in 17 of 31, 55%) and patients with SCLC without neurologic illness (median 5 [mean 4.3] cycles of chemotherapy; thoracic radiotherapy in 125 of 279, 45%). Concurrent thoracic radiotherapy with carboplatin/cisplatin was used in 5 of 31 patients with LEMS-SCLC (16%) and 33 of 279 patients with SCLC (12%) (p = 0.56). Prophylactic cranial irradiation was given to 11 of 19 patients with LEMS-SCLC (58%) with limited-stage SCLC and 72 of 96 patients with SCLC (75%) with limited-stage SCLC (p = 0.16).
Patient demographic data
All 31 patients with LEMS-SCLC had raised VGCC antibody titers (mean 589 pmol/L, 67–1368 pmol/L) at significantly higher levels than the VGCC antibodies found in patients with SCLC without clinical or neurophysiologic evidence of LEMS or other neurologic disease (11 of 279, 4%; mean 145 pmol/L, 65–288 pmol/L) (p = 0.0007). SOX2 antibodies were found more frequently in patients with LEMS-SCLC (24 of 31, 78%) compared with patients with SCLC without neurologic illness (73 of 279, 26%) (p < 0.0001).
The median Dutch-English LEMS Tumor Association Prediction SCLC prediction score in the 29 patients with LEMS-SCLC whose neurologic symptoms appeared before the cancer was raised at 5 (range 2–6), indicating a high (96.6%) probability of underlying SCLC at the time of LEMS diagnosis.27 However, 5 of 29 patients with LEMS-SCLC had only a modestly raised Dutch-English LEMS Tumor Association Prediction score of 2 of 6, and 2 of these 5 patients had a prolonged delay before SCLC was subsequently diagnosed (17 and 19 months).
In univariate analysis, median survival of all 31 patients with LEMS-SCLC from time of SCLC diagnosis (18 months) was significantly greater than survival in patients with SCLC without neurologic disease (9.5 months) (p = 0.0005, log rank) (figure 1). There was no correlation between circulating VGCC antibody titers and survival (R2 = 0.019, p = 0.48). When only patients who presented with extensive disease at SCLC diagnosis were compared, median survival (12 months) in patients with LEMS-SCLC was greater than survival in patients with SCLC with no PNS (8 months), but this did not reach significance (p = 0.09, log rank) (figure e-1). Median survival of the 11 patients with SCLC who had low levels of VGCC antibodies but without clinical LEMS was similar to that of patients with SCLC with no neurologic disease at 12.25 months (p = 0.47). Median survival in the 21 patients with SCLC and a PNS other than LEMS (13 months) was not significantly better than that of patients with SCLC alone (9.5 months, p = 0.096 Mann-Whitney U) (tables e-1 and e-2).
Patients with Lambert-Eaton myasthenic syndrome with small-cell lung cancer (LEMS-SCLC) show improved tumor survival compared to patients with SCLC alone (log-rank test, p = 0.0005).
Allowing for known SCLC survival prognostic factors of disease extent, age, sex, performance status, and sodium values, we performed a multivariate Cox regression analysis and showed that the presence of LEMS with SCLC conferred a significant survival advantage independently of the other prognostic variables (p = 0.011, table 2). For direct analysis between local LEMS-SCLC and control patients with SCLC, we performed further Cox regression analysis of survival just in the 13 Nottingham Trent patients with LEMS-SCLC, and the survival advantage remained independently statistically significant (p = 0.037, table e-3).
Cox regression analysis of prognostic factors for survival in SCLC
DISCUSSION
Previous studies have suggested that survival from SCLC is improved in patients with coexistent LEMS.3,–,5,14 However, these studies were, at least in part, retrospective, lacking full multivariate analysis to eliminate bias from prognostic confounders of survival. It has been proposed that the findings of improved SCLC survival in patients with coexistent LEMS could be due to lead-time bias, in that the patients more often present with limited, nonmetastatic disease, conferring a survival advantage after cancer treatment.
Survival in SCLC is related to stage of disease at onset. Data from the UK National Lung Cancer Audit reported a median survival of 11.4 months for patients with limited disease compared to 4 months for extensive-stage SCLC.28 Studies of survival in patients with LEMS-SCLC have consistently shown that a greater proportion of patients present with limited SCLC (50%–65%)4,14 compared to patients with SCLC without neurologic disease (typically ≈35%). Figures from England's National Cancer Data Repository have highlighted that only 1% of patients with SCLC undergo lung resection but that in those patients 5-year survival improves from 3% to 31%.29 Although the majority of our patients with LEMS-SCLC presented with limited disease (61%) and 6 of 31 (19%) had stage 1A or 1B SCLC that might have been amenable to surgery, only one patient (3%) underwent lung resection. Nevertheless, even when accounting for an excess of cases with limited SCLC at diagnosis, using Cox regression analysis, we found that patients with coexistent LEMS still had significantly better survival from SCLC.
Almost all patients with SCLC died as a result of their tumor, and >90% had died by the end of the study. In addition, not one LEMS patient died as a result of their neurologic symptoms. As a result, with complete follow-up data, we were able to use death as a single unbiased endpoint in multivariate analysis.
We attempted to limit bias in this study, compared to previously reported data, in several ways. First, recruitment of all patients was prospective and unselected and represented complete LEMS case ascertainment over the study period from the Trent region.15 Although there may have been case selection in the patients with LEMS referred to us from the UK-BNSU, where patients with poor cancer prognosis may not have been referred, the survival data remained significant when only the all-inclusive, unselected Trent LEMS-SCLC subgroup data were analyzed (table e-3). Second, the control group of patients with SCLC without neurologic disease were enrolled contemporaneously to the patients with LEMS. All patients with SCLC were examined by one of us (P.M.) and followed up until death (in 93% of cases) or the end of the study for surveillance of neurologic symptoms and signs. Approximately 70% of all eligible patients with newly diagnosed SCLC were recruited to our study, and patient demographics were representative of general populations of patients with SCLC in terms of age, sex, and tumor extent.23 The proportion of patients with LEMS in our SCLC cohort (4.2%) was comparable to that published previously (3%)3,26 with an incidence of LEMS-SCLC similar to that reported from a similar-sized Dutch population.30 Both the control SCLC population and the patients with LEMS-SCLC received similar standard chemotherapeutic and radiotherapy treatment regimens in the same oncology unit, in line with national guidelines,31 and although SCLC survival rates vary throughout different regions of the United Kingdom, data from the UK National Lung Cancer Audit (2005–2009) showed that survival from SCLC in the Trent region of Nottingham was in the midrange with a mortality hazard ratio close to 1.0 compared with other regions.32
In terms of lead-time bias, most of our patients with LEMS-SCLC were diagnosed promptly with SCLC, within 5 months of LEMS diagnosis in the majority of patients, despite few pulmonary symptoms suggestive of cancer. Several of our patients with LEMS-SCLC were diagnosed with extensive SCLC with widespread metastases despite tumor surveillance, yet survival was still improved in these patients compared to SCLC control patients with extensive disease. Interestingly, 17 of 21 patients (81%) with PNS other than LEMS were seen before SCLC diagnosis (table e-1), yet the survival advantage in this group was not as significant as that seen in the patients with LEMS-SCLC, indicating that lead-time bias may not be the only factor responsible for improved survival in the LEMS-SCLC group.
The mechanisms for improved survival are not established. LEMS immunoglobulin G has been shown to reduce whole-cell Ca2+ currents in SCLC and neuroblastoma cells and can specifically downregulate P-type Ca2+ channels and inhibit exocytosis in SCLC cells.33,34 It is therefore possible that serum factors such as VGCC antibodies or other soluble components from patients with LEMS-SCLC are capable of reducing SCLC tumor cell proliferation via mediation of mitogenic signal transduction from neuropeptides, perhaps as a result of cell-surface binding to VGCCs, and that this effect could be, at least partially, responsible for the improved survival seen in patients with LEMS-SCLC.
It is noteworthy that in both our current study and other published series, patients with SCLC who harbor VGCC antibodies (often of low titers) yet do not manifest with LEMS clinically fail to show prolonged survival compared to patients with SCLC without VGCC antibodies.3,35 This suggests that only pathogenically functional VGCC antibodies are able to act on channels on SCLC cells, reducing tumor proliferation, and that other detectable VGCC antibodies in patients with SCLC may be directed toward intracellular epitopes of the VGCC and remain nonfunctional. In our study patients with LEMS-SCLC, VGCC antibody titer was unrelated to survival, which may be due to the fact that the radioimmunoassay used does not detect solely VGCC antibodies to extracellular domains of the calcium channel.24 Proliferation experiments with other Ca2+ channel blockers such as conotoxins could be undertaken to investigate this hypothesis more fully.
In part because of the original clinical observations of improved survival in LEMS-SCLC4 and the findings of enhanced activated macrophage and lymphocyte tumor infiltration in these patients36 with reduced CTLA-4 expression and regulatory T-cell dysfunction in the peripheral circulation,37 there is increasing interest in the use of immune therapies for patients with SCLC. Adjunctive therapy with ipilimumab, an anti–CTLA-4 monoclonal antibody, has been shown to improve immune-related progression-free survival in extensive-stage SCLC in phase 2 clinical trials,38 and phase 3 trials of ipilimumab and nivolumab (an anti–program cell death 1 protein monoclonal antibody) are ongoing.
Our findings of improved survival in patients with LEMS-SCLC and the beneficial effects of immunotherapies designed to promote antitumor immunity suggest that potential future therapeutic targets for SCLC may involve regulation of calcium channel influx, perhaps via immune signaling pathways.
AUTHOR CONTRIBUTIONS
Professor Maddison conceptualized and designed the study, collected the data, analyzed and interpreted the data, and wrote the manuscript. Dr. Gozzard collected data and edited the manuscript. Dr. Grainge performed statistical data analysis and interpretation and edited the manuscript. Dr. Lang collected data, performed antibody assays, and edited the manuscript.
STUDY FUNDING
The study was funded by the Myasthenia Gravis Association, United Kingdom (Myaware), through a Clinical Research Fellowship to P.G.
DISCLOSURE
The authors report no disclosures relevant to the manuscript. Go to Neurology.org for full disclosures.
ACKNOWLEDGMENT
The authors are grateful to Dr. Caroline Chapman, University of Nottingham, UK, for SOX2 antibody analysis.
Footnotes
Go to Neurology.org for full disclosures. Funding information and disclosures deemed relevant by the authors, if any, are provided at the end of the article.
Supplemental data at Neurology.org
- Received September 29, 2016.
- Accepted in final form January 17, 2017.
- © 2017 American Academy of Neurology
REFERENCES
Letters: Rapid online correspondence
REQUIREMENTS
If you are uploading a letter concerning an article:
You must have updated your disclosures within six months: http://submit.neurology.org
Your co-authors must send a completed Publishing Agreement Form to Neurology Staff (not necessary for the lead/corresponding author as the form below will suffice) before you upload your comment.
If you are responding to a comment that was written about an article you originally authored:
You (and co-authors) do not need to fill out forms or check disclosures as author forms are still valid
and apply to letter.
Submission specifications:
- Submissions must be < 200 words with < 5 references. Reference 1 must be the article on which you are commenting.
- Submissions should not have more than 5 authors. (Exception: original author replies can include all original authors of the article)
- Submit only on articles published within 6 months of issue date.
- Do not be redundant. Read any comments already posted on the article prior to submission.
- Submitted comments are subject to editing and editor review prior to posting.
You May Also be Interested in
More Online
Topics Discussed
Alert Me
Recommended articles
Long‐term follow‐up, quality of life, and survival of patients with Lambert‐Eaton myasthenic syndrome
SOX1 antibodies are markers of paraneoplastic Lambert–Eaton myasthenic syndrome
Updated Diagnostic Criteria for Paraneoplastic Neurologic Syndromes
Lambert-Eaton myasthenic syndrome: Electrodiagnostic findings and response to treatment