Documented frequency of Lambert-Eaton myasthenic syndrome among patients with small cell lung cancer in a large United States health system

Background Lambert-Eaton myasthenic syndrome (LEMS) is the paraneoplastic syndrome most often associated with small cell lung cancer (SCLC). Four published series in which unselected patients with SCLC were examined neurologically report LEMS in 1.3–3.8%; analyses of claims data report under 0.2%. We measured how often LEMS is documented among patients with SCLC in a large United States health system. Methods In this retrospective cross-sectional study at MedStar Health, a 10-hospital academic health system, patients with SCLC were identified between May 24, 2019 and May 2025 by structured ICD-10 and chemotherapy-based queries (Arm 1A) and, because ICD-10 has no code specific to SCLC, by classifying oncology notes containing “small cell” with a large language model, every positive classification then confirmed against the note text (Arm 1B). Patients with LEMS were identified without date restriction using diagnosis codes, amifampridine orders, and text search, then adjudicated by chart review. Results We identified 436 patients with SCLC (425 by text classification, 11 by structured query only) and 9 meeting a LEMS case definition (2 definite, 2 probable, 5 possible). No patient with SCLC diagnosed during the study period had a documented LEMS diagnosis (0/436; 95% CI 0–0.84%). One further patient with SCLC and possible LEMS was found only by the unrestricted search and died before the study period began. Seven of the nine had an associated malignancy (three NSCLC, one SCLC, one lymphoma, one astrocytoma, one breast) and four had died. A neurology encounter or electrodiagnostic study was mentioned in the index note of 18 of the 425 patients. Conclusion The documented frequency of LEMS in SCLC in routine care was zero, with an upper confidence bound below the lowest estimate from series that examined patients systematically. This is consistent with underrecognition, though the design cannot separate it from true absence of disease, care elsewhere, or incomplete documentation. Clinicians should consider LEMS when a patient with SCLC develops unexplained proximal weakness, hyporeflexia, autonomic symptoms, or bulbar complaints. The absence of a histology-specific code for SCLC, alongside a discrete code for LEMS, is itself a barrier to studying this question at scale.