Abstract
BACKGROUND: Paraneoplastic Lambert-Eaton myasthenic syndrome (pLEMS) is well-established in small-cell lung cancer (SCLC), but data on other malignancies are limited. We aimed to define the clinical phenotype of pLEMS in non-SCLC cancers (non-SCLC-pLEMS) relative to SCLC-associated LEMS (SCLC-pLEMS) and autoimmune LEMS (aiLEMS). METHODS: Retrospective analysis was conducted to compare patients with SCLC-pLEMS, aiLEMS and non-SCLC-pLEMS from the European LEMS registry, and further non-SCLC-pLEMS cases were identified by a systematic review following the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines. RESULTS: The registry included 72 aiLEMS, 12 SCLC-pLEMS, and 11 non-SCLC-pLEMS patients. LEMS preceded cancer diagnosis in 33% of SCLC-pLEMS (median 2 months) and 30% of non-SCLC-pLEMS (median 15 months), was concurrent in 25% and 30% and followed tumor diagnosis in the remainder. At study enrollement, Quantitative Myasthenia Gravis (QMG) scores were higher in SCLC-pLEMS (median 12, range 1-24) with recent tumor therapy initiation (median 3 months), and lower in non-SCLC-pLEMS (median 6, range 0-19), with longer (median 12 months) or completed tumor therapy, and aiLEMS (median 5, range 0-23). During follow-up, QMG improved with tumor therapy, and worsened with recurrence/progression in pLEMS groups. After completion of cancer treatment, QMG values in SCLC-pLEMS (median 6, range 0-19) and non-SCLC-pLEMS (median 5, range 1-22) were comparable to each other and to aiLEMS (median 7, range 0-29). Ataxia was significantly more frequent in SCLC-pLEMS (64%) and non-SCLC-pLEMS (55%) than in aiLEMS (19%, p = 0.006 and p = 0.024). Another 115 literature-reported non-SCLC-pLEMS cases were identified (total n = 126, comprising 137 tumors). Most common were non-small cell lung cancer (NSCLC) (n = 25, 18%), Merkel cell carcinoma (n = 18, 13%) and lymphoproliferative disorders (n = 15, 11%). In 52 literature-reported LEMS patients with outcome data, 88% partially or fully recovered after tumor therapy, leaving the paraneoplastic origin uncertain in many. CONCLUSIONS: Our results suggest that, beyond SCLC, other tumors can trigger pLEMS. Compared with aiLEMS, non-SCLC-pLEMS and SCLC-pLEMS showed a higher frequency of ataxia, and LEMS severity tended to reflect tumor treatment status, while disease severity becomes comparable across subtypes after cancer therapy. The frequent improvement of symptoms with tumor-directed treatment supports extended screening beyond SCLC and timely management.