Patterns and incidence of brain metastases in high-grade neuroendocrine carcinomas.

Journal article

Reem Dergham, Yongdong Ouyang, Megan Wansart, Renuka Iyer, Prantesh Jain
Journal of Clinical Oncology, 2026

Cite

APA Click to copy
Dergham, R., Ouyang, Y., Wansart, M., Iyer, R., & Jain, P. (2026). Patterns and incidence of brain metastases in high-grade neuroendocrine carcinomas. Journal of Clinical Oncology.

Chicago/Turabian Click to copy
Dergham, Reem, Yongdong Ouyang, Megan Wansart, Renuka Iyer, and Prantesh Jain. “Patterns and Incidence of Brain Metastases in High-Grade Neuroendocrine Carcinomas.” Journal of Clinical Oncology (2026).

MLA Click to copy
Dergham, Reem, et al. “Patterns and Incidence of Brain Metastases in High-Grade Neuroendocrine Carcinomas.” Journal of Clinical Oncology, 2026.

BibTeX Click to copy

@article{reem2026a,
  title = {Patterns and incidence of brain metastases in high-grade neuroendocrine carcinomas.},
  year = {2026},
  journal = {Journal of Clinical Oncology},
  author = {Dergham, Reem and Ouyang, Yongdong and Wansart, Megan and Iyer, Renuka and Jain, Prantesh}
}

Abstract

e14025

Background: High-grade neuroendocrine carcinomas (HGNEC) are aggressive malignancies with early dissemination. While small cell lung cancer (SCLC) is highly neurotropic, the incidence of brain metastases (BM) in large cell neuroendocrine carcinoma (LCNEC) and extrapulmonary NEC (EP-NEC) remains poorly characterized. We evaluated baseline and cumulative BM incidence and overall survival (OS) across HGNEC subtypes. Methods: Patients with HGNEC (SCLC, LCNEC, EP-NEC) treated at Roswell Park Comprehensive Cancer Center (2000–2024) were retrospectively reviewed. Baseline BM were intracranial lesions on imaging within 90 days of diagnosis; subsequent BM were new lesions >90 days post-diagnosis. Cumulative BM incidence was estimated using competing-risks methodology with death as the competing event. Multivariable Fine–Gray and Cox models assessed BM risk and OS. Results: Of 1665 patients (SCLC n=1215; EP-NEC n=366; LCNEC n=84), 270 (16.2%) had baseline BM. Baseline BM prevalence differed by histology (SCLC 19.5%, LCNEC 17.9%, EP-NEC 4.9%; p<0.001). Among 1394 patients without baseline BM, 24-month cumulative BM incidence was highest in SCLC (26.5%) versus LCNEC (10.1%) and EP-NEC (6.3%; p<0.001). SCLC had significantly higher subsequent BM risk versus EP-NEC (sHR 3.77, 95% CI 2.40–5.93; p<0.001). Median OS in BM-free patients was 15.8 months (SCLC), 20.2 months (LCNEC), and 13.9 months (EP-NEC; p=0.07); among patients with BM, median OS was 14.7, 14.1, and 10.3 months, respectively (p=0.054). Post-2018 diagnosis was associated with higher BM incidence (sHR 1.37, p=0.022) but improved OS (HR 0.79, p<0.001). Conclusions: In the largest single-institution HGNEC cohort reported to date, BM risk varies markedly by histology. SCLC demonstrates the highest baseline and cumulative BM incidence, followed by LCNEC, while EP-NEC shows the lowest CNS tropism. These findings support routine CNS surveillance for pulmonary HGNEC and a symptom-directed approach for EP-NEC.

Histology SCLC (n=1215) EP-NEC (n=366) LCNEC (n=84)

Baseline BM (%) 19.5 4.9 17.9

BM CI (%) 6/12/18/24 mo 3.6/17.7/23.7/26.5 2.0/4.6/6.0/6.3 1.4/7.2/10.1/10.1

OS, mo (BM-/BM+) 15.8/14.7 13.9/10.3 20.2/14.1

CI=cumulative incidence; OS=overall survival (months).