nature.com

Revealing neuroendocrine transformation in gynecological cancers through genomic analysis

Abstract

Neuroendocrine transformation (NT) in cancers, typically observed under the selective pressure of targeted therapies, involves lineage plasticity where adenocarcinomas adopt neuroendocrine characteristics while retaining the molecular alterations of their original histology. This phenomenon, well-documented in prostate and lung cancers, has not been observed in gynecological malignancies until now. We present two pivotal cases involving primary ovarian and uterine cancers that developed neuroendocrine carcinomas post-treatment. Initially presumed to be independent primaries, comprehensive next-generation sequencing technologies, including UW-OncoPlex and BROCA panels, were used to establish a clonal relationship between the primary tumors and their respective neuroendocrine metastases. This report provides the first documented instances of NT in gynecological cancers, indicating that it may be a more widespread resistance mechanism than previously recognized. Routine re-biopsy and early integration of advanced molecular diagnostics into clinical practice will identify NT and provide insights into pathogenesis and eventual therapeutic options.

Introduction

Neuroendocrine transformation (NT) refers to the transformation of an adenocarcinoma’s original histology into a neuroendocrine phenotype - a phenomenon distinct from de novo neuroendocrine/small cell carcinomas, mixed neuroendocrine-non-neuroendocrine neoplasms (MiNEN), or carcinomas with focal neuroendocrine features. This transformation often occurs under the selective pressure of targeted therapy, such as anti-androgen, and anti-EGFR treatment in prostate and lung adenocarcinoma, respectively1,2,3. The transformed cancers no longer rely on the original driver pathway, and therefore, NT is likely a form of drug resistance. Neuroendocrine-transformed cancers exhibit many features of de novo neuroendocrine carcinomas, including histology, specific immunohistochemical (IHC) markers, and frequent RB1 and TP53 mutations. Initial speculation could not pinpoint the source of this change - whether NT involved the emergence of independent primary neuroendocrine cancers, evolutionary selection of sub-clones, or a direct transformation from adenocarcinoma4. Currently, it is recognized as a direct transformation and is considered an example of lineage plasticity—the ability of cancer cells to switch developmental pathways5,6. Evidence of this is that these transformed neuroendocrine cancers retain the molecular characteristics of their original adenocarcinomas, including somatic driver mutations. These cancers show aggressive growth, and poor clinical outcomes, with very few treatment options despite exceptional responses to initial platinum-based chemotherapy due to the transient nature of such a response.

To our knowledge, there are no published examples of NT in gynecological cancers. In this case report, we present two distinct cases: one involving primary ovarian cancer and the other a uterine cancer, each initially displaying classic endometrioid histology. Remarkably, several years post initial management, both cases developed distant metastases with neuroendocrine pathological features, initially suggestive of new primaries. Employing comprehensive next-generation sequencing (NGS), we established a clonal relationship between the primary tumors and their neuroendocrine metastases, thereby confirming neuroendocrine transformation. These findings mark some of the first documented instances of NT in these malignancies. It also highlights the pivotal role of genomic profiling, molecular pathologists, genomic experts, and the deliberations of molecular tumor boards in determining the true origin of a tumor and guiding precision medicine strategies.

Results

Case presentation #1

Patient 1 was diagnosed at age 60 with Stage IC, Grade 2 ovarian endometrioid adenocarcinoma. IHC analysis revealed strong expression of estrogen receptor (ER) and paired box gene 8 (PAX8) (Fig. 1A, upper panel), consistent with a cancer of Mullerian origin. Following staging laparotomy, she was treated with adjuvant chemotherapy without bevacizumab and remained disease-free for ten years. At recurrence, a biopsy of a pelvic mass maintained endometrioid morphology with classic glandular features. Her subsequent treatment included platinum-based doublet chemotherapy, followed by palliative radiation to the pelvic mass (28 Gy) and a prolonged course of single-agent carboplatin. Two and a half years after pelvic recurrence, she presented with lumbar pain; imaging studies revealed rapid disease progression with the emergence of new pulmonary, hepatic, and pelvic lesions, along with spinal pathological fractures due to metastasis. Given the extent of spinal involvement, she underwent T11-L1 posterior spinal fusion, laminectomies, and tumor resection. The spinal specimen exhibited classic morphological characteristics of neuroendocrine carcinoma (Fig. 1A, middle panel), a complete absence of ER and PAX8 staining, and positive staining for neuroendocrine markers, including chromogranin, synaptophysin, and Insulinoma associated protein 1 (INSM1) (Fig. 1A, lower panel). Given these findings, the spinal specimen was initially deemed to be a separate primary not arising from the Mullerian tract. To establish the clonal relationship between primary ovarian cancer and the newly diagnosed neuroendocrine carcinoma, both tumors underwent NGS using OncoPlex version 7 (OPXv7). This analysis, complemented by data from the BROCA version 12 NGS-panel used on the prior pelvic recurrence, indicated all three specimens were clonally related, with overlapping copy-number alterations and shared mutations including a pathogenic mutation in PTEN (p.T319*), which was absent from the germline specimen (Fig. 1B). Additionally, both the primary ovarian and spinal metastatic tumors shared an activating mutation in CTNNB1 (p.G34V), with further acquisition of a SMAD4 mutation (p.E538del) in the pelvic recurrence and neuroendocrine metastasis. The neuroendocrine metastasis also acquired multiple additional mutations, including a mutation in TP53 (p.T211Ffs*4), indicative of tumor heterogeneity and/or divergent evolution (Fig. 1C). The patient received palliative radiation to the vertebral metastases and enrolled in hospice with no further systemic therapy.

Fig. 1: Comparison of primary and metastatic lesions in Patient 1.

figure 1

A Histologic examination of the primary ovarian neoplasm (top panel) demonstrates complex branching glands composed of the pseudostratified back-to-back columnar epithelium (shown at 10X and 40X magnification) positive for ER and PAX8 (shown at 20X magnification) consistent with low-grade endometrioid carcinoma, while the spinal metastasis (middle panel) consists of sheets of poorly differentiated cells with high nuclear-to-cytoplasmic ratio and nuclear molding (10 and 40X magnification), negative for ER and PAX8 and (lower panel) with variable positivity for CK7, chromogranin, synaptophysin, and INSM1 indicating neuroendocrine differentiation (all at 20X magnification). B Next-generation sequencing demonstrates the overlap of several copy alterations with a notable divergence between the specimens, consistent with tumor heterogeneity and/or clonal evolution. C Comparison of the single nucleotide variants and indels (and their variant allele fractions, VAFs) between the primary ovarian neoplasm, pelvic recurrence, and neuroendocrine spinal metastasis reveals multiple shared somatic findings, confirming their clonal relationship (shared variants are highlighted in yellow).

Full size image

Case presentation #2

Patient 2 was diagnosed at age 69 with Stage IA, FIGO Grade 1 endometrioid endometrial adenocarcinoma (Fig. 2A, upper panel), and underwent surgery with no subsequent adjuvant therapy. Pathology confirmed 29% myometrial invasion, absence of lympho-vascular invasion, and negative sentinel lymph nodes; IHC for mismatch repair (MMR) proteins was not performed. The patient was followed with active surveillance guidelines set by the NCCN. After remaining disease-free for 5 years, she presented with cough, small-volume hemoptysis, weight loss, and right upper extremity neuropathy. Imaging demonstrated a large 13.6 cm right pulmonary mass (Pancoast tumor); enlarged supraclavicular, prevascular, and precarinal lymph nodes; and a 2.8 cm right inferior pelvic mass. A CT-guided biopsy of the pulmonary mass showed a high-grade neuroendocrine carcinoma with an overlap between large-cell and small-cell features. (Fig. 2A, middle panel). IHC demonstrated positivity for cytokeratin 7 (CK7), with focal staining for chromogranin, patchy CDX2, and rare cell positivity for p40 and p63, and absence of synaptophysin, TTF-1, Napsin-A, CK20, GATA3, CK5/6, and PAX8 (Fig. 2A, lower panel). Initially treated as a primary pulmonary neoplasm with a cycle of carboplatin and pemetrexed, followed by palliative radiation (45 Gy) to the Pancoast tumor, subsequent NGS using OPXv7 revealed the lung, and primary endometrial tumors arose from the same clonal process. Both neoplasms harbored an elevated tumor mutational burden (TMB) secondary to microsatellite instability (MSI-High), which was subsequently determined to be the result of MLH1 promoter hypermethylation. The presence of multiple shared pathogenic mutations in tumor suppressors including mutations predicted to result in early termination and/or loss of function in PTEN (p.V133_C136del), CHEK2 (p.L380Hfs*14), PIK3R1 (1. p.M563_S565del and 2. p.Y607*), and PMS2 (p.E109Kfs*3) supported that the two samples were genetically related and likely from the same clonal source (Fig. 2B). The presence of additional unshared mutations in both samples was consistent with tumor heterogeneity and/or divergent evolution. Following these insights, the patient was started on single-agent pembrolizumab with an exceptional response and no evidence of progression for over two years.

Fig. 2: Comparison of primary and metastatic lesions in Patient 2.

figure 2

A Histologic examination of the primary uterine neoplasm (top panel) shows complex branching glands composed of pseudostratified back-to-back columnar epithelium (shown at 10X and 40X magnification) consistent with low-grade endometrioid carcinoma. Histologic examination of the lung neoplasm (middle panel) demonstrates sheets of cells with a high nuclear-to-cytoplasmic ratio and nuclear molding (shown at 10X and 40X magnification). Lung tumor cells are negative for PAX8, but positive for chromogranin (lower panel), indicating neuroendocrine differentiation (shown at 20X). B Comparison of the single nucleotide variants and indels (and their variant allele fractions, VAFs) between the uterine neoplasm and neuroendocrine metastasis revealed multiple shared somatic findings, confirming their clonal relationship (shared variants are highlighted in yellow).

Full size image

Discussion

This molecular case report highlights primary ovarian and uterine cancers that initially presented with endometrioid histology but later developed distant neuroendocrine metastases, illustrating the transformative impact of precision medicine in oncology. De-novo neuroendocrine gynecological carcinomas and mixed carcinomas with neuroendocrine features are rare but recognized entities. Previous case reports have described the coexistence of a neuroendocrine carcinoma with endometrial serous carcinoma and its recurrence as a pure neuroendocrine carcinoma7. However, our findings document the first instances of true neuroendocrine transformation (NT) in gynecological cancers, a phenomenon thought to be mostly confined to prostate and lung adenocarcinomas. Genomic profiling in our cases has established a clonal relationship between the neuroendocrine metastases and the primary cancers. For instance, Patient 1 exhibited shared driver mutations in PTEN and CTNNB1 across primary and metastatic sites, while Patient 2 showed microsatellite instability in both tissue samples. These findings align with observations in lung and prostate adenocarcinomas, where neuroendocrine-transformed cancers retain notable molecular features of the original adenocarcinomas. For example, the detection of microsatellite instability in Patient 2’s neuroendocrine carcinoma prompted a shift from chemotherapy to pembrolizumab monotherapy, resulting in an exceptional and sustained response. MSI is a common phenomenon in endometrial but not in neuroendocrine carcinomas. Currently, platinum-etoposide combinations are the treatments for NT, but if these tumors retain their original drivers, these could be targets for therapy.

Interestingly, the neuroendocrine cancer in Patient 1 acquired a TP53 mutation, and demonstrated loss of CDKN2A which is in the RB1 pathway and also associated with NT8. Both patients had PTEN loss, all of these are commonly associated with NT.

The practice of re-biopsy and NGS at the time of metastatic recurrence is crucial for identifying additional cases of NT, which probably occur in a variety of cancers. Reports suggest that NT is not limited to adenocarcinomas and has been described in squamous cell carcinomas of the head and neck9. A neuroendocrine gene expression signature linked to poor prognosis has been identified across various epithelial malignancies, suggesting that NT might be a more widespread mechanism of resistance than previously understood10. Identifying and understanding these cases is critical to understanding pathogenesis and developing unique therapeutic strategies.

The challenges of differentiating metastases from metachronous and synchronous cancers often hinge on clinical, radiological, and histopathological characteristics. Molecular diagnostics, however, are reshaping these traditional diagnostic approaches, enhancing our understanding of cancer pathogenesis and evolution. Our report also underscores the direct benefit of conducting repeated NGS on metastatic lesions

In summary, this case report not only presents the first known instances of NT in gynecological cancers but also demonstrates how early integration of molecular diagnostics into oncological practice can help identify such cases promptly, improving our understanding of cancer pathogenesis, tailoring treatments, and ultimately enhancing patient outcomes in complex clinical scenarios.

Methods

Immunohistochemistry and histopathology

Formalin-fixed paraffin-embedded (FFPE) tissue blocks were stained with hematoxylin and eosin at our institution for in-house cases (uterus and metastasis in case #1 and metastasis only in case #2). IHC stains were performed on 4-µm thick unstained slides obtained also from formalin-fixed paraffin-embedded (FFPE) tissue blocks. The IHC slides were stained on Leica Bond III automated platform (Leica Microsystems, Buffalo Grove, IL) using Leica Bond Refine Detection kit. Epitope retrieval protocols included either HIER (Heat-Induced Epitope Retrieval)using ER2 (pH 9.0) or ER1 buffer (pH 6.0) and/or enzymatic pretreatment. After staining was completed, slides were counterstained, dehydrated, and cleared through several changes of graded alcohol, xylene, and cover-slipped. Positive controls were performed on the same run to evaluate for appropriate staining.

UW-OncoPlex Version 7 (OPXv7)

OPXv7 is a comprehensive DNA-based next-generation sequencing (NGS) panel that evaluates for single nucleotide variants, insertions and deletions (indels), copy number alterations, and select gene fusions across 376 cancer-relevant genes. This panel also provides assessments of microsatellite instability (MSI) and tumor mutational burden (TMB). Consistent with established protocols, DNA is extracted from FFPE tissue blocks and sheared before library preparation, as detailed in previous studies11. Sequencing is conducted using an Illumina NextSeq500 system (Illumina Inc, San Diego, CA). Analyzed data is processed through a custom-designed bioinformatics pipeline developed at the UW NGS Analytics Laboratory and reviewed by a Genetics and Solid Tumors Lab director.

BROCA Version 12 (BROv12)

The BROv12 panel is a targeted NGS panel that sequences all exons and selects non-repeating introns and promoter regions of 92 genes; it also evaluates MSI as previously described. DNA samples, both germline and FFPE tumors, undergo shearing followed by library preparation, as reported in prior research12. Sequencing is performed on an Illumina HiSeq 2500 system (Illumina Inc, San Diego, CA). Similar to OPXv7, the resulting data is processed through a dedicated bioinformatics pipeline at the UW NGS Analytics Laboratory, with final evaluation by a Genetics and Solid Tumors Lab director.

MLH1 promoter methylation analysis

Analysis of MLH1 promoter hypermethylation is conducted on tumor DNA following sodium bisulfite conversion using the EZ DNA Methylation-Lightning Kit (Zymo Research, Irvine CA, Cat. No. D5030). The assessment is based on a fluorescence-based, real-time quantitative PCR methodology as previously documented13, with subsequent evaluation by a Genetics and Solid Tumors Lab director.

UW/FHCC Gynecologic Oncology Precision Medicine Tumor Board

This multidisciplinary tumor board includes molecular and anatomic pathologists, gynecological oncologists, medical oncologists, radiation oncologists, laboratory scientists, and genetic counselors, as needed. Recommendations are formulated following a comprehensive review and discussion of each case, considering the clinical presentation, pathology, imaging studies, and molecular diagnostic results.

Data availability

Data sharing is not applicable to this article as no datasets were generated or analyzed during the current study.

References

Sequist, L. V. et al. Genotypic and histological evolution of lung cancers acquiring resistance to EGFR inhibitors. Sci. Transl. Med. 3, 75ra26 (2011).

ArticlePubMedPubMed CentralGoogle Scholar

Lee, J. K. et al. Clonal history and genetic predictors of transformation into small-cell carcinomas from lung adenocarcinomas. J. Clin. Oncol. 35, 3065–3074 (2017).

ArticleCASPubMedGoogle Scholar

Beltran, H. et al. Divergent clonal evolution of castration-resistant neuroendocrine prostate cancer. Nat. Med. 22, 298–305 (2016).

ArticleCASPubMedPubMed CentralGoogle Scholar

Quintanal-Villalonga, Á. et al. Lineage plasticity in cancer: a shared pathway of therapeutic resistance. Nat. Rev. Clin. Oncol. 17, 360–371 (2020).

ArticlePubMedPubMed CentralGoogle Scholar

Gardner, E. E. et al. Lineage-specific intolerance to oncogenic drivers restricts histological transformation. Science 383, eadj1415 (2024).

ArticleCASPubMedPubMed CentralGoogle Scholar

Davies, A., Zoubeidi, A., Beltran, H. & Selth, L. A. The transcriptional and epigenetic landscape of cancer cell lineage plasticity. Cancer Discov. 13, 1771–1788 (2023).

ArticleCASPubMedPubMed CentralGoogle Scholar

Karpathiou, G. et al. Ovarian neuroendocrine carcinoma of metastatic origin: clues for diagnosis. Hum. Pathol. 85, 309–312 (2019).

ArticleCASPubMedGoogle Scholar

Huang, J. et al. Genomic and transcriptomic analysis of neuroendocrine transformation in ALK-rearranged lung adenocarcinoma after treatments with sequential ALK inhibitors: a brief report. JTO Clin. Res. Rep. 3, 100338 (2022).

PubMedPubMed CentralGoogle Scholar

Schartinger, V. H. et al. Neuroendocrine differentiation in head and neck squamous cell carcinoma. J. Laryngol. Otol. 126, 1261–1270 (2012).

ArticleCASPubMedGoogle Scholar

Balanis, N. G. et al. Pan-cancer convergence to a small-cell neuroendocrine phenotype that shares susceptibilities with hematological malignancies. Cancer Cell 36, 17–34.e7 (2019).

ArticleCASPubMedPubMed CentralGoogle Scholar

Kuo, A. J. et al. Validation and implementation of a modular targeted capture assay for the detection of clinically significant molecular oncology alterations. Pract. Lab. Med. 19, e00153 (2020).

ArticlePubMedPubMed CentralGoogle Scholar

Walsh, T. et al. Detection of inherited mutations for breast and ovarian cancer using genomic capture and massively parallel sequencing. Proc. Natl Acad. Sci. USA 107, 12629–12633 (2010).

ArticleCASPubMedPubMed CentralGoogle Scholar

Bosch, D. E. et al. Isolated MLH1 loss by immunohistochemistry because of benign germline MLH1 polymorphisms. JCO Precis. Oncol. 6, e2200227 (2022).

ArticlePubMedPubMed CentralGoogle Scholar

Download references

Acknowledgements

This study received no funding.

Author information

Authors and Affiliations

Division of Gynecological Oncology, University of Washington, Seattle, WA, USA

Ann Oluloro, Tiffany Luu & Renata R. Urban

Department of Laboratory Medicine and Pathology, University of Washington, Seattle, WA, USA

David L. Wells, Charles K. Childers, Eric Q. Konnick & Vera A. Paulson

Clinical Research Division, Fred Hutchison Cancer Center, Seattle, WA, USA

Keith D. Eaton & Kalyan Banda

Division of Medical Oncology, University of Washington, Seattle, WA, USA

Keith D. Eaton & Kalyan Banda

Authors

Ann Oluloro

View author publications