Abstract
Differentiation between G3 well-differentiated neuroendocrine tumor (WD-NET) and poorly differentiated neuroendocrine carcinoma can be challenging. The distinction is based on the presence of certain features, such as high-grade cytologic atypia, solid growth pattern, necrosis, a very high Ki67 index, and certain molecular mutation profiles. Such distinctions have significant clinical implications, as neuroendocrine tumor and neuroendocrine carcinoma have different management and clinical outcomes for these diseases. Here, we report a stomach neuroendocrine carcinoma patient with ambiguous morphology and molecular alterations.
Case Report
The patient was a 76-year-old woman who presented with abdominal pain. An upper GI endoscopy was conducted to show severe mucosal changes, including bleeding, nodularity, and altered texture in the gastric antrum. H&E sections revealed an infiltrative proliferation in a solid or vague trabecular arrangement. The neoplastic cells were small- to medium-sized, with slightly pleomorphic nuclei and fine granular chromatin and without nuclear molding. There were 9 mitotic figures/2 mm2. The Ki67 index was 95% in the hotspot area. Scattered apoptotic figures were noted, and no necrosis or atypical mitotic figures were noted. A diagnosis of high-grade neuroendocrine neoplasm was rendered, and the patient received systemic therapy consisting of carboplatin and etoposide for 6 months.
Post-chemotherapy PET-CT showed stomach residual disease and focal duodenum uptake. The lesion was re-biopsied, which demonstrated similar histology compared to the previous biopsy (Figure 1A and B). The neoplastic cells were positive for CAM5.2 (Figure 1C), mCEA, synaptophysin (Figure 1D), and chromogranin and negative for TTF1, CDX2, and PAX8. The lesion also demonstrated retained RB1 protein and an abnormal, null expression pattern in p53 (Figure 1E). The Ki67 index was 90% (Figure 1F) in the hotspot area. Considering a lack of unequivocal morphology for a diagnosis of neuroendocrine carcinoma and the limitation of a biopsy specimen, a high-grade neuroendocrine neoplasm was reported to umbrella the possibility of a high-grade Grade 3 neuroendocrine tumor and neuroendocrine carcinoma.

(A, B). Tumor cells were arranged in nests and trabeculae, negative for necrosis. H&E, 100× and 200×. (C). Tumor cells were positive for CAM5.2, 200×. (D) Tumor cells demonstrated neuroendocrine differentiation by synaptophysin (shown) and chromogranin stain, 200X. (E) Tumor cells demonstrated null pattern for p53 protein, 200×. (F) Tumor cells demonstrated a high, 90% Ki67 index, 200×.
Meanwhile, patient specimens were submitted to multiple commercialized molecular diagnostic panels. The NeoGenomic Solid Tumor FISH panel detected MYC amplification and PTEN deletion. Tempus xG + NGS panel detected NTHL1 heterozygous mutation (c.859, C > T). Guardant360 cell-free DNA panel detected 2 significant mutations: NOTCH1 p. Q2415* and TP53 p. R273H. FoundationOne CDx NGS panel demonstrated CCNE1 amplification. The NeoGenomic NeoTYPE discovery profile did not identify pathologic mutations. Variants of unknown significance (VUS) identified included ATRX (p. S519F), DAXX (p. 103 A > V), AMER1, CIC, FLT4, PIK3CB, PMS2, SLIT2, and TET2. The molecular work-up is summarized in Table 1.
Summarization of Clinical Important Molecular Findings of the Current Tumor.
After the second biopsy, the patient received systemic therapy FOLFIRI, which was stopped due to poor tolerance. She was put on palliative radiotherapy for gastrointestinal bleeding. Afterward, she was put on nivolumab/ipilimumab, on which her disease progressed. Subsequently, the patient was put on lurbinectedin, on which a new liver metastasis was discovered. The patient opted for hospice care and passed away soon afterward.
Discussion
Based on the 5th WHO Classification of Tumors of the Digestive System, neuroendocrine neoplasms are divided into well-differentiated neuroendocrine tumors and neuroendocrine carcinomas. By definition, a grade 3 neuroendocrine tumor requires a Ki67 index > 20% and a mitotic count > 20/2 mm2. 1 Molecularly, NETs are defined by MEN1, ATRX, and DAXX mutations. NECs, on the other hand, are poorly differentiated tumors that, based on morphology, can be further divided into small-cell type and large-cell type. Molecularly, neuroendocrine carcinomas are defined by mutations in TP53 or RB1. 1 Compared to NET, NECs usually demonstrate high-grade atypia and tumor necrosis. Ki67 proliferation index of NECs is usually much higher than the 20% cut-off for a G3-NET. However, ambiguous patients with features overlapping between NEC and NET are well reported among pancreatic NENs.2,3
Many studies have been conducted in the past 15 years on neuroendocrine neoplasms.2,4–7 Using a cohort of 19 NEC patients and 11 WD-NET patients and a combined IHC and sequencing panel, Shinchi et al in 2012 demonstrated WD-NETs are genetically different from NECs. 4 DAXX and ATRX labelling were intact in NECs, and they were lost in NETs. P53 aberrant expression was seen in 18 out of 19 NEC patients and loss of RB1 in 14 out of 19. Meanwhile, BCL2 overexpression is associated with higher count and Ki67 index. A study in 2016 3 by Tang et al demonstrated that a small proportion WD-NET shows both areas with conventional NET histological features and with high-grade features, including confluent growth pattern, necrosis, and increased mitotic activity. The cohort of the so-called “transformed NET” included 21 patients whose origin of disease was the pancreas, small bowel, bile duct, and rectum. These patients showed an average mitotic count of 20, Ki67 of 50% and median survival between patients with conventional WD-NETs and NECs. 4 patients were investigated by IHC, which demonstrated a profile like NET. These findings led to the understanding that there exists an extinct high-grade/ G3 NETs that are genetically like WD-NETs, yet with higher histological grade and worse prognosis.
However, differentiating between G3 WD-NETs and NECs still poses a significant challenge diagnostically, and expert pathologists don’t agree with each other all the time.8,9 The 2023 study by Umetsu et al 2 proposed the usage of combined molecular profiling and IHC for RB1, p16, p53, and 1 or 2 acinic cell carcinoma (ACC) markers to work up these ambiguous patients. The inclusion of the ACC markers is based on their observation that 17% (8/47) of high-grade neoplasms with extensive neuroendocrine expression turned out to be mixed ACC-NEC. These may be misclassified as either G3 pancreatic NET or pancreatic NEC if ACC markers are not performed. 2 The authors demonstrated an alteration of TP53 in 35% of G3 NET patients, higher than 3% to 10% that were reported previously. Lastly, CCNE1 amplification was seen in 1 (1/34) NEC patient, who also had a concurrent TP53 mutation.
The CCNE1 gene encodes the protein cyclin E, which, upon activation, binds and activates CDK2 to form the cyclin E-CDK2 complex. This complex completes the phosphorylation of RB1, leading to the synthesis of proteins required for entering S from the G1 phase. 10 Upregulation of cyclin E can be driven by CCNE1 gene amplification, increased activity of c-MYC or other transcription factors, or inactivation of the F box protein, FVXW7 (CDC4), a protein responsible for its degradation. 10 The former 2 mechanisms are observed in the current tumor.
Clinically, upregulation of cyclin E is seen in various cancer types, including ovarian, breast, and liver cancers.11–14 It is also associated with sarcomas like osteosarcoma 15 and chordoma. 16 In breast cancer, a truncated, low-molecular-weight isoform of cyclin E is found to be associated with cytoplasmic relocation, higher affinity with CDK2, chromosomal instability, and more aggressive tumor phenotypes. 17 High level of cyclin E is associated with poor prognosis in ovarian cancer, osteosarcoma, and chordoma.14–16 CCNE1 amplifications in gastrointestinal NECs have been increasingly recognized in recent years. The 2023 study by Shinichi et al 18 demonstrated CCNE1 amplification in 9 of 38 GI NECs and in 1 of 24 pancreatic NECs. These patients also demonstrated TP53 mutations with preserved RB1 IHC. Loss of RB1 in IHC seems to be a mutually exclusive event with MYC/CCNE1 amplification. 18 Cubiella et al demonstrated PCDHGC3 hypermethylation is associated with reduced overall survival in GI NEC patients. The authors also claimed a superior predictive value of PCDHGC3 hypermethylation for discriminating GI NEC from GI NET. 4 Chiara et al explored the diagnostic and prognostic role of DLL3 IHC expression in GI and Pancreatic NENs. DLL3 expression is associated with RB1-loss and a poor prognosis. 5
In summary, we report a patient with stomach neuroendocrine carcinoma with features overlapping between G3 WD-NET and NEC. The tumor demonstrated a nested and/or solid growth pattern. The overall cytology of tumor cells was bland, and focal atypia was interpreted as neuroendocrine atypia. No typical morphologic features of poorly differentiated adenocarcinoma or neuroendocrine carcinoma were appreciated. The CCNE1 and MYC amplification has been recently recognized in GI-NECs. 6 Clinically, the disease demonstrated aggressive behavior, which led to patient demise within 18 months. We want to report this TP53-mutated and CCNE1/MYC amplified gastrointestinal neuroendocrine carcinoma so that similar patients would be better recognized in the future, given that target therapy showed promising results in other CCNE1-amplified carcinomas. 7
Footnotes
Ethical Approval and Informed Consent Statements
Waived. Our institution's IRB does not require ethical approval for a case report.
Author Contributions
Both authors, Z.X. and Q.C, reviewed the specimen and wrote the manuscript.
Funding
The authors received no financial support for the research, authorship, and/or publication of this article.
Declaration of Conflicting Interests
The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Data Availability Statement
Please contact correspondent author Q.C. for data request.
