Colorectal neuroendocrine carcinoma (CR-NEC) is a rare and poorly understood subtype of colorectal cancer, accounting for less than 1% of all colorectal malignancies. Unlike the far more common colorectal adenocarcinoma (CR-AC), which arises from glandular epithelial cells, CR-NEC originates from neuroendocrine cells and behaves in a far more aggressive manner.
These tumors are classified as poorly differentiated, meaning the cancer cells have lost their normal structure and function. They typically show a very high proliferation rate, reflected in a Ki-67 index that often exceeds 75% to 90%. Most patients present with advanced or metastatic disease at the time of diagnosis, leaving palliative chemotherapy as the main treatment option.
CR-NEC frequently shares genetic driver mutations with colorectal adenocarcinoma, including alterations in BRAF, KRAS, APC, and TP53. Many CR-NEC tumors also contain an adenocarcinoma component alongside the neuroendocrine component. When each component makes up at least 30% of the tumor, it is classified as a mixed neuroendocrine-non-neuroendocrine neoplasm (MiNEN). This shared molecular landscape has led researchers to hypothesize that CR-NEC and CR-AC may share a common clonal origin.
Despite these connections, the biological reasons behind CR-NEC's dramatically worse prognosis remain unclear. Understanding the molecular and clinical features of CR-NEC, and how they compare to CR-AC, is essential for developing better treatments for this aggressive disease.
This study drew its CR-NEC cohort from the NORDIC NEC Registry, a prospective multi-center registry enrolling patients diagnosed with high-grade neuroendocrine neoplasms across nine Scandinavian hospitals from 2013 to 2017, with follow-up through 2021. Initially, 290 colorectal cases were enrolled, but 65 were excluded following re-evaluation by three expert neuroendocrine pathologists using updated 2019 WHO classification criteria.
Of the 225 confirmed CR-NEC cases with available pathology, 195 had metastatic disease. Among these, 32 patients did not receive palliative chemotherapy due to poor performance status or advanced age, leaving a final cohort of 163 patients who received first-line treatment and formed the primary study population.
The comparative CR-AC cohort was drawn from the Scandinavian Prospective Colorectal Cancer Registration (SPCRC), a population-based registry from three Scandinavian hospitals during 2003 to 2006. To minimize misdiagnosis, immunohistochemical staining for the neuroendocrine marker synaptophysin was performed, and 263 synaptophysin-negative patients with metastatic disease receiving first-line palliative chemotherapy were included for comparison.
Molecular data including BRAF, KRAS, APC, TP53, RB1, and MSI status were collected using multiple methods, including targeted next-generation sequencing panels and plasma-based ctDNA sequencing. Treatment response was assessed by RECIST 1.1 criteria, and statistical analyses used Kaplan-Meier survival curves, log-rank tests, and Cox regression models.
Patients with metastatic CR-NEC were older at diagnosis (median age 68 vs. 64 years) and more frequently had poor performance status. Only 31% of CR-NEC patients had a fully active performance status (PS 0) at treatment start, compared to 51% of CR-AC patients. Smoking history was notably more common in the CR-NEC group (51% vs. 26%).
The distribution of primary tumor sites differed markedly between the two groups. Rectal primaries were far more common in CR-NEC (48%) than in CR-AC (26%), while left colon primaries were less frequent in CR-NEC (14% vs. 39%). Right colon primaries were similarly distributed between the groups (38% vs. 34%). Large cell morphology predominated in colonic CR-NEC (76%), while large cell and small cell morphologies were equally common in rectal CR-NEC.
The median Ki-67 proliferation index in CR-NEC was 90%, with only 13 patients having Ki-67 of 55% or below. Synchronous metastases were present at diagnosis in 86% of CR-NEC vs. 56% of CR-AC patients. Liver metastases were more frequent in CR-NEC (79% vs. 68%), as were bone metastases (15% vs. 5%), while peritoneal metastases were paradoxically less common in CR-NEC (4% vs. 20%).
CR-NEC patients also reported substantially higher symptom burden at the start of treatment. Cancer-related pain was reported by 59% of CR-NEC patients vs. 37% of CR-AC patients, and anorexia was present in 45% vs. 27%, both differences statistically significant. This heavier symptom burden at baseline reflects the more aggressive biology of CR-NEC.
Among the 163 CR-NEC patients, 83% received first-line platinum-etoposide (EP) chemotherapy, the standard extrapolated from small-cell lung cancer. The response rate was 24%, with 19% achieving stable disease, while 46% experienced immediate disease progression as their best response. The median number of treatment cycles was only three.
Outcomes in CR-NEC were dramatically worse than in CR-AC across all measures. The median progression-free survival (PFS) was 2.4 months for CR-NEC vs. 7.7 months for CR-AC (p less than 0.001), and median overall survival (OS) was 6.7 months vs. 16.8 months (p less than 0.001). Even patients with the best performance status (PS 0) achieved only 12.2 months median OS in CR-NEC, compared to 23.0 months in PS 0 CR-AC patients. Two-year survival rates were 9% in CR-NEC vs. 37% in CR-AC.
Poor performance status was the strongest prognostic factor in CR-NEC: PS 0 patients achieved OS of 12.2 months vs. 4.2 months for PS 2 patients and only 0.9 months for PS 3 patients. Elevated alkaline phosphatase (ALP) and liver metastasis were also independently associated with shorter survival. Notably, patients whose Ki-67 was 55% or below who received EP had no disease control at all.
Second and third-line treatments showed uniformly poor results, with response rates of 14% and 9% respectively, and PFS of only 2.0 months at both stages. No significant difference was found between irinotecan-based, oxaliplatin-based, or CAPTEM regimens in the second and third-line settings. Overall, these outcomes confirm that CR-NEC responds poorly to all currently available chemotherapy approaches and that new therapeutic strategies are urgently needed.
BRAF mutations were found in 26% of CR-NEC patients, occurring predominantly in right-sided primaries (54% of right colon cases). BRAF mutations were mutually exclusive with KRAS mutations in CR-NEC. Patients with BRAF-mutated tumors had a significantly shorter overall survival compared to those with BRAF/KRAS double wild-type status (4.8 months vs. 10.7 months). BRAF mutation frequency was similar between CR-NEC and CR-AC overall (26% vs. 20%), but was notably higher in right-sided CR-NEC than right-sided CR-AC (54% vs. 34%).
KRAS mutations were found in 34% of CR-NEC patients, less frequently than in CR-AC (45%, p equals 0.041). In CR-NEC, KRAS mutations were associated with significantly shorter OS across all primary sites (7.1 vs. 10.7 months), with the effect driven particularly by rectal primaries (5.7 vs. 11.6 months). Interestingly, KRAS mutations did not influence survival in CR-AC, suggesting a different functional role for this mutation in neuroendocrine vs. adenocarcinoma contexts.
RB1 mutations were more frequent in CR-NEC than CR-AC (49% vs. 14%), consistent with CR-NEC's neuroendocrine nature and paralleling what is seen in small-cell lung cancer. APC and TP53 mutations were similarly distributed between CR-NEC and CR-AC and did not impact survival in either group. Microsatellite instability-high (MSI-H) status was rare in both groups, approximately 4% in CR-NEC and 6% in CR-AC.
Tumor mutation burden (TMB) was available in 61 CR-NEC patients and was low, with a median of 3.3 mutations per megabase, showing no impact on survival. The shared driver mutations between CR-NEC and CR-AC, combined with the striking difference in clinical behavior, suggest that these mutations alone do not explain CR-NEC's aggressive phenotype. Other, yet undiscovered, molecular factors likely drive the different biology.
Despite sharing key driver mutations such as BRAF, KRAS, APC, and TP53, metastatic CR-NEC and metastatic CR-AC are clinically distinct diseases with fundamentally different behavior. CR-NEC presents with earlier and more widespread metastasis, higher symptom burden, and worse performance status at the time of treatment initiation, all of which independently contribute to the inferior outcomes observed.
The disparity in treatment outcomes is stark. The disease control rate on first-line treatment was 43% in CR-NEC vs. 74% in CR-AC, and 46% of CR-NEC patients had immediate progressive disease vs. only 15% of CR-AC patients. This means nearly half of all patients with CR-NEC derived no clinical benefit from the standard first-line treatment. The rates of secondary radical metastatic surgery also differed dramatically: 2% in CR-NEC vs. 11% in CR-AC, reflecting fewer opportunities for curative-intent intervention.
Even when CR-NEC and CR-AC patients were compared within the same good-performance-status group (PS 0), CR-NEC patients survived significantly shorter, indicating that the difference in prognosis is not merely a reflection of patients being sicker at diagnosis but reflects intrinsic disease biology. The lack of peritoneal metastases in CR-NEC, despite otherwise more widespread disease, remains an interesting and unexplained biological observation.
These comparative data support the hypothesis that CR-NEC and CR-AC, though potentially sharing a common cellular origin, have diverged into clinically distinct entities. Future research must focus on the unique molecular pathways driving CR-NEC's aggressive behavior to identify actionable therapeutic targets beyond the currently ineffective platinum-based regimen.
This large prospective study confirms that metastatic CR-NEC carries an extremely poor prognosis, with a median overall survival of 6.7 months and only 9% two-year survival. These outcomes have not improved compared to data from earlier retrospective studies conducted a decade ago, highlighting that the current standard first-line treatment with platinum-etoposide offers limited benefit to most patients with this disease.
The high rate of immediate disease progression (46%) on first-line platinum-etoposide raises fundamental questions about whether this treatment strategy, extrapolated from small-cell lung cancer without CR-NEC-specific trial data, is appropriate for these patients. Alternative adenocarcinoma-type regimens such as fluorouracil-based combinations were used in only a small number of patients and showed no significant survival advantage in this analysis, though the sample was too small to draw firm conclusions.
The ongoing FOLFIRINEC randomized trial may help clarify whether mFOLFIRINOX represents a superior first-line option. Targeted therapies such as BRAF inhibitors, already established for BRAF-mutated CR-AC, have not yet been evaluated in CR-NEC and represent a priority for investigation, given that BRAF mutations occur in 26% of patients and are associated with significantly shorter survival. Similarly, the role of immune checkpoint inhibitors in the small subset of MSI-H CR-NEC patients warrants prospective study.
This study provides one of the most comprehensive prospective datasets available on metastatic CR-NEC and establishes a baseline for future clinical trials. Identifying the molecular mechanisms that differentiate CR-NEC's aggressive behavior from CR-AC despite shared driver mutations will be critical to developing effective, targeted therapies for this devastating cancer.