Neuroendocrine neoplasms (NENs) generally refer to all heterogeneous tumors originating from peptidergic neurons and neuroendocrine cells, which can occur in many organs and tissues of the body. The 2019 World Health Organization (WHO) classification of Tumors of the digestive system updated the digestive system NENs classification (1). NENs were divided into neuroendocrine tumors (NET) G1, NET G2, NET G3, neuroendocrine carcinoma (NEC) and mixed neuroendocrine-non-neuroendocrine neoplasm (MiNEN). In recent years, the incidence of gastroenteropancreatic neuroendocrine neoplasms (GEP-NENs) has gradually increased, among which the incidence of colorectal NENs (CRNENs) has increased 10 times (2, 3). It may be due to the popularity of abdominal radiography and endoscopy (3, 4). Most CRNENs are nonfunctional (5), whose clinical symptoms are similar to those of typical colorectal cancer.

CRNENs are more common at stage I, which can often be treated by endoscopy with a good prognosis (6, 7). The European Neuroendocrine Tumor Society (ENETS) and Union for International Cancer Control/American Joint Committee on Cancer (UICC/AJCC) guidelines classify locally advanced and metastatic NENs as stage II-IV. A previous study identified 24.8% of patients with CRNENs at stage II-IV (8). Even though they constitute only a small percentage of diagnosed CRNENs, they pose high mortality risk due to their aggressiveness and high malignancy. Among those patients with a median overall survival (OS) of 9 months to 13.2 months, the 3-year OS rate ranged from 5.9%-36.2%, and the 5-year OS rate ranged from 4.1%- 29.8% (9–11). For CRNENs, the AJCC TNM staging system of 2010 and the WHO’s histological classification system were crucial prognostic systems (12, 13). Nevertheless, there are many other important prognostic factors that may also influence individual outcomes with CRNENs, such as age at diagnosis, sex, race, tumor location, grade, treatment and so on (14–17). Therefore, it is a clinical demand to establish more capable prognostic predictive models for stages II-IV CRNENs. Nomograms are currently widely used for individualized prediction of patient outcomes, and online tools of established nomograms expanded their convenience for clinical application. Different predictors and determinants can be combined to create a nomogram that can generate individuals’ digital probabilities of clinical events, thus meeting our requirements for biological and clinical integrated models and personalized medicine (18).

As there are few CRNENs-specific reports, this study examined the incidence trend for all stage CRNEN patients and identified independent prognostic factors for stage II-IV patients diagnosed from Surveillance, Epidemiology, and End Results (SEER) cancer registry. Following that, we aimed to create prognostic nomograms to provide patients and clinicians with valuable prognostic information about stage II-IV CRNENs and risk factors increasing CRNENs-related death.

The CRNENs are heterogeneous neoplasms characterized by neuroendocrine secretory granules in the cytoplasm of the cells, but its pathogenesis is still unclear. As little information is available on the incidence and survival of CRNENs, it is difficult to assess its prognosis. In this study, we assessed the incidence and survival outcomes of patients with stage II-IV CRNENs from both the SEER database and external real-world cohort as well as developed relevant prognostic nomograms and convenient online tools for OS and DSS prediction.

In our study of the SEER dataset, even though it is a rare disease, the incidence of CRNENs increased from 0.31 to 1.69 parts per million over the past 44 years. 20.7% of patients with CRNENs were 75 years old or older. We observed that the prognosis of OS in elderly patients was poor, while the impact of age on CRNENs’ specific death failed to show significance. In the SEER database, the grade is classified as highly differentiated (G I), moderately differentiated (G II), poorly differentiated (G III), and undifferentiated (G IV), which are determined primarily according to the level of ki-67 index and are deemed as important prognostic factors of OS and DSS (19, 20). NET G3 and NEC were highly invasive and had a poor prognosis. More than half of the patients had distant metastasis at the time of diagnosis (9, 21). In this study, we also explored the prognostic value of specific metastatic location for CRNENs, which showed that brain and liver metastasis were independent risk indicators of shorter OS outcome and would increase the risk of CRNENs-specific death, while bone and lung metastasis failed to show significance in multivariate analysis. Brain metastasis in neuroendocrine tumors is rare with an incidence of 1.5-5% (22). In our analysis, brain metastasis was only observed in 1.0% of patients with stage II-IV CRNENs. While liver metastasis was the most common metastatic organ which was consistent with our study. In another study, liver metastasis was observed in 72.84% of stage IV gastrointestinal neuroendocrine neoplasm and was associated with a higher risk of concurrent, while the significance of brain metastasis failed to be observed (23). We also found that 65.7% of the stage II-IV CRNENs occurred in the colon, but the location of the tumor had no correlation with the prognosis, which was different from the results previously reported (24, 25). Many patients were initially diagnosed with CRNENs in an advanced stage, our study showed that 45.4% of patients were in stage IV, which was similar to the previous study (26). Our study also confirmed the robust prognostic ability of the TNM staging system in CRNENs.

Recently, competing risk analyses have received increasing higher popularity in the clinical research of cancer-specific death (27). Compared to traditional Cox regression analysis, it introduced the concept of competitive risk and avoid the impact of death due to other causes to the final result (28). In this study, the competing risk model was used to analyze the prognostic factors of death specifically due to CRNENs. Our results showed that, except age, all other risk factors of OS including grade, stage, surgery, brain metastasis, and liver metastasis were found to increase the risk of CRNENs-specific death. The nomogram of DSS and relative online tools were also established successfully.

In terms of treatment, surgery can improve the prognosis of stage II-IV stage CRNENs. Although radiotherapy and chemotherapy have always been the treatment for patients with unresectable primary tumors (29), our analysis showed that they did not significantly prolong the OS outcomes of patients with CRNENs or decrease the risk of CRNENs-specific death. The treatment of CRNENs is determined according to the degree of differentiation and staging of the tumor. CRNENs with a diameter of less than 1 cm, being limited to the submucosa and without distant metastasis, can be resected locally under an endoscope or through the anus (30, 31). As a result of the risk of regional lymph node metastasis, radical surgery is recommended for CRNENs with locally advanced stages (32). However, the therapeutic value of surgery is still debated in patients with advanced tumors. In previous studies, palliative resection has been suggested to be a viable option for patients experiencing obstructions, bleeding, perforations, etc (31). Nevertheless, many studies suggested that palliative surgery can improve survival rates of CRNENs by reducing tumor burden (11, 21, 33). Moreover, simultaneous surgical treatment is recommended for patients with liver metastasis in order to improve long-term survival (29, 34). These are consistent with our findings. The following are some possible reasons. To begin with, the efficacy of chemotherapy for advanced G1, G2, G3 and NEC is unsatisfying with an objective response rate of less than 50%, and the options for chemotherapeutic agents were rather limited (29, 35). Secondly, the use of radiotherapy is only recommended for patients with locally advanced and advanced rectal neuroendocrine tumors (36, 37). It is believed that conventional radiotherapy can reduce the local recurrence rate but will not improve survival rates (37). Thirdly, confounding factors existed inevitably in this study, due to the inclusion of all pathological classifications, grades, and stages of CRNENs, the sensitivity of radiotherapy and chemotherapy might vary according to the stage, grade, and location of the tumor. In this study, the detailed radiation modality is unavailable for us to perform further analysis. In previous studies, the effectiveness of peptide receptor radionuclide therapy (PPRT) has been reported to have an impact on the prognosis of CRNENs (38, 39).

In recent years, some novel treatments such as somatostatin analogue (SSA) and targeted therapy are found to show efficacy in the treatment of CRNENs. SSA has been proven to inhibit tumor growth and prolong tumor progression-free survival (40). The effect of SSA in the treatment of NENs is mainly to control symptoms and stabilize tumor progression. SSA acts on somatostatin receptors on the surface of tumor cells to inhibit functional NENs secreting peptides and growth hormone (41, 42). Although the mammalian target of rapamycin (mTOR) inhibitor everolimus and multi-targeted kinase inhibitor sunitinib has been approved for locally advanced and metastatic pancreatic NENs in clinical practice (43), the effect of its application in colorectal neuroendocrine tumors still needs to be observed (44).

Some limitations remained in this study. Firstly, the SEER database lacks neuroendocrine biomarkers information like chromogranin A (CgA), synaptophysin (Syn), and CD56, which are important prognostic factors (45). In addition, apart from the basic treatment records, the SEER database does not contain any additional information about operation mode, chemotherapy regimen, radiation dose, tumor shape, various health statuses, or socioeconomic factors that might influence survival (35, 46). Since these parameters could not be evaluated and integrated into our nomogram, further studies should try to include these valuable factors and assess their values. Secondly, due to the retrospective nature of this study, there is the possibility of selection bias among the patients we enrolled. Larger prospective studies are warranted to verify our results.

Over the past 40 years, the incidence of CRNENs increased steadily, along with improved survival outcomes. Grade III-IV, higher TNM stage, liver metastasis, brain metastasis, and without receiving surgery were found to be associated with worse OS and DSS. Advanced age was a risk factor for OS but not DSS. Nomograms for patients with stage II-IV stage CRNENs are capable of predicting the 1-, 3-, and 5-year OS and DSS with high accuracy, and realize risk stratification. Besides, two online tools regarding OS and DSS prediction were established, facilitating nomogram score calculation, risk group determination, as well as survival prediction for each individual patient.

The original contributions presented in the study are included in the article/supplementary material. Further inquiries can be directed to the corresponding author.

The studies involving human participants were reviewed and approved by the ethics committee of the Cancer Hospital of the Chinese Academy of Medical Sciences & Peking Union Medical College. The patients/participants provided their written informed consent to participate in this study. Written informed consent was obtained from the individual(s) for the publication of any potentially identifiable images or data included in this article.

QL contributed to the conception and design of the study. FZ, LH, and ZW analyzed the data and drafted the manuscript. FW and TX checked all data and results. All authors contributed to the article and approved the submitted version.