Palliative primary tumor resection may not offer survival benefits for patients with unresectable metastatic colorectal neuroendocrine neoplasms, one multicenter retrospective cohort study

Background The efficacy of palliative primary tumor resection (PTR) in improving prognosis for patients with unresectable metastatic colorectal neuroendocrine neoplasms (NENs) has not been fully explored. Methods We performed one retrospective cohort study and recruited 68 patients with unresectable metastatic colorectal NENs from two Chinese medical centers between 2000 and 2022. All patients were assigned to PTR group and no PTR group. The clinicopathological manifestation data were carefully collected, and the survival outcomes were compared between the two groups using Kaplan–Meier methods. Propensity score matching (PSM) was conducted to minimize confounding bias. Univariate and multivariate Cox proportional hazards regression analyses were performed to identify prognostic factors. Results A total of 32 patients received PTR, and the other 36 patients did not. The median progression-free survival (PFS) and overall survival (OS) times were 4 and 22 months in the whole cohort, respectively. For patients who received no PTR, the median OS was 16 months, and the 1-year OS rate and 3-year OS rate were 56.4% and 39.6%, respectively. For patients who received PTR, the median OS was 24 months, and the 1-year OS rate and 3-year OS rate were 67.9% and 34.1%, respectively. However, the Kaplan–Meier survival curves and log-rank test demonstrated no significant survival difference between the two groups (P = 0.963). Moreover, palliative PTR was also not confirmed as a prognostic factor in subsequent univariable and multivariable Cox proportional hazards regression analyses in both the original and matched cohorts. Only histological differentiation was identified as an independent prognostic factor affecting PFS [hazard ratio (HR) = 1.86, 95% confidence interval (CI): 1.02–3.41, P = 0.043] and OS [HR = 3.70, 95% CI: 1.09–12.48, P = 0.035] in the original cohort. Conclusions Palliative PTR may not offer survival benefits for patients with unresectable metastatic colorectal NENs.


Introduction
Colorectal neuroendocrine neoplasms (NENs) are rare diseases arising from the diffuse neuroendocrine system of the colon and rectum.Although colorectal NENs constitute less than 1% of all colorectal tumors, a rapid increase in their incidence has been observed in recent years, owing to the popularization of colorectal cancer coloscopy programs [1,2].Colorectal NENs are a heterogeneous group of diseases with varying clinical manifestations and malignant potentials, ranging from diminutive, indolent, and early NENs with favorable prognosis to huge, highly aggressive, and metastatic NENs with dismal oncological outcomes that depend on their histological grade and differentiation [3,4].
The incidence rate of distant metastasis was approximately 11-14% for all colorectal NENs at initial diagnosis [5,6].However, the probability of metastatic disease varies widely from G1 to G3 NENs.For G1 and G2 NENs, they were only 0.3% and 6.3%, respectively [7].However, for G3 NENs, more than half can have distant metastasis at diagnosis [8,9].Tumor grade, histological differentiation, depth of tumor invasion, and size were risk factors for distant metastasis [10,11].The liver is the most common organ of metastases, followed by distant lymph nodes, peritoneum, lung, and bone [8,9].More than half of metastatic patients show multiple metastases that cannot be resected radically [12].
Numerous previous studies have explored the efficacy of primary tumor resection (PTR) in prolonging survival for unresectable colorectal adenocarcinomas, leading to controversial conclusions.Most retrospective studies have suggested that PTR may offer survival benefits in carefully selected patients [13][14][15].However, a randomized clinical trial from Japan showed that PTR followed by chemotherapy did not offer any survival benefit over chemotherapy alone [16].However, there is a paucity of data on the therapeutic effects of PTR for the management of unresectable metastatic colorectal NENs, and optimal therapeutic strategies have not been well established for these patients [17].This retrospective cohort study aims to explore whether PTR can offer survival benefits for patients with uncurable metastatic colorectal NENs.

Patients
Our study received approval from the ethics committee of the National Cancer Center and followed the rules of the Helsinki Declaration of the World Medical Association.Informed consent was obtained from all participants.A total of 68 consecutive patients were included between 2000 and 2022, with 60 from the National Cancer Center/Cancer Hospital, Chinese Academy of Medical Sciences, and the other 8 from the Strategic Support Force Medical Center.All patients were histologically diagnosed with colorectal NENs through pathological reviews and immunohistochemical examinations.All patients were definitely diagnosed with metastatic NENs that could not be resected with curative intent through imaging examinations at the initial date of diagnosis.We designed a retrospective cohort study; 32 patients received palliative resection of primary colorectal NENs, and the other 36 patients did not undergo palliative surgery.Data including demographic information, clinicopathological features, and survival outcomes were obtained through medical records and telephone calls.The last date of follow-up was July 30, 2023.The primary outcomes of interest were progression-free survival (PFS) time and overall survival (OS) time.The PFS was calculated between the data of initial diagnosis and cancer progression confirmed by imaging evaluation.OS was obtained between the initial diagnosis and cancer-specific mortality data.

Statistical analysis
All statistics were performed using the Statistical Package for the Social Sciences (SPSS version 24.0,IBM Corp., Armonk, NY, United States).Quantitative data are presented as the means ± standard deviations (SD) and were compared using t tests if they followed a normal distribution.Quantitative data that did not follow the normal distribution are shown as medians and ranges and were compared through Mann-Whitney U tests.Qualitative and ordinal data were described as frequencies with percentages and were compared through χ 2 tests for qualitative variables and Mann-Whitney U tests for ordinal variables.Propensity score matching (PSM) was performed by fitting a logistic regression model and setting the caliper at 0.1.One-to-one pair matching was performed without replacement, and 26 matched pairs were selected.The covariates included gender, age, Eastern Cooperative Oncology Group (ECOG) score, the primary tumor location, grade, differentiation, TNM T stage, TNM N stage, site of metastases, cycles of chemotherapy.PFS and OS rates were calculated from the Kaplan-Meier survival curves and compared using the log-rank test.Univariable and multivariable Cox proportional hazards regression analyses were performed to identify the independent risk factors affecting clinical outcomes.
The differences in the clinicopathological variables between patients received PTR and those who did not receive PTR before and after PSM are presented in Table 2. Statistical analysis demonstrated that the two groups were comparable with regard to the demographics and clinicopathological variables in both the original cohort and the matched cohort.

Survival outcomes
A median follow-up of 15.5 months (range 1-96 months) was achieved in our study.Of the 68 cases, 6 were lost to follow-up due to loss of communication or unexpected death from other accidents, resulting in a follow-up rate of 91.2%.In the entire cohort, the median PFS was 4 months, and the 1-year PFS rate and 3-year PFS rate were 25.7% and 13.8% in the cohort, respectively (Fig. 1A).With regard to OS, the median OS was 22 months, and the 1-year OS rate and 3-year OS rate were 61.9% and 36.6%,respectively (Fig. 1B).For patients who received PTR, the median PFS was 4 months, with 1-year and 3-year PFS rates of 23.6% and 17.7%, respectively.The median OS was 24 months, with 1-year and 3-year OS rates of 67.9% and 34.1%, respectively.For patients who did not receive PTR, the median PFS were also 4 months, with 1-year and 3-year PFS rates of 20.5% and 10.2%, respectively.The median OS was 16 months, with 1-year and 3-year OS rates of 56.4% and 39.6%, respectively.The Kaplan-Meier survival curves showed that there was no significant difference in terms of PFS (P = 0.545) and OS (P = 0.963) between the patients who underwent PTR and those who did not (Fig. 1C and D).Kaplan-Meier survival curves were then plotted in the matched cohort, and no survival benefit for PFS (P = 0.585) and OS (P = 0.983) was observed for patients who received PTR (Fig. 1E and  F).
Given the histological differences between NET and NEC, we further performed subgroup analysis based on histological differentiation.In the NET group, the median PFS was 9 months, with 1-year and 3-year PFS rates of 38.0% and 22.5%, respectively.The median OS was 62 months, with 1-year and 3-year OS rates of 96.3% and 78.8%, respectively.The NEC subgroup had significantly worse PFS (P = 0.005) and OS (P < 0.001) than NET subgroup.In the NEC group, the median PFS was only 2 months, with 1-year and 3-year PFS rates of 10.5% and 7.0%, respectively.The median OS was only 9 months, with 1-year and 3-year OS rates of 40.4% and 9.3%, respectively (Fig. 2A and B).The Kaplan-Meier survival curves showed that there was no significant difference in terms of PFS and OS between the patients who underwent PTR and those who did not both in the NET subgroup (Fig. 2C and D) and the NEC subgroup (Fig. 2E  and F).

Univariable and multivariable analysis of prognostic factors
Univariable and multivariable Cox proportional hazards regression analyses were conducted in both the original cohort and the matched cohort to identify significant variables affecting PFS and OS (Tables 3 and 4).In the original cohort, the univariable and multivariable analysis indicated that histological differentiation [hazard ratio (HR) = 1.86, 95% confidence interval (CI): 1.02-3.41,P = 0.043] was the only prognostic factor for PFS.In terms of OS, univariable analysis showed that patients with primary colonic NENs (P = 0.011), G3 NENs (P < 0.001) and NECs (P < 0.001) had lower median OS than patients with primary rectal NENs, G1 and G2 NENs, and NETs.However, only histological differentiation was further verified as an independent risk factor for OS [HR = 3.70, 95% CI:

Discussion
Based on the current results of clinical studies, systematic chemotherapy is the mainstay of treatment for patients with metastatic colorectal cancer (mCRC) and the most reliable choice to prolong OS [18].Surgical resection with curative intent is only performed for patients with solitary metastasis, which account for only 10-15% of all mCRC [19].For other mCRC, surgery is only indicated in dealing with emergent accidents of obstruction, bleeding and perforation.In recent years, many studies have explored the value of palliative PTR in improving survival.Although their results were controversial, most of these studies reported prolonged OS obtained from palliative PTR [20][21][22][23].However, nearly all data concerning PTR of colorectal cancer were from studies of colorectal adenocarcinoma, as it is the main histological type of colorectal tumors.There remain few data exploring the survival benefits of palliative PTR for metastatic colorectal NENs.
Colorectal NENs are a group of heterogenous disease, their clinical manifestations and outcomes varied widely from G1 to G3 neoplasms.Although G1 and G2 NENs are regarded as well-differentiated and indolent disease with favorable prognosis, unresectable distant metastasis can still be found in some patients, even if the primary tumors are only small and T1 lesions [10].For colorectal NENs of G3 grade and poor differentiation, over a half of patients presented with distant metastasis at the initial diagnosis, and most of these metastatic lesions cannot be resected with curable intent [9,24].
Understanding how to prolong the OS of patients with metastatic colorectal NENs is thus an urgent issue.The median OS of all metastatic colorectal NENs were 24.8 months.However, for poorly differentiated NENs, the median OS dropped to only 8.7-10 months in literature reports [8,9,25].In our study, it was 24 months in the entire cohort and 10 months for poorly differentiated G3 NENs, which was consistent with previous reports.PFS: progression free survival, OS: overall survival, HR: hazards ratio, CI: confidence interval, NET: neuroendocrine tumor, NEC: neuroendocrine carcinoma, PTR: primary tumor resection surgical resection had offered better survival than those who had not, but whether these patients only received resection of the primary tumor, or both the primary and metastatic tumor was not detailed in this report [32].However, Smith et al. reviewed 126 cases of colorectal NECs from a single American institution and found that resection of the primary tumor had no influence on OS for metastatic disease [8].Additionally, another research from China argued against PTR for stage IV colorectal NECs [33].
To our knowledge, our study is thus far the first report specially focused on palliative PTR in colorectal NENs.Unlike prior reports that supported the decision of palliative PTR in the management of metastatic NENs, we didn't observe a correlation between PTR and improved survival.PTR may not bring superior survival for these patients.Therefore, palliative PTR should be considered carefully to avoid delay of systematic chemotherapy, especially for asymptomatic cases.However, all current results were concluded from retrospective studies, prospective randomized control trials with more sample sizes still needed to explore the efficacy of PTR in improving survival.Moreover, we observed that tumor histological differentiation was the main prognostic factor affecting patients' survival; colorectal poorly differentiated NECs presented significant worse prognosis than well-differentiated NETs.Given the significant difference in degree of malignancy, aggressiveness and clinical outcomes between NETs and NENs, it may be more reasonable to further explore the role of PTR in NETs and NECs patients alone in future reports.
Our study has several limitations.Firstly, it is retrospective in nature, as we included patients over a 20-year period, and thus, bias from patient selection and information collection cannot be entirely avoided.Secondly, although we collected patients from two Chinese medical centers, the sample size was still small, owing to the rarity of colorectal NENs.Thirdly, the majority of patients received PTR for relief of obstruction and bleeding, making it challenging to conclude whether PTR can offer a survival benefit for asymptomatic patients.

Conclusions
In conclusion, palliative PTR may lack benefits for patients with unresectable metastatic colorectal NENs, the decision to do so should be made carefully for these patients, especially for asymptomatic patients.

Fig. 1 Fig. 2
Fig. 1 Kaplan-Meier survival analyses of patients before and after PSM.(A) PFS of the whole cohort, (B) OS of the whole cohort, (C) PFS of patients who received PTR and those who did not in the original cohort, (D) OS of patients who received PTR and those who did not in the original cohort, (E) PFS of patients who received PTR and those who did not in the matched cohort, (F) OS of patients who received PTR and those who did not in the matched cohort.PTR: primary tumor resection, PSM: propensity score matching, PFS: progression free survival, OS: overall survival 1.09-12.48,P = 0.035].Palliative resection demonstrated

Table 1
Demographic and clinicopathological manifestations

Table 2
Clinicopathological manifestations between patients who received PTR and those who did not before and after PSM PTR: primary tumor resection, PSM: propensity score matching, SD: standard deviation, ECOG: Eastern Cooperative Oncology Group, NET: neuroendocrine tumor, NEC: neuroendocrine carcinoma SD: standard deviation, ECOG: Eastern Cooperative Oncology Group, NET: neuroendocrine tumor, NEC: neuroendocrine carcinoma, PTR: primary tumor resection Table 1 (continued)

Table 3
Univariable and multivariable Cox proportional hazards regression analysis for PFS and OS in the original cohort

Table 4
Univariable and multivariable Cox proportional hazards regression analysis for PFS and OS in the matched cohort