Appendiceal Neuroendocrine Tumors: Does Colon Resection Improve Outcomes?

Abstract

Background

Appendiceal neuroendocrine tumors (A-NETs) are rare neoplasms of the GI tract. They are typically managed according to tumor size; however, the impact of surgical strategy on the short- and long-term outcomes is unknown.

Methods

All patients who underwent resection of A-NET at 8 institutions from 2000 to 2016 were analyzed retrospectively. Patient clinicopathologic features and outcomes were stratified according to resection type.

Results

Of 61 patients identified with A-NET, mean age of presentation was 44.7 ± 16.0 years and patients were predominantly Caucasian (77%) and female (56%). Mean tumor size was 1.2 ± 1.3 cm with a median of 0.8 cm. Thirty-one patients (51%) underwent appendectomy and 30 (49%) underwent colonic resection. The appendectomy group had more T1 tumors (87% vs 42%, p < 0.01) than the colon resection group. Of patients in the colon resection group, 27% had positive lymph nodes and 3% had M1 disease. R0 resections were achieved in 90% of appendectomy patients and 97% of colon resection patients. Complications occurred with a higher frequency in the colon resection group (30%) compared with those in the appendectomy group (6%, p = 0.02). The colon resection group also had a longer length of stay, higher average blood loss, and longer average OR time. Median RFS and OS were similar between groups.

Conclusion

A-NET RFS and OS are equivalent regardless of surgical strategy. Formal colon resection is associated with increased length of stay, OR time, higher blood loss, and more complications. Further study is warranted to identify patients that are likely to benefit from more aggressive surgery.

Introduction

Appendiceal neuroendocrine tumors (A-NETs) are a rare subset of neoplasms of the gastrointestinal tract; however, they are the most common neoplasm of the appendix.¹ They are the second most common gastrointestinal NETs following gastric NETs and comprise approximately a fifth of all gastrointestinal NETs.² Patients with A-NETs tend to present at a younger age than patients with NETs originating elsewhere within the gastrointestinal tract.³ Although neuroendocrine tumors have the potential to secrete serotonin and other bioactive agents, A-NETs are rarely associated with carcinoid syndrome and are usually diagnosed as an incidental finding following appendectomy.⁴

A-NETs comprise a range of clinical entities including well-differentiated neuroendocrine tumors (carcinoids), mixed adenoneuroendocrine carcinomas, and poorly differentiated neuroendocrine carcinomas. Each tumor type is characterized by diverse histologic features and clinical behavior.⁵ Well-differentiated NETs typically follow an indolent clinical course and carry a favorable prognosis, though both regional and distant metastases have been documented.^{6, 7} For small and well-differentiated tumors, appendectomy is considered curative and there is a limited role for post-operative surveillance.^{8, 9}

Unlike A-NETs, mixed adenoendocrine carcinomas contain both adenocarcinoma and neuroendocrine components with malignant histology.^{5, 10} In addition to having higher rates of lymphovascular invasion and perineural invasion, they behave more aggressively than their well-differentiated carcinoid counterparts.^{11, 12} Because they have a similar prognosis to colonic adenocarcinomas, the National Comprehensive Cancer Network (NCCN) recommends treating these tumors in accordance with the published guidelines for colon cancer.^{12, 13} Poorly differentiated neuroendocrine carcinomas are the most rare subtype and carry the highest rate of metastasis and are associated with the poorest prognosis.^{11, 14, 15}

Patients thought to be at an increased risk of regional lymph node involvement are counseled to undergo a right hemicolectomy.⁴ This recommendation has been traditionally based on tumor size, although more recently there has been significant debate regarding prognostic factors and the appropriate surgical strategy for A-NETs.⁴ Consequently, National and international societies including the North American Neuroendocrine Tumors Society (NANETS), European Neuroendocrine Tumor Society (ENETS), and NCCN have published guidelines regarding A-NET management.^{13, 16, 17} These societies advocate offering right hemicolectomy to patients with tumors measuring ≥ 20 mm, as well as those with lymphovascular invasion, positive resection margins, involvement of the mesoappendix, and atypical or mixed histology.^{13, 16, 17}

There has been increased scrutiny regarding these recommendations with multiple groups advocating for more limited resections, especially for patients with well-differentiated tumors, as their prognosis is excellent regardless of regional lymph node involvement.¹ Additionally, recent studies have suggested that even for mixed tumors, more extensive surgery with lymphadenectomy does not impact survival.¹ As a result, the benefit of these more aggressive surgical resections with regional lymphadenectomy on survival in patients with A-NET remains unclear. Given the potential for increased morbidity of colectomy, particularly when performed as a second operation following index appendectomy, we sought to determine its clinical impact.

Methods

The US Neuroendocrine Tumor Study Group (US-NETSG) is a collaborative effort between 8 institutions: Virginia Mason Medical Center (Seattle, WA), Emory University (Atlanta, GA), The Ohio State University (Columbus, OH), Stanford University (Palo Alto, CA), Vanderbilt University (Nashville, TN), University of Michigan (Ann Arbor, MI), University of Wisconsin (Madison, WI), and Washington University in St. Louis (St. Louis, MO). The study group was established with the goal of assessing clinicopathologic and outcome data on the full spectrum of resected neuroendocrine tumors using data from academic and high-volume referral centers. Institutional Review Board (IRB) approval was obtained at each institution.

All patients with A-NETwho underwent surgical resection of a primary tumor between January 1, 2000, and December 31, 2015, were included. Investigators from each institution collected baseline demographic, perioperative, and pathologic data through a careful review of the medical record. Tumor staging was assigned according to the American Committee on Cancer (AJCC) 7th edition guidelines.¹⁸ Recurrence and survival data were obtained from the medical record, and survival data were verified though Social Security Death Index and publicly available obituaries. Patients were stratified by resection type: appendectomy (laparoscopic or open) and formal colon resection (ileocecectomy or right hemicolectomy, laparoscopic or open). Those patients who underwent extended colectomies (total abdominal with or without proctectomy) for reasons other than A-NET were excluded from further analysis as these were considered cases where the A-NET was likely incidentally identified or patient had other pathology driving extents of resection. In addition, those undergoing non-curative or debulking procedures were excluded.

Statistical analysis was conducted using the MEDCALC software version 18.0.0. Categorical variables were compared using chi-squared analysis, and 2-tailed t tests were used to compare continuous variables. Primary outcomes for survival analysis consisted of recurrence-free survival (RFS) and overall survival (OS). RFS was measured from time from curative-intent resection to disease recurrence, death, or last documented follow-up in patients who had macroscopically negative (R0 or R1) final resection margins. OS was measured from the time of resection to death or last follow-up. Kaplan-Meier survival plots for RFS and OS were constructed to compare patients according to resection strategy using a log-rank test. A p value ≤ 0.05 was considered statistically significant.

Results

A total of 61 of the 2182 patients in the US-NETSG database underwent curative-intent resection of an A-NET. Table 1 details demographic data for this cohort. Females comprised 55.7% (n = 34) of the cohort and 77.0% were Caucasian (n = 47). Average age was 44.7 (± 16.0) years. The majority of patients presented with abdominal pain at presentation (n = 40, 65.6%).

Table 1.

Demographic characteristics of 61 patients with A-NET

Baseline variable	n (%)
Age (years), mean ± STD	44.7 ± 16.0
Female	34 (55.7)
BMI, mean ± STD	27.8 ± 15.8
Race
White	47 (77.0)
Black	11 (18.0)
Other	3 (4.9)
Functional status
Independent	58 (95.1)
Partially dependent	3 (4.9)
Smoking history	6 (9.8)
Diabetic	1 (1.6)
Presented with pain	40 (65.6)
Type of operation
Appendectomy	31 (50.8)
Open	13 (21.3)
Laparoscopic	18 (29.5)
Colon resection	30 (49.2)
Ileocecectomy	4 (6.6)
Open	2 (3.3)
Laparoscopic	2 (3.3)
Right hemicolectomy	26 (42.6)
Open	11 (18.0)
Laparoscopic	15 (24.6)

Thirty-one patients (50.8%) underwent appendectomy (AP); 18 were laparoscopic and 13 were performed open. Thirty (49.2%) underwent formal colon resection (CR). The colon resection group comprised 4 ileocecectomies (2 laparoscopic, 2 open) and 26 right hemicolectomies, including 11 open operations and 15 laparoscopic operations. A-NETs were incidental findings in the majority of cases in both groups (AP n = 27, 90.3%, vs CR n = 23, 76.7%, p = 0.34). Six of the right hemicolectomies were performed as a second operation following an appendectomy; these patients were included in the formal colon resection group only.

Body mass index was similar between groups (AP 27.4 ± 7.2 vs CR 29.3, p = 0.44). Gender distribution was also similar between groups with males comprising 38.2% (n = 12) of AP patients versus 50.0% (n = 15) of CR patients (p = 0.44). The AP group had more T1 tumors, defined as ≤ 20 mm (AP n = 27, 87.1%, vs CR n = 13, 41.9%, p < 0.01); however, the median tumor size was similar (AP 0.6 cm (IQR 0.40–1.04) vs CR 0.95 cm (IQR 0.42–2.5), p = 0.24).

The majority of patients in both groups had well-differentiated tumors (AP n = 28, 90.3%, vs CR n = 24, 80.0%, p = 0.15). The CR group had a higher percentage of mixed tumors and poorly differentiated tumors (n = 6, 20.0%, and n = 1, 3.3%, respectively) compared with the AP group (n = 2, 6.5%, and n = 1, 3.2%, respectively; p= 0.18). Of patients in the CR group, 8 (26.7%) had nodal disease present. Of the 8 patients with nodal disease, 1 had a mixed tumor and 7 had well-differentiated tumors. One tumor was T1, 4 were T2, 1 was T3, and 1 was T4. Rate of nodal disease by T stage was 8.3% for T1, 57.1% for T2, 20.0% for T3, and 100.0% for T4, with an overall nodal disease rate of 42.9% for tumors classified as T2 or greater.

R0 resections were achieved in 90.3% of AP patients and 96.7% of CR patients. The remaining 9.7% of AP patients had R1 resections whereas one CR patient had R2 disease secondary to positive resection margins on the mesentery. All 6 patients who underwent colon resection following appendectomy achieved R0 resections. Four of them were found to have stage 2 disease on final pathology and the remaining 3 had stage 3 disease. Of the patients in the AP group who had known tumors prior to surgery, all 3 had R0 resections and had T1 tumors on final pathology. The 8 patients in the CR group who had A-NETs diagnosed preoperatively had R0 resections. Two of these patients were found to have stage 3 mixed tumors. The remaining 6 patients had well-differentiated NETs, 2 patients had stage 3 disease, and 4 patients had stage 2 disease (Table 2).

Table 2.

Clinicopathologic characteristics of patients with A-NET

Clinical factor	Appendectomy, n = 31	Colon resection N = 30	p
Age	40.4 ± 14.7	48.9 ± 16.0	0.04
T1 tumor	87.1%	50.0%	< 0.01
Male gender	38.7%	50.0%	0.44
BMI (kg/m2)	27.4 ± 7.2	28.2 ± 10.2	0.75
Size (cm)	0.9 ± 0.8	1.6 ± 1.7	0.27
ANET incidental finding	90.3%	76.7%	0.34
LVI	23.1%	40.0%	0.18
PNI	40.0%	37.5%	> 0.99
Well-differentiated	90.3%	80.0%	0.15
Mixed tumor	6.5%	18.7%	0.26
Poorly differentiated	3.2%	2.9%	> 0.99
Nodal disease present	N/A	26.7%
M1 disease	0%	6.3%	0.49
Resection status
R0 resection	90.3%	96.7%
R1 resection	9.7%	0%	0.49
R2 resection	0%	3.3%
Complications	6.5%	30.0%	0.02
Readmission	6.5%	3.3%	> 0.99
Reoperation	3.2%	3.3%	> 0.99
Median length of stay (days)	2	4	< 0.01
Estimated blood loss (cm3)	75 ± 91	153 ± 150	0.05
Operating time (min)	101 ± 61	177 ± 86	< 0.01

Blood loss was significantly higher in the CR group compared with that in the AP group (153 ± 150 cm³ vs 75 ± 91 cm³, p = 0.05). Operating time was significantly longer in the CR group (177 ± 86 min vs 101 ± 61 min, p < 0.01). Perioperative complications occurred in a total of 11 patients with an overall complication rate of 19.7%. Surgical complications occurred with a higher frequency in the CR group (CR 9 patients, 30.0%, vs AP 2 patients, 6.5%, p = 0.02). Both complications in the AP group were from open appendectomies. Table 3 details the Clavien-Dindo classification of complications in both groups.

Table 3.

Clavien-Dindo classification of perioperative complications

Complication grade	Appendectomy (n = 31)	Colon Resection (n = 30)
	n (%)	n (%)
I	0	1 (3.3)
II	0	3 (10.0)
IIIa	0	2 (6.7)
IIIb	1 (3.2)	1 (3.3)
IV	1 (3.2)	2 (6.7)
Total	2 (6.5)	9 (30.0)

Readmission rates were similar with 2 patients in the AP group requiring readmission and 1 patients in the CR group requiring readmission within 30 days of surgery (p > 0.99). One patient in each group (3%) required reoperation secondary to surgical complications (p > 0.99). All patients in both groups were discharged to home. Median length of stay was significantly longer in the CR group (4 days vs 2 days, p < 0.01).

Median RFS and OS were identical and not different between groups (106 months vs 124 months, p = 0.72). (Fig. 1) One patient in the CR group with a T3 mixed tumor with direct extension into the mesentery was treated with adjuvant therapy. No somatostatin analogues were used in either the neoadjuvant or adjuvant setting.

Fig. 1.

Kaplan-Meier survival curves for patients with appendiceal carcinoids. There was no difference between the appendectomy and colectomy groups (106 months vs 124 months, p = 0.72)

Discussion

The suggested treatment for ANETs has been aggressive surgery with lymphadenectomy for patients thought to have a higher risk of regional lymph node involvement.⁴ However, even in the setting of positive lymph nodes and atypical histology, evidence of additional benefit from formal colon resection with lymphadenectomy in the treatment of A-NET remains elusive.¹

In 1987, Moertel et al. suggested that in addition to being adequate surgical management of small A-NETs, appendectomy may be appropriate for tumors especially inthe elderly and in higher risk operative candidates.⁸ Based on findings of no recurrences or metastases in patients with A-NET after a median follow-up of over 26 years, they proposed that right hemicolectomy should be reserved for young patients with higher risk tumors, such as those with vascular involvement and invasion of the mesoappendix.

Rault-Petit et al. evaluated a large multi-center cohort of 403 patients with A-NETand demonstrated that A-NETs with size ≥ 2 cm, lymphovascular invasion, and perineural invasion harbored the greatest risk of regional lymph node positivity.⁴ Additionally, they called into question some of the other factors that had previously been thought to be associated with a higher risk of lymph node involvement, including tumor grade and invasion of the mesoappendix. Despite presence of positive lymph nodes, right hemicolectomy did not appear to improve survival.

Similarly, using the SEER database, Mehrvarz Sarshekeh et al. demonstrated excellent prognosis in patients with node-positive, well-differentiated A-NET.¹ Moreover, there was no difference in cancer-specific survival in patients with node-positive A-NET regardless of resection strategy, suggesting that formal colon resection with regional lymphadenectomy may not necessarily improve survival. Additionally, they found that mixed adenoneuroendocrine tumors with lymph node metastases had worse survival despite more aggressive surgical strategies. As a result, they concluded that right hemicolectomy did not improve outcomes and cautioned against reflexively offering right hemicolectomy to patients even in the setting of locally advanced tumors with large size or regional lymph node metastases. Instead, they advocated for a more conservative approach with careful weighing of the potential risks and benefits of each therapeutic approach.

Consistent with the current body of literature, we found that patients had similar RFS and OS regardless of surgical strategy. Multiple studies have documented that mixed tumors are more aggressive than well-differentiated carcinoid tumors, and behave more similarly to colon cancers.^{11, 12} It may be that patients with these tumors represent a population of patients that benefit from colon resection rather than appendectomy. In our study, both groups consisted of a dominant majority of well-differentiated tumors and although the colon resection group had more mixed tumors, this difference did not reach statistical significance. Given the small sample size of 61 patients, it is possible that our study was underpowered to demonstrate this difference. Additionally, the colon resection group included a statistically significantly smaller number of T1 tumors than did the appendectomy group. As a result, our similar long-term outcomes may reflect that more extensive surgery benefitted patients with mixed tumors and those with larger tumors. Additionally, Mehrvarz Sarshekeh et al. advocated for studies evaluating the role of adjuvant chemotherapy in the treatment of locally advanced and biologically aggressive tumors.1 The mixed tumors and neuroendocrine carcinomas comprised a minority of A-NETs in our study, and it is possible that patients with these neoplasms would have benefitted from systemic chemotherapy rather than from upfront resection.

A-NET was an incidental finding in the majority of patients in both groups. Of the 30 patients in the colon resection group, 6 were patients who had incidental findings of A-NET on appendectomy and subsequently went on to have a right hemicolectomy; unfortunately, the database did not contain the margin status of the appendectomy specimens. All 6 of those patients ultimately achieved R0 resections. The R1 resection rate in the appendectomy group for our study was 10%, similar to the 8% found in the large French national study.⁴ None of the R1 patients in the appendectomy group subsequently underwent a formal colon resection, whereas 72% of patients with R1 resections in the French study underwent a second operation. Given our findings of similar RFS and OS between colon resection and appendectomy groups, it is unclear whether a second operation would have added appreciable benefit to long-term outcomes in these patients.

Just under half of patients in our study underwent colon resection. In addition to the 26 right hemicolectomies, there were 4 ileocecectomies that were performed. All 4 of them had incidental findings of A-NET and the database lacked the granularity to determine the indication for ileocecectomy. Ileocecectomy is a documented strategy for addressing R1 A-NET resections; however, Rault-Petit et al. cautions against this approach as it does not allow lymphadenectomy while still subjecting the patient to a similar morbidity profile.⁴

There were some differences in short-term outcomes between groups. Although the readmission and reoperation rates were low, and similar between groups, there were notable differences in other clinical outcomes. Colon resections were associated with longer operative times, increased blood loss, and longer hospital stays. Additionally, there was an almost 5-fold higher rate of perioperative complications in the colon resection group which reached statistical significance. Our findings suggest that a more conservative surgical approach may provide similar survival outcomes while offering a more favorable complication and morbidity profile.

This study has a number of limitations. First, it is limited by a small sample size despite the multi-institutional design including 8 tertiary academic centers. Additionally, there was no standardization of the evaluation and management of patients with A-NET. The dataset does not provide the intent of the clinical reasoning or factors that influenced the decision to offer appendectomy versus colon resection. Finally, the retrospective design introduces challenges associated with capturing disease-specific survival.

In conclusion, management strategies for A-NET include appendectomy and formal colon resection with lymphadenectomy. In concordance with the current literature, our findings affirm the findings that long-term outcomes are not improved with more radical surgical resection strategies. Rather, we advocate for an individualized approach to patients with A-NET, carefully balancing potential treatment morbidity with the risk of recurrence and mortality. Further studies with help to elucidate what population of patients stand to benefit from implementation of more aggressive surgical strategies in the management of A-NET.