Appendiceal Well-Differentiated Neuroendocrine Tumors: A Single-Center Experience and New Insights into the Effective Use of Immunohistochemistry

Abstract

Background. Appendiceal well-differentiated neuroendocrine tumor is the most common histological type of appendiceal tumor. The majority of tumors are found incidentally at the tip of the appendix, with few exceptions. Due to its primarily indolent nature, this entity presents unique pathological challenges, particularly in the appropriate use of immunohistochemistry which this study aims to clarify. Patients and methods. Patients diagnosed at University Health Network (Canada) between 2005–2019 were selected and reviewed. Results. We identified 70 patients and sex distribution was female 60%; median age 36.5 years. Among them, 63 patients underwent appendectomy, and seven had initial right hemicolectomy for non-appendix lesions. Mean tumor size was 5.0 mm. Tumor extent was submucosa (15%); muscularis propria (34%); subserosa or mesoappendix (42%); visceral peritoneum (8%). All were clinically non-functional and negative for nodal and distant metastasis. Ninety percent of tumors were WHO Grade 1; 10% were WHO Grade 2. Immunohistochemically, an average of six stains were performed per patient. Nearly all tumors were positive for chromogranin A, synaptophysin, CAM5.2, and CDX2. MIB-1 staining was < 3% in 58/63 tumors. Other immunohistochemical stainings performed were hormonal markers (serotonin, glucagon, pancreatic peptide, peptide YY). Subsequent right hemicolectomy was performed on five patients. All were followed up (median 4 years 8 months), and all were alive without recurrence except for one patient who died of another comorbidity. Conclusion. Tumors that are small, localized, and of low grade can be reasonably exempt from an extensive immunohistochemical panel in the absence of non-typical clinical and morphological features.

Introduction/Background

Appendectomy is the most frequently performed surgery today, ¹ which typically follows an emergency room visit with acute abdominal pain. An appendectomy may also be performed on patients without acute appendicitis-related symptoms, in conjunction with gynecological surgery procedures and right hemicolectomies for cecal or right colon lesions. The incidence of appendiceal tumors in an appendectomy specimen is low, and most of these tumors are clinically undetected. ² As a result, many patients diagnosed with incidental appendiceal tumors lack clinical staging information such as size and lymph node/distant metastasis at initial diagnosis, and the treatment decision weighs heavily on the initial pathology report.

Appendiceal well-differentiated neuroendocrine tumor accounts for more than half of appendiceal tumors and is the most common histological type. ³ Most are incidental, small, and negative for hormonal symptoms and have an excellent 5-year and long-term prognosis. ⁴ Major international guidelines have listed several risk factors for lymph node metastasis. European Neuroendocrine Tumour Society (ENETS) and the North American Neuroendocrine Tumour Society (NANETS) guidelines state tumor size as the most crucial factor, with tumors > 2 cm at increased risk of lymph node metastasis and should be followed by subsequent right hemicolectomy.^5,6 Other risk factors include the extent of the tumor, pathological tumor stage, presence of lymphovascular invasion, and tumor grade.^7,8 Given the general indolent nature and the younger demographic of appendiceal well-differentiated neuroendocrine tumors compared to other common sites of well-differentiated gastropancreas-intestinal neuroendocrine tumors (pancreas, rectum, small intestine, stomach), ⁹ aggressive treatment options should be limited to select patients at high-risk.

While many case reports and single-center pathology reviews have contributed to recognizing this incidental lesion,^10–15 including the most recent pathology review by Noor et al, ⁸ the practical immunohistochemical panel based on risk assessment has not been thoroughly examined to date.

Here, we reviewed clinicopathological information among 70 consecutive appendiceal well-differentiated neuroendocrine tumors in a single tertiary institution and aimed to provide practical information on the utility of immunohistochemistry.

Material and Methods

An institutional review board (University Health Network, REB 21-5299) approved this study. Two pathologists performed a retrospective review of pathology and clinical charts. Pathology data were retrieved from CoPath, a pathology information system, and clinical data from EPR, an electronic medical records system. All patients diagnosed with appendiceal well-differentiated neuroendocrine tumors from 2005 to 2019 at University Health Network were included. Patients were either treated with appendectomy for clinically diagnosed appendicitis or as appendectomy as part of surgery for other reasons. All patients were anonymized. Poorly-differentiated neuroendocrine tumors and neuroendocrine carcinoma were not included in our series. Patients with concurrent appendiceal carcinoma and goblet cell adenocarcinoma were also excluded from our study. Patient sex, age, clinical presentation, diagnostic imaging, operative details, surgical pathology, and clinical outcomes were reviewed.

Statistical Analysis

Statistical analysis was performed using GraphPad Prism 7.00. Briefly, comparisons of categorical data were performed using Fisher's exact test in 2 × 2 contingency tables. The metric variables such as age (years) and tumor size (cm) were described using median and range, and between-group comparisons were performed using the nonparametric Mann–Whitney U test. All statistical tests were two-sided; p values below 0.05 were considered significant. * P < 0.05.

Results

The number of appendectomies performed at our institution from 2005 to 2019 was 4110, and the proportion of appendix containing appendiceal well-differentiated neuroendocrine tumors was 1.7%. Clinicopathological features of 70 appendiceal well-differentiated neuroendocrine tumors are summarized in Table 1. More than half of the patients were female (60%). Patients’ median age was 36.5 (range 5−87) years; seven were children aged 5, 14, 15, 15, 17, 17, 17 years. Initially, 63 patients were treated with appendectomy, and seven patients were treated with upfront right hemicolectomy for the following reasons; ascending colon cancer in 4 patients, cecal cancer in 1 patient, Crohn’s disease in 1 patient, and recurrent diverticulitis in 1 patient. Forty-seven patients were clinically suspected appendicitis; for eight patients, the reason for the appendectomy was not documented and for 15 patients was other reasons including ovarian cyst/tumor (5), diverticulitis (1), a-colon ca (4), cecal ca (1), uterine carcinosarcoma (1), endometrial ca (1), kidney/pancreas transplant (1), and Crohn’s disease (1). Among the 63 simple appendectomies, five patients underwent subsequent right hemicolectomy following a well-differentiated neuroendocrine tumor diagnosis.

Table 1.

Clinicopathological and Imaging Features of Appendiceal Well-Differentiated Neuroendocrine Tumors.

Variable	N (%)
Sex
Female	42 (60)
Male	28 (40)
Age at diagnosis (years)	Mean 40.6, Median 36.5​, Range 5-87
Initial procedure
Appendectomy	63 (90)
Right hemicolectomy	7 (10)
Reason for appendectomy (62)
Clinically suspected appendicitis	47 (75)
Other	15 (24)
Preoperative diagnostic imaging​
CT scan	31 (44)
Ultrasound	16 (23)
Not performed (Clinical diagnosis only)	23 (33)
Imaging follow up period (months)​ (n = 35)	Mean 39, Median 28​, Range 2-129
Imaging follow up status​ (n = 35)
Alive with no recurrence of AWNET	35 (100)
Alive with other cancer​	4 (11)
Final follow up period (months)​	Mean 65, Median 58​, Range 1-197
Final status
Alive with no recurrence of AWNET	69 (99)
Dead from other cause	1 (1)
Macroscopic lesion	Yes 23/70 (33%)
Location (n = 63)
Distal	54 (86)
Mid	5 (8)
Proximal	2 (3)
Diffuse	2 (3)
Length of the appendix (cm)	Mean 6.0, Median 5.7, Range ​0.7-10
Diameter of the appendix (cm)	Mean 0.9, Median 0.8, Range 0.1-2.0
Size of tumor (cm)	Mean 0.7, Median 0.5, Range 0.05-2.9
Size of tumor (n = 67)
< 1 cm	49 (73)
1-2 cm	14 (21)
> 2 cm	4 (6)
Tumor extent (n = 65)
SM	11 (17)
MP	22 (34)
SS/Meso	27 (42)
VP	5 (8)
Lymphovascular invasion	Yes 10/70 (14%)
Perineural invasion	Yes 7/70 (10%)
Mitoses (count/ 2 mm2) (n = 65)
< 2	63 (97)
2-20	2 (3)
Resection margin	Positive 2/70 (3%)
Rupture site	Yes 3/70 (4%)
T stage (n = 66)
pT1	34 (52)
pT2	0
pT3	27 (41)
pT4	5 (8)
WHO grade (n = 67)
G1	62 (93)
G2	5 (7)

Abbreviations: AWNET, appendiceal well-differentiated neuroendocrine tumor.

Diagnostic imaging was performed in 47 patients (67%), using abdominal CT in 31/70 patients (44%) and abdominal ultrasound in 16/70 patients (23%). CT showed a thickened appendiceal wall in 15 patients; thickened appendiceal wall and peri-appendiceal inflammation in 3; distended appendix in 3 patients; non-significant appendiceal findings in ten patients. The ten patients with non-significant appendiceal findings received appendectomy for other reasons described above. Ultrasound showed thickened appendiceal wall in 15 patients; distended appendix in one patient.

The median size of the tumor was 0.5 cm (range 0.05−2.9 cm), 49 tumors measured < 1.0 cm, 14 tumors measured 1.0–2.0 cm, and four tumors > 2.0 cm. Three tumors had no information on size. A macroscopic lesion at grossing was identified in 23 patients (33%). The macroscopic lesion was described as lumen occlusion (8, 35%), nodule (4, 17%), thickening (2, 9%), tumor (1, 4%), and not specified (7, 30%). One patient had concurrent low-grade appendiceal mucinous neoplasm (LAMN), which presented as mucin material at grossing. The appendix was submitted in total for pathological analysis in 60 patients (86%) following the discovery of well-differentiated neuroendocrine tumors.

Description of microscopic growth pattern was documented in only 18 tumors (26%); which were described as acini (1), cords (2), glands/tubules (4), gyriform (1), microacini (1), nest (11), ribbon (1), sheet (2), single file (2), trabecular (4) with some tumors showing multiple growth patterns (Figure 1). No tumor was described as a tubular neuroendocrine tumor.

Figure 1.

Representative pictures of appendiceal well-differentiated neuroendocrine tumors (hematoxylin and eosin staining) show (A) solid and nested, (B) nested and focal glandular architecture. Tumor cells are associated with reactive stromal fibrosis. Tumor cells are monotonous with salt and pepper nuclear chromatin.

Pathological tumor extent was submucosa; SM (11 tumors, 17%), muscularis propria; MP (22 tumors, 34%), subserosa; SS/ Mesoappendix; Meso (27 tumors, 42%), visceral peritoneum; VP (5 tumors, 8%) and no information available (5 tumors, 8%). Lymphovascular invasion was observed in 10 tumors (14%) and perineural invasion in 7 tumors (10%). The resection margin was positive in 2 tumors (both radial margins); tumor rupture site was observed in 3 tumors. Based on mitotic counts and MIB-1 index, most tumors were classified as WHO Grade 1 (63, 93%) and the remaining seven tumors as WHO Grade 2. No tumor was classified as WHO Grade 3.

Inflammatory change within the appendix was seen in the form of acute appendicitis (33 patients), subacute appendicitis (3 patients), and serositis/peritonitis (13 patients). Concurrent lesions within the appendix include villous adenoma with low-grade dysplasia (1), sessile serrated adenoma (1), Low-grade appendiceal mucinous neoplasms (LAMN) (1), and diverticulum (1).

The median number of immunohistochemical staining performed in one tumor was six antibodies (range 0−15). Representative pictures of immunohistochemical markers are shown in Figure 2. Among the antibodies, chromogranin A, synaptophysin, CAM5.2, and CDX2 were positive in > 90% of tumors. At least one chromogranin A or synaptophysin was performed in 60 tumors (86%). Both markers were selected in 43 tumors, and when only one of the two was stained (17 tumors, 24%), most pathologists chose chromogranin A (16 tumors). CAM5.2 was stained in 36 tumors (51%), and all were diffusely positive (36/36, 100%). CDX2 was positive in all but one tumor stained for this marker (32/33 tumors, 97%). Serotonin was positive in 34/47 tumors (79%); glucagon in 9/25 tumors (36%); PP in 8/30 tumors (27%); PYY in 17/32 tumors (53%). Based on immunohistochemistry, 18 tumors were from EC-cell origin, 12 tumors from L-cell origin, 6 tumors from both EC and L-cell origin. 10 more tumors contained EC-cells but could not be fully classified due to the incomplete immunohistochemical panel. MIB-1 staining was performed in 63 out of 70 tumors (90%); < 3% 58 tumors, 3-20% 5 tumors. No tumor had a MIB-1 index over 20%.

Figure 2.

Representative immunohistochemical staining of appendiceal well-differentiated neuroendocrine tumors show diffuse positivity for (A) chromogranin A, (B) synaptophysin, (C) CAM5.2, (D) CDX2, (E) serotonin, and (F) peptide YY.

Based on prior studies,^5,6 tumor size over 2 cm and tumor stage pT3/T4 were considered patients at high risk for nodal metastasis, and our 70 patients were divided into high risk and low risk based on these two criteria. Patients that did not have information on tumor size or pathological stage were excluded at this point. As a result, 32 patients were considered high risk; 32 were considered low-risk. Clinicopathological characteristics based on two-tier risk classification are summarized in Table 2. Briefly, there was no significant difference in sex distribution and age between the low and high-risk groups. Differences in the mitotic count and MIB-1 index were also statistically insignificant. No patient in the low-risk group received subsequent right hemicolectomy - five out of 29 high-risk patients treated with appendectomy initially went on to receive subsequent hemicolectomy (17%). No patient in the low-risk group had lymphovascular/perineural invasion. Among hormonal immunohistochemical markers, serotonin was more frequently positive in the high-risk group (p = 0.0212); PYY positive in the low-risk group (p = 0.0060). No significant difference was seen in the expressions of glucagon and PP.

Table 2.

Comparison of Clinicopathological Factors Between low Risk and High Risk Appendiceal Well-Differentiated Neuroendocrine Tumors.

	Low risk (n = 32)	High risk (n = 32)	P
Size of tumor (cm)	Median 0.3 (0.05-1.5)	Median 1.0 (Range 0.5-2.9)	n/a
Tumor extent	SM (11/32), MP (21/32)	SS/Meso (27/32), VP (5/32)	n/a
pT (TNM)	pT1 (32/32)	pT3 (27/32), pT4 (5/32)	n/a
Sex	Female 16, Male 16	Female 22, Male 10	P = 0.2028
Age at diagnosis (years)	36 (Range 5-87)	Median 38 (Range 14-83)	P = 0.4399
Location	Distal (26/30) Proximal/mid (4/30)	Distal (24/30) Proximal/mid/diffuse (6/30)	P > 0.9999
Macroscopic lesion	Yes (9/32)	Yes (13/32)	P = 0.4302
Lymphovascular invasion	Yes (0/32)	Yes (10/32)	P = 0.0009
Perineural invasion	Yes (0/32)	Yes (7/32)	P = 0.0108
Mitoses (count/ 2 mm2)	< 2 (32/32), 2-20 (0/32)	< 2 (29/31), 2-20 (2/31)	P = 0.2381
MIB-1 labeling index (%)	< 3 (29/30), 3-20 (1/30)	< 3 (26/30), 3-20 (4/30)	P = 0.3533
Serotonin	Positive (12/22)	Positive (21/24)	P = 0.0212
Glucagon	Positive (6/13)	Positive (3/12)	P = 0.4110
PP	Positive (4/13)	Positive (4/17)	P = 0.6976
PYY	Positive (12/15)	Positive (5/17)	P = 0.0060​

Abbreviations: SM, Submuccosa; MP, Muscularis Propria; SS, Subserosa; Meso, Mesoappendix; VP, Visceral Peritoneum; PP, pancreatic peptide; PYY, peptide YY.

Significant (P < 0.05) values are shown in bold italic.

Imaging follow-up (CT: 31, ultrasound: 3, PET 1 patient) was available in 35 patients with a median of 2 years 3 months (range 2 months – 10 years 7 months). All patients were followed up for a median of 4 years 8 months (1 month – 16 years 2 months). One patient died of uterine carcinosarcoma; all other 69 patients are alive without recurrence.

Discussion

This retrospective study identified 70 patients of incidental appendiceal well-differentiated neuroendocrine tumors over 15 years. Sixty percent of our patients were female, in keeping with past reports of slight female predominance.^8,16 Our series affected seven children (10%). The mean age was 40 years, consistent with younger age distribution compared to other neuroendocrine tumors of other sites. ⁹ All 70 tumors were incidental, and the primary reason for the appendectomy was clinically diagnosed appendicitis. No patient was diagnosed as appendiceal tumor preoperatively. No patient exhibited symptoms other than acute appendicitis-associated symptoms. Hormone-related symptoms are extremely rare in non-metastatic appendiceal well-differentiated neuroendocrine tumors, ¹⁷ and no patient in our series presented with them.

Appendiceal well-differentiated neuroendocrine tumors arise from neuroendocrine cells located within the lamina propria/submucosa of the appendiceal wall, and the tip of the appendix is where most of these cells are located.^18,19 In our series, the tumor's location was predominantly the tip or distal appendix (86%).

The median size was 5 mm, and 67% of the tumors were smaller than 1.0 cm. Among four tumors with a maximum dimension of > 2 cm, one received appendectomy with hysterectomy for endometrial carcinoma. Despite the large tumor size, no further treatment was considered due to comorbidity and age (> 80 years). The remaining three tumors received subsequent right hemicolectomy following the diagnosis of appendiceal well-differentiated neuroendocrine tumor, and none presented with residual tumor or lymph node metastasis. Only one other tumor < 2.0 cm received subsequent right hemicolectomy. This tumor was a G1 well-differentiated neuroendocrine tumor in a 22-year-old male, with tumor size 0.8 cm, extending into the mesoappendix, negative for lymphovascular invasion. The reason for subsequent colon resection is unknown; however, we speculate that the younger age and extent of the tumor (mesoappendix) may have prompted the procedure at the time of diagnosis (2005). Regardless of the status of subsequent colectomy, no patient in our series presented with recurrence/metastasis nor died of disease.

Tumors > 2 cm are usually rare, comprising 6% of our series. Tumors > 2 cm have an increased risk of metastasis, and subsequent colectomy is recommended.^5,6 Conversely, tumors < 1 cm have a low risk of lymph node metastasis, and the demerit of surgical complications weighs heavier than benefit. ⁵ In the most recent paper by Noor et al, ⁸ the study recommended a renewed nomenclature for the smallest tumors (size < 5 mm) that did not invade subserosa/mesoappendix and suggested no synoptic is needed for those tumors. Our study included 30 tumors of size < 5 mm, and among them, only two tumors extended into the subserosa/mesoappendix. Only one tumor was G2. None received subsequent hemicolectomy, and none presented with recurrence nor died of disease. This result is similar to their study. Management of tumors of size 1-2 cm is still of debate, and Rault-Petite et al, suggested that lymphovascular invasion and the extent of the tumor may be potential factors that prompt resection in this group. ²⁰ In our study, none of the 1-2 cm sized tumors presented with LVI; therefore, LVI was not a significant factor in determining the prognosis of this group. Based on our findings, the simple two-tier risk classification using size (< 2 cm) and extent of the tumor (absence of subserosa/mesoappendix extension) was sufficient.

The practical immunohistochemical panel for appendiceal well-differentiated neuroendocrine tumors has not been established. The median of six immunostains performed for pathological diagnosis in our series is quite large for a primarily indolent tumor, and some stains may be omitted without negatively impacting the patient. Immunohistochemistry for neuroendocrine tumors can be divided into five groups. ²¹ The first group is markers for neuroendocrine differentiation (chromogranin A and Synaptophysin), the antibody group most frequently ordered by the signing pathologist. All appendiceal well-differentiated neuroendocrine tumors stained positive for at least one of the two neuroendocrine markers in our series. While demonstrating neuroendocrine differentiation using immunohistochemistry for routine diagnosis of neuroendocrine tumors is considered standard and practical, ²² some experts argue that it is not mandatory for morphologically typical neuroendocrine tumors with an indolent behavior. ²³ Although we agree with this point, immunohistochemistry for chromogranin A and Synaptophysin is undoubtedly reasonable and, in most cases, the minimum requirement considering the rarity of this tumor.

The second group, CAM5.2, is used to demonstrate the epithelial nature of the tumor and is positive in almost all appendiceal well-differentiated neuroendocrine tumors, ²² including our series where all tumors stained positive (36/36, 100%). While the positive results are reassuring, routine staining of CAM5.2 is also redundant for diagnosing typical appendiceal well-differentiated neuroendocrine tumors and is usually only necessary when there is unusual morphology, including ganglion-like cells indicating the possibility of a paraganglioma. ²³

The third group is markers for tumor proliferation, namely Ki-67/MIB-1. 92% of our series had MIB-1 index < 3%. The remaining tumors had MIB-1 index 3-20%, and none were > 20%. Most appendiceal well-differentiated neuroendocrine tumors seem to be MIB-1 index < 3% or G1, and immunohistochemistry for MIB-1 may not even be necessary for small tumors, although certainly reasonable as some tumors may show unexpected higher MIB-1 index despite low mitotic counts. ⁸

The fourth group is markers for the site of origin. CDX2 is useful in determining midgut origin ²⁴ and was positive in all but one tumor stained for this marker (32/33 tumors) in our series. While this marker is helpful in patients presenting with metastatic disease to determine the site of origin, routine CDX2 is unnecessary in most tumors as negative CDX2 does not provide additional information about the tumor.

The fifth group is hormonal markers. Several hormonal markers were stained in our study, including serotonin, glucagon, somatostatin, PP, and PYY. While staining for hormonal markers does not provide additional clinical information for most non-functioning appendiceal well-differentiated neuroendocrine tumors, ²³ our retrospective study found novel findings related to some of the markers. In our study, when the patients were divided into two groups based on the risk of lymph node metastasis, serotonin was more positive in the high-risk group (p = 0.0212), while PYY was more positive in the low-risk group (p = 0.0060). Therefore, immunohistochemical staining for serotonin may be applied in the high-risk appendiceal well-differentiated neuroendocrine tumors; nevertheless, it may not be clinically relevant in the absence of hormonal symptoms. Also, hormonal markers may be utilized to distinguish EC-cell and L-cell neuroendocrine tumors. Based on available immunohistochemical results, 12 cases were classified as L-cell neuroendocrine tumors (serotonin-negative; glucagon or PP or PYY-positive). According to our study's limited morphological data, all four tumors with a characteristic trabecular growth pattern were classified as L-cell tumors using immunohistochemistry. A detailed morphological examination may aid in the identification of the less common L-cell tumors; nevertheless, hormonal markers are not mandatory in appendiceal well-differentiated neuroendocrine tumors, as it lacks significance in prognostic value. ²⁰

Our study has its limitations; most of all, the retrospective study design. This design has resulted in the variation of immunohistochemical panels between patients, as the immunohistochemistry has been ordered at the discretion of different pathologists at different time periods under different guidelines and available antibodies. Also, this study was a single-tertiary center study, which may not truly reflect the incidence of appendiceal well-differentiated neuroendocrine tumors as many appendectomies are performed in community hospitals. Also, because of its indolent and incidental presentation, the patient's age at diagnosis may not represent the actual occurrence of this lesion.

In conclusion, appendiceal well-differentiated neuroendocrine tumors are incidental findings in an appendectomy specimen with an excellent prognosis. Typical appendiceal well-differentiated neuroendocrine tumors occur in younger (< 40 years) females, are small (< 1 cm), G1, with no lymph node and distant metastasis. Our study shows evidence that the comprehensive immunohistochemical panel used in other neuroendocrine tumors may be excessive for the clinically and pathologically typical appendiceal well-differentiated neuroendocrine tumors. Immunohistochemistry should be chosen carefully at the pathologist's discretion, and morphological examination of tumor size and extent remains the most critical factor.