Context.—

Although several neuroendocrine cell types constitute gastroenteropancreatic neuroendocrine tumors (NETs), the clinical and prognostic implications of the expression of multiple peptide hormones have not been comprehensively evaluated in rectal NETs.

Objective.—

To identify the clinicopathologic characteristics and prognostic impact of peptide hormone expression.

Design.—

We evaluated the expression of peptide YY (PYY), glucagon, somatostatin, serotonin, insulin, and gastrin using immunolabeling in 446 endoscopically or surgically resected rectal NETs.

Results.—

PYY, glucagon, serotonin, somatostatin, insulin, and gastrin were expressed in 261 of 389 (67.1%), 205 of 446 (46.0%), 36 of 446 (8.1%), 33 of 446 (7.4%), 2 of 446 (0.4%), and 1 of 446 cases (0.2%), respectively. Immunoreactivity to any peptide hormone was present in 345 of 446 cases (77.4%). Tumors expressing serotonin or somatostatin were associated with lymphovascular invasion, chromogranin A expression, and shorter disease-free survival (DFS). Rectal NETs were classified as L-cell, enterochromaffin-cell, D-cell, null-expression, or mixed-expression type based on peptide hormonal expression status. Patients with D-cell NET had the shortest DFS (10-year DFS, 54.5%), followed by those with enterochromaffin-cell NET (89.5%), null expression (97.0%), L-cell NET (99.6%), and mixed-expression NET (100%; P < .001). Multivariable analyses revealed that somatostatin expression was an independent indicator of poor prognosis with respect to DFS in rectal NETs (P = .001).

Conclusions.—

Somatostatin expression is a poor prognostic indicator in patients with rectal NETs. Therefore, additional peptide hormonal immunolabeling, including somatostatin, serotonin, and PYY, in rectal NETs can provide more information regarding DFS.

The incidence of gastroenteropancreatic (GEP) neuroendocrine tumors (NETs) has increased substantially during the past few decades. They are currently the second most common digestive system tumor, after colorectal cancer.1,2  Among the GEP-NETs, incidence of rectal NETs is increasing most rapidly,1  and they account for the largest proportion in the United States, Korea, Taiwan, and Japan.1,36  GEP-NETs are a heterogeneous group of tumors that exhibit various biological and clinical behaviors depending on the types of neuroendocrine cells constituting the tumor and the primary location.4,79  Many types of neuroendocrine cells constitute the GEP system, and their distribution affects the cell type of GEP-NETs in a specific organ.7  Normally, scattered endocrine cells in the colorectal mucosa are distributed at the surface or crypt of the glandular epithelial cells and include somatostatin-producing D cells, vasoactive intestinal peptide (VIP)–producing VIP cells, serotonin-producing enterochromaffin (EC) cells, glucagon-like peptide (GLP)– and pancreatic polypeptide (PP)/peptide YY (PYY)–producing L cells, and ghrelin-producing P/D1 cells.7  In contrast, endocrine cells in a normal pancreas are distributed as well-circumscribed nests, islets of Langerhans, and consist of insulin-producing beta cells, glucagon-producing alpha cells, somatostatin-producing D cells, PP-producing PP cells, and VIP-producing VIP cells.7  NETs derived from the embryonic midgut, including the jejunum, ileum, appendix, cecum, and ascending colon, consist mainly of serotonin-producing EC cells. In contrast, NETs derived from the embryonic hindgut, including the distal colon and rectum, consist mainly of GLP- and PP/PYY-producing L cells and account for 80% of NETs from the colorectum.7,1013 

Previously, rectal NETs comprising a specific L-cell type were considered to exhibit a favorable clinical behavior, and the rectal L-cell NETs were classified as tumors with uncertain malignant potential in the previous 2010 World Health Organization (WHO) classification.11,14,15  However, they were reclassified in the 2019 WHO classification system as malignant neoplasms, similar to other rectal NETs, as they exhibited occasional nodal and distant metastasis, as well as muscularis propria invasion.11,13,14,16  Therefore, none of these cell types or their peptide hormone expression is currently being studied with respect to clinicopathologic behavior and prognosis.

In pancreatic NETs, however, the prognostic significance of peptide hormone expression has been demonstrated.17  Patients presenting with pancreatic NETs with insulin or GLP1 expression were associated with better survival, whereas those with gastrin expression in pancreatic NETs were associated with worse survival.17  In rectal NETs, chromogranin A expression was associated with poorer clinical behavior and a non–L-cell phenotype.11,18  Hence, we hypothesized that the expression of certain peptide hormones, or subtypes that reflect normal cell differentiation in rectal NETs, could affect clinical behavior and prognosis, and that they may be associated with chromogranin A expression status. Here, we analyzed the expression profiles of peptide hormones, including PYY, glucagon, serotonin, somatostatin, insulin, and gastrin, and compared their expression with various clinicopathologic factors, including disease-free survival (DFS).

Case Selection

This study was approved by the Institutional Review Board of the Asan Medical Center (Seoul, Republic of Korea) with a waiver of consent from the patients (approval number 2014-0580). We selected 446 well-differentiated grade 1 and 2 NET cases with available formalin-fixed, paraffin-embedded blocks that had been endoscopically or surgically resected between 2006 and 2015. Data acquired from the review of electronic medical records included age, sex, the procedure needed for treatment, and survival outcome.

Pathologic Evaluation

Data acquired from the surgical pathology reports included the primary diagnosis, location, and size of the tumors. All the hematoxylin and eosin–stained slides were reviewed, and the pathologic features, including invasion depth, lymphovascular and perineural invasion, resection marginal status, and nodal and distant metastases were reevaluated. Invasion depth was assessed according to the 8th edition of the American Joint Committee on Cancer staging scheme.19  Evaluating invasion depth was not applicable in endoscopically resected NETs, as endoscopically resected (endoscopic mucosal resection or endoscopic submucosal dissection) specimens did not contain entire rectal wall layers. Therefore, we considered those cases inapplicable for invasion depth. However, lymphovascular or perineural invasion was assessed regardless of the procedure. Nodal status was determined by pathologic evaluation. In endoscopically or surgically resected NETs without lymph node dissection, nodal status was regarded as not applicable. Tumor grading was assigned based on mitotic counts per field in 2 mm2, and the Ki-67 labeling index was in accordance with the 2019 WHO classification system.16 

Tissue Microarray Construction and Immunohistochemistry

Tissue microarrays (TMAs) were constructed from archived formalin-fixed, paraffin-embedded tissue blocks using a manual tissue microarrayer (Uni TMA Co Ltd, Seoul, Korea), the details of which have been previously described.11,18  A single representative tissue core of 2.0-mm diameter was punched from each tumor block and was transferred into the recipient block in parallel. Of the peptide hormones that normal endocrine cells produce in the colorectal mucosa and pancreas, we selected several antibodies that are currently used in clinical practice, including gastrin, glucagon, insulin, serotonin, somatostatin, and PYY. Immunohistochemical staining was performed on 4-μm-thick serial tissue sections from TMA blocks. Tissue sections were deparaffinized and hydrated by sequential immersion in xylene and a graded ethanol series. Endogenous peroxidase was blocked by incubation in 3% H2O2 for 10 minutes, followed by heat-induced antigen retrieval. The staining procedure was performed on automated immunostainers (Benchmark XT; Ventana Medical Systems, Tucson, Arizona) with an Optview DAB IHC Detection Kit (Ventana Medical Systems), in accordance with the manufacturer's instructions. Tissue sections were incubated at room temperature for 32 minutes with primary antibodies against insulin (1:100; 2D11-H5, Novocastra, Newcastle, United Kingdom), glucagon (1:400; 295A-15, Cell Marque, Rocklin, California), serotonin (1:200; 5HTH209, DAKO, Glostrup, Denmark), somatostatin (1:4000; A0566, DAKO), gastrin (prediluted; A0568, DAKO), PYY (1:1000; rabbit, polyclonal, ab131246, Abcam, Cambridge, United Kingdom), synaptophysin (1:200; rabbit, MRQ-40; Cell Marque), chromogranin A (1:1600; mouse, DAK-A3, DAKO), and Ki-67 (1:200; mouse, MIB1, DAKO). Following light counterstaining with hematoxylin, the immunostained sections were serially dehydrated through a series of ethanol and xylene and mounted. Immunohistochemical expression was evaluated by 2 pathologists (J.S.K. and S.M.H.) blinded to the clinicopathologic information. Immunostaining for synaptophysin, chromogranin A, and 6 peptide hormones, including insulin, glucagon, serotonin, somatostatin, gastrin, and PYY, was interpreted as positive when the proportion of tumor cells showing cytoplasmic staining was 5% or greater and as negative when the proportion was less than 5%, as described in our previous studies.11,18 

Outcome Measures

DFS was defined as the duration from tumor resection to locoregional recurrence, and was selected as the primary outcome. Two cases with synchronous distant metastasis were excluded from the survival analysis. All patients were followed up with periodic evaluations of colonoscopy, sigmoidoscopy (with or without computed tomography), magnetic resonance imaging, or 68Ga-DOTATOC positron emission tomography every 3 to 12 months based on their tumor characteristics. However, no strict protocol was set for follow-ups. Locoregional recurrence was defined as the histologic confirmation of reemergence of NET at the tumor bed of the prior resection site, rectal lumen, or regional lymph nodes.

Statistical Analysis

For comparison of continuous variables, an independent 2-sample t test or Mann-Whitney U test was employed. For comparison of categorical variables, the χ2 test or Fisher exact test was used. Survival curves were plotted using the Kaplan-Meier method and were compared using the log-rank test. Univariable Cox regression analysis was conducted for identifying clinicopathologic factors associated with DFS. Multivariable Cox regression analysis was performed for the evaluation of independent prognostic significance. All statistical tests were performed using SPSS version 21.0 (IBM Corp, Armonk, New York) and R software version 4.0.3 (R Foundation, Vienna, Austria). Statistical significance was set at a 2-tailed P value < .05.

Baseline Characteristics of Patients With Rectal NETs

The clinicopathologic features of patients with rectal NETs are summarized in Table 1. The mean patient age was 48.4 ± 11.9 years (range, 19–80 years) with a male to female ratio of 1.4; 395 of 446 cases (88.6%) were NET grade 1 and 51 of 446 cases (11.4%) grade 2. Grade 3 cases were not included. The mean tumor size was 0.6 ± 0.4 cm. When the tumor size was dichotomized, sizes 1 cm or smaller were identified in 407 of 446 cases (91.3%) and sizes larger than 1 cm were identified in 39 of 446 cases (8.7%). According to the T category of the 8th edition of the American Joint Committee on Cancer staging system,19  36 of 45 cases (80.0%) were pT1, 5 of 45 (11.1%) were pT2, 2 of 45 (4.5%) were pT3, and 2 of 45 (4.4%) were pT4 in surgically resected NETs. A total of 401 endoscopically resected NETs were not applicable for proper invasion depth, as endoscopically resected specimens contained mucosa and submucosa only. Lymphovascular and perineural invasion were identified in 20 of 446 cases (4.5%) and 7 of 446 cases (1.6%), respectively. Involvement of resection margin was noted in 62 of 446 cases (13.9%). Nodal metastasis was identified in 6 of 18 cases (33.3%) with nodal dissection, and synchronous distant metastasis in 2 of 446 cases (0.4%). Endoscopic resection was performed in 401 of 446 cases (89.9%) and surgical resection including transanal excision and radical surgery in 45 of 446 cases (10.1%). The median follow-up period for patients included in the survival analysis was 50.7 months (range, 1–143.4 months).

Table 1

Baseline Characteristics of Patients With Rectal Neuroendocrine Tumors and the Correlations Between Peptide Hormone Expression Status and Clinicopathologic Factorsa

Baseline Characteristics of Patients With Rectal Neuroendocrine Tumors and the Correlations Between Peptide Hormone Expression Status and Clinicopathologic Factorsa
Baseline Characteristics of Patients With Rectal Neuroendocrine Tumors and the Correlations Between Peptide Hormone Expression Status and Clinicopathologic Factorsa

Synaptophysin and Chromogranin A Expression

Immunohistochemical expression for chromogranin A was identified in 57 of 446 cases (12.8%). In contrast, synaptophysin expression was detected in all 446 rectal NETs (100%) with a diffuse expression pattern.

Peptide Hormonal Expression and Clinicopathologic Correlation in Patients with Rectal NETs

Representative images showing immunohistochemical expression of each peptide hormone in rectal NETs are illustrated in Figure 1, A through L. Immunolabeling for PYY (Figure 1, A and B), glucagon (Figure 1, C and D), serotonin (Figure 1, E and F), somatostatin (Figure 1, G and H), insulin (Figure 1, I and J), and gastrin (Figure 1, K and L), was present in 261 of 389 (67.1%), 205 of 446 (46.0%), 36 of 446 (8.1%), 33 of 446 (7.4%), 2 of 446 (0.4%), and 1 of 446 cases (0.2%), respectively. Overall, 345 of 446 cases (77.4%) of rectal NETs showed immunoreactivity to a peptide hormone.

Figure 1

Representative images of rectal neuroendocrine tumors (NETs) with the expression of individual peptide hormones. Rectal NETs with expression of peptide YY (PYY) (A and B), glucagon (C and D), serotonin (E and F), somatostatin (G and H), insulin (I and J), and gastrin (K and L) (hematoxylin-eosin, original magnification ×200 [A, C, E, G, I, and K]; PYY, original magnification ×200 [B]; glucagon, original magnification ×200 [D]; serotonin, original magnification ×200 [F]; somatostatin, original magnification ×200 [H]; insulin, original magnification ×200 [J]; gastrin immunolabeling, original magnification ×200 [L]).

Figure 1

Representative images of rectal neuroendocrine tumors (NETs) with the expression of individual peptide hormones. Rectal NETs with expression of peptide YY (PYY) (A and B), glucagon (C and D), serotonin (E and F), somatostatin (G and H), insulin (I and J), and gastrin (K and L) (hematoxylin-eosin, original magnification ×200 [A, C, E, G, I, and K]; PYY, original magnification ×200 [B]; glucagon, original magnification ×200 [D]; serotonin, original magnification ×200 [F]; somatostatin, original magnification ×200 [H]; insulin, original magnification ×200 [J]; gastrin immunolabeling, original magnification ×200 [L]).

Close modal

The associations between the expression status of each peptide hormone and clinicopathologic factors are compared in Table 1. Glucagon expression was associated with younger ages (≤50 years, P = .005), a lack of proper muscle invasion (P = .02), and tumors with a clear or indeterminate resection margin (P = .04). Serotonin expression was associated with male sex (P = .03) and tumors with lymphovascular and perineural invasion (P = .003 and P = .01, respectively). Somatostatin expression was associated with lymphovascular invasion (P = .01). PYY expression was associated with grade 1 (P = .04), small tumor size (≤1 cm; P = .02), lack of proper muscle invasion (P = .005), and an absence of lymphovascular (P = .001) and perineural (P < .001) invasion. Chromogranin A expression was significantly associated with serotonin and somatostatin expression (both P < .001). In contrast, chromogranin A expression was associated with a lack of glucagon (P = .02) and PYY (P < .001) expression. The association between insulin or gastrin expression and clinicopathologic factors was not evaluated because of the small number of cases with insulin (2 cases) and gastrin (1 case) expression.

DFS Outcomes Based on the Expression Status of Peptide Hormones

The overall 10-year DFS of the entire patient group with rectal NETs was 97.6%. Patients presenting with rectal NETs with somatostatin expression (10-year DFS, 76.0%) exhibited worse DFS than those without somatostatin expression (99.1%, log-rank test, P < .001; hazard ratio [HR], 21.158; 95% CI, 4.719–94.857; Cox regression, P < .001; Figure 2, A; Table 2). Patients presenting with rectal NETs with serotonin expression (10-year DFS, 89.1%) exhibited worse DFS than those without serotonin expression (98.3%, log-rank test, P < .001; HR, 9.472; 95% CI, 2.118–42.360; Cox regression, P = .003; Figure 2, B; Table 2). However, there was no difference in survival according to glucagon (Figure 2, C; Table 2) expression status. Patients presenting with rectal NETs with PYY expression (10-year DFS, 99.5%) exhibited better DFS than those without PYY expression (92.4%; log-rank test, P = .002; HR, 0.078; 95% CI, 0.009–0.645; Cox regression, P = .02; Figure 2, D; Table 2). Survival analyses according to insulin and gastrin expression status were not performed because of the small number of cases with insulin (2 cases) and gastrin (1 case) expression. Meanwhile, patients presenting with rectal NETs with chromogranin A expression (10-year DFS, 92.9%) showed worse survival than those without chromogranin A expression (98.3%, log-rank test, P = .01; HR, 5.555; 95% CI, 1.243–24.820; Cox regression, P = .02, Figure 2, E; Table 2).

Figure 2

Disease-free survival (DFS) rates in rectal neuroendocrine tumor (NET) patients based on expression of each peptide hormone. A, Patients presenting with rectal NETs with somatostatin expression showed worse 10-year DFS than those without serotonin expression (76.0% versus 99.1%; P < .001). B, Those with serotonin expression showed worse 10-year DFS than those without serotonin expression (89.1% versus 98.3%; P < .001). C, DFS was not statistically different based on glucagon expression status (99.5% versus 96.2%; P = .11). D, Those with peptide YY (PYY) expression showed better 10-year DFS than those without PYY expression (99.5% versus 92.4%; P = .002). E, Chromogranin A expression showed worse 10-year DFS than those without chromogranin A expression (92.9% versus 98.3%; P = .01).

Figure 2

Disease-free survival (DFS) rates in rectal neuroendocrine tumor (NET) patients based on expression of each peptide hormone. A, Patients presenting with rectal NETs with somatostatin expression showed worse 10-year DFS than those without serotonin expression (76.0% versus 99.1%; P < .001). B, Those with serotonin expression showed worse 10-year DFS than those without serotonin expression (89.1% versus 98.3%; P < .001). C, DFS was not statistically different based on glucagon expression status (99.5% versus 96.2%; P = .11). D, Those with peptide YY (PYY) expression showed better 10-year DFS than those without PYY expression (99.5% versus 92.4%; P = .002). E, Chromogranin A expression showed worse 10-year DFS than those without chromogranin A expression (92.9% versus 98.3%; P = .01).

Close modal
Table 2

Univariable Analysis for the Association of Clinicopathologic Factors With Recurrence in Patients With Rectal Neuroendocrine Tumors

Univariable Analysis for the Association of Clinicopathologic Factors With Recurrence in Patients With Rectal Neuroendocrine Tumors
Univariable Analysis for the Association of Clinicopathologic Factors With Recurrence in Patients With Rectal Neuroendocrine Tumors

Univariable Survival Analysis of Other Clinicopathologic Factors in Rectal NET

The impact of other clinicopathologic factors on DFS is summarized in Table 2. Factors associated with adverse DFS were grade 2 (HR, 9.877; 95% CI, 2.209–44.171; P = .003), proper muscle invasion (HR, 13.305; 95% CI, 2.575–68.739; P = .002), lymphovascular invasion (HR, 8.490; 95% CI, 1.641–43.920; P = .01), perineural invasion (HR, 11.559; 95% CI, 1.387–96.310; P = .02), lymph node metastasis (HR, 10.340; 95% CI, 1.243–86.002; P = .03), and surgical resection (HR, 21.680; 95% CI, 4.203–111.825; P < .001; Table 2).

Multivariable Survival Analysis Based on the Expression Status of Peptide Hormones in Rectal NETs

Because the significance of serotonin, somatostatin, and PYY expression on survival was demonstrated in univariable Cox regression, they were incorporated in multivariable Cox analysis as individual variables. Other clinicopathologic factors that showed significance in univariable Cox regression analysis were also adjusted for multivariable Cox regression analysis. Somatostatin expression in rectal NETs (HR, 137.568; 95% CI, 7.891–999.999; P = .001) was an independent poor prognostic indicator for DFS in our model (Table 3). The necessity of surgical resection for rectal NETs (HR, 94.150; 95% CI, 3.903–999.999; P = .005) was also associated with poor DFS in the multivariable survival analysis (Table 3).

Table 3

Multivariable Analysis for the Association of Clinicopathologic Factors with Recurrence in Patients With Rectal Neuroendocrine Tumors

Multivariable Analysis for the Association of Clinicopathologic Factors with Recurrence in Patients With Rectal Neuroendocrine Tumors
Multivariable Analysis for the Association of Clinicopathologic Factors with Recurrence in Patients With Rectal Neuroendocrine Tumors

Rectal NET Subtypes According to Hormonal Expression Status

In contrast to serotonin-producing enterochromaffin (EC)-cell NETs and GLP- and PP/PYY-producing L-cell NETs,15,16  somatostatin-producing D-cell NETs have not been well characterized in rectal NETs. However, based on the independent prognostic significance of somatostatin-expressing rectal NETs, we classified cases showing dominant somatostatin expression into D-cell NETs separately from L-cell and EC-cell NETs. Accordingly, our classification of rectal NETs included L-cell, EC-cell, and D-cell subtypes based on PYY, serotonin, and somatostatin expression status. NETs showing dominant PYY expression were classified as L-cell NETs (Figure 3, A through C); NETs showing dominant serotonin expression were classified as EC-cell NETs (Figure 3, D through F); NETs showing dominant somatostatin expression were classified as D-cell NETs (Figure 3, G through I); NETs showing no immunoreactivity to serotonin, somatostatin, or PYY were classified as null-expression NETs (Figures 3, J through L); and NETs showing codominant expression among serotonin, somatostatin, and PYY were classified as mixed-expression NETs (Figures 3, M through O). We observed 1 case of rectal NET with mixed serotonin and somatostatin expression (Figure 4, A) and 3 cases of rectal NETs with mixed somatostatin and PYY expression (Figure 4, B), and confirmed the mixed-expression pattern of 2 peptide hormones with dual immunohistochemical staining. The cases showing mixed serotonin/somatostatin and somatostatin/PYY expression were classified as mixed-expression NETs. L-cell, EC-cell, and D-cell NETs were identified in 258 of 390 cases (57.8%), 24 of 390 cases (5.4%), and 9 of 390 cases (2.0%). Null-expression and mixed-expression NETs were identified in 95 of 390 (21.3%) and 4 of 390 cases (0.9%), respectively. The associations among the NET subtypes and clinicopathologic factors are summarized in Table 4. L-cell NETs were associated with smaller tumor size (≤1 cm; P = .04), absence of proper muscle (P = .005), lymphovascular (P <.001) and perineural (P = .003) invasion, and an absence of chromogranin A expression (P < .001; Table 4). In contrast, EC-cell NETs were associated with larger tumor size (>1 cm) (P = .04); proper muscle (P = .005), lymphovascular (P < .001), and perineural (P = .003) invasion, and chromogranin A expression (P < .001; Table 4). D-cell NETs were associated with lymphovascular invasion (P < .001) and chromogranin A expression (P < .001; Table 4). Null-expression rectal NETs were associated with an absence of lymphovascular invasion (P < .001) and an absence of chromogranin A expression (P < .001; Table 4). The mixed-expression group was associated with lymphovascular invasion (P < .001; Table 4).

Figure 3

Representative images of rectal neuroendocrine tumor (NET) subtypes according to serotonin, somatostatin, and peptide YY (PYY) expression status. A through C, PYY-expressing L-cell type NET. D through F, Serotonin-expressing enterochromaffin-cell type NET. G through I, Somatostatin-expressing D-cell type NET. J through L, Null-expression NET case without any serotonin, somatostatin, or PYY expression. M through O, Mixed-expression NET case showing codominant immunoreactivity of somatostatin and PYY peptide hormones (serotonin immunolabeling, original magnification ×200 [A, D, G, J, and M]; somatostatin immunolabeling, original magnification ×200 [B, E, H, K, and N]; PYY immunolabeling, original magnification ×200 [C, F, I, L, and O]).

Figure 3

Representative images of rectal neuroendocrine tumor (NET) subtypes according to serotonin, somatostatin, and peptide YY (PYY) expression status. A through C, PYY-expressing L-cell type NET. D through F, Serotonin-expressing enterochromaffin-cell type NET. G through I, Somatostatin-expressing D-cell type NET. J through L, Null-expression NET case without any serotonin, somatostatin, or PYY expression. M through O, Mixed-expression NET case showing codominant immunoreactivity of somatostatin and PYY peptide hormones (serotonin immunolabeling, original magnification ×200 [A, D, G, J, and M]; somatostatin immunolabeling, original magnification ×200 [B, E, H, K, and N]; PYY immunolabeling, original magnification ×200 [C, F, I, L, and O]).

Close modal
Figure 4

Representative image of rectal neuroendocrine tumors coexpressing (A) serotonin (magenta) and somatostatin (brown) and (B) peptide YY (PYY, brown) and somatostatin (magenta) (serotonin and somatostatin dual immunolabeling, original magnification ×400 [A]; PYY and somatostatin dual immunolabeling, original magnification ×400 [B]).

Figure 4

Representative image of rectal neuroendocrine tumors coexpressing (A) serotonin (magenta) and somatostatin (brown) and (B) peptide YY (PYY, brown) and somatostatin (magenta) (serotonin and somatostatin dual immunolabeling, original magnification ×400 [A]; PYY and somatostatin dual immunolabeling, original magnification ×400 [B]).

Close modal
Table 4

Baseline Characteristics and Clinicopathologic Factors of Rectal Neuroendocrine Tumors According to Subtypes Reflecting Hormonal Expression

Baseline Characteristics and Clinicopathologic Factors of Rectal Neuroendocrine Tumors According to Subtypes Reflecting Hormonal Expression
Baseline Characteristics and Clinicopathologic Factors of Rectal Neuroendocrine Tumors According to Subtypes Reflecting Hormonal Expression

DFS Outcomes Based on the NET Subtypes According to Hormonal Expression Status

DFS outcomes were analyzed according to NET subtypes. Patients with D-cell NET exhibited the worst DFS rate (10-year DFS, 54.5%), followed by those with EC-cell NET (89.5%), null-expression NET (97.0%), L-cell NET (99.6%), and mixed-expression NET (100%; P < .001; Figure 5). Pairwise comparisons revealed patients with EC-cell (Bonferroni corrected log-rank test, P = .004) and D-cell NETs (P < .001) showed worse DFS than those with L-cell NETs (Figure 5). Similarly, patients with D-cell NETs showed worse DFS than those with null-expression NETs (P = .01, Figure 5). However, there was no difference in DFS among the remaining subtypes, including L cell versus null expression, L-cell versus mixed expression, EC-cell versus D cell expression, EC cell versus null expression, EC cell versus mixed expression, D cell versus mixed expression, and null expression versus mixed expression (all, P > .99; Figure 5).

Figure 5

Disease-free survival (DFS) rate in rectal neuroendocrine tumor (NET) patients based on peptide hormone (peptide YY, serotonin, and somatostatin) expression status. Patients presenting with D-cell NET showed the worst DFS rate (10-year DFS, 54.5%), followed by those with enterochromaffin-cell NET (89.5%), null-expression NET (97.0%), L-cell NET (99.6%), and mixed-expression NET (100%; P < .001). In pairwise comparison, patients presenting with EC-cell and D-cell NETs showed worse DFS than those with L-cell NETs (Bonferroni corrected log-rank test, P = .004 and P < .001, respectively). Patients with D-cell NETs also showed worse DFS than those with null-expression NETs (P = .01). But no statistical difference was observed among the remaining subtypes, including L cell versus null expression, L cell versus mixed expression, EC cell versus D cell, EC cell versus null expression, EC cell versus mixed expression, D cell versus mixed expression, and null expression versus mixed expression (all, P > .99).

Figure 5

Disease-free survival (DFS) rate in rectal neuroendocrine tumor (NET) patients based on peptide hormone (peptide YY, serotonin, and somatostatin) expression status. Patients presenting with D-cell NET showed the worst DFS rate (10-year DFS, 54.5%), followed by those with enterochromaffin-cell NET (89.5%), null-expression NET (97.0%), L-cell NET (99.6%), and mixed-expression NET (100%; P < .001). In pairwise comparison, patients presenting with EC-cell and D-cell NETs showed worse DFS than those with L-cell NETs (Bonferroni corrected log-rank test, P = .004 and P < .001, respectively). Patients with D-cell NETs also showed worse DFS than those with null-expression NETs (P = .01). But no statistical difference was observed among the remaining subtypes, including L cell versus null expression, L cell versus mixed expression, EC cell versus D cell, EC cell versus null expression, EC cell versus mixed expression, D cell versus mixed expression, and null expression versus mixed expression (all, P > .99).

Close modal

The main findings of the present study are that (1) three-quarters (77%; 345 of 446) of rectal NET cases express a peptide hormone; (2) the commonly expressed peptide hormones in rectal NETs in order of frequency are PYY (67%; 261 of 389), glucagon (46%; 205 of 446), serotonin (8%; 36 of 446), and somatostatin (7%; 33 of 446), whereas insulin and gastrin expression are rare in rectal NETs; (3) rectal NET patients with serotonin or somatostatin expression presented short DFS; (4) rectal NETs with somatostatin expression were an independent indicator of poor prognosis with respect to DFS in endoscopically or surgically resected rectal NETs; and (5) D-cell NETs showed worse DFS among NET subtypes based on dominant peptide hormonal expression among PYY, serotonin, and somatostatin.

To the best of our knowledge, this study is the first to demonstrate the clinical significance of somatostatin expression and D-cell NET in rectal NETs. This might be because somatostatin expression and D-cell NET are not common in rectal NETs; 7% (33 of 446) of rectal NETs exhibited somatostatin expression and 2% (9 of 390) of the rectal NET cases were classified as D-cell NET (defined by dominant somatostatin expression) in the present study. Somatostatin expression and D-cell subtype in rectal NETs were associated with lymphovascular invasion. Lymphovascular invasion was observed in 15% (5 of 33) of rectal NETs with somatostatin expression and 22% (2 of 7) of rectal D-cell NET cases; this incidence is higher than the incidence of lymphovascular invasion (4.5%; 20 of 446) in the entire rectal NET cohort in the present study. Common lymphovascular invasion of rectal NETs with somatostatin expression and D-cell NET exhibits an increased chance of lymph node or distant metastasis, and it may ultimately contribute to short DFS.

Somatostatin-producing D-cell NETs have been reported mainly in the periampullary regions, including the duodenum, ampulla of Vater, and pancreas.16,2023  Capella et al20  reported that most ampullary D-cell NETs showed deep infiltration into periampullary tissue in 5 of 6 cases (83%) and para-duodenal lymph node metastasis in 4 of 6 cases (67%). Frequent lymph node metastasis (33%–69%) has been reported in duodenal D-cell NETs.2123  Distant liver metastasis was reported in up to 8% (1 of 13) of cases.22  Recurrence was reported in 20% of duodenal D-cell NET cases with nodal metastasis.22  Similarly, pancreatic NETs with somatostatin expression frequently harbored regional lymph node metastasis (6 of 16 cases; 38%) and distant metastasis to liver (5 of 16 cases; 31%).22  Psammoma bodies were reported in 50% (12 of 24 cases) of somatostatin-producing periampullary NETs.22  However, we did not detect any psammoma bodies in rectal NETs with somatostatin expression.

Approximately 8% (36 of 446) of rectal NET cases exhibited serotonin expression and 5% (24 of 390) the rectal NET cases were classified as EC-cell NET (defined by dominant serotonin expression). Serotonin expression and EC-cell subtypes in rectal NETs were associated with frequent lymphovascular and perineural invasions and low DFS in the present study. EC-cell subtypes were also associated with large tumor size (>1 cm). No previous study has evaluated clinicopathologic aspects of serotonin expression or EC-cell subtypes in rectal NETs because of the rarity of these cases. However, serotonin-expressing NETs in other GEP organs exhibited aggressive clinical behaviors. For example, most of the serotonin-producing EC-cell NETs in the jejunum and ileum demonstrated insular growth patterns and tumor invasion into or beyond the muscularis propria in 77% of cases; lymph nodes and/or liver metastasis was observed in one-third of the reported cases at the time of diagnosis.24,25  The 5-year overall survival rate in patients with jejunoileal NETs, most of which expressed serotonin (86%), was 58%.25  In small intestinal NETs, an increased urinary level of 5-hydroxyindoleacetic acid (5-HIAA; more than 3.7 times the upper limit of normal), a byproduct of the serotonin metabolism, was associated with poor overall survival; however, the prognostic implication of serotonin immunolabeling has not been demonstrated.26  Pancreatic serotonin-producing NETs were frequently of large size, and have trabecular architecture, fibrotic stroma, and large pancreatic duct involvement. However, lymph node metastasis was less commonly observed than in serotonin-negative pancreatic NETs, and the serotonin expression status did not relate to any difference in overall survival.27,28  Vascular and perineural invasion are frequently encountered in pancreatic serotonin-producing NETs.29  In the present study, EC-cell NETs were associated with a large tumor size (>1 cm) in line with cecal and pancreatic EC-cell NETs, which exhibited a larger tumor size.24,29 

From the multivariable Cox regression, somatostatin expression was shown to be an independent prognostic factor for adverse DFS after adjusting for clinicopathologic variables, including grade, proper muscle invasion, lymphovascular and perineural invasions, lymph node metastasis, and treatment procedure, as well as PYY, serotonin, and chromogranin A expression. Our previous study showed that chromogranin A expression was an independent indicator of poor prognosis18 ; however, in the present model this was not the case, according to multivariable Cox analysis performed with PYY, serotonin, and somatostatin expression as covariates. This result may suggest that somatostatin expression is prognostically more important than chromogranin A, PYY, and serotonin expression in patients with rectal NETs. In addition to synaptophysin, chromogranin A, and Ki-67 immunolabeling for the histopathologic diagnosis of rectal NETs, we recommend adding PYY, serotonin, and somatostatin immunolabeling for the stratification of DFS in patients with rectal NETs.

In our previous study, we speculated that there was a strong association between rectal NETs with chromogranin A expression and non–L-cell-type immunophenotypes; most of these NETs may exhibit the aggressive EC-cell phenotype, which could explain the lowered DFS of patients presenting with rectal NETs with chromogranin A expression.11,18  In the present study on rectal NETs, we confirmed that non–L-cell-type immunophenotype NETs, including serotonin expressing EC-cell NETs and somatostatin-expressing D-cell NETs, were strongly associated with chromogranin A expression. The association between serotonin expression and chromogranin A expression has also been reported in other GEP-NETs, including jejunoileal and pancreatic NETs.29,30  Similarly, the association between somatostatin expression and chromogranin A expression was demonstrated in gastric and pancreatic NETs; however, ampullary and pancreatic D-cell NETs are less likely to express chromogranin A.16,22 

Our study has several limitations. The first is that the study was a retrospective investigation in a single institution. Second, assessment of lymphovascular and perineural invasion could have been another limiting factor, as endoscopically resected specimens did not retain the complete thickness of rectal wall layers. Third, as we used single-core TMAs for immunohistochemical staining of peptide hormones, there may exist a potential issue of undersampling for intratumoral heterogeneity. However, previous studies have demonstrated that 1-core and 2- to 4-core TMAs represent concordance rates of more than 90% and 95%–97% with immunolabeling of whole sectioned slides, respectively.31,32  Fourth, L-cell phenotype was based only on PYY expression, and GLP-1 and PP immunostaining were not performed for this study although they are as important as PYY expression, because L-cell type was reported in up to 79% of rectal NETs when evaluated with all 3 peptide hormones, including GLP-1, PYY, and PP.14  The most substantial positive aspect of our study is that we evaluated a large series of rectal NET cases (446) with an adequate follow-up period (median, 50.7 months). Further prospective multi-institutional studies would enable practical utilization of peptide hormone expression as a prognostic indicator in rectal NETs. In addition, recruitment of a large number of surgically resected NETs would further validate the association between peptide hormonal expressions and lymphovascular and perineural invasion.

Serotonin or somatostatin expression is observed in a small subset (approximately 15%) of rectal NETs and is associated with an increase in aggressive clinical behavior, including shorter DFS. In contrast, PYY expression is associated with better clinical behavior, including longer DFS. Somatostatin expression is an indicator of poor prognosis with respect to increased recurrence risk in patients with an endoscopically or surgically resected rectal NET. Therefore, additional peptide hormonal immunolabeling, including somatostatin, serotonin, and PYY, could provide survival and/or prognostic information with an endoscopically or surgically resected rectal NET.

1.
Lawrence
B,
Gustafsson
BI,
Chan
A,
Svejda
B,
Kidd
M,
Modlin
IM.
The epidemiology of gastroenteropancreatic neuroendocrine tumors
.
Endocrinol Metab Clin North Am
.
2011
;
40
(1)
:
1
18
,
vii.
2.
Chauhan
A,
Kohn
E,
Del Rivero
J.
Neuroendocrine tumors—less well known, often misunderstood, and rapidly growing in incidence
.
JAMA Oncol
.
2020
;
6
(1)
:
21
22
.
3.
Leoncini
E,
Boffetta
P,
Shafir
M,
Aleksovska
K,
Boccia
S,
Rindi
G.
Increased incidence trend of low-grade and high-grade neuroendocrine neoplasms
.
Endocrine
.
2017
;
58
(2)
:
368
379
.
4.
Cho
MY,
Kim
JM,
Sohn
JH,
et al
Current trends of the incidence and pathological diagnosis of gastroenteropancreatic neuroendocrine tumors (GEP-NETs) in Korea 2000–2009: multicenter study
.
Cancer Res Treat
.
2012
;
44
(3)
:
157
165
.
5.
Chang
JS,
Chen
LT,
Shan
YS,
Chu
PY,
Tsai
CR,
Tsai
HJ.
An updated analysis of the epidemiologic trends of neuroendocrine tumors in Taiwan
.
Sci Rep
.
2021
;
11
(1)
:
7881
.
6.
Masui
T,
Ito
T,
Komoto
I,
Uemoto
S.
Recent epidemiology of patients with gastro-entero-pancreatic neuroendocrine neoplasms (GEP-NEN) in Japan: a population-based study
.
BMC Cancer
.
2020
;
20
(1)
:
1104
.
7.
Kim
JY,
Hong
SM.
Recent updates on neuroendocrine tumors from the gastrointestinal and pancreatobiliary tracts
.
Arch Pathol Lab Med
.
2016
;
140
(5)
:
437
448
.
8.
Klöppel
G,
Couvelard
A,
Perren
A,
et al
ENETS consensus guidelines for the standards of care in neuroendocrine tumors: towards a standardized approach to the diagnosis of gastroenteropancreatic neuroendocrine tumors and their prognostic stratification
.
Neuroendocrinology
.
2009
;
90
(2)
:
162
166
.
9.
Jung
ES,
Kang
YK,
Cho
MY,
et al
Update on the proposal for creating a guideline for cancer registration of the gastrointestinal tumors (I-2)
.
Korean J Pathol
.
2012
;
46
(5)
:
443
453
.
10.
Fiocca
R,
Rindi
G,
Capella
C,
et al
Glucagon, glicentin, proglucagon, PYY, PP and proPP-icosapeptide immunoreactivities of rectal carcinoid tumors and related non-tumor cells
.
Regul Pept
.
1987
;
17
(1)
:
9
29
.
11.
Kim
JY,
Kim
KS,
Kim
KJ,
et al
Non-L-cell immunophenotype and large tumor size in rectal neuroendocrine tumors are associated with aggressive clinical behavior and worse prognosis
.
Am J Surg Pathol
.
2015
;
39
(5)
:
632
643
.
12.
O'Briain
DS,
Dayal
Y,
DeLellis
RA,
Tischler
AS,
Bendon
R,
Wolfe
HJ.
Rectal carcinoids as tumors of the hindgut endocrine cells: a morphological and immunohistochemical analysis
.
Am J Surg Pathol
.
1982
;
6
(2)
:
131
142
.
13.
Sohn
JH,
Cho
MY,
Park
Y,
et al
Prognostic significance of defining L-cell type on the biologic behavior of rectal neuroendocrine tumors in relation with pathological parameters
.
Cancer Res Treat
.
2015
;
47
(4)
:
813
822
.
14.
Lee
SH,
Kim
BC,
Chang
HJ,
et al
Rectal neuroendocrine and L-cell tumors: diagnostic dilemma and therapeutic strategy
.
Am J Surg Pathol
.
2013
;
37
(7)
:
1044
1052
.
15.
Bosman
FT,
Carneiro
F,
Hruban
RH,
Theise
ND.
WHO Classification of Tumours of the Digestive System. 4th ed
.
Lyon, France
:
International Agency for Research on Cancer;
2010
.
WHO Classification of Tumours; vol 3.
16.
WHO Classification of Tumours Editorial Board.
Digestive System Tumours. 5th ed
.
Lyon, France
:
International Agency for Research on Cancer;
2019
.
WHO Classification of Tumours; vol 1.
17.
Kim
JY,
Kim
MS,
Kim
KS,
et al
Clinicopathologic and prognostic significance of multiple hormone expression in pancreatic neuroendocrine tumors
.
Am J Surg Pathol
.
2015
;
39
(5)
:
592
601
.
18.
Kim
J,
Kim
JY,
Oh
EH,
et al
Chromogranin A expression in rectal neuroendocrine tumors is associated with more aggressive clinical behavior and a poorer prognosis
.
Am J Surg Pathol
.
2020
;
44
(11)
:
1496
1505
.
19.
Amin
MB ES,
Greene
F,
Byrd
DR,
et al
AJCC Cancer Staging Manual. 8th ed
.
Chicago, IL
:
Springer International Publishing;
2017
.
20.
Capella
C,
Riva
C,
Rindi
G,
et al
Histopathology, hormone products, and clinicopathological profile of endocrine tumors of the upper small intestine: a study of 44 cases
.
Endocr Pathol
.
1991
;
2
(2)
:
92
110
.
21.
Dayal
Y,
Tallberg
KA,
Nunnemacher
G,
DeLellis
RA,
Wolfe
HJ.
Duodenal carcinoids in patients with and without neurofibromatosis: a comparative study
.
Am J Surg Pathol
.
1986
;
10
(5)
:
348
357
.
22.
Garbrecht
N,
Anlauf
M,
Schmitt
A,
et al
Somatostatin-producing neuroendocrine tumors of the duodenum and pancreas: incidence, types, biological behavior, association with inherited syndromes, and functional activity
.
Endocr Relat Cancer
.
2008
;
15
(1)
:
229
241
.
23.
Burke
AP,
Sobin
LH,
Shekitka
KM,
Federspiel
BH,
Helwig
EB.
Somatostatin-producing duodenal carcinoids in patients with von Recklinghausen's neurofibromatosis: a predilection for black patients
.
Cancer
.
1990
;
65
(7)
:
1591
1595
.
24.
Levy
AD,
Sobin
LH.
From the archives of the AFIP: gastrointestinal carcinoids: imaging features with clinicopathologic comparison
.
Radiographics
.
2007
;
27
(1)
:
237
257
.
25.
Burke
AP,
Thomas
RM,
Elsayed
AM,
Sobin
LH.
Carcinoids of the jejunum and ileum: an immunohistochemical and clinicopathologic study of 167 cases
.
Cancer
.
1997
;
79
(6)
:
1086
1093
.
26.
Bergestuen
DS,
Aabakken
L,
Holm
K,
Vatn
M,
Thiis-Evensen
E.
Small intestinal neuroendocrine tumors: prognostic factors and survival
.
Scand J Gastroenterol
.
2009
;
44
(9)
:
1084
1091
.
27.
McCall
CM,
Shi
C,
Klein
AP,
et al
Serotonin expression in pancreatic neuroendocrine tumors correlates with a trabecular histologic pattern and large duct involvement
.
Hum Pathol
.
2012
;
43
(8)
:
1169
1176
.
28.
Kawamoto
S,
Shi
C,
Hruban
RH,
et al
Small serotonin-producing neuroendocrine tumor of the pancreas associated with pancreatic duct obstruction
.
AJR Am J Roentgenol
.
2011
;
197
(3)
:
W482
W488
.
29.
La Rosa
S,
Franzi
F,
Albarello
L,
et al
Serotonin-producing enterochromaffin cell tumors of the pancreas: clinicopathologic study of 15 cases and comparison with intestinal enterochromaffin cell tumors
.
Pancreas
.
2011
;
40
(6)
:
883
895
.
30.
Portela-Gomes
GM,
Grimelius
L,
Stridsberg
M.
Secretogranin III in human neuroendocrine tumours: a comparative immunohistochemical study with chromogranins A and B and secretogranin II
.
Regul Pept
.
2010
;
165
(1)
:
30
35
.
31.
Camp
RL,
Charette
LA,
Rimm
DL.
Validation of tissue microarray technology in breast carcinoma
.
Lab Invest
.
2000
;
80
(12)
:
1943
1949
.
32.
Watanabe
A,
Cornelison
R,
Hostetter
G.
Tissue microarrays: applications in genomic research
.
Expert Rev Mol Diagn
.
2005
;
5
(2)
:
171
181
.

Author notes

This work was supported by a grant (number FRD2021-15, to Kim) from the Gachon University Gil Medical Center, Incheon, Republic of Korea.

The authors have no relevant financial interest in the products or companies described in this article.

Presented in part at the 19th Annual European Neuroendocrine Tumor Society (ENETS) Conference; March 10–11, 2022; Barcelona, Spain.