- Email: [email protected]
Prognostic significance of neuroendocrine components in gastric carcinomas
European Journal of Cancer (2014) 50, 2802– 2809
Available at www.sciencedirect.com
ScienceDirect journal homepage: www.ejcancer.com
Original Research
Prognostic significance of neuroendocrine components in gastric carcinomas Ji Y. Park a, Min-Hee Ryu b, Young Soo Park a,⇑, Hye Jin Park b, Baek-Yeol Ryoo b, Min Gyu Kim c, Jeong Hwan Yook d, Byung Sik Kim d, Yoon-Koo Kang b a Department of Pathology, Asan Medical Center, University of Ulsan College of Medicine, 88, Olympic-Ro 43-Gil, Songpa-Gu, Seoul 138-736, Republic of Korea b Department of Oncology, Asan Medical Center, University of Ulsan College of Medicine, 88, Olympic-Ro 43-Gil, Songpa-Gu, Seoul 138-736, Republic of Korea c Department of Surgery, Hanyang University Guri Hospital, 153, Gyeongchun-ro, Guri, Gyeonggi-do 471-701, Republic of Korea d Department of Surgery, Asan Medical Center, University of Ulsan College of Medicine, 88, Olympic-Ro 43-Gil, Songpa-Gu, Seoul 138-736, Republic of Korea Received 5 March 2014; received in revised form 6 July 2014; accepted 5 August 2014 Available online 4 September 2014
KEYWORDS Gastric carcinoma Neuroendocrine Component Proportion Prognosis
Abstract Background: Gastric neuroendocrine carcinomas (NECs) and mixed adenoneuroendocrine carcinomas (MANECs) are aggressive tumours but the prognostic significance of a neuroendocrine component in <30% of the tumour remains unclear. Here, the implication of neuroendocrine components in gastric carcinomas was assessed according to proportion. Methods: Surgically resected primary gastric carcinomas with neuroendocrine morphology (NEM; n = 88) from 2000 to 2012 at Asan Medical Center were retrospectively reviewed. Neuroendocrine differentiation (NED) was defined as immunopositivity for one of three neuroendocrine markers (synaptophysin, chromogranin or CD56) within the NEM area. To validate the prognostic significance of NED, these cases were compared with 650 randomly selected gastric adenocarcinomas without NEM from the same time period. Results: Gastric carcinomas with NEM were reclassified as NEC (P70% NED, n = 47), MANEC (30–70% NED, n = 10), gastric carcinoma with 10–30% NED (GCNED, n = 8) and carcinoma with <10% NED (n = 23). The survival rates of patients with P10% NED were significantly poorer than those with <10% NED but no survival difference was observed between NEC and MANEC. In univariate analyses, older age (P60 years), larger tumour size (P4 cm), advanced stage group, P10% NED and lymphovascular or perineural invasion were indicative of a poor prognosis. Stage group and P10% NED remained as independent prognostic factors by multivariate analysis.
⇑ Corresponding author: Tel.: +82 2 3010 5608; fax: +82 2 472 7898.
E-mail address: [email protected] (Y.S. Park). http://dx.doi.org/10.1016/j.ejca.2014.08.004 0959-8049/Ó 2014 Elsevier Ltd. All rights reserved.
J.Y. Park et al. / European Journal of Cancer 50 (2014) 2802– 2809
2803
Conclusions: A minor proportion (10–30%) of NED should not be overlooked in gastric carcinomas with NEM. NED should be carefully evaluated to predict patient outcomes and plan optimal additional therapies. Ó 2014 Elsevier Ltd. All rights reserved.
1. Introduction The recent 2010 World Health Organisation (WHO) classification classified neuroendocrine neoplasm of the stomach into three categories: neuroendocrine tumour, neuroendocrine carcinoma (NEC) and mixed adenoneuroendocrine carcinoma (MANEC) [1]. Neuroendocrine tumours are well-differentiated neoplasms with less than 20 mitotic counts per 10 high-power fields and/or less than a 20% Ki67 index. Grade 1 (carcinoid) and grade 2 neuroendocrine tumours are categorised in this group. NECs are poorly differentiated, high-grade malignant neopla sms that encompass small cell and large cell types. MANECs show both malignant exocrine and neuroendocrine components, with at least 30% of either component [1]. While the behaviour of neuroendocrine tumour is considered to be that of a low to intermediate malignancy [2], NECs, which account for 6–16% of all gastric neuroendocrine neoplasms [1,2], are extremely malignant [3,4].The prognosis for gastric NECs is considered to be poorer than conventional adenocarcinomas because they tend to present in advanced clinical stages and often metastasise to the lymph nodes or the liver even in the early stages of the disease [4,5]. As for the coexistence of NEC and adenocarcinoma, mixed adenocarcinoma components in NECs [5] or combined neuroendocrine components in gastric adenocarcinoma are frequently observed [1], but their documentation in the literature is uncommon [6–9]. Although the WHO defines MANEC as harbouring a more than 30% neuroendocrine component and states that a minor (<30%) neuroendocrine component should not prevent its classification as a conventional adenocarcinoma [1], the 30% cut-off is somewhat arbitrary and not enough data on the prognostic significance of the neuroendocrine component are available. Identification of the neuroendocrine component usually depends on both histological features of neuroendocrine neoplasms and immunohistochemical (IHC) positivity for neuroendocrine markers, such as synaptophysin (SYN), chromogranin A (CGA), neuron cell adhesion molecule (NCAM or CD56), neuron-specific enolase and Leu7 (CD57) [10]. Neuroendocrine marker positivity in gastric carcinomas without regarding histological features of neuroendocrine neoplasm has been reported [11–13], but these studies do not provide enough information on whether neuroendocrine marker positivity should be considered a prognostic indicator. In this study, we investigated the prognostic significance of the neuroendocrine component in surgically
resected primary gastric carcinomas, taking its proportion into account. Clinicopathological features and IHC profiles of neuroendocrine markers in gastric carcinomas with neuroendocrine morphology (NEM) were evaluated. We compared these results with those of conventional gastric carcinomas (with no NEM), according to the proportion of neuroendocrine components, to provide a general understanding of the prognostic influence of neuroendocrine components in gastric carcinomas. 2. Patients and methods 2.1. Selection of patients Gastrectomy cases with the word ‘neuroendocrine’ in the diagnosis that presented between 2000 and 2012 were retrieved from the database of the pathology department of Asan Medical Center. Only primary gastric carcinomas with no prior treatment (chemotherapy and/or radiotherapy) were primarily collected. Cases with grade 1 or grade 2 neuroendocrine components and those with distant metastasis at the time of surgery were excluded. Ki67 index was evaluated in cases with <20 mitotic counts, which were all >20%. A total of 88 cases were finally selected. Clinicopathologic parameters including age, sex, tumour location, depth of invasion, nodal status, lymphovascular and perineural invasion, tumour necrosis, postoperative chemotherapy, recurrence and/or metastasis and survival outcome were obtained by reviewing medical records and haematoxylin and eosin (H&E) stained slides. The pathological tumour-node-metastasis (pTNM) stage was given according to the American Joint Committee on Cancer (AJCC) staging manual for gastric adenocarcinoma. This study adhered to the guidelines established by the Declaration of Helsinki and was approved by the institutional review board. 2.2. Histopathological evaluation and IHC staining From the review of H&E slides, we were able to identify NEM in all 88 cases, as defined by the WHO classification and recent reports [1,4,5]: (1) organoid architectures such as solid nests, sheets, broad trabeculae or rosette formation; (2) nuclear features manifested by hyperchromatic nuclei with finely to coarsely granular, but evenly distributed, chromatin and (3) cytoplasmic features with a scant to moderate amount of
2804
J.Y. Park et al. / European Journal of Cancer 50 (2014) 2802– 2809
slightly eosinophilic, finely granular cytoplasm and indistinct cellular membranes (Fig. 1A). IHC stainings for SYN (clone SP11; NeoMarkers, Fremont, CA), CGA (clone DAK-A3; Dako, Glostrup, Denmark) and CD56 (clone 1B6; Novocastra, Newcastle, United Kingdom (UK)) were performed using a BenchMark XT automated system (Ventana Medical System, Tucson, AZ) on representative sections showing NEM. Briefly, 4-lm-thick sections on adhesive slides were dried at 62 °C for 30 min, followed by standard heat epitope retrieval for 60 min in ethylene diamine tetra acetic acid, pH 8.0. After incubation with primary antibodies, the sections were incubated with UltraView Universal DAB kit (Ventana Medical Systems) and counterstained with Harris haematoxylin. 2.3. Definition and evaluation of neuroendocrine differentiation
was applied after plotting receiver operating characteristic (ROC) curves for predicting the event of patient death (Fig. 1D). Although the cut-off point calculated by the point of convergence that maximised both sensitivity and specificity was 8.5%, we adopted 10% (area under the curve = 0.646, P = 0.019). 2.4. Selection of gastric carcinoma without NEM for comparison To validate the prognostic significance of NED we collected gastric carcinomas without NEM during the same period of time. Of 15,090 primary gastric carcinomas, 650 cases (10 times the cases that showed more than 10% NED) were randomly selected. Cases with prior treatment or initial metastasis were excluded before selection. 2.5. Statistical analyses
Neuroendocrine differentiation (NED) was defined as immunopositivity of one of the three neuroendocrine markers (SYN, CGA and CD56) within the area of the NEM (Fig. 1B and C). Scattered positive cells identified in the non-NEM area did not qualify for this definition (Fig. 1C). The proportion of NED was semiquantified by counting 20 high-power fields (400) with a grid in the eyepiece using the IHC slide of the marker that stained the largest area. After evaluating the proportion of NED, the cut-off of 10% immunopositivity
A chi-square test or Fisher’s exact test was used to evaluate the correlation between the clinicopathological parameters and the presence of NED. Recurrence-free survival (RFS), disease-specific survival (DSS) and overall survival (OS) were calculated from the time of surgery. The survival curves were analysed by the Kaplan–Meier method and the differences between the survival curves were estimated by a log-rank test. Univariate and multivariate Cox proportional hazards anal-
Fig. 1. Representative histological and immunophenotypic features of gastric carcinomas with neuroendocrine differentiation (NED) (A–C) and receiver operating characteristic (ROC) curve according to the proportion of NED (D). The tumour shows both neuroendocrine (left) and glandular (right) morphology (A, original magnification, 100). Tumour cells with neuroendocrine morphology (NEM) are immunopositive for synaptophysin (SYN) (B, 100) and chromogranin A (CGA) (C, 100). The point at which both sensitivity and specificity for the prediction of patient death were maximised is just below 10% (D, asterisk).
J.Y. Park et al. / European Journal of Cancer 50 (2014) 2802– 2809
yses were performed to evaluate prognostic significance for the RFS and OS. All statistical analyses were performed using SPSS software (Chicago, IL). 3. Results 3.1. Classification and IHC staining pattern of gastric carcinomas with NEM After a histological and IHC staining review of 88 cases with NEM, we were able to categorise each case into one of four groups: (1) NEC (n = 47), P70% NED; (2) MANEC (n = 10), 30–70% NED; (3) gastric carcinoma with NED (GCNED, n = 8), 10–30% NED and (4) gastric carcinoma with <10% NED (n = 23). NEC included pure small cell (n = 7), pure large cell (n = 21) and mixed cases with minor components (<30%) of non-NEM carcinomas (n = 19). All mixed cases were large cell NECs. In MANEC, the mixed components were poorly differentiated (n = 8), moderately differentiated (n = 1) and signet-ring cell carcinoma (n = 1). The mixed components in GCNED were poorly differentiated (n = 5), moderately differentiated (n = 2) and signet-ring cell carcinoma (n = 1). The mixed components in gastric carcinoma with <10% NED were poorly differentiated (n = 20), moderately differentiated (n = 2) and signet-ring cell carcinoma (n = 1). When the IHC staining pattern of 65 cases harbouring P10% NED (NEC, MANEC and GCNED) was analysed, SYN was positive in 61 cases (93.8%), CGA was positive in 56 (86.2%) and CD56 was positive in 49 (75.4%). The immunopositivity of each neuroendocrine marker is depicted in Fig. 2. Expression for all three markers was recognised in 38 cases (58.5%), two markers in 25 (38.5%) and single marker in 2 (3.0%; SYN and CD56). 3.2. Clinicopathological features of gastric carcinomas with NED The clinicopathologic parameters of each group according to the proportion of NED are summarised in Table 1. Additional 650 cases of gastric carcinoma without NEM were randomly selected for comparison. Significant proportions (19.1%) of NEC were located in the cardia or fundus, whereas the other types of carcinomas were concentrated in the body or antrum. Lymphovascular invasion was more frequently identified in
2805
carcinomas with NED. Perineural tumour invasion was significantly increased in GCNED. Recurrence or metastasis was more commonly observed in the NEC, MANEC and GCNED than in carcinomas with <10% NED or without NEM. Since this was a simple comparison without considering the time of follow-up, we also compared survival curves by Kaplan–Meier method (Fig. 3). The most common metastatic site was the liver, followed by the lymph nodes and peritoneum. Other parameters such as age, sex, tumour size, pT, pN stage and adjuvant chemotherapy status were not significantly different. More than two-thirds of the patients received adjuvant chemotherapy in all groups. In 65 cases with P10% NED, 44 patients (67.7%) received adjuvant chemotherapy, which were 5-fluorouracil (5-FU) alone (n = 14) or 5-FU plus platinum-based chemotherapy (n = 20) in 34 patients and etoposide plus cisplatin (EP) in 10 patients. The EP regimen was used only in NECs. Detailed regimens of adjuvant chemotherapy in each patient group and stage are given in the supplementary Table 1. 3.3. Prognostic significance of NED in gastric carcinomas We then analysed patient outcomes (RFS, DSS and OS) according to the proportion of NED. The median follow-up period was 43.5 months (range, 1.5– 159.1 months). During the follow-up period, 21 patients (44.7%) in NEC, 5 (50.0%) in MANEC, 6 (75.0%) in GCNED and 1 (4%) in gastric carcinoma with <10% NED died. The causes of death were disease-related in 23 cases (69.7%) and uncertain in 10 cases (30.3%). In the gastric carcinoma without NEM group, 84 patients (12.9%) expired; 59 (70.2%) of which were diseaserelated, 16 (19.0%) were uncertain and 9 (10.7%) were other causes. The trends in the Kaplan–Meier survival curves were identical for RFS, DSS as well as OS (Fig. 3 and supplementary figures). Carcinomas with <10% NED (88.9% 5-year OS) and gastric carcinomas without NEM (85.2% 5-year OS) showed significantly better RFS and OS than the other groups (Fig. 3A and C). The 5year OS rates for NEC, MANEC and GCNED were 59.1%, 53.5% and 16.7%, respectively. Therefore, we grouped the patients into two groups using 10% NED as a cut-off. As expected, RFS and OS were significantly better in patients with <10% NED (85.6% 5-year OS)
Fig. 2. An illustration of immunohistochemical staining patterns in gastric carcinomas with P10% neuroendocrine differentiation (NED) . Immunopositivity is indicated by differently coloured boxes: black (triple expression), dark grey (double expression) and light grey (single expression). Abbreviations: S, SYN; C, CGA; D, CD56.
2806
J.Y. Park et al. / European Journal of Cancer 50 (2014) 2802– 2809
Table 1 Clinicopathological features of gastric carcinomas with or without neuroendocrine differentiation (NED). Parameters
Neuroendocrine carcinomas (NEC) (n = 47)
Adenoneuroendocrine carcinoma (MANEC) (n = 10)
Gastric carcinoma with 10–30% NED (GCNED) (n = 8)
<10% NED (n = 23)
GC without NEM (n = 650)
Age (years) [median (range)] Gender M F Location C+F B Ant Diameter (cm) [median (range)] pT stage T1 T2 T3 T4 pN stage N0 N1 N2 N3 LVI – + PNI – + Stage group I II III Adjuvant CTx – + Follow-up (months) [median (range)] Recurrence or – metastasis + Liver LN Peritoneum Lung Others
64.0 (40–79) 34 (72.3%) 13 (27.7%) 9 (19.1%)a 20 (42.6%)a 18 (38.3%)a 5.0 (1.7–13.0) 7 (14.9%) 11 (23.4%) 21 (44.7%) 8 (17.0%) 19 (40.4%) 10 (21.3%) 12 (25.5%) 6 (12.8%) 5 (10.6%)a 42 (89.4%)a 30 (63.8%) 17 (36.2%) 12 (25.5%) 19 (40.5%) 16 (34.0%) 16 (34.0%) 31 (66.0%) 52.6 (4.2–108.7)
65.5 (55–70) 10 (100%) 0 (0.0%) 0 (0.0%) 2 (20.0%) 8 (80.0%) 5.6 (1.0–12.0) 2 (20.0%) 1 (10.0%) 5 (50.0%) 2 (20.0%) 2 (20.0%) 4 (40.0%) 2 (20.0%) 2 (20.0%) 1 (10.0%)a 9 (90.0%)a 5 (50.0%) 5 (50.0%) 3 (30.0%) 3 (30.0%) 4 (40.0%) 3 (30.0%) 7 (70.0%) 50.4 (16.3–126.8)
56.5 (35–77) 5 (62.5%) 3 (37.5%) 0 (0.0%) 5 (62.5%) 3 (37.5%) 6.0 (4.2–9.0) 0 (0.0%) 1 (12.5%) 5 (62.5%) 2 (25.0%) 1 (12.5%) 2 (25.0%) 3 (37.5%) 2 (25.0%) 1 (12.5%)a 7 (87.5%)a 2 (25.0%)a 6 (75.0%)a 1 (12.5%) 2(25.0%) 5 (62.5%) 2 (25.0%) 6 (75.0%) 20.1 (6.3–76.4)
62.0 (28–96) 395 (60.8%) 255 (39.2%) 47 (7.2%) 348 (53.5%) 255 (39.3%) 5.0 (1.0–33.5) 120 (18.5%) 156 (24.0%) 225 (34.6%) 149 (22.9%) 272 (41.8%) 115 (17.7%) 124 (19.1%) 139 (21.4%) 327 (50.3%) 323 (49.7%) 448 (68.9%) 202 (31.1%) 188 (28.9%) 220 (33.8%) 242 (37.2%) 219 (33.7%) 431 (66.3%) 41.5 (1.5–159.1)
31 (64.6%)a
6 (60.0%)a
4 (50.0%)a
60.0 (32–82) 15 (65.2%) 8 (34.8%) 0 (0.0%) 13 (56.5%) 10 (43.5%) 5.6 (2.5–17.0) 4 (17.4%) 8 (34.8%) 8 (34.8%) 3 (13.0%) 11 (47.8%) 5 (21.7%) 5 (21.7%) 2 (8.8%) 5 (21.7%)a 18 (78.3%)a 15 (65.2%) 8 (34.8%) 9 (39.1%) 9 (39.1%) 5 (21.8%) 8 (34.8%) 15 (65.2%) 62.6 (3.2– 119.2) 22 (96%)
17 (35.4%)a 8 5 5 0 1b
4 (40.0%)a 3 1 0 1b 2b
4 (50.0%)a 1 0 3 0 0
1 (4%) 1 0 0 0 0
65 (10.0%) 7 7 27 4b 21
585 (90.0%)
Abbreviations: GC, gastric carcinoma; M, male; F, female; C, cardia; F, fundus; B, body; Ant, antrum; LVI, lymphovascular invasion; PNI, perineural invasion; CTx, chemotherapy; LN, lymph node. a Parameters with statistically significant differences (P < 0.05) compared with GC without NED are indicated in bold letters. b Some patients recurred in multiple sites.
than in those harbouring P10% NED (53.3% 5-year OS; Fig. 3B and D). This prognostic significance of NED was also valid in all stage groups. 3.4. Prognostic factors in gastric carcinomas according to NED We subsequently evaluated other factors associated with RFS, DSS or OS in all gastric carcinomas. In univariate analyses, old age (P60 years), larger tumour size (P4 cm), advanced stage group, presence of NED, lymphovascular and perineural invasion were poor prognostic factors for both RFS and OS (Tables 2 and 3). Patient’s sex, tumour location, tumour necrosis, type and number of positive neuroendocrine markers were not associated with patient survival. Adjuvant chemotherapy or the type of adjuvant chemotherapy did not exert prognostic effects on patient outcome.
After multivariate analysis, stage group and the P10% NED remained as independent prognostic factors for OS, RFS and DDS (Tables 2 and 3 and supplementary Table 2). The adjusted hazard ratio for OS in patients with P10% NED was significantly higher compared to those of perineural invasion or the stage group. 4. Discussion Although many previous studies have already elucidated the poor prognosis of NEC in the stomach [14,15], few studies have examined the prognostic significance of neuroendocrine components in gastric carcinomas, even if the coexistence of NEC and adenocarcinoma is frequently observed [16,17]. In our study, we looked for the NEM in gastric carcinomas and semiquantified the proportion of immunopositivity
J.Y. Park et al. / European Journal of Cancer 50 (2014) 2802– 2809
2807
Fig. 3. Kaplan–Meier survival curves grouped by proportion of neuroendocrine differentiation (NED) in gastric carcinomas. Gastric carcinomas with less than 10% NED and gastric carcinomas without neuroendocrine morphology (NEM) show better survival than neuroendocrine carcinoma (NEC), adenoneuroendocrine carcinoma (MANEC) and gastric carcinoma with 10–30% NED (GCNED) in recurrence-free survival (RFS) and overall survival (OS) (A and C). The prognostic significance of NED is more clearly demonstrated when the cut-off of 10% NED is applied (B and D).
of neuroendocrine markers (as the NED percentage). We used three commonly used neuroendocrine markers (SYN, CGA and CD56) in combination because profiles of various neuroendocrine markers have shown inconsistent results [10,14,18]. The most notable finding of our study was that NEC, MANEC and GCNED showed poorer outcome than gastric carcinomas with <10% NED or without NEM (Fig. 3A and C). Although the survival rates of GCNED were worse than those of NEC and MANEC, this finding can be explained by the fact that 62.5% of patients with GCNED were stage III, whereas only 34% and 40% of patients in NEC and MANEC were at that stage (Table 1). These results agree with the previously reported observation that there was no significant difference in survival between NEC and MANEC [15]. The results of this study clearly demonstrate that gastric carcinomas with P10% NED should be differentiated from conventional adenocarcinomas since they are similar to NEC not only in terms of prognosis, but also in some clinicopathological parameters. The current WHO classification [1] provides 30% of neuroendocrine components as a cut-off and states that minor
components can be overlooked, whereas Jiang et al. [4] suggested a 20% cut-off while leaving some room for further studies to validate the use of 10% or less as cut-off. To the best of our knowledge, this is the first large-scale study to provide a neuroendocrine proportion cut-off that can be used to distinguish from conventional gastric adenocarcinomas. We have also found that P10% NED is an independent poor prognostic factor, although this result needs to be verified in other independent patient groups. Surgical respectability has been reported to be the most important prognostic factor in gastric NECs [15]. Therefore, radical gastrectomy with appropriate lymph node dissection would be the treatment of choice in localised diseases. It has recently been established that adjuvant S-1 [19] or capecitabine plus oxaliplatin [20] is effective in gastric carcinomas after D2 surgery. However, the adjuvant treatments of NECs or gastric carcinomas with NED are not standardised. The EP regimen, which is used for small cell carcinoma of the lung, is the most widely used treatment in NECs [21] but the biological similarity between small cell carcinoma of the lung and gastric NEC is controversial [22]
2808
J.Y. Park et al. / European Journal of Cancer 50 (2014) 2802– 2809
Table 2 Univariate and multivariate analyses for overall survival (OS). Clinicopathologic variables
Age (years) Diameter (cm) Stage group
NED LVI PNI
<60 versus P60 <4 versus P4 I II III <10% versus P10% – versus + – versus +
Univariate analysis
Multivariate analysis
HR
95% CI
P value
1.987 3.200 1 3.965 10.613 3.920 2.339 1.915
1.336–2.953 1.913–5.352 Reference 1.836–8.566 5.114–22.025 2.610–5.890 1.735–3.152 1.323–2.773
0.001 <0.0001 <0.0001 <0.0001 <0.0001 <0.0001 0.001
HR
95% CI
P value
1 5.200 7.456 13.459
Reference 1.181–22.905 1.696–32.775 1.824–99.302
0.029 0.008 0.011
2.294
1.111–4.739
0.025
Abbreviations: HR, hazard ratio; CI, confidence interval; NED, neuroendocrine differentiation; LVI, lymphovascular invasion; PNI, perineural invasion.
Table 3 Univariate and multivariate analyses for recurrence-free survival (RFS). Clinicopathologic variables
Age (years) Diameter (cm) Stage group
NED LVI PNI
<60 versus P60 <4 versus P4 I II III <10% versus P10% – versus + – versus +
Univariate analysis
Multivariate analysis
HR
95% CI
P value
HR
95% CI
P value
1.597 3.103 1 3.985 10.189 3.196 2.438 1.886
1.152–2.218 2.011–4.790 Reference 2.067–7.684 5.466–18.993 2.197–4.649 1.734–3.426 1.376–2.584
0.005 <0.0001
1.447
1.041–2.011
0.028
1 3.878 10.154 3.060
Reference 2.011–7.480 5.447–18.928 2.097–4.464
<0.0001 <0.0001 <0.0001
<0.0001 <0.0001 <0.0001 <0.0001 <0.0001
Abbreviations: HR, hazard ratio; CI, confidence interval; NED, neuroendocrine differentiation; LVI, lymphovascular invasion; PNI, perineural invasion.
and other regimens have also been reported to be effective [23,24]. Some authors recommend adjuvant chemotherapy even in the early stage [25]. Since most previous studies have evaluated heterogeneous tumours, commonly lacking standardised diagnostic criteria the proportion of the non-NED (epithelial) component could have affected the response [23]. Therefore, multi-institutional or multinational prospective studies that consider the proportion of NED should be designed for the development of proper guidelines for the treatment of these rare but grave diseases. The major limitation of our study was that we were unable to validate the 10% cut-off of NED in other cohorts because it was a single centre study. Another was lack of data on expression of neuroendocrine markers in the control group. It has been reported that neuroendocrine marker positivity is associated with higher stage and poor survival in undifferentiated colorectal carcinomas [26,27] and gastric carcinomas that express neuroendocrine markers without NEM might also have different clinical outcomes. Therefore, significance of neuroendocrine marker positivity in gastric carcinoma regardless of NEM should be evaluated in future studies. Nevertheless, our results suggest that there is no survival benefit caused by a small proportion of NED and that even a minor proportion (10–30%) of NED should not be overlooked.
In summary, we evaluated the clinicopathologic features of gastric carcinomas with NED and discovered that there are no significant survival differences among NEC, and GCNED. Therefore, gastric carcinomas harbouring P10% NED should be differentiated from conventional adenocarcinomas. The P10% NED cut-off may be practically useful and may serve as an informative parameter for predicting patient outcome and establishing optimal therapeutic guidelines for gastric carcinomas with NED. Conflict of interest statement None declared. Acknowledgements This study was supported by the Proteogenomic Research Program through a National Research Foundation of Korea grant funded by the Korea government (Ministry of Science, ICT & Future Planning). Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/ 10.1016/j.ejca.2014.08.004.
J.Y. Park et al. / European Journal of Cancer 50 (2014) 2802– 2809
References [1] Solcia E, Arnold R, Capella C, et al. Neuroendocrine neoplasms of the stomach. In: Bosman FT, Jaffe ES, Lakhani SR, Ohgaki H, editors. WHO classification of tumors of the digestive system. Lyon: International Agency for Research on Cancer; 2010. p. 64–8. [2] Rindi G, Azzoni C, La Rosa S, et al. ECL cell tumor and poorly differentiated endocrine carcinoma of the stomach: prognostic evaluation by pathological analysis. Gastroenterology 1999;116:532–42. [3] Matsui K, Jin XM, Kitagawa M, Miwa A. Clinicopathologic features of neuroendocrine carcinomas of the stomach: appraisal of small cell and large cell variants. Arch Pathol Lab Med 1998;122:1010–7. [4] Jiang SX, Mikami T, Umezawa A, Saegusa M, Kameya T, Okayasu I. Gastric large cell neuroendocrine carcinomas: a distinct clinicopathologic entity. Am J Surg Pathol 2006;30:945–53. [5] Kubota T, Ohyama S, Hiki N, Nunobe S, Yamamoto N, Yamaguchi T. Endocrine carcinoma of the stomach: clinicopathological analysis of 27 surgically treated cases in a single institute. Gastric Cancer 2012;15:323–30. [6] Kim JJ, Kim JY, Hur H, Cho YK, Han SU. Clinicopathologic significance of gastric adenocarcinoma with neuroendocrine features. J Gastric Cancer 2011;11:195–9. [7] Miguchi M, Iseki M, Shimatani K. Advanced gastric neuroendocrine carcinoma with an adenocarcinoma component. Case Rep Gastroenterol 2012;6:52–7. [8] Lee JH, Kim HW, Kang DH, Choi CW, Park SB, Kim SH. A gastric composite tumor with an adenocarcinoma and a neuroendocrine carcinoma: a case report. Clin Endosc 2013;46:280–3. [9] Terada T, Maruo H. Simultaneous large cell neuroendocrine carcinoma and adenocarcinoma of the stomach. World J Gastroenterol 2011;17:4831–4. [10] Erickson LA, Lloyd RV. Practical markers used in the diagnosis of endocrine tumors. Adv Anat Pathol 2004;11:175–89. [11] Blumenfeld W, Chandhoke DK, Sagerman P, Turi GK. Neuroendocrine differentiation in gastric adenocarcinomas. An immunohistochemical study. Arch Pathol Lab Med 1996;120:478–81. [12] Qvigstad G, Qvigstad T, Westre B, Sandvik AK, Brenna E, Waldum HL. Neuroendocrine differentiation in gastric adenocarcinomas associated with severe hypergastrinemia and/or pernicious anemia. APMIS 2002;110:132–9. [13] Eren F, Celikel C, Gulluoglu B. Neuroendocrine differentiation in gastric adenocarcinomas; correlation with tumor stage and expression of VEGF and p53. POR 2004;10:47–51. [14] La Rosa S, Inzani F, Vanoli A, et al. Histologic characterization and improved prognostic evaluation of 209 gastric neuroendocrine neoplasms. Hum Pathol 2011;42:1373–84.
2809
[15] Ishida M, Sekine S, Fukagawa T, et al. Neuroendocrine carcinoma of the stomach: morphologic and immunohistochemical characteristics and prognosis. Am J Surg Pathol 2013;37:949–59. [16] Waldum HL, Aase S, Kvetnoi I, et al. Neuroendocrine differentiation in human gastric carcinoma. Cancer-Am Cancer Soc 1998;83:435–44. [17] Nishikura K, Watanabe H, Iwafuchi M, Fujiwara T, Kojima K, Ajioka Y. Carcinogenesis of gastric endocrine cell carcinoma: analysis of histopathology and p53 gene alteration. Gastric Cancer 2003;6:203–9. [18] Rayhan N, Sano T, Qian ZR, Obari AK, Hirokawa M. Histological and immunohistochemical study of composite neuroendocrine-exocrine carcinomas of the stomach. J Med Invest 2005;52:191–202. [19] Sakuramoto S, Sasako M, Yamaguchi T, et al. Adjuvant chemotherapy for gastric cancer with S-1, an oral fluoropyrimidine. N Engl J Med 2007;357:1810–20. [20] Bang YJ, Kim YW, Yang HK, et al. Adjuvant capecitabine and oxaliplatin for gastric cancer after D2 gastrectomy (CLASSIC): a phase 3 open-label, randomised controlled trial. Lancet 2012;379:315–21. [21] Okita NT, Kato K, Takahari D, et al. Neuroendocrine tumors of the stomach: chemotherapy with cisplatin plus irinotecan is effective for gastric poorly-differentiated neuroendocrine carcinoma. Gastric Cancer 2011;14:161–5. [22] Hainsworth JD, Spigel DR, Litchy S, Greco FA. Phase II trial of paclitaxel, carboplatin, and etoposide in advanced poorly differentiated neuroendocrine carcinoma: a Minnie Pearl Cancer Research Network Study. J Clin Oncol 2006;24:3548–54. [23] Fazio N, Spada F, Giovannini M. Chemotherapy in gastroenteropancreatic (GEP) neuroendocrine carcinomas (NEC): a critical view. Cancer Treat Rev 2013;39:270–4. [24] Meyer T, Qian W, Caplin ME, et al. Capecitabine and streptozocin ± cisplatin in advanced gastroenteropancreatic neuroendocrine tumours. Eur J Cancer 2014;50:902–11. [25] Fukuda T, Ohnishi Y, Nishimaki T, Ohtani H, Tachikawa S. Early gastric cancer of the small cell type. Am J Gastroenterol 1988;83:1176–9. [26] Grabowski P, Schindler I, Anagnostopoulos I, et al. Neuroendocrine differentiation is a relevant prognostic factor in stage III–IV colorectal cancer. Eur J Gastroenterol Hepatol 2001;13:405–11. [27] Grabowski P, Schonfelder J, Ahnert-Hilger G, et al. Expression of neuroendocrine markers: a signature of human undifferentiated carcinoma of the colon and rectum. Virchows Arch 2002;441:256–63.
Available at www.sciencedirect.com
ScienceDirect journal homepage: www.ejcancer.com
Original Research
Prognostic significance of neuroendocrine components in gastric carcinomas Ji Y. Park a, Min-Hee Ryu b, Young Soo Park a,⇑, Hye Jin Park b, Baek-Yeol Ryoo b, Min Gyu Kim c, Jeong Hwan Yook d, Byung Sik Kim d, Yoon-Koo Kang b a Department of Pathology, Asan Medical Center, University of Ulsan College of Medicine, 88, Olympic-Ro 43-Gil, Songpa-Gu, Seoul 138-736, Republic of Korea b Department of Oncology, Asan Medical Center, University of Ulsan College of Medicine, 88, Olympic-Ro 43-Gil, Songpa-Gu, Seoul 138-736, Republic of Korea c Department of Surgery, Hanyang University Guri Hospital, 153, Gyeongchun-ro, Guri, Gyeonggi-do 471-701, Republic of Korea d Department of Surgery, Asan Medical Center, University of Ulsan College of Medicine, 88, Olympic-Ro 43-Gil, Songpa-Gu, Seoul 138-736, Republic of Korea Received 5 March 2014; received in revised form 6 July 2014; accepted 5 August 2014 Available online 4 September 2014
KEYWORDS Gastric carcinoma Neuroendocrine Component Proportion Prognosis
Abstract Background: Gastric neuroendocrine carcinomas (NECs) and mixed adenoneuroendocrine carcinomas (MANECs) are aggressive tumours but the prognostic significance of a neuroendocrine component in <30% of the tumour remains unclear. Here, the implication of neuroendocrine components in gastric carcinomas was assessed according to proportion. Methods: Surgically resected primary gastric carcinomas with neuroendocrine morphology (NEM; n = 88) from 2000 to 2012 at Asan Medical Center were retrospectively reviewed. Neuroendocrine differentiation (NED) was defined as immunopositivity for one of three neuroendocrine markers (synaptophysin, chromogranin or CD56) within the NEM area. To validate the prognostic significance of NED, these cases were compared with 650 randomly selected gastric adenocarcinomas without NEM from the same time period. Results: Gastric carcinomas with NEM were reclassified as NEC (P70% NED, n = 47), MANEC (30–70% NED, n = 10), gastric carcinoma with 10–30% NED (GCNED, n = 8) and carcinoma with <10% NED (n = 23). The survival rates of patients with P10% NED were significantly poorer than those with <10% NED but no survival difference was observed between NEC and MANEC. In univariate analyses, older age (P60 years), larger tumour size (P4 cm), advanced stage group, P10% NED and lymphovascular or perineural invasion were indicative of a poor prognosis. Stage group and P10% NED remained as independent prognostic factors by multivariate analysis.
⇑ Corresponding author: Tel.: +82 2 3010 5608; fax: +82 2 472 7898.
E-mail address: [email protected] (Y.S. Park). http://dx.doi.org/10.1016/j.ejca.2014.08.004 0959-8049/Ó 2014 Elsevier Ltd. All rights reserved.
J.Y. Park et al. / European Journal of Cancer 50 (2014) 2802– 2809
2803
Conclusions: A minor proportion (10–30%) of NED should not be overlooked in gastric carcinomas with NEM. NED should be carefully evaluated to predict patient outcomes and plan optimal additional therapies. Ó 2014 Elsevier Ltd. All rights reserved.
1. Introduction The recent 2010 World Health Organisation (WHO) classification classified neuroendocrine neoplasm of the stomach into three categories: neuroendocrine tumour, neuroendocrine carcinoma (NEC) and mixed adenoneuroendocrine carcinoma (MANEC) [1]. Neuroendocrine tumours are well-differentiated neoplasms with less than 20 mitotic counts per 10 high-power fields and/or less than a 20% Ki67 index. Grade 1 (carcinoid) and grade 2 neuroendocrine tumours are categorised in this group. NECs are poorly differentiated, high-grade malignant neopla sms that encompass small cell and large cell types. MANECs show both malignant exocrine and neuroendocrine components, with at least 30% of either component [1]. While the behaviour of neuroendocrine tumour is considered to be that of a low to intermediate malignancy [2], NECs, which account for 6–16% of all gastric neuroendocrine neoplasms [1,2], are extremely malignant [3,4].The prognosis for gastric NECs is considered to be poorer than conventional adenocarcinomas because they tend to present in advanced clinical stages and often metastasise to the lymph nodes or the liver even in the early stages of the disease [4,5]. As for the coexistence of NEC and adenocarcinoma, mixed adenocarcinoma components in NECs [5] or combined neuroendocrine components in gastric adenocarcinoma are frequently observed [1], but their documentation in the literature is uncommon [6–9]. Although the WHO defines MANEC as harbouring a more than 30% neuroendocrine component and states that a minor (<30%) neuroendocrine component should not prevent its classification as a conventional adenocarcinoma [1], the 30% cut-off is somewhat arbitrary and not enough data on the prognostic significance of the neuroendocrine component are available. Identification of the neuroendocrine component usually depends on both histological features of neuroendocrine neoplasms and immunohistochemical (IHC) positivity for neuroendocrine markers, such as synaptophysin (SYN), chromogranin A (CGA), neuron cell adhesion molecule (NCAM or CD56), neuron-specific enolase and Leu7 (CD57) [10]. Neuroendocrine marker positivity in gastric carcinomas without regarding histological features of neuroendocrine neoplasm has been reported [11–13], but these studies do not provide enough information on whether neuroendocrine marker positivity should be considered a prognostic indicator. In this study, we investigated the prognostic significance of the neuroendocrine component in surgically
resected primary gastric carcinomas, taking its proportion into account. Clinicopathological features and IHC profiles of neuroendocrine markers in gastric carcinomas with neuroendocrine morphology (NEM) were evaluated. We compared these results with those of conventional gastric carcinomas (with no NEM), according to the proportion of neuroendocrine components, to provide a general understanding of the prognostic influence of neuroendocrine components in gastric carcinomas. 2. Patients and methods 2.1. Selection of patients Gastrectomy cases with the word ‘neuroendocrine’ in the diagnosis that presented between 2000 and 2012 were retrieved from the database of the pathology department of Asan Medical Center. Only primary gastric carcinomas with no prior treatment (chemotherapy and/or radiotherapy) were primarily collected. Cases with grade 1 or grade 2 neuroendocrine components and those with distant metastasis at the time of surgery were excluded. Ki67 index was evaluated in cases with <20 mitotic counts, which were all >20%. A total of 88 cases were finally selected. Clinicopathologic parameters including age, sex, tumour location, depth of invasion, nodal status, lymphovascular and perineural invasion, tumour necrosis, postoperative chemotherapy, recurrence and/or metastasis and survival outcome were obtained by reviewing medical records and haematoxylin and eosin (H&E) stained slides. The pathological tumour-node-metastasis (pTNM) stage was given according to the American Joint Committee on Cancer (AJCC) staging manual for gastric adenocarcinoma. This study adhered to the guidelines established by the Declaration of Helsinki and was approved by the institutional review board. 2.2. Histopathological evaluation and IHC staining From the review of H&E slides, we were able to identify NEM in all 88 cases, as defined by the WHO classification and recent reports [1,4,5]: (1) organoid architectures such as solid nests, sheets, broad trabeculae or rosette formation; (2) nuclear features manifested by hyperchromatic nuclei with finely to coarsely granular, but evenly distributed, chromatin and (3) cytoplasmic features with a scant to moderate amount of
2804
J.Y. Park et al. / European Journal of Cancer 50 (2014) 2802– 2809
slightly eosinophilic, finely granular cytoplasm and indistinct cellular membranes (Fig. 1A). IHC stainings for SYN (clone SP11; NeoMarkers, Fremont, CA), CGA (clone DAK-A3; Dako, Glostrup, Denmark) and CD56 (clone 1B6; Novocastra, Newcastle, United Kingdom (UK)) were performed using a BenchMark XT automated system (Ventana Medical System, Tucson, AZ) on representative sections showing NEM. Briefly, 4-lm-thick sections on adhesive slides were dried at 62 °C for 30 min, followed by standard heat epitope retrieval for 60 min in ethylene diamine tetra acetic acid, pH 8.0. After incubation with primary antibodies, the sections were incubated with UltraView Universal DAB kit (Ventana Medical Systems) and counterstained with Harris haematoxylin. 2.3. Definition and evaluation of neuroendocrine differentiation
was applied after plotting receiver operating characteristic (ROC) curves for predicting the event of patient death (Fig. 1D). Although the cut-off point calculated by the point of convergence that maximised both sensitivity and specificity was 8.5%, we adopted 10% (area under the curve = 0.646, P = 0.019). 2.4. Selection of gastric carcinoma without NEM for comparison To validate the prognostic significance of NED we collected gastric carcinomas without NEM during the same period of time. Of 15,090 primary gastric carcinomas, 650 cases (10 times the cases that showed more than 10% NED) were randomly selected. Cases with prior treatment or initial metastasis were excluded before selection. 2.5. Statistical analyses
Neuroendocrine differentiation (NED) was defined as immunopositivity of one of the three neuroendocrine markers (SYN, CGA and CD56) within the area of the NEM (Fig. 1B and C). Scattered positive cells identified in the non-NEM area did not qualify for this definition (Fig. 1C). The proportion of NED was semiquantified by counting 20 high-power fields (400) with a grid in the eyepiece using the IHC slide of the marker that stained the largest area. After evaluating the proportion of NED, the cut-off of 10% immunopositivity
A chi-square test or Fisher’s exact test was used to evaluate the correlation between the clinicopathological parameters and the presence of NED. Recurrence-free survival (RFS), disease-specific survival (DSS) and overall survival (OS) were calculated from the time of surgery. The survival curves were analysed by the Kaplan–Meier method and the differences between the survival curves were estimated by a log-rank test. Univariate and multivariate Cox proportional hazards anal-
Fig. 1. Representative histological and immunophenotypic features of gastric carcinomas with neuroendocrine differentiation (NED) (A–C) and receiver operating characteristic (ROC) curve according to the proportion of NED (D). The tumour shows both neuroendocrine (left) and glandular (right) morphology (A, original magnification, 100). Tumour cells with neuroendocrine morphology (NEM) are immunopositive for synaptophysin (SYN) (B, 100) and chromogranin A (CGA) (C, 100). The point at which both sensitivity and specificity for the prediction of patient death were maximised is just below 10% (D, asterisk).
J.Y. Park et al. / European Journal of Cancer 50 (2014) 2802– 2809
yses were performed to evaluate prognostic significance for the RFS and OS. All statistical analyses were performed using SPSS software (Chicago, IL). 3. Results 3.1. Classification and IHC staining pattern of gastric carcinomas with NEM After a histological and IHC staining review of 88 cases with NEM, we were able to categorise each case into one of four groups: (1) NEC (n = 47), P70% NED; (2) MANEC (n = 10), 30–70% NED; (3) gastric carcinoma with NED (GCNED, n = 8), 10–30% NED and (4) gastric carcinoma with <10% NED (n = 23). NEC included pure small cell (n = 7), pure large cell (n = 21) and mixed cases with minor components (<30%) of non-NEM carcinomas (n = 19). All mixed cases were large cell NECs. In MANEC, the mixed components were poorly differentiated (n = 8), moderately differentiated (n = 1) and signet-ring cell carcinoma (n = 1). The mixed components in GCNED were poorly differentiated (n = 5), moderately differentiated (n = 2) and signet-ring cell carcinoma (n = 1). The mixed components in gastric carcinoma with <10% NED were poorly differentiated (n = 20), moderately differentiated (n = 2) and signet-ring cell carcinoma (n = 1). When the IHC staining pattern of 65 cases harbouring P10% NED (NEC, MANEC and GCNED) was analysed, SYN was positive in 61 cases (93.8%), CGA was positive in 56 (86.2%) and CD56 was positive in 49 (75.4%). The immunopositivity of each neuroendocrine marker is depicted in Fig. 2. Expression for all three markers was recognised in 38 cases (58.5%), two markers in 25 (38.5%) and single marker in 2 (3.0%; SYN and CD56). 3.2. Clinicopathological features of gastric carcinomas with NED The clinicopathologic parameters of each group according to the proportion of NED are summarised in Table 1. Additional 650 cases of gastric carcinoma without NEM were randomly selected for comparison. Significant proportions (19.1%) of NEC were located in the cardia or fundus, whereas the other types of carcinomas were concentrated in the body or antrum. Lymphovascular invasion was more frequently identified in
2805
carcinomas with NED. Perineural tumour invasion was significantly increased in GCNED. Recurrence or metastasis was more commonly observed in the NEC, MANEC and GCNED than in carcinomas with <10% NED or without NEM. Since this was a simple comparison without considering the time of follow-up, we also compared survival curves by Kaplan–Meier method (Fig. 3). The most common metastatic site was the liver, followed by the lymph nodes and peritoneum. Other parameters such as age, sex, tumour size, pT, pN stage and adjuvant chemotherapy status were not significantly different. More than two-thirds of the patients received adjuvant chemotherapy in all groups. In 65 cases with P10% NED, 44 patients (67.7%) received adjuvant chemotherapy, which were 5-fluorouracil (5-FU) alone (n = 14) or 5-FU plus platinum-based chemotherapy (n = 20) in 34 patients and etoposide plus cisplatin (EP) in 10 patients. The EP regimen was used only in NECs. Detailed regimens of adjuvant chemotherapy in each patient group and stage are given in the supplementary Table 1. 3.3. Prognostic significance of NED in gastric carcinomas We then analysed patient outcomes (RFS, DSS and OS) according to the proportion of NED. The median follow-up period was 43.5 months (range, 1.5– 159.1 months). During the follow-up period, 21 patients (44.7%) in NEC, 5 (50.0%) in MANEC, 6 (75.0%) in GCNED and 1 (4%) in gastric carcinoma with <10% NED died. The causes of death were disease-related in 23 cases (69.7%) and uncertain in 10 cases (30.3%). In the gastric carcinoma without NEM group, 84 patients (12.9%) expired; 59 (70.2%) of which were diseaserelated, 16 (19.0%) were uncertain and 9 (10.7%) were other causes. The trends in the Kaplan–Meier survival curves were identical for RFS, DSS as well as OS (Fig. 3 and supplementary figures). Carcinomas with <10% NED (88.9% 5-year OS) and gastric carcinomas without NEM (85.2% 5-year OS) showed significantly better RFS and OS than the other groups (Fig. 3A and C). The 5year OS rates for NEC, MANEC and GCNED were 59.1%, 53.5% and 16.7%, respectively. Therefore, we grouped the patients into two groups using 10% NED as a cut-off. As expected, RFS and OS were significantly better in patients with <10% NED (85.6% 5-year OS)
Fig. 2. An illustration of immunohistochemical staining patterns in gastric carcinomas with P10% neuroendocrine differentiation (NED) . Immunopositivity is indicated by differently coloured boxes: black (triple expression), dark grey (double expression) and light grey (single expression). Abbreviations: S, SYN; C, CGA; D, CD56.
2806
J.Y. Park et al. / European Journal of Cancer 50 (2014) 2802– 2809
Table 1 Clinicopathological features of gastric carcinomas with or without neuroendocrine differentiation (NED). Parameters
Neuroendocrine carcinomas (NEC) (n = 47)
Adenoneuroendocrine carcinoma (MANEC) (n = 10)
Gastric carcinoma with 10–30% NED (GCNED) (n = 8)
<10% NED (n = 23)
GC without NEM (n = 650)
Age (years) [median (range)] Gender M F Location C+F B Ant Diameter (cm) [median (range)] pT stage T1 T2 T3 T4 pN stage N0 N1 N2 N3 LVI – + PNI – + Stage group I II III Adjuvant CTx – + Follow-up (months) [median (range)] Recurrence or – metastasis + Liver LN Peritoneum Lung Others
64.0 (40–79) 34 (72.3%) 13 (27.7%) 9 (19.1%)a 20 (42.6%)a 18 (38.3%)a 5.0 (1.7–13.0) 7 (14.9%) 11 (23.4%) 21 (44.7%) 8 (17.0%) 19 (40.4%) 10 (21.3%) 12 (25.5%) 6 (12.8%) 5 (10.6%)a 42 (89.4%)a 30 (63.8%) 17 (36.2%) 12 (25.5%) 19 (40.5%) 16 (34.0%) 16 (34.0%) 31 (66.0%) 52.6 (4.2–108.7)
65.5 (55–70) 10 (100%) 0 (0.0%) 0 (0.0%) 2 (20.0%) 8 (80.0%) 5.6 (1.0–12.0) 2 (20.0%) 1 (10.0%) 5 (50.0%) 2 (20.0%) 2 (20.0%) 4 (40.0%) 2 (20.0%) 2 (20.0%) 1 (10.0%)a 9 (90.0%)a 5 (50.0%) 5 (50.0%) 3 (30.0%) 3 (30.0%) 4 (40.0%) 3 (30.0%) 7 (70.0%) 50.4 (16.3–126.8)
56.5 (35–77) 5 (62.5%) 3 (37.5%) 0 (0.0%) 5 (62.5%) 3 (37.5%) 6.0 (4.2–9.0) 0 (0.0%) 1 (12.5%) 5 (62.5%) 2 (25.0%) 1 (12.5%) 2 (25.0%) 3 (37.5%) 2 (25.0%) 1 (12.5%)a 7 (87.5%)a 2 (25.0%)a 6 (75.0%)a 1 (12.5%) 2(25.0%) 5 (62.5%) 2 (25.0%) 6 (75.0%) 20.1 (6.3–76.4)
62.0 (28–96) 395 (60.8%) 255 (39.2%) 47 (7.2%) 348 (53.5%) 255 (39.3%) 5.0 (1.0–33.5) 120 (18.5%) 156 (24.0%) 225 (34.6%) 149 (22.9%) 272 (41.8%) 115 (17.7%) 124 (19.1%) 139 (21.4%) 327 (50.3%) 323 (49.7%) 448 (68.9%) 202 (31.1%) 188 (28.9%) 220 (33.8%) 242 (37.2%) 219 (33.7%) 431 (66.3%) 41.5 (1.5–159.1)
31 (64.6%)a
6 (60.0%)a
4 (50.0%)a
60.0 (32–82) 15 (65.2%) 8 (34.8%) 0 (0.0%) 13 (56.5%) 10 (43.5%) 5.6 (2.5–17.0) 4 (17.4%) 8 (34.8%) 8 (34.8%) 3 (13.0%) 11 (47.8%) 5 (21.7%) 5 (21.7%) 2 (8.8%) 5 (21.7%)a 18 (78.3%)a 15 (65.2%) 8 (34.8%) 9 (39.1%) 9 (39.1%) 5 (21.8%) 8 (34.8%) 15 (65.2%) 62.6 (3.2– 119.2) 22 (96%)
17 (35.4%)a 8 5 5 0 1b
4 (40.0%)a 3 1 0 1b 2b
4 (50.0%)a 1 0 3 0 0
1 (4%) 1 0 0 0 0
65 (10.0%) 7 7 27 4b 21
585 (90.0%)
Abbreviations: GC, gastric carcinoma; M, male; F, female; C, cardia; F, fundus; B, body; Ant, antrum; LVI, lymphovascular invasion; PNI, perineural invasion; CTx, chemotherapy; LN, lymph node. a Parameters with statistically significant differences (P < 0.05) compared with GC without NED are indicated in bold letters. b Some patients recurred in multiple sites.
than in those harbouring P10% NED (53.3% 5-year OS; Fig. 3B and D). This prognostic significance of NED was also valid in all stage groups. 3.4. Prognostic factors in gastric carcinomas according to NED We subsequently evaluated other factors associated with RFS, DSS or OS in all gastric carcinomas. In univariate analyses, old age (P60 years), larger tumour size (P4 cm), advanced stage group, presence of NED, lymphovascular and perineural invasion were poor prognostic factors for both RFS and OS (Tables 2 and 3). Patient’s sex, tumour location, tumour necrosis, type and number of positive neuroendocrine markers were not associated with patient survival. Adjuvant chemotherapy or the type of adjuvant chemotherapy did not exert prognostic effects on patient outcome.
After multivariate analysis, stage group and the P10% NED remained as independent prognostic factors for OS, RFS and DDS (Tables 2 and 3 and supplementary Table 2). The adjusted hazard ratio for OS in patients with P10% NED was significantly higher compared to those of perineural invasion or the stage group. 4. Discussion Although many previous studies have already elucidated the poor prognosis of NEC in the stomach [14,15], few studies have examined the prognostic significance of neuroendocrine components in gastric carcinomas, even if the coexistence of NEC and adenocarcinoma is frequently observed [16,17]. In our study, we looked for the NEM in gastric carcinomas and semiquantified the proportion of immunopositivity
J.Y. Park et al. / European Journal of Cancer 50 (2014) 2802– 2809
2807
Fig. 3. Kaplan–Meier survival curves grouped by proportion of neuroendocrine differentiation (NED) in gastric carcinomas. Gastric carcinomas with less than 10% NED and gastric carcinomas without neuroendocrine morphology (NEM) show better survival than neuroendocrine carcinoma (NEC), adenoneuroendocrine carcinoma (MANEC) and gastric carcinoma with 10–30% NED (GCNED) in recurrence-free survival (RFS) and overall survival (OS) (A and C). The prognostic significance of NED is more clearly demonstrated when the cut-off of 10% NED is applied (B and D).
of neuroendocrine markers (as the NED percentage). We used three commonly used neuroendocrine markers (SYN, CGA and CD56) in combination because profiles of various neuroendocrine markers have shown inconsistent results [10,14,18]. The most notable finding of our study was that NEC, MANEC and GCNED showed poorer outcome than gastric carcinomas with <10% NED or without NEM (Fig. 3A and C). Although the survival rates of GCNED were worse than those of NEC and MANEC, this finding can be explained by the fact that 62.5% of patients with GCNED were stage III, whereas only 34% and 40% of patients in NEC and MANEC were at that stage (Table 1). These results agree with the previously reported observation that there was no significant difference in survival between NEC and MANEC [15]. The results of this study clearly demonstrate that gastric carcinomas with P10% NED should be differentiated from conventional adenocarcinomas since they are similar to NEC not only in terms of prognosis, but also in some clinicopathological parameters. The current WHO classification [1] provides 30% of neuroendocrine components as a cut-off and states that minor
components can be overlooked, whereas Jiang et al. [4] suggested a 20% cut-off while leaving some room for further studies to validate the use of 10% or less as cut-off. To the best of our knowledge, this is the first large-scale study to provide a neuroendocrine proportion cut-off that can be used to distinguish from conventional gastric adenocarcinomas. We have also found that P10% NED is an independent poor prognostic factor, although this result needs to be verified in other independent patient groups. Surgical respectability has been reported to be the most important prognostic factor in gastric NECs [15]. Therefore, radical gastrectomy with appropriate lymph node dissection would be the treatment of choice in localised diseases. It has recently been established that adjuvant S-1 [19] or capecitabine plus oxaliplatin [20] is effective in gastric carcinomas after D2 surgery. However, the adjuvant treatments of NECs or gastric carcinomas with NED are not standardised. The EP regimen, which is used for small cell carcinoma of the lung, is the most widely used treatment in NECs [21] but the biological similarity between small cell carcinoma of the lung and gastric NEC is controversial [22]
2808
J.Y. Park et al. / European Journal of Cancer 50 (2014) 2802– 2809
Table 2 Univariate and multivariate analyses for overall survival (OS). Clinicopathologic variables
Age (years) Diameter (cm) Stage group
NED LVI PNI
<60 versus P60 <4 versus P4 I II III <10% versus P10% – versus + – versus +
Univariate analysis
Multivariate analysis
HR
95% CI
P value
1.987 3.200 1 3.965 10.613 3.920 2.339 1.915
1.336–2.953 1.913–5.352 Reference 1.836–8.566 5.114–22.025 2.610–5.890 1.735–3.152 1.323–2.773
0.001 <0.0001 <0.0001 <0.0001 <0.0001 <0.0001 0.001
HR
95% CI
P value
1 5.200 7.456 13.459
Reference 1.181–22.905 1.696–32.775 1.824–99.302
0.029 0.008 0.011
2.294
1.111–4.739
0.025
Abbreviations: HR, hazard ratio; CI, confidence interval; NED, neuroendocrine differentiation; LVI, lymphovascular invasion; PNI, perineural invasion.
Table 3 Univariate and multivariate analyses for recurrence-free survival (RFS). Clinicopathologic variables
Age (years) Diameter (cm) Stage group
NED LVI PNI
<60 versus P60 <4 versus P4 I II III <10% versus P10% – versus + – versus +
Univariate analysis
Multivariate analysis
HR
95% CI
P value
HR
95% CI
P value
1.597 3.103 1 3.985 10.189 3.196 2.438 1.886
1.152–2.218 2.011–4.790 Reference 2.067–7.684 5.466–18.993 2.197–4.649 1.734–3.426 1.376–2.584
0.005 <0.0001
1.447
1.041–2.011
0.028
1 3.878 10.154 3.060
Reference 2.011–7.480 5.447–18.928 2.097–4.464
<0.0001 <0.0001 <0.0001
<0.0001 <0.0001 <0.0001 <0.0001 <0.0001
Abbreviations: HR, hazard ratio; CI, confidence interval; NED, neuroendocrine differentiation; LVI, lymphovascular invasion; PNI, perineural invasion.
and other regimens have also been reported to be effective [23,24]. Some authors recommend adjuvant chemotherapy even in the early stage [25]. Since most previous studies have evaluated heterogeneous tumours, commonly lacking standardised diagnostic criteria the proportion of the non-NED (epithelial) component could have affected the response [23]. Therefore, multi-institutional or multinational prospective studies that consider the proportion of NED should be designed for the development of proper guidelines for the treatment of these rare but grave diseases. The major limitation of our study was that we were unable to validate the 10% cut-off of NED in other cohorts because it was a single centre study. Another was lack of data on expression of neuroendocrine markers in the control group. It has been reported that neuroendocrine marker positivity is associated with higher stage and poor survival in undifferentiated colorectal carcinomas [26,27] and gastric carcinomas that express neuroendocrine markers without NEM might also have different clinical outcomes. Therefore, significance of neuroendocrine marker positivity in gastric carcinoma regardless of NEM should be evaluated in future studies. Nevertheless, our results suggest that there is no survival benefit caused by a small proportion of NED and that even a minor proportion (10–30%) of NED should not be overlooked.
In summary, we evaluated the clinicopathologic features of gastric carcinomas with NED and discovered that there are no significant survival differences among NEC, and GCNED. Therefore, gastric carcinomas harbouring P10% NED should be differentiated from conventional adenocarcinomas. The P10% NED cut-off may be practically useful and may serve as an informative parameter for predicting patient outcome and establishing optimal therapeutic guidelines for gastric carcinomas with NED. Conflict of interest statement None declared. Acknowledgements This study was supported by the Proteogenomic Research Program through a National Research Foundation of Korea grant funded by the Korea government (Ministry of Science, ICT & Future Planning). Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/ 10.1016/j.ejca.2014.08.004.
J.Y. Park et al. / European Journal of Cancer 50 (2014) 2802– 2809
References [1] Solcia E, Arnold R, Capella C, et al. Neuroendocrine neoplasms of the stomach. In: Bosman FT, Jaffe ES, Lakhani SR, Ohgaki H, editors. WHO classification of tumors of the digestive system. Lyon: International Agency for Research on Cancer; 2010. p. 64–8. [2] Rindi G, Azzoni C, La Rosa S, et al. ECL cell tumor and poorly differentiated endocrine carcinoma of the stomach: prognostic evaluation by pathological analysis. Gastroenterology 1999;116:532–42. [3] Matsui K, Jin XM, Kitagawa M, Miwa A. Clinicopathologic features of neuroendocrine carcinomas of the stomach: appraisal of small cell and large cell variants. Arch Pathol Lab Med 1998;122:1010–7. [4] Jiang SX, Mikami T, Umezawa A, Saegusa M, Kameya T, Okayasu I. Gastric large cell neuroendocrine carcinomas: a distinct clinicopathologic entity. Am J Surg Pathol 2006;30:945–53. [5] Kubota T, Ohyama S, Hiki N, Nunobe S, Yamamoto N, Yamaguchi T. Endocrine carcinoma of the stomach: clinicopathological analysis of 27 surgically treated cases in a single institute. Gastric Cancer 2012;15:323–30. [6] Kim JJ, Kim JY, Hur H, Cho YK, Han SU. Clinicopathologic significance of gastric adenocarcinoma with neuroendocrine features. J Gastric Cancer 2011;11:195–9. [7] Miguchi M, Iseki M, Shimatani K. Advanced gastric neuroendocrine carcinoma with an adenocarcinoma component. Case Rep Gastroenterol 2012;6:52–7. [8] Lee JH, Kim HW, Kang DH, Choi CW, Park SB, Kim SH. A gastric composite tumor with an adenocarcinoma and a neuroendocrine carcinoma: a case report. Clin Endosc 2013;46:280–3. [9] Terada T, Maruo H. Simultaneous large cell neuroendocrine carcinoma and adenocarcinoma of the stomach. World J Gastroenterol 2011;17:4831–4. [10] Erickson LA, Lloyd RV. Practical markers used in the diagnosis of endocrine tumors. Adv Anat Pathol 2004;11:175–89. [11] Blumenfeld W, Chandhoke DK, Sagerman P, Turi GK. Neuroendocrine differentiation in gastric adenocarcinomas. An immunohistochemical study. Arch Pathol Lab Med 1996;120:478–81. [12] Qvigstad G, Qvigstad T, Westre B, Sandvik AK, Brenna E, Waldum HL. Neuroendocrine differentiation in gastric adenocarcinomas associated with severe hypergastrinemia and/or pernicious anemia. APMIS 2002;110:132–9. [13] Eren F, Celikel C, Gulluoglu B. Neuroendocrine differentiation in gastric adenocarcinomas; correlation with tumor stage and expression of VEGF and p53. POR 2004;10:47–51. [14] La Rosa S, Inzani F, Vanoli A, et al. Histologic characterization and improved prognostic evaluation of 209 gastric neuroendocrine neoplasms. Hum Pathol 2011;42:1373–84.
2809
[15] Ishida M, Sekine S, Fukagawa T, et al. Neuroendocrine carcinoma of the stomach: morphologic and immunohistochemical characteristics and prognosis. Am J Surg Pathol 2013;37:949–59. [16] Waldum HL, Aase S, Kvetnoi I, et al. Neuroendocrine differentiation in human gastric carcinoma. Cancer-Am Cancer Soc 1998;83:435–44. [17] Nishikura K, Watanabe H, Iwafuchi M, Fujiwara T, Kojima K, Ajioka Y. Carcinogenesis of gastric endocrine cell carcinoma: analysis of histopathology and p53 gene alteration. Gastric Cancer 2003;6:203–9. [18] Rayhan N, Sano T, Qian ZR, Obari AK, Hirokawa M. Histological and immunohistochemical study of composite neuroendocrine-exocrine carcinomas of the stomach. J Med Invest 2005;52:191–202. [19] Sakuramoto S, Sasako M, Yamaguchi T, et al. Adjuvant chemotherapy for gastric cancer with S-1, an oral fluoropyrimidine. N Engl J Med 2007;357:1810–20. [20] Bang YJ, Kim YW, Yang HK, et al. Adjuvant capecitabine and oxaliplatin for gastric cancer after D2 gastrectomy (CLASSIC): a phase 3 open-label, randomised controlled trial. Lancet 2012;379:315–21. [21] Okita NT, Kato K, Takahari D, et al. Neuroendocrine tumors of the stomach: chemotherapy with cisplatin plus irinotecan is effective for gastric poorly-differentiated neuroendocrine carcinoma. Gastric Cancer 2011;14:161–5. [22] Hainsworth JD, Spigel DR, Litchy S, Greco FA. Phase II trial of paclitaxel, carboplatin, and etoposide in advanced poorly differentiated neuroendocrine carcinoma: a Minnie Pearl Cancer Research Network Study. J Clin Oncol 2006;24:3548–54. [23] Fazio N, Spada F, Giovannini M. Chemotherapy in gastroenteropancreatic (GEP) neuroendocrine carcinomas (NEC): a critical view. Cancer Treat Rev 2013;39:270–4. [24] Meyer T, Qian W, Caplin ME, et al. Capecitabine and streptozocin ± cisplatin in advanced gastroenteropancreatic neuroendocrine tumours. Eur J Cancer 2014;50:902–11. [25] Fukuda T, Ohnishi Y, Nishimaki T, Ohtani H, Tachikawa S. Early gastric cancer of the small cell type. Am J Gastroenterol 1988;83:1176–9. [26] Grabowski P, Schindler I, Anagnostopoulos I, et al. Neuroendocrine differentiation is a relevant prognostic factor in stage III–IV colorectal cancer. Eur J Gastroenterol Hepatol 2001;13:405–11. [27] Grabowski P, Schonfelder J, Ahnert-Hilger G, et al. Expression of neuroendocrine markers: a signature of human undifferentiated carcinoma of the colon and rectum. Virchows Arch 2002;441:256–63.