Introduction

Pancreatic neuroendocrine neoplasms (PNENs) are rare, accounting for less than 5 % of all pancreatic tumors [1, 2]. They are generally divided into two main, clinically, histologically, and biologically very different entities with a common neuroendocrine histiogenesis—the pancreatic neuroendocrine tumors (PNETs) and the high-grade pancreatic neuroendocrine carcinomas (hgPNECs).

hgPNECs are defined as PNENs with poorly differentiated morphology and a higher proliferation rate than in well-differentiated PNETs [3]. They represent about 5 % of all PNENs [4] and are characterized by a highly aggressive behavior and dismal prognosis [5••]. In Norway, the incidence of hgPNEC has remained stable throughout the past two decades with an incidence rate of approximately 0.04 per 100,000 person-years [4]. Most hgPNEC are diagnosed in patients around 60 years of age and with a male predilection (male/female ratio around 1.5) and predominance of tumors located in the pancreatic head (65–70 %) [5••, 6••].

Throughout the last decade, a notable progress has occurred in basic, translational, and clinical research on PNETs [7, 8]. At the same time, there has been a general trend towards both more aggressive and minimally invasive surgery of PNETs [9–13]. In contrast, hgPNECs have not gained much attention although basic and clinical research on hgPNEC has recently started to develop. In addition to clinical trials that focus on cytoreductive treatment, the pathology and surgical treatment of this diverse subgroup of neuroendocrine neoplasms have received increasing interest.

Robust knowledge of the histologic characteristics and defining criteria of hgPNEC is a prerequisite for the understanding of oncologic outcomes after surgery. The review will therefore provide an overview of the histologic diagnostic and staging criteria of PNENs, including hgPNEC, followed by a discussion of the surgical treatment of hgPNEC.

Surgical Pathology

As hgPNECs are morphologically and biologically heterogeneous [14••], thorough and standardized histopathologic reporting is of great importance for treatment planning and prognostic evaluation of patients. Due to the low incidence of hgPNEC and ensuing risk of misdiagnosis, cases should be reviewed by pathologists with expertise in the evaluation of gastroenteropancreatic (GEP) NENs [5••]. Histopathologic characterization of hgPNEC follows the general classification systems that apply for PNENs. The latter are diagnosed based on morphologic appearance and immunohistochemistry, graded and classified according to the tumor-node-metastasis (TNM) staging system. While the characterization of neuroendocrine cell morphology and evaluation of immunohistochemistry in hgPNEC remain the domain of pathologists [15] that will not be further discussed in this review, pancreatic surgeons should have a thorough understanding of the grading and TNM staging of PNENs, including hgPNECs.

Grading

According to the 2010 World Health Organization (WHO), grading of GEP NENs, including hgPNEC, is based on the mitotic rate (number of mitotic figures per 10 high-power fields) and/or the Ki-67 index (percentage of tumor cells with nuclear staining for Ki-67) [3]. A mitotic rate of <2 and/or Ki-67 index of ≤2 corresponds to a NET G1, and a mitotic rate of 2–20 and/or Ki-67 index of 2.5–20 characterizes a NET G2, while a mitotic rate and/or Ki-67 index of >20 defines a NEC G3 (Table 1). NECs are further subdivided into a small and large cell subtype based on the morphology of the tumor. The genetics of hgPNEC differ from those of PNETs (G1 and G2) [16, 17••], indicating that hgPNEC does not usually develop by genetic progression of G1–G2 PNET [5••, 17••]. Small and large cell morphology is not a prognostic factor for patients with hgPNEC [6••]. It is important to note that grading of a neuroendocrine neoplasm is determined by the highest mitotic rate or Ki-67 index, irrespective of whether this is found in the primary tumor or a metastatic deposit.

Table 1 WHO 2010 grading system for pancreatic neuroendocrine neoplasms
Full size table

A discrepancy in grading defined by mitotic rate and Ki-67 has been observed in up to 44 % of PNENs [18]. PNENs with a mitotic rate within the G2 range and a Ki-67 index corresponding to G3 have been described [14••, 19, 20••]. Such “grade-discordant” PNENs were found to have better prognosis compared with true hgPNECs (median survival 54 versus 11 months), but a worse outcome compared with “grade-concordant” PNENs (median survival 54 versus 68 months) [14••]. A further recent observation that exemplifies the heterogeneity of hgPNEC is the difference in response rate to first-line platinum-based systemic chemotherapy among patients with a GEP NEC depending on whether they had a Ki-67 index above or below 55 %. Interestingly, the response rate correlated with the Ki-67 value (response rate 42 % with Ki-67 above 55 % versus 14 % with Ki-67 below 55 %) [21••]. Both observations, i.e., the existence of a “grade-discordant” group of hgPNEC with unique clinical features and the association between the Ki-67 index and effect of platinum-based chemotherapy of hgPNEC, imply the need for modification of the current WHO 2010 grading system for PNENs [22].

TNM Staging

PNENs, including hgPNECs, are classified according to their tumor-node-metastasis (TNM) pattern, as defined by validated TNM staging systems [23]. There is no separate TNM classification for hgPNECs. There are currently two TNM staging systems for staging of PNENs. The first classification was recommended by the European Neuroendocrine Tumor Society (ENETS) in 2006 [24] and is predominant in Europe. This was followed by the classification suggested by the American Joint Cancer Committee and International Union for Cancer Control (AJCC/UICC) in 2009 [25], which is now widely used in the North American region. The ENETS and AJCC/UICC classification systems for PNENs differ in their definition of the T stage as shown in Table 2.

Table 2 Comparison of the criteria for the T category in the ENETS and AJCC/UICC TNM classifications of pancreatic neuroendocrine neoplasms
Full size table

There is an ongoing debate as to which of the two staging systems is the most precise in terms of prognostic stratifications, with some studies demonstrating similar strength [26] and others indicating superiority of the ENETS over the AJCC/UICC TNM staging system [27, 28].

As two different TNM staging systems are currently being used, publications should report the tumor features on which the T stage classifications are based, such that translation between ENETS and AJCC/UICC classifications can be made [29]. This is all the more important, in view of the likely future adaptions of current staging systems when more clinical follow-up data have been gathered [28].

Future Directions

Upcoming trials on surgical treatment of PNENs, including hgPNEC, should report histopathological features as outlined above. A modification of the WHO 2010 grading system of PNENs with inclusion of the “grade-discordant” group of hgPNEC as a new subgroup is likely. Moreover, there is a need for one internationally accepted TNM staging system in order to avoid confusion and misinterpretation among researchers and clinicians. There is also a need to clarify whether staging should be the same for PNETs (G1–G2) and hgPNECs (G3) as clinical behavior of these PNEN subgroups is very diverse. Future studies should report Ki-67 values of both primary and metastatic diseases, if available, as discordance of histologic grade between primary and metastatic PNENs is sometimes observed [30]. This is probably more relevant for metachronous than synchronous metastases, as there is often an increase in Ki-67 over time [30]. Multicenter studies should include central slide review by pathologists with expertise in the field. The use of digital image analysis should be further explored [31]. With the increasing knowledge about the molecular and genetic aberrations in hgPNEC, current classification systems are likely to change.

Surgical Treatment

Surgical resection is the only curative treatment option for pancreatic NETs (G1 and G2) and remains the therapeutic cornerstone, even for patients with advanced disease [12, 13]. In contrast, the role of surgery in the treatment of hgPNEC remains unclear [6••]. This may be explained by the common presence of synchronous metastatic disease and rapid progression of hgPNECs, which traditionally has supported a treatment choice of palliative systemic chemotherapy [21••]. However, less than half of patients with hgPNEC respond to such treatment regimens and alternative treatment options are urgently needed. A novel interest in surgical treatment of hgPNEC has recently been noted [6••, 32••, 33, 34], as shown in Table 3, although the underlying clinical evidence for a surgical approach is still scarce and prospective trials on surgery for hgPNEC are lacking. Nevertheless, efforts should be made to increase the attention of surgeons to the treatment of hgPNEC as a possible measure to improve patient survival. In the following sections, different aspects of surgical treatment for hgPNEC will be discussed.

Table 3 Summary of the current surgical series for patients with high-grade pancreatic neuroendocrine carcinoma
Full size table

Localized Disease

Surgery of localized nonmetastatic disease, defined as T stage T1 or T2, combined with chemotherapy seems to improve overall survival despite the presence of a recurrent disease, compared with chemotherapy alone (median survival 23 versus 13 months) [6••]. Surgical resection of localized nonmetastatic disease has also been associated with a relatively high 5-year survival of 43 % [33]. For patients with localized hgPNEC ≤2 cm who underwent resection, median survival was 29 months, while patients who were left to best supporting care had a median survival of 5 months [34]. The case number of patients with localized hgPNEC undergoing surgery in these studies was generally low ranging from 20 to 26.

According to the current ENETS consensus guidelines for high-grade GEP NECs, combination of postoperative platinum-based chemotherapy with local treatment consisting of surgery, radiotherapy, or both probably offers the greatest likelihood of long-term survival for patients with localized disease irrespective of the exact site of the primary [35••]. As prognosis of GEP NECs correlates with the Ki-67 index, tumors with a very high Ki-67 are more aggressive than tumors with a Ki-67 just above 20 % [21••, 36]. At the same time, GEP NECs with a high Ki-67 (≥55 %) do respond better on systemic platinum-based chemotherapy than GEP NECs with a low Ki-67 (<55 %) [21••]. Based on these observations, we would recommend upfront radical surgery for hgPNECs with a Ki-67 <55 % and neoadjuvant systemic platinum-based chemotherapy for hgPNECs with a Ki-67 ≥55 %.

Peripancreatic Lymphadenectomy

The risk of peripancreatic lymph node metastasis in patients with PNENs correlates with increasing tumor size and tumor grading [37]. Presence of regional lymph node metastasis is independently associated with decreased disease-specific survival in patients with PNENs, including patients with hgPNEC [37, 38]. There is an ongoing debate as to whether or not peripancreatic lymphadenectomy should be performed routinely in patients with PNENs [38–43]. It is unknown if peripancreatic lymphadenectomy improves survival in patients with hgPNEC since no trial has evaluated this. However, as lymph node stage predicts prognosis, peripancreatic lymphadenectomy should probably be performed routinely in patients undergoing surgery for hgPNEC.

Locally Advanced Disease

The impact of surgery on the prognosis of patients with locally advanced hgPNEC, defined as T-stage T3 or T4, has not been evaluated yet. However, there are indications that surgery may improve survival in locally advanced disease when compared with chemotherapy alone, as shown by a recent report on 19 patients with stage T3 or T4 disease [6••]. The current North American Neuroendocrine Tumor Society (NANETS) consensus guidelines provide only an expert opinion on this matter, which supports radical surgery, if the risk of morbidity is low and the risk of intestinal obstruction is high [44]. Our experiences are in accordance with this recommendation. Furthermore, we would suggest surgery for radically resectable locally advanced nonmetastatic hgPNEC for selected patients, despite a higher risk for a margin-positive resection. This is supported by the observed improved survival after R0/R1 resections of hgPNEC, compared with R2 resections [32••]. As mentioned earlier, we would recommend neoadjuvant systemic platinum-based chemotherapy for hgPNECs with a Ki-67 ≥55 %.

Preoperative and Postoperative Chemotherapy

There are no systematic studies on the use of preoperative chemotherapy in patients with localized hgPNEC. Most localized hgPNECs will recur or metastasize within 1 year after resection [6••]. This might suggest the presence of occult metastases at diagnosis. Thus, postoperative platinum-based chemotherapy should always be considered for patients with hgPNEC, regardless of the stage of the disease and provided the treatment is tolerated [6••, 35••]. It seems that >4 courses of postoperative chemotherapy is a significant factor of improved survival compared with 1–4 courses [6••]. Based on the limited clinical evidence available, we recommend upfront radical surgery of localized disease with postoperative platinum-based systemic chemotherapy of more than four courses [6••]. This is also supported by an expert opinion reported in the current NANETS guidelines [44]. Given the early manifestation of a recurrent disease after surgery for localized nonmetastatic hgPNEC, it remains to be established whether chemotherapy should also be given in a neoadjuvant setting.

Metastatic Disease

Most patients with hgPNEC develop distant metastases, which is often already present at the time of diagnosis [45]. This reduces the prospects for long-term survival. The current ENETS and NANETS guidelines do not recommend surgery for distant metastasis or debulking surgery for hgPNEC [35••, 44]. This is also supported by the guidelines of the Nordic Neuroendocrine Tumor Group (NNTG) from 2014 [46] and by a recent international consensus conference on the treatment of neuroendocrine liver metastasis [47].

However, there is evidence that surgery of metastatic hgPNEC may improve survival. In one study, overall survival was 24 months with a 5-year survival of 21 % among 13 patients who underwent surgery for liver metastasis [48]. In another study, overall survival after surgery of metastases was 29 months with a 3-year survival of 69 % among 12 patients [6••].

Resection of the primary tumor is an independent prognostic factor for improved survival in patients with hgPNEC at different disease stages [6••]. This has also recently been demonstrated for PNENs across all stages of disease [13, 49] and suggests that resection of the primary tumor should always be considered, even in patients with metastatic disease. Aggressive locoregional treatment of liver metastases (debulking surgery, radiofrequency ablation, and liver-directed intra-arterial intervention) of hgPNEC may improve survival [50••]. Another important and recent finding is that surgical treatment combined with systemic chemotherapy may improve the survival of patients with metastatic hgPNEC compared with chemotherapy alone [6••].

As the current guidelines are exclusively based on expert opinions and considering the latest clinical data on this matter that diverge from the guidelines' recommendations, we could expect that the expert opinions may be modified in the near future as more clinical data on surgery of hgPNEC become available.

Future Directions

The therapeutic approach for localized hgPNEC is at present neither consistent nor uniform [46]. Future studies on surgical treatment of hgPNEC should focus on the establishment of standardized sequences of treatment for hgPNEC, especially the combined use of platinum-based chemotherapy pre- and postoperatively [51]. High-grade PNECs should be studied as a separate entity with precise reporting of their characteristics in future trials [52]. Moreover, initiatives should be undertaken to carry out to plan and conduct prospective multicenter studies. We propose the need for a prospective trial on neoadjuvant platinum-based chemotherapy for nonmetastatic hgPNEC with a Ki-67 index above 55 %. We would also suggest the need for a prospective study on radical surgical treatment combined with platinum-based chemotherapy of resectable metastatic disease versus platinum-based chemotherapy alone.

Prognosis and Surveillance

Median survival for patients with hgPNECs receiving best supportive care is 2 months, 13 months for those receiving chemotherapy, and 23 months if treated with combined chemotherapy and surgery [6••].

There is no evidence as to how the follow-up of patients with hgPNECs should be planned and conducted. Hence, recommendations concerning surveillance are based on expert opinion. According to the current recommendations of the NNTG, radically treated patient should undergo surveillance with a CT scan of the abdomen and chest, alternatively MRI, every 3 months [46]. The most recent ENETS guidelines support this scheme and recommend that if the patient is in remission 2–3 years postoperatively, surveillance intervals can be prolonged to 6–12 months during the following 5 years [35••]. Surveillance of patients with metastatic disease should be decided individually with a general recommendation of CT/MRI every 2–3 months while undergoing treatment. Recurrent disease is typically distant and not local [6••].

Conclusion

High-grade PNECs represent a rare and diverse group of PNENs, both histopathologically and clinically, which warrants special attention from pancreatic surgeons. Recent novel findings support the notion of surgery as a principle of treatment for patients with hgPNEC. As treatment options are limited, initiatives should be made to further evaluate the role of surgery in this aggressive group of malignancies.