Pancreatic neuroendocrine tumors (PNET) are increasingly diagnosed with advancements in imaging technology.1 The natural history of PNET is different from the more common type of pancreatic cancer, pancreatic ductal adenocarcinoma, where survival outcomes are poor. PNET biologic behavior can vary greatly, with some patients experiencing slow progression of the disease while others may have rapid progression leading to death.2 Owing to the low incidence of PNET and longer survival rates in most patients, it is challenging to conduct prospective studies to understand their prognosis and survival. Therefore, the current data guiding PNET treatment is largely retrospective and the high variability in practice patterns between centers limits meaningful comparison of the factors driving PNET survival outcomes.

Functional PNET (F-PNET) results when these tumors secrete active hormones, causing a clinical syndrome. In malignant functional PNET, lymph node metastasis (LNM) commonly occurs, and the removal of both the primary tumor and affected lymph nodes is often necessary to address the hormonal imbalance.3,4 However, in nonfunctional PNET (NF-PNET), where no clinically active hormone is produced, the impact of regional LND is more uncertain. Generally, small (< 2 cm) low-grade nonfunctional PNET are thought to have a benign course and a low risk of regional metastasis. This has led some experts to recommend observation or limited tumor removal without sampling the lymph nodes for small nonfunctional PNET in specialized medical centers.5,6 However, other studies report more aggressive biology and LNM in some small NF-PNET, suggesting that lymph node dissection may be beneficial.7,8 Even in larger (≥ 2 cm) low-grade NF-PNET, it is unclear how lymph node metastasis affects survival outcomes. Some studies indicate that patients without lymph node involvement have better survival rates,9,10 while others suggest that distant metastasis and higher tumor grade are more significant prognostic factors, with LMN having limited relevance.2,11 The National Comprehensive Cancer Network (NCCN) guidelines recommend removing both the tumor and the affected lymph nodes for all PNET larger than 2 cm, and considering lymph node removal for tumors between 1 and 2 cm [12].

Therefore, the role of routine LND in low-grade non-functional PNET is still uncertain. For PNET patients where the expected survival often exceeds a decade even in cases of advanced disease at diagnosis, the balance between maximizing treatment benefits and minimizing surgical risks needs to be carefully considered. This systematic review aims to determine whether LNM has prognostic significance in low-grade NF-PNET and, if so, whether LND improves survival.

Patients and Methods

Design and Literature Search Strategy

A systematic literature search of studies published between 1990 and 2021 was performed using PubMed, EMBASE, and the Cochran Library. The following Medical Subject Heading (MeSH) terms were used: pancreatic neuroendocrine tumor, islet cell carcinoma, pancreas neuroendocrine neoplasm, lymph node, metastasis, and surgery. This systemic review was considered exempt on the basis of the Medical College of Wisconsin institutional review board policies and followed the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines for systemic reviews.13

Selection of Studies

Non-English text, nonhuman, and pediatric studies were excluded from analysis. Studies of patients with widely metastatic PNET, studies with greater > 50% representation of F-PNET, and studies of other neuroendocrine primary tumors (e.g., gastric, small bowel, or duodenal) were also excluded. Literature reviews without unique content, case reports, and studies where lymph node metastasis and/or survival outcomes were not reported were excluded. The remaining studies that evaluated the impact of positive lymph nodes or LND on recurrence-free survival (RFS) or overall survival (OS) for patients with PNET were selected for analysis.

Data Extraction

Study variables, including patient demographics, representation of F-PNET and NF-PNET, rate of LNM, disease-free-survival, and overall survival were extracted. Extracted data were independently reviewed and verified by two authors (C.C. and E.W.) for accuracy.

Results

Overview of Literature Search

A total of 426 unique articles were identified using the following search terms: pancreatic neuroendocrine tumor, islet cell carcinoma, pancreas neuroendocrine neoplasm, lymph node, metastasis, and/or surgery (Fig. 1). In total, 62 of the original 426 manuscripts were removed, including: not written in English (n = 36), not involving human adults (n = 15), or published before 1990 (n = 11). The remaining 364 articles were then screened on the basis of the title and abstract content. An additional 270 articles were excluded owing to inapplicability of neuroendocrine tumor subtype (duodenal, small bowel, or F-PNET only), metastasis in NF-PNET, and classification as a literature review with no unique content or case report. The 94 remaining manuscripts were then included in a full review of the entire manuscript and reapplication of inclusion criteria. Thereafter, 42 articles were further excluded owing to overrepresentation of F-PNET in the study cohort, lack of lymph node status, or lack of LND. A final total of 52 manuscripts were identified as meeting the full inclusion criteria for this study. According to the Grades of Recommendation, Assessment, Development, and Evaluation (GRADE) system, each study was scored for quality of evidence.14 Findings are reported in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA).13

Fig. 1
figure 1

Preferred reporting items for systemic reviews and meta-analyses (PRISMA) chart

Full size image

Reported Rates of Positive Lymph Nodes

Overall, 50 studies reported the rate of positive lymph nodes for NF-PNET (Fig. 2). The median rate of LNM was 24.5% (range 7.1–63.5%). Several studies stratified the rate of positive LNM on the basis of size. For NF-PNET > 2 cm, the median rate of positive LNM was 37.5% (range 29–47%, n = seven studies). For tumors < 2 cm, the median rate of positive LNM was 11.3% (range 5.9–51%, n = 14 studies). Overall, several studies reported increased rates of positive lymph nodes with higher grade and larger tumors, and some reported more LNM for tumors in the head of the pancreas.

Fig. 2
figure 2

Reported rates of positive lymph node for non-functional pancreatic neuroendocrine tumors

Full size image

Impact of Lymph Node Status on Recurrence-Free Survival

In total, 29 studies retrospectively analyzed the impact of LNM or LND on recurrence-free survival (RFS) for NF-PNET (Table 1); 66% (n = 19) of studies found an association between LNM and worse RFS after curative resection.

Table 1 Studies reporting the impact of lymph node metastasis (LNM) on recurrence-free survival in NF-PNET
Full size table

In summary, most studies reported that LNM was associated with shorter RFS and resection with regional LND may mitigate the need for additional surgery compared with margin-negative resection without LND. The benefit is most significant in tumors > 2cm and with G2 and G3 tumors.

Impact of Lymph Node Status on Overall Survival

Of the 35 studies that retrospectively analyzed the impact of LNM or LND on OS (Table 2), 60% (n = 21) of studies observed an association between LNM and worse OS after curative resection of PNET.

Table 2 Studies reporting the impact of lymph node metastasis (LNM) on overall survival in PNET
Full size table

Among the 35 studies that retrospectively analyzed the impact of LNM or LND on OS, 40% (n = 14) reported no association between LNM and OS after curative pancreatectomy for PNET. With the exception of the previously discussed SEER data analysis by Li et al.,15 it should be noted that most of these studies consist of smaller cohort studies.8,9,16,17,18,19,20,21,22,23,24,25,26

Discussion

Although LND is widely supported for malignant F-PNET, significant controversy remains regarding the benefit of LND for patients with NF-PNET, especially small NF-PNET. Current literature concerning this topic are all retrospective reviews of single institutions, multi-institution collaborations, or large national cancer databases. In our systematic review, only two studies (n = 2) were found that directly evaluated if LND provides a survival benefit for NF-PNET. One study found that LND improved RFS, while the other study failed to find any difference in RFS with the addition of LND. Wu et al. found the therapeutic index for LND to be greatest in patients with tumors ≥ 2 cm, moderately or poorly differentiated, Ki-67 ≥ 3%, and located in the head of the pancreas.27 Likely because of the very limited data that exist on LND for NF-PNET, most studies (n = 54) assess the prognostic impact of LNM and its effect on survival to then suggest if LND could improve survival outcomes. Thus, they evaluated the incidence and impact of LNM on RFS and OS.

Most of the currently available literature supports the idea that LNM rates can be significant, even for small NF-PNET. Most studies found LNM to be associated with worse survival outcomes. The reported rates of LNM varied significantly between cohorts; generally, rates of LNM were lower for tumors < 2 cm.28,29,30 While the median rate of LNM for the entire cohort was 25%, reported rates of LNM were as high as 65%. While only seven studies stratified LNM rates by tumor size, the median rate of LNM in these studies for tumors > 2 cm was three times the median rate of LNM for tumors of all sizes. Reported rates of LNM for patients with tumors < 2 cm ranged from 5.1% to 51% [28,31]. Several studies also found that higher Ki-67, larger tumor size, and higher grade are associated with a greater risk of LNM.18,32,33 Overall, 66% of studies (n = 19), found LNM to be associated with worse RFS, and 60% of studies (n = 21) found an association with worse OS. In addition to LNM, tumors that were >2 cm, symptomatic from disease burden, and had Ki-67 > 3% were found to correlate with worse RFS.34 In conclusion, current available literature supports that LNM is prognostic of worse survival; pancreatic resection with LND may improve outcomes compared with margin-negative resection without LND, especially for NF-PNET > 2 cm.

NCCN guidelines recommend completion of LND for all NF-PNET > 2 cm and consideration of LND for tumors between 1 and 2 cm;35 North American Neuroendocrine Tumor Society (NANETS) guidelines agree that LND should be considered for all NF-PNET over 2 cm.36,37

Currently, standard resection for patients with tumors > 2 cm requires a pancreatectomy in which a LND is integrated as part of the underlying operation. Given that current consensus statements recommend a formal pancreatectomy for all patients with > 2 cm lesions, the question regarding the need to complete LND for all NF-PNET has the greatest potential impact on patients with small tumors where enucleation without LND remains an option. Owing to the risk of morbidity associated with pancreatic resections, enucleation has become a viable option that is offered in attempt to reduce the risks of exocrine and endocrine insufficiency.8,38 Enucleation is often cited to have a shorter operative time, less blood loss, lower morbidity, and shorter hospital stays compared with a standard pancreatectomy.39 However, some studies have cited similar complication rates and increased rates of clinically significant pancreatic fistulas.5,39,40 Given the relatively high rates of LNM reported for tumors < 2 cm and the potential marginal improvement in operative morbidity with enucleation, the risks of leaving LNM behind must be strongly considered, even for small NF-PNET.

Failure to complete an LND while operating on a NF-PNET risks leaving residual disease behind, which may prove to be difficult to monitor for progression and challenging to address later. Unlike other disease sites, such as melanoma, the location of these lymph nodes makes them impossible to monitor based on a physical exam alone and challenging to monitor on imaging. Thus, the option of close surveillance for missed LNM and progressive disease is limited. In addition, any significant progression of residual LNM not removed at initial resection has the potential to challenge removal of symptomatic local compression in a previously operated field. Although formal pancreatectomy has its own risks, the risks of leaving LNM behind are not insignificant. These risks must be weighed against the benefits of more limited upfront surgery.

LND has multiple potential benefits for patients with NF-PNET, including potential complete resection of all cancer and accurate staging/prognosis based on the presence of LNM. Given the significant impact LNM has on both DFS and OS in most of the reports, accurate staging may help to inform decisions on the frequency of surveillance and may help to guide decisions regarding additional future therapies. Although LND may provide important prognostic information for some patients, it is also important to note that major pancreatectomy with LND may overtreat small, low-grade PNET without benefitting the patient.41 Unfortunately, at this time, we do not have a great way to delineate who will have positive nodes except LND. Unlike in melanoma and breast cancer, the variability of lymph drainage has limited the efficacy of a sentinel lymph node mapping approach for pancreas tumors.42 Thus, given the relatively high rates of LNM at this time and the prognosis, even in small NF-PNET, LND should be considered for all patients when comorbidities allow for formal pancreatectomy. There are several limitations to this study; primarily, as the entirety of the literature included in this systematic review is retrospective, each study likely has associated biases. Selection bias was likely common among the studies as patients were not randomized to LND. In addition, although each paper focused on NF-PNET, heterogeneity within the cohorts existed. Cohorts with an over representation of F-PNET were excluded. These exclusions were in the case of studies only including specific subtypes of NF-PNET, including tumors < 2 cm or cystic tumors < 2 cm. In addition, as some of the large studies were based on national cancer databases, they included overlapping cohorts. Lastly, very few (only two) of the studies specifically looked at the benefit of LND. Thus, our conclusions that LND may provide benefits to patients with LNM are based on assumptions that surgical removal of LNM may mitigate the worse RFS and OS associated with LNM. Unfortunately, until improved techniques are developed to predict which patients will have LNM, the only way to identify these patients will be through LND.

The definition of adequate lymphadenectomy in patients with NF-PNET is still being debated. In a National Cancer Database (NCDB) study of 999 patients who underwent surgical resection for PNET, 72.8% of patients had regional lymphadenectomy performed with a median of eight LNs examined.43 Zhang et al. also sought to determine the appropriate number of lymph nodes necessary to accurately stage PNET using Surveillance, Epidemiology, and End Results (SEER) data.44 They found that LMN was associated with worse RFS and that discriminatory power was highest when more than eight lymph nodes were examined. As such, we believe the best threshold for the number of LNs to be examined for PNET appears to be ≥ eight.

In summary, we believe that oncological resection with regional LND should be offered to all patients with NF-PNET that are > 2 cm. Significant consideration of the risks and benefits of LND for small NF-PNET should be made as rates of LNM for small NF-PNET are not insignificant and are associated with worse survival outcomes. Additional studies are warranted to validate that removal of LNM through LND mitigates the survival outcomes associated with LNM for NF-PNET.

Conclusions

On the basis of the available evidence, we strongly recommended that patients diagnosed with low-grade NF-PNET tumors larger than 2 cm should undergo surgical resection with routine regional lymphadenectomy. For low-grade NF-PNET smaller than 2 cm, careful consideration should be given to preoperative imaging, patient symptoms, and Ki-67 levels when deciding between suitability for observation versus surgical resection. However, it is important to acknowledge that the quality of evidence supporting these recommendations is currently low. Further research and prospective studies are needed to provide higher-quality evidence and a more comprehensive understanding of the role of lymphadenectomy in the management of NF-PNET.