Background
Although pNETs are uncommon and their prognosis is better than that of pancreatic cancer, the incidence of pNETs is increasing [
1,
2]. PNETs are heterogeneous neoplasms and can be divided into functional and non-functional according to hormone secretion [
3]. Nonfunctional pNETs account for 80% of cases and radical surgery is the only way to cure resectable pNETs [
3]. Unlike the obvious symptoms of functional pNETs, nonfunctional pNETs are either incidentally discovered by abdominal computed tomography or when symptoms associated with tumor compression or invasion become evident [
4]. Although pNETs show indolent tumor biology, the 5-year survival rate of pNETs ranges from 15 to 100%. A subset of patients with aggressive tumors still has poor outcomes [
5,
6]. Thus, identifying high risk factors for OS and making more appropriate management for these patients is urgent. Muscogiuri et al. aimed to exam how gender shapes risk factors with the hope of providing gender-tailored strategy [
7]. Recently, family history of non-neuroendocrine gastroenteropancreatic (GEP) cancer, type 2 diabetes mellitus and obesity have been identified as independent risk factors for GEP-NENs by Feola et al. [
8]. And in our study, we tried to explore the potential prognostic value in LNM.
At present, there are two international clinical staging systems for pNETs: the American Joint Committee on Cancer (AJCC) and European Neuroendocrine Tumor Society (ENETS) [
9]. Although the two clinical staging systems are contradictory and inconsistent in data comparison, LNM is considered to be an important prognostic indicator by both of them and LN status includes N0 (no regional LNM) and N1 (regional LNM) [
10]. However, recommendations for lymphadenectomy in pNETs are still inconsistent. The National Comprehensive Cancer Network (NCCN) guidelines recommend regional lymphadenectomy for tumors of 1–2 cm due to the risk of LNM while performing routine lymphadenectomy blindly is not advocated in all tumors of < 2 cm [
11].
The literature addressing the significance of LNM in the management of nonfunctional pNETs remains conflicting. Several studies have demonstrated that LNM was not independently correlated with survival [
12,
13] while others have suggested that LNM is associated with poorer overall survival (OS) or disease-free survival (DFS) [
14,
15]. Therefore, a more accurate classification of these patients based on LNM is needed. The Union for International Cancer Control and AJCC tumor, node, metastasis (TNM) classifications divide LNM into N1 (1–3 positive LNs) and N2 (≥ 4 positive LNs) for high-grade pNET [
16,
17]. Nevertheless, the distinction is not valid for well-differentiated pNETs. The accuracy in prognostication of a TNM staging system based on positive LNs for well and intermediately differentiated pNETs is unknown.
Given that prognosis of pNETs with LNM varies widely and the importance of differentiation for LNM in TNM classification, we evaluated the predictive value of LNM for prognosis based on tumor grade. Moreover, we studied the preoperative predictive factors for LNM to guide surgical procedures and avoid unnecessary lymphadenectomy.
Discussion
Recently, systemic treatment of metastatic and advanced pNETs has made progress, although surgical resection remains the only radical therapy and represents the mainstay of treatment for resectable pNET [
20‐
22]. However, pancreatic surgery for pNETs is associated with potential morbidity [
23]. Thus, the optimal management for pNET currently remains controversial and the therapeutic strategies range from observation to surgery. The surgery varies from formal resection including pancreaticoduodenectomy or distal pancreatectomy to tumor enucleation with or without lymphadenectomy [
23‐
25]. The NCCN guidelines advocate formal resection with lymphadenectomy in tumors > 2 cm, but there is no firm consensus for smaller tumors. The guidelines suggest radiographic surveillance, formal resection, or enucleation with or without lymphadenectomy in smaller tumors, while lymphadenectomy is recommended for tumors of 1–2 cm in consideration of the risk of LNM [
26]. Despite these local surgical procedures historically showing short-term benefits, there is no conclusion on whether these surgical innovations have compromised long-term outcomes because of the indolent nature of these tumors. Given the lack of definitive treatment guidelines, physicians must evaluate multiple factors including LN status when determining the surgical procedure. Several studies have demonstrated that 30–40% of patients with nonfunctional pNET were diagnosed with LNM [
27,
28]. Thus, it is important to recognize preoperatively patients at high risk of LNM, who may benefit from lymphadenectomy.
Therefore, we first evaluated whether LNM was correlated with OS. Next, we sought to identify the related factors predicting LNM to guide clinical therapeutic decisions and avoid more aggressive therapies in low-risk patients. We noted that the guidelines may prefer to harvest LND ≥ 12. However, in the real world, surgeons generally remove fewer lymph nodes for neuroendocrine tumors than for pancreatic cancer because of inconsistencies in understanding the importance of lymph node dissection. Therefore, we chose the median number of lymph nodes harvested in our group for grouping in our study. Thus, grouped patients with LNM undergoing harvest LND 1–6 or > 6 lymph nodes.
The results on the prognostic value of LNM in pNET have been conflicting [
29,
30]. For example, some researchers have found that patients with LNM have a poor outcome [
31,
32], while others have reported that LNM does not decrease survival [
33,
34]. Our study indicated that LNM was a predictor of OS in grade 2/3 patients, while there was no such association in grade 1 patients. The data from our centric series confirmed the positive significance of regional lymphadenectomy in grade 2/3 patients, while adequate lymphadenectomy is not recommended for grade 1 patients because LNM shows key prognostic information about survival. Additionally, the univariate analysis suggested that LNM was a prognostic factor for grade 2/3 patients. However, prognostic factors for OS for grade 1 patients did not include LNM. Although LNM or not was no longer related to prognosis for patients with LND ≤ 6, LNM was significantly correlated with a poorer prognosis for patients with LND > 6. Moreover, there was no significant correlation between LNM and OS for grade 1 patients with LND > 6. While there was a significant correlation between LNM and poorer OS for grade 2/3 patients with LND > 6. When LND is sufficient, the conclusion that LNM predicts a poor prognosis for grade 2/3 patients with pNETs is still valid. Currently, the number of harvested lymph nodes for pNETs has not yet been properly addressed. Although, based on pancreatic cancer, it is recommended at least 12 lymph nodes should be removed for pNETs. Most surgeons generally believe that the role of LND is not very important and it is not necessary to expand the dissection to obtain more lymph nodes, so we have some patients with a relatively small number of LND [
35]. Therefore, regional lymphadenectomy may not be necessary for grade 1 patients and it is reasonable to make more selective decisions. Patients with more adverse tumor biology may benefit from removing occult nodal diseases.
We found that tumor grade and tumor size were associated with LNM. Given that it is often possible to get these two factors, we mainly focused on the tumor grade and size, which can be determined before surgery. Aguiar et al. found an increased prevalence of LNM in nonfunctional pNETs > 2 cm. However, they reported that 9% of patients with tumor size < 2 cm had LNM [
36,
37]. In contrast, Parekh et al. found that tumor size could not significantly predict LNM, although 31% of patients with LNM had tumors < 3 cm [
29]. Haynes et al. reported that factors positively associated with progression or metastasis of the disease also included tumor size (> 2 cm). However, among patients with tumor size < 2 cm, 8% of patients had metastasis [
38]. Our data indicated that tumors > 4 cm were almost twofold as likely to have LNM compared with tumors < 4 cm. However, 12.4% of patients with 1–4 cm tumors had LNM.
Grade 1 was correlated with a significantly low risk of LNM. Additionally, we found that LNM reliably predicted OS based on grade. Thus, clinical decisions may benefit from the classification of tumor grade, which usually depends on accurate pathological examination. Consequently, preoperative pathological evaluation can be performed using EUS combined with FNA. Piani et al. reported that Ki-67 expression on histological sections had good agreement with Ki-67 expression measured in cytological samples after EUS, in which the Ki-67 value was consistent in 89% and 78% of patients for Ki-67 values of 2% and of 2%–10%, respectively [
39]. Hasegawa et al. reported a 90% concordance rate for surgical histopathology with EUS–FNA-evaluated tumor grade using > 2000 cells (74% of patients) [
40]. The high concordance and reproducibility of EUS–FNA-determining Ki-67 values were further demonstrated by Weynand and colleagues [
41]. EUS–FNA is usually performed only at highly experienced centers. Preoperative examination of tumor grade, combined with tumor size may guide surgeons to choose the best surgical procedure and whether regional lymphadenectomy should be performed.
Additionally, regional lymphadenectomy may lead to the inclusion of splenectomy, increased blood loss, longer operating time and hospital stay, and increased lymphocele development. Thus, the benefits and risks of lymphadenectomy should be evaluated carefully. Our study demonstrated that there was no difference in OS between grade 1 patients with LNM and those without LNM. The benefits of lymphadenectomy in patients with grade 1 tumors remain unclear and more clinical trials and high-quality clinical data are needed to deal with the problem. Moreover, it is not clear if lymphadenectomy should be omitted for small nonfunctional pNETs because of low rates of LNM, and better prognosis compared with larger tumors. Gratian et al. reported that whether lymphadenectomy was performed did not significantly affect the 5-year OS in 1854 operated patients with nonfunctional pNETs ≤ 2 cm [
33]. Rui Mao et al. reported that lymphadenectomy did not show any survival benefit in patients undergoing resection for pNETs [
27]. Based on the current research on LND, many surgeons routinely carry out functional sparing surgery, such as spleen preservation, enucleation, middle pancreatectomy and so on [
27,
32]. Such surgical decision-making mainly depends on the size of the tumor in which functional sparing surgery is generally chosen for smaller tumors. However, there is still a certain risk in oncology. The present study revealing that the LNM was not associated with the prognosis of patients with grade 1 pNETs may provide some theoretical basis. Lymphadenectomy may not be performed routinely in patients with grade 1 and/or small tumors.
There were several limitations to the present study that should be considered in the interpretation of the data. First, the data collected from the SHPCI series were retrospective, thus well-designed clinical trials need to be performed to verify the results. Second, the database is from a single center and the sample size is not very large, therefore, the subgroup analysis was not sufficient. In addition, the low amount of LNM seen in grade 1 (which is expected with lower-grade tumors) may not provide enough power for survival. Finally, some patients from the SHPCI series lacked detailed follow-up data including postoperative imaging or care.
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.