Comparative Analysis of Complications Following Distal Resections for Neuroendocrine Tumors and Adenocarcinoma: Unveiling the Differences

Introduction

The incidence of pancreatic adenocarcinomas is 4–5/100 000 population [1] and it is one of the most common pancreatic tumors [1]. Pancreatic neuroendocrine tumors (pNETs) are much less common than adenomas and account for only 1–2% of pancreatic tumors [2]. Their incidence is 1–2/100 000 people and has been increasing in recent years [3]. Adenocarcinomas are more than twice as common in the head of the pancreas than in the body and tail [1], and neuroendocrine tumors occur twice as often in distal locations (the body and tail) compared to the head of the pancreas [4].

Adenocarcinomas and neuroendocrine tumors have different prognoses [5–7]. Adenocarcinomas are associated with a much shorter 5-year survival rate compared to pancreatic NETs (around 5.0% for adenocarcinomas) [5], while the 5-year survival rate of patients with pancreatic neuroendocrine tumors is 51.3% [5]. In addition, the differences between the above-mentioned histological types of pancreatic tumors also concern morphological parameters such as the frequency of metastases to lymph nodes and distant metastases, as well as the presence of intraepithelial neoplasia (PanIN) [6,8].

In the case of tumor localization in the body or tail of pancreas, distal resection with or without splenectomy is a standard procedure [9]. The decision to perform splenectomy depends, among other things, on the local advancement of the tumor [10]. Pancreatic fistulas are one of the most common postoperative complications after distal resections [9]. Risk factors for this type of complication include pancreatic texture and main pancreatic duct diameter [9,11]. Soft pancreas and pancreatic duct diameter <3 mm are risk factors for postoperative pancreatic fistula [9,11]. The texture of the pancreas is related to the histopathological type of the tumor [12]. Soft pancreas is more often found in neuroendocrine tumors compared to adenocarcinomas [13]. In addition, adenocarcinomas are more likely to have dilation of the Wirsung duct >3 mm [14]. The Fistula Risk Score (FRS), used to assess the risk of postoperative pancreatic fistula, considers histopathological diagnosis “other than adenocarcinoma or inflammatory tumor” as a risk factor for this type of complication [11].

Few studies have compared pancreatic adenocarcinomas and pancreatic neuroendocrine tumors located in the body or tail; therefore, the present study focused on the comparative analysis of these 2 histopathological types of pancreatic neoplasms. The aim of this study was a comparative analysis of preoperative data, morphological parameters of adenocarcinomas and neuroendocrine tumors of the body and tail of the pancreas, as well as the assessment of postoperative complications after resection.

Material and Methods

ETHICS STATEMENT:

The was a retrospective study of patients’ medical records. All data were fully anonymized. Written informed consent was obtained from all participants. All procedures were performed in accordance with the Helsinki Declaration on Human Medical Research and its subsequent amendments or comparable ethical standards. Our retrospective analysis of patients’ medical records fell into a category exempt from Institutional Review Board (IRB) approval according to local law.

STATISTICAL ANALYSIS:

A descriptive analysis was carried out. A 95% confidence interval was used. The distribution of quantitative variables was analyzed. Normally distributed variables are expressed as means and their standard deviations, while non-normally distributed variables are given as medians and their interquartile ranges. An analysis of differences between the groups of patients divided according to the histological type of cancer (NET or pancreatic adenocarcinoma) was performed. The following factors were taken into account in the analysis: age, sex, ASA-PS grade, histological grade (G1–G3), PanIN, presence of lymph node and distant metastases, adipose tissue and nerve trunk infiltration, presence of vascular occlusions, diameter of the pancreatic duct, texture of the pancreas, occurrence of pancreatic fistulas, risk group according to the FRS scale, and the severity of postoperative complications according to the Clavien-Dindo scale. The level of statistical significance was set at P<0.05. Homogeneity of variance was tested using Levene’s test. Quantitative variables were analyzed using the Mann-Whitney U test. Analysis of nominal variables was performed using the chi-square test or Fischer’s exact test when necessary. The correlation strength was calculated using the Phi coefficient for 2×2 matrices or Cramer’s V for larger matrices. A regression analysis showing the odds ratio along with the lower and upper confidence intervals was performed. All calculations and statistical analyzes were performed in IBM SPSS Statistics 26 (Armonk, New York, USA).

Results

The study group consisted of 61 men (46.6%) and 70 women (53.4%) with a median age of 63 years (IQR 51–69). Seventy-eight patients (59.5%) were ≥60 years old; while 30 patients (22.9%) were ≥70 years old. Seventy-seven patients (58.8%) had a neuroendocrine tumor, and 54 (41.2%) had an adenocarcinoma. Most patients were ASA-PS grade II (84 patients – 64.1%). The characteristics of the patients in the study population are shown in Table 1.

Morphological parameters are shown in Table 1. Tumors were classified according to the WHO classification into 3 groups: G1, G2, and G3 [21]. The G1 group included 48 (36.6%) patients, G2–70 (53.4%), and G3–13 (9.9%) patients. Among patients, PanIN 1 was the most common (59 patients – 45%), followed by 34 patients (30%) in both the PanIN 1–3 group and those without pancreatic intraepithelial neoplasia. PanIN 1–2 was the least common (4 patients – 3%). Nineteen (14.5%) patients had pancreatic duct diameter ≥3 mm. Soft consistency of the pancreas was found in 35 (26.7%) cases. Tumor metastases to local nodes were present in 65 (49.6%) patients. Distant metastases, mainly to the liver, were found in 19 patients (14.5%). Other parameters, such as perineural invasion, adipose tissue, and vascular occlusion, are also shown in Table 1.

In the analyzed population, the most common finding was no fistula or type A pancreatic fistula, in 113 (86.3%) patients. Type B fistulas affected 13 patients (10%), including 10 patients in the group with NET and 3 in the group of patients with adenocarcinoma. Type C fistula was found only in the NET group of patients, with 5 cases (3.8%). Most patients belonged to the moderate risk group according to the FRS scale (96 patients – 73.3%). Postoperative complications were presented using the Clavien-Dindo classification [24]. The most common severity of complications in patients with a diagnosis of neuroendocrine tumor was grade I in 64 (48.9%) patients. In contrast, among the group of patients with a diagnosis of adenocarcinoma, the most common was grade II in 25/54 (46.3%). The above data are presented in Table 1.

Patients with adenocarcinoma were significantly older than patients with NET (median age 68.5 and 58 years, respectively, P<0.001; OR=1.085). Patients with adenocarcinoma were more likely to have a higher ASA-PS class than patients with NET (P<0.001; OR (ASA-PS III/II)=4.071). More patients with adenocarcinoma had a histopathological diagnosis of G3 (12 vs 1 NET patient; P<0.001; OR (G3/G1)=24.0). Lymph node metastasis (P<0.001; OR=5.943), infiltration of adipose tissue (P=0.016; OR=2.625), and nerve trunks (P=0.045; OR=2.667) were more frequent in the group with adenocarcinoma as well as the pancreatic duct diameter ≥3 mm (P=0.045; OR=2.857). Clinically significant pancreatic fistulas were more common in patients with neuroendocrine tumors (P=0.044), and this group of patients also had a higher risk of complications in the form of postoperative pancreatic fistula as assessed by the FRS score (P<0.001; OR=5.943). Correlations between the study groups are shown in Table 1. The odds ratio of statistically significant results along with their lower and upper 95% confidence intervals are shown in Table 2. There was a strong statistical trend concerning the histopathological diagnosis G3 compared to G1 in adenocarcinomas (odds ratio of 24.0) (Table 2). A similar significant result was found between ASA-PS Scale III in comparison to ASA-PS Scale II. ASA-PS Scale III are more than 4 times likely to occur in adenocarcinomas than in NETs in comparison to ASA-PS Scale II (odds ratio 4.071). Lymph node metastasis was almost 6 times more common in adenocarcinomas than in NETs (odds ratio 5.943). These results are shown in Figures 1 and 2.

Discussion

AGE:

Chen et al reported that 38.4% of pancreatic neuroendocrine tumors occur in patients over age 60 years [30]. This is consistent with the results of our analysis. In our study, 33/77 (42.9%) of NET patients were over 60 years old, while only 10/77 (13%) were over 70 years old. patient age is an important issue affecting the course of postoperative complications and overall survival after NET resection, as demonstrated by Liu et al, who found that younger patients had a better long-term prognosis compared to older patients [31]. Similar report concerns patients with adenocarcinoma [32], although Cai et al found no relationship between long survival and patient age after resection of adenocarcinoma [33] because adenocarcinomas have a more aggressive course than pancreatic neuroendocrine tumors [33]. In the analyzed group, patients with adenocarcinoma were older (P<0.001). In this group 45/54 (83.3%) of patients were over 60 years old and 20/54 (37%) were over 70 years old. In our study, the median age of patients with adenocarcinoma was 68.5 years, which is comparable to that reported in other studies [32,33]. Partelli et al found a significantly older average age of patients with adenocarcinoma compared to patients with NET (64 vs 56 years, P<0.0001) [34].

SEX:

In the analyzed groups, there was no significant difference in the occurrence of NETs and PDAC according to sex. Males accounted for 48.1% of neuroendocrine tumors and 44.4% of adenocarcinomas. Some authors reported a higher percentage of men among patients with pancreatic NETs, amounting to approximately 55% [35]. According to the literature, the incidence of adenocarcinoma by sex is similar to that found in our analysis [18]. In both analyzed types of pancreatic tumors, long-term survival is lower in men compared to women [18,36].

ASA-PS:

Nell et al showed that most patients with NETs had a low ASA-PS grade (89% in ASA-PS I or II) [37]. Adenocarcinomas occur a lower percentage (approximately 75%) of patients with ASA-PS I and II compared to neuroendocrine tumors [36], probably due to age differences. There are differences in the frequency of ASA-PS grades between different studies, but most indicate a higher percentage of patients with ASA-PS III or higher in the group with adenocarcinoma compared to NET, which is consistent with the results of our study. In our analysis, most NET patients were ASA-PS I or II (62/77, 80.5%), while half of the adenocarcinoma patients were ASA-PS III (27/54, 50%); which shows a significant difference in the ASA-PS scale between patients in both groups (P<0.001; OR=4.071).

GRADING:

According to the literature, pancreatic neuroendocrine tumors G1 account for 40–50% of surgically treated NETs [38,39]. In our analysis, G1 tumors accounted for 41.6% of all surgically treated NETs. According to the literature, in patients with adenocarcinomas, G2 and G3 tumors are more common than G1 tumors [5]. In our study, G2 and G3 adenocarcinomas accounted for 48.1% and 22.2% of cases, respectively. On the other hand, G1 tumors were found in 29.6% of cases. The above results agree with the analysis of Yadav et al, which showed that neuroendocrine tumors have a lower grade of histological malignancy compared to adenocarcinomas [5].

PANIN:

Pancreatic intraepithelial neoplasia (PanIN) is one of the precursor lesions in pancreatic adenocarcinomas [40]. It is most common in this histopathological type of pancreatic tumors, but it is also found in neuroendocrine tumors and chronic pancreatitis [8,40]. In neuroendocrine tumors, PanIN 1 predominates, with an incidence of 60–75% [8,40]. However, PanIN 1–3 is more common in adenocarcinomas, occurring in about 75% of cases [41]. In the analyzed group, PanIN 1 predominated in the group of patients with neuroendocrine tumors (59.7%). Among patients diagnosed with adenocarcinoma, PanIN 1–3 was found more often than in patients with neuroendocrine tumors (57.4%).

According to the literature, PanIN 1–3 is twice as common in patients over 60 years old [40]. The more frequent occurrence of PanIN 1–3 in our study group may be related to the fact that among patients diagnosed with adenocarcinoma, the median age was higher compared to patients with neuroendocrine tumors. Hassid et al showed that as the grade of intraepithelial neoplasia increases, the percentage of tumors that have metastasized to the lymph nodes increases [42]. In our analysis, PanIN 1–3 was more common in patients with adenocarcinoma, which was more likely to metastasize to lymph nodes.

INFILTRATION OF STRUCTURES:

The presence of adipose tissue and perineural infiltration and vascular occlusions is associated with a greater aggressiveness of the tumor and worsens the long-term prognosis [6]. In the case of perineural infiltrations, they are more common in patients with adenocarcinomas than neuroendocrine tumors [6]. In the analyzed population, a statistically significant difference was found between the study groups in terms of adipose tissue and perineural infiltration. In the group of patients with adenocarcinoma, perineural and adipose tissue infiltration were 1.5 times more frequent than in patients with neuroendocrine tumors. Takahashi et al showed that vascular occlusions occur twice as often in the case of adenocarcinomas compared to neuroendocrine tumors [6]. We found a higher percentage of patients with NETs (58.4%) with vascular occlusions compared to reports in the literature showing that they occur in about 40% of patients [6]. This may be the reason for not observing statistically significant differences between the groups in the presence of vascular occlusions.

LYMPH NODES AND SISTANT METASTASES:

In patients with pancreatic neuroendocrine tumors, lymph node metastases occur 9–38% of cases [6,43]. In the analyzed group, nodal metastases were found in 32.5% of cases. Compared to pancreatic neuroendocrine tumors, ductal adenocarcinomas metastasize more often, metastasizing to lymph nodes in up to 75% of cases [6]. In our study, nodal metastases occurred in 74.1% of patients with adenocarcinomas, which is consistent with the literature [6].

According to Dasari et al, distant metastases occur in about 15% of patients with neuroendocrine tumors [44]. The present study found that distant metastases occurred in 16.9% of patients with NETs. According to the available literature, pancreatic adenocarcinomas have distant metastases at the time of diagnosis in about 15% of cases [45]. In our study, distant metastases occurred in 11.1% of patients with in adenocarcinomas. We found more distant metastases in patients with neuroendocrine tumors compared to adenocarcinomas, but the difference was not statistically significant. In patients with pancreatic adenocarcinomas, most distant metastases occur metachronously [45], but in patients with neuroendocrine tumors, most distant metastases occur synchronously [46]. Our analysis included only synchronous distant metastases due to the retrospective nature of the analysis. Therefore, the number of metachronous metastases in the analyzed group is underestimated.

WIRSUNG DIAMETER:

In distal pancreatic resections, complications often arise from the difficulty of visualizing the pancreatic duct and closing it safely [47]. In our study, pancreatic duct diameter ≥3 mm was present in 7 (9.1%) patients with NET and in 12 (22.2%) patients with adenocarcinoma (P=0.045, OR=2.857). Also, authors of other papers comparing the incidence of dilated pancreatic duct in patients with NET and in those with adenocarcinoma reported that they were significantly more common (2–3 times) in patients with adenocarcinomas [14,48]. Jin et al reported a 22.6% incidence of pancreatic duct ≥3 mm in patients with neuroendocrine tumors [49], but it was a study discussing the dilatation of the diameter of the pancreatic duct based on preoperative magnetic resonance imaging. The diameter of the pancreatic duct may differ in preoperative and intraoperative studies because it is often measured at different locations. We found that a dilated pancreatic duct was significantly more common in patients with adenocarcinoma than in patients with NET (P<0.01). According to Baxi et al, pancreatic duct dilatation occurs up to 8 times more often in patients with adenocarcinomas compared to those with neuroendocrine tumors, regardless of the location of the lesion in the pancreas [14]. This is due to the greater “hardness” of the adenocarcinoma tumor [14], and this is another factor contributing to the higher incidence of pancreatic fistulas in patients with neuroendocrine tumors [50].

PANCREATIC TEXTURE:

Assessment of the texture of the pancreas is a subjective evaluation, made by the operating surgeon [12], but for an experienced surgeon, it is a method that is recognized in research [12]. In the group we analyzed, soft pancreas was present in 23 (29.9%) patients with NET and in 12 (22.2%) patients with adenocarcinomas, but the difference was not statistically significant (P=0.423). In neuroendocrine tumors but not in adenocarcinomas, there is no desmoplastic reaction in the pancreas. This results in less frequent pancreatic fibrosis in neuroendocrine tumors compared to other neoplastic lesions and chronic inflammatory tumors of the pancreas [11]. As a result, the soft consistency of the pancreas is more often found in patients with neuroendocrine tumors compared to adenocarcinomas [13]. According to Partelli et al, soft pancreas is 2 times more common in patients with NET compared to adenocarcinomas [34].

PANCREATIC FISTULAS AND FISTULA RISK SCORE:

The overall incidence of pancreatic fistulas after pancreatic resection surgery was 5–30% without histological grade differentiation [50,51]. Clinically significant pancreatic fistulas are more common in distal resections of neuroendocrine tumors compared to adenocarcinomas [52–54]. According to the literature, pancreatic fistulas are 1.5 times more common in pancreatic patients with NETs [53,55]. The analysis showed that type B and C pancreatic fistulas were found in 13.5% and 6.0% of neuroendocrine tumor cases, respectively. Type B pancreatic fistulas occurred in only 6.0% of cases. No case of type C pancreatic fistula was found in patients with adenocarcinoma.

The histopathological diagnosis of a tumor “other than adenocarcinoma” is also a risk factor for postoperative pancreatic fistula on the FRS scale [11,19]. The results calculated according to the FRS scale were similar in the 2 groups. Twenty-two (28.6%) patients with a neuroendocrine tumor were classified as being at high risk for postoperative pancreatic fistula. In contrast, among patients with adenocarcinoma, only 3.7% were found to be at high risk of pancreatic fistula.

CLAVIEN-DINDO COMPLICATIONS:

In our study, clinically significant (≥III) complications on the Clavien-Dindo scale were 16/77 (21%) for NET and 12/54 (22%) for adenocarcinoma. Heidsma et al presented the “textbook outcome” (TO) of patients operated on for neuroendocrine tumors [52]. TO was defined as no significant postoperative complications (Clavien-Dindo ≥III), no 90-day postoperative mortality, no prolonged hospital stay, no rehospitalization within 90 days, and no R0 resection [52]. For distal TO pancreatectomy, it was 56.7% for neuroendocrine tumors [52]. According to the literature, clinically significant complications occur in 20–30% of patients operated on for NET [37,39]. Beek et al points to the extent of resection as a key risk factor for major complications on the Clavien-Dindo scale after resections of neuroendocrine tumors [56]. Researchers indicate a significantly higher risk of clinically significant complications when pancreatoduodenectomy is performed as compared to distal resections (OR: 3.13; CI 95%: 1.32–7.41) [56]. On the other hand, for adenocarcinomas, postoperative complications of grade ≥III on the Clavien-Dindo scale affect 19–26% of patients [20,57]. The literature contains many reports of the higher incidence of significant postoperative complications after resections of adenocarcinomas compared to neuroendocrine tumors. In the population we analyzed concerning distal resections, clinically significant postoperative complications occurred with similar frequency in both groups, although the overall incidence of postoperative complications was higher for adenocarcinomas than neuroendocrine tumors. Patients with NETs tended to be younger than those with adenocarcinoma, which may influence the frequency of postoperative complications. However, there are also other factors such as soft pancreas and non-dilated pancreatic duct in NETs, which may have influenced our results.

SURGICAL TECHNIQUE:

Modifications of surgical technique aimed at reducing the rate of complications after pancreatic resection procedures are subject to ongoing discussion [58–60]. The use of minimally invasive surgery in distal pancreatic resections improves some short-term results [58,59]. According to Karunakaran et al, the use of minimally invasive methods clearly shortened the length of hospital stay [61]. However, there are reports showing the rate of clinically significant complications after minimally invasive procedures does not differ from open surgery [55,61], but Adams et al found a higher incidence of pancreatic fistulas after laparoscopic distal pancreatic resections [58]. This may be due to the use of a mechanical suture closing the cross-section of the remaining pancreatic parenchyma [58]. However, Tieftrunk et al showed the advantage of mechanical suture over manual suture in the prevention of postoperative pancreatic fistula after distal pancreatic resections [62].

The impact of the use of sealants, such as fibrin glue or fibrin patch, on the occurrence of postoperative complications, including pancreatic fistula, remains unclear [60,63]. According to Serra et al, the use of sealants in pancreatic surgery requires further validation [63]. Each modification of the surgical technique during pancreatic resection procedures should be individualized and adapted to the specific clinical case [61]. We did not analyze the surgical technique in the context of postoperative complications because there were no differences in the technique in our patients – all patients in the study population were operated on using the open method via laparotomy. The pancreatic duct was closed with 2–3 single sutures and the cross-section of the pancreas was closed with single sutures in 2 layers.

The undoubted limitation of the study is the lack of follow-up, which results, among other things, in an underestimation of the number of metachronous distant metastases in both groups, which prevents a complete characterization of the studied tumor types. Another limitation is the retrospective nature of the analysis, but, taking into account the small number of publications comparing adenocarcinomas and neuroendocrine tumors of the pancreas with distal localization, the present study has high didactic utility.

Conclusions

Patients undergoing distal resections tend to be older and have a higher ASA-PS score for adenocarcinomas compared to those with neuroendocrine tumors. Adenocarcinomas in the pancreatic body and tail are more aggressive than NETs and tend to have a higher grade of histological malignancy. They cause more frequent infiltration of adipose tissue and perineural invasion, and more often metastasize to lymph nodes. Synchronous distant metastases occurred at a similar rate in both groups.

A pancreatic duct ≥3 mm in diameter is more commonly found in pancreatic adenocarcinomas. In contrast, the softer texture of the pancreas and a less dilated pancreatic duct were more common in patients with neuroendocrine tumors. This may contribute to the higher incidence of clinically significant pancreatic fistulas in patients with NETs, although postoperative complications of Clavien-Dindo scale grade III and above occurred to a similar degree in both groups.