Evaluation of Outcomes Following Surgery for Locally Advanced Pancreatic Neuroendocrine Tumors

Key Points Question What are the outcomes of patients who undergo aggressive resection of locally advanced pancreatic neuroendocrine tumors (PNETs)? Findings In this case series, 99 patients with locally advanced (T3/T4) PNETs and no known distant metastatic disease who underwent resection had a recurrence rate of 35%, but their quality of life remained high, and the overall 5-year survival was 91%. Meaning These findings suggest that locally advanced PNETs warrant aggressive surgical resection, including local organs and blood vessels, if necessitated by tumor invasion.


Introduction
Pancreatic neuroendocrine tumors (PNETs) are a heterogeneous group of epithelial neoplasms that comprise 2% to 7% of all pancreatic tumors. 1,2 As many as 10% of PNETs may arise in association with a hereditary syndrome, most commonly multiple endocrine neoplasia type 1 (MEN1); however, the majority of PNETs occur sporadically. 3,4 PNETs can be classified as either functional or nonfunctional depending on whether they secrete a hormone that usually causes a clinical syndrome. Given that most PNETs (60%-90%) are nonfunctional and therefore are not associated with secretion of hormones, patients are unfortunately usually diagnosed with advanced disease for which surgery may not be possible. 5,6 Complete surgical resection of a PNET has been suggested to be the only potentially curative treatment. 7 Recently, the European Neuroendocrine Tumor Society identified that assessment of the effects of surgical treatment for locally advanced PNETs is a major unmet need. 1 This study attempts to address this unmet need by evaluating the outcome of patients who underwent surgery for locally advanced PNETs without distant metastases (including liver metastases). The outcome of surgery for liver metastases from PNETs has been extensively studied previously. [8][9][10][11][12][13][14][15] However, little is known about surgery for locally advanced PNETs without liver (or other distant) metastases 16 because it has not, to our knowledge, been previously addressed, likely secondary to the rarity of this presentation. It is important because major surgery to resect locally advanced, possibly nodal metastatic PNETs may be morbid and have life-threatening complications.
Further, this aggressive surgery may not improve symptoms and long-term survival compared with newer treatments, such as somatostatin analogues, everolimus, sunitinib, interferon α, and peptide receptor radionuclide therapy (PRRT). 17,18 We set out to determine the association of resection of locally advanced PNETs with recurrence and mortality using a single academic institution database. We have attempted to define important variables in these patients that are associated with outcomes as defined by recurrence risk, 5-year disease-free survival, and overall survival.

Methods
This case series study of a single academic institution database of consecutive patients who underwent surgical resection of a locally advanced nondistant metastatic PNET was approved by the Stanford University School of Medicine institutional review board. The requirement for informed patient consent was waived because it was determined that this study did not directly involve human participants.

Study Population
Patients were included if they had undergone surgical resection of a locally advanced (>4 cm, T3/T4) PNET without distant metastases and with or without lymph node metastases. 19 Patients were treated between 2003 and 2018. The patients were followed up until death or censored at the cutoff date of November 31, 2019. The records were reviewed by 2 clinicians who were trained to perform standardized medical record abstraction on eligible procedures to supplement the databases.
Demographic characteristics, clinical data, and pathological data were retrospectively reviewed and recorded. Age, sex, imaging details, operative findings, tumor characteristics, and postoperative course were recorded. Functionality was assessed according to the presence of a detectable elevated serum level of the relevant hormone associated with a clinical syndrome. Mortality at last follow-up was obtained through a combined review of the medical record, social security index, and the California Cancer Registry. The outcomes measured were overall survival, disease-related survival, disease-free status, time to tumor recurrence, and quality of life as assessed by an Eastern Cooperative Oncology Group (ECOG) score. Recurrence was defined as clear identification of tumor recurrence on computed tomography, magnetic resonance imaging, or gallium 68 DOTATOC scan.

Statistical Analysis
In the time-to-event analyses, cumulative mortality was estimated by Kaplan-Meier approach and was presented as a curve increasing with time since surgery. Cox proportional hazard models were used to assess risk of recurrence after surgical resection of a locally advanced PNET after adjustment of multiple important demographic characteristics and clinical factors. We performed a stepwise approach for covariate selection and then added additional variables with potential clinical relevance (independent of significance) for inclusion in the model to maximize the risk adjustment. In our stepwise selection process, we used a 2-sided χ 2 test. All analyses were conducted using Stata version 15 (StataCorp). Statistical significance was defined as P < .05, and testing was 2-tailed. Data analysis was performed in August 2019.

Results
There were 249 patients with PNETs identified. After applying cohort selection criteria, the final study consisted of 99 patients (39.8%) with T3/T4 tumors and no distant metastatic disease (eFigure in the Supplement). Of these 99 patients, the mean (SEM) age was 57 (1.4) years and 56 (57%) were men. Table 1 highlights that the mean (SEM) duration of follow-up was 5.3 (0.06) years and the mean (SEM) follow-up after surgery was 5.1 (0.4) years.
Each patient underwent advanced imaging studies, most of which were computed tomography scans (88 [89%]) that demonstrated T3/T4 disease without distant metastatic disease (Figure 1 and Table 1). At the time of imaging, the mean (SEM) tumor size was 4 (0.2) cm. The pancreatic tail had the most tumors, which was found in 42 patients (42%). Twenty-five blood vessels demonstrated involvement on imaging. Tumors in 9 patients (9%) had imaging evidence of invasion to surrounding structures (Table 1). There was a mean (SEM) delay of 3 (1.2) months from presentation to surgery.

Discussion
To our knowledge, this study is unique in its evaluation of patients with locally advanced PNETs without distant metastatic disease. Most other publications have focused on either resection of liver or distant metastases to determine whether resection improves survival for those patients [8][9][10][11][12][13][14] or on the resection of tumors less than 2 cm (T1/T2 tumors), 20 but this study focused on large, locally advanced tumors. In patients with locally advanced disease, aggressive surgical excision of tumor remains controversial. 1,[21][22][23][24] Most patients in this series were felt to be unresectable by other surgeons and were referred to our center for a second opinion for that reason. A recent study reported that the use of surgery can vary by 4-fold between various centers treating patients with advanced PNETs. This large variation in the use of surgery in these patients was largely because of a lack of information from specific studies in this group of patients and is the primary reason that this area was identified as an important unmet need. 25 The patient pictured in Figure 1A was diagnosed as unresectable, given the need to resect and reconstruct the portal vein, but he was able to have his tumor completely excised. This study demonstrates that despite previous relative contraindications, there is a role for aggressive surgery in these patients. Patients had an excellent overall survival at 5 years of 91%, with an associated good quality of life, as indicated by low ECOG scores, and an overall recurrence rate of only 35%. Our findings suggest that surgical resection of locally advanced PNETS without distant metastatic disease is indicated.
Factors associated with tumor recurrence as well as overall survival have not been previously established for locally advanced tumors without metastatic disease. Locally advanced tumors may invade venous structures like the superior mesenteric vein and portal vein, and in these instances they require resection and reconstruction, but in many instances the tumor only abuts the vein without invasion and can be dissected off with blunt and sharp dissection. This was demonstrated by the fact that 25% of patients in our study had preoperative imaging findings suggestive of vascular involvement, but only 17% required resection and reconstruction. Vascular involvement was also not associated with an increased risk of recurrence, further suggesting that vascular resection and reconstruction is warranted. However, there are several retrospective small cohort studies that suggest the contrary. 20,[26][27][28][29] Interestingly, in a study done by US Neuroendocrine Tumor Study Group investigating 873 patients with nonfunctional localized (no distant metastatic disease) PNET that underwent surgical resection, recurrence-free survival was found to be associated with a multitude of factors, including tumor burden and lymph node involvement but not vascular involvement. 26 The 5-year disease-free survival in patients found to have locally advanced tumor was 60%, which was similar to our findings. 26 However, the current study demonstrated that male sex, lymph node metastases, and resection of additional organs (but not blood vessels) were associated with an increased probability of tumor recurrence. Interestingly, functional tumors had a lower incidence of   differences have yet to be fully elucidated.
Since the beginning of this study in 2003, new options for medical management of these patients have been developed, with the most recent being PRRT. 33 In addition to PRRT, recent phase 3 trials have led to the approval of everolimus, 34 sunitinib, 35 and somatostatin analogues 36 to treat patients with advanced PNETs. Each of these treatments has been shown to prolong progressionfree survival. Although they have not been shown to extend overall survival, each are now being increasingly used in patients with advanced PNETs. Among others, 37,38 these previous studies indicate that upfront therapy with either PRRT or the chemotherapy listed does reduce tumor burden and therefore may improve the ability to perform a successful resection on patients with locally advanced PNET. Prior to the availability of these newer nonsurgical treatments, there were few treatment alternatives.
The recent approval of these other nonsurgical treatments has increased the relevance of this study, which defines the beneficial role of surgery in patients with locally advanced PNETs. It provides the information necessary for the clinician to better select an appropriate antitumor treatment.

Limitations
This study has several limitations. Given that all patients included in the present study underwent surgery, there is no control group of similar patients who did not undergo surgery. There is also a selection bias toward operative treatment. We elected to include only patients who underwent surgery because having complete pathological analysis of the extent of tumor makes the survival data more meaningful. Additionally, tumor grade was not available for 14 patients, which may have skewed our multivariable model's finding regarding the lack of association of grade with recurrence risk, which has been well-established in the literature and was observed with our univariate analysis of patients with G2 tumors compared with patients with G1 tumors.

Conclusions
This study's findings suggest that patients with locally advanced PNETs without liver or distant metastases can have the tumor excised with acceptable rates of disease-free progression and mortality. Furthermore, 61 patients (61%) remained disease-free at 5 years postsurgery, and 91 patients (91%) were alive with excellent quality of life. These suggest that aggressive surgical resection can provide excellent outcomes for these patients.