- Email: [email protected]
Resection versus expectant management of small incidentally discovered nonfunctional pancreatic neuroendocrine tumors
ARTICLE IN PRESS
Resection versus expectant management of small incidentally discovered nonfunctional pancreatic neuroendocrine tumors Alex M. Rosenberg, BA,a Patricia Friedmann, MS,a Jaydira Del Rivero, MD,a,b Steven K. Libutti, MD,a,b and Amanda M. Laird, MD,a,b Bronx, NY
Background. Sporadic, nonfunctional pancreatic neuroendocrine tumors (NF-PNETs) are diagnosed with increasing frequency. We compared the risk of tumor growth, metastasis, and mortality between patients treated versus those treated expectantly. Method. A retrospective study of patients seen at our institution with sporadic NF-PNETs, with $12 months of follow-up. Kaplan-Meier analysis was performed. Results. Between 1999 and 2014, 35 patients with an incidentally discovered nonfunctional PNET were identified. Twenty underwent resection and 15 were followed with imaging. In the operative group, 8 had NF-PNETs < 2 cm, while 12 had NF-PNETs $ 2 cm. In the nonoperative expectant management by serial imaging group, 10 had NF-PNETs < 2 cm while 5 had NF-PNETs $ 2 cm. Small NF-PNETs (<2 cm) in either the operative or nonoperative groups demonstrated no evidence of progression or metastasis (median follow-up of 27.8 months). Morbidity in the operative group was 35% with pancreatic pseudocyst the most common. Conclusion. Incidentally discovered NF-PNETs < 2 cm in size can be observed safely with serial imaging. (Surgery 2015;j:j-j.) From the Albert Einstein College of Medicinea and Montefiore Medical Center,b Bronx, NY
PANCREATIC NEUROENDOCRINE TUMORS (PNETs) are rare neoplasms of islet cells in the pancreas. The reported incidence is 1–5/100,000.1,2 Detection is increasing owing to the widespread use of axial imaging.2 Presentation of these lesions depends on whether they are functional or nonfunctional (NF). Functional PNETs can secrete hormones such as insulin, glucagon, gastrin, vasoactive intestinal polypeptide, or somatostatin.3 In contrast, NF-PNETs do not cause hormonal imbalance and are typically asymptomatic. Diagnosis is most often incidental; the majority of NF-PNETs are discovered as a result of routine imaging or in the evaluation of nonspecific abdominal pain.3,4 Although
Presented at the 36th Annual Meeting of the American Association of Endocrine Surgeons, May 2015, Nashville, Tennessee. Accepted for publication October 14, 2015. Reprint requests: Dr Amanda M. Laird, MD, Montefiore Medical Center, 3400 Bainbridge Avenue, Bronx, NY 10461. E-mail: [email protected]. 0039-6060/$ - see front matter Ó 2015 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.surg.2015.10.013
resection is recommended generally for functional lesions, controversy regarding management of small NF-PNETs persists.5,6 Management of NFPNETs with resection has demonstrated a positive impact on mortality7; however, the benefit of resecting small NF-PNETs is limited, and management guidelines are mixed. Small tumors (<2 cm) have a more indolent course, and many studies suggest these NF-PNETs may be monitored safely through imaging.7 Management guidelines regarding NF-PNETs vary from institution to institution. Data from the Massachusetts General Hospital recommend resection of all NF-PNETs regardless of their size; a study by Haynes et al8 of 139 patients, all managed by surgical resection, demonstrated improved outcomes regardless of tumor size. These data suggested that NF-PNETs of all sizes should be resected to minimize disease progression; however, a nonoperatively managed control group was not used for comparison. In addition, data from the National Cancer Database demonstrated that patients small NF-PNETs managed by resection have an overall lesser mortality rate.9 Other data contradict these findings. Data from the Mayo Clinic SURGERY 1
ARTICLE IN PRESS 2 Rosenberg et al
support observation for small NF-PNETs (<2 cm), because these neoplasms have limited growth; observation had no impact on survival or development of metastases.5 Furthermore, these conclusions were based on outcomes from both an operative group and a nonoperative comparison group. Other groups have made similar findings.10 Patient selection is a critical issue in many of the published studies. The foregoing studies were retrospective, which resulted in selection bias in treatment groups. Randomizing intervention among patients with NF-PNETs would remove selection bias, but because NF-PNETs are rare, recruitment at a single institution for a randomized controlled trial is difficult. The morbidity rate for pancreatic resection of these lesions can be as great as 40%, which includes pancreatic leak and abscess formation.11-13 It is unclear whether all NF-PNETs need to be managed by resection, given the potential for postoperative complications. The present study sought to address the management of these lesions by evaluating outcomes with respect to distant metastases, tumor growth or recurrence, and overall rates of mortality, progression-free survival (PFS), and disease-free survival (DFS) in patients with NF-PNETs < 2 cm managed either operatively or nonoperatively by expectant management. Our secondary aim was to assess operative-related morbidity. We evaluated the extent to which smaller NF-PNETs (<2 cm) can be managed nonoperatively. We further hypothesized that larger NF-PNETs ($2 cm) are biologically distinct and that, as a result, outcomes may be improved with operative management. METHODS Patients who were diagnosed with pancreatic islet cell tumors at a single academic institution from January 1999 to November 2014 were identified from preexisting institutional registries using the Clinical Looking Glass, a health care, quality improvement software at Montefiore Medical Center.14 Patients were followed for $12 months. NFPNETs were selected based on pathology reports, computed tomography (CT), and MRI. Patients were divided into 2 cohorts based on whether they were managed operatively or observed. Patients deemed unresectable were included in the observational cohort. Within each cohort, patients were grouped based on a primary tumor size of >2 or <2 cm. Records were reviewed to catalog data including patient demographics, method of diagnosis, pathologic findings, Ki-67 index, imaging findings, operative interventions, follow-up imaging, tumor
Surgery j 2015
recurrence for those patients managed operatively, tumor growth for those patients managed nonoperatively, postoperative complications, and death. Tumor progression was determined using the Response Evaluation Criteria in Solid Tumors. Tumor growth was considered clinically relevant if the lesion demonstrated an increase in size of $20% from baseline and a 5-mm absolute increase in size from baseline.15 Postoperative complications were graded using the Clavien-Dindo classification system.16 Diagnosis of NF-PNET was based on classic and well-established CT or MRI findings or cytopathology; nonfunctionality was established by history of symptoms and laboratory analysis. Patients managed nonoperatively were followed in an outpatient setting. Every 6 months, these patients were evaluated using imaging studies including CT or MRI. Exclusion criteria included (1) pathologic signs of high grade features, including a Ki-67 of $20% index, (2) history of a familial syndrome associated with PNETs, including multiple endocrine neoplasia or Von Hippel Lindau, (3) existing metastases at the time of diagnosis, (4) patients #17 years, (5) or patients with <12 months of follow-up. Patient clinical characteristics, including survival and disease progression data, were cataloged from patient medical records. Survival in the nonoperative and operative groups was calculated from date of diagnosis to date of death or to date of last contact for patients who were still alive at the time of their last follow-up. Patients who were alive at the date of last contact were censored at this date. Outcomes including metastases, tumor growth/recurrence, overall mortality, and PFS for nonoperatively managed patients and DFS for operatively managed patients were compared. Time-to-event data were described using the Kaplan-Meier method. Survival curves were compared using the log-rank test. Categorical variables were analyzed using the Chi-square and Fisher’ exact tests. Continuous variables were analyzed using the non-parametric Wilcoxon rank-sum test. The Institutional Review Board at Albert Einstein College of Medicine approved this study. RESULTS Overall, 35 patients met the inclusion criteria. Of those, 20 underwent operative resection and 15 were managed nonoperatively. Median duration of follow-up for all patients was 30 months (range, 12–143), with 34 months (range, 12–143) for operatively managed patients and 28 months (range, 19–113) for nonoperatively managed
ARTICLE IN PRESS Rosenberg et al 3
Surgery Volume j, Number j
Table I. Patient and tumor characteristics Nonoperative group (n = 15) Median age, y (range) Sex, n (%) Male Female Ethnicity, n (%) African American Latino or Hispanic Caucasian Asian Multiracial Other/unknown Clinical status Neoplasm baseline size, cm (range) Neoplasms of baseline size < 2 cm, n (%) Neoplasms of baseline size $ 2 cm, n (%) Neoplasm end size, cm (range) Follow-up time, months (range) Tumor grade Ki-67 < 2%, n (%) Ki-67 3–20%, n (%) No Ki-67 data available, n (%) Location of neoplasm, n (%) Head Tail Neck Body Uncinate process Indeterminate Indication for abdominal imaging, n (%) Abdominal pain Other cancer surveillance Other disease surveillance Back Pain Other Unspecified
61 (31–70) 8 (53) 7 (47)
Operative group (n = 20) 61 (44–77)
P value .845* .42*
10 (50) 10 (50) .775y
7 3 2 1 2 0 1.4 10 5 1.6 28
(47) (20) (13) (7) (13) (0) (0.8–11.0) (67) (33) (0.8–12) (19–113)
0 (0) 4 (27) 11 (73)
7 6 3 0 2 2
(35) (30) (15) (0) (10) (10)
2.3 (0.5–13.5) 8 (40) 12 (60) N/A 34 (12–143)
.113* .118* .118* z .842*
13 (65) 6 (30) 1 (5) .428y
4 (27) 5 (33) 2 (13) 1 2 (13) 1
3 6 3 5 0 3
(15) (30) (15) (25) (0) (15)
6 (40) 3 (20) 3 (20) 0 1 2 (13.3)
8 (40) 1 5 (25) 1 1 4 (20)
.842y
*Chi-square test. yFisher’s exact test. zInsufficient data for comparison. N/A, Not applicable.
patients. Median age was 61 years (range, 31–70) in the nonoperative group and 61 years (range, 44–77) in the operative group. One-half of the operatively managed patients and 47% of nonoperatively managed patients were women. Our patient pool was diverse racially. Overall, 14 patients were African American (40%), 9 were Hispanic (26%), 5 were Caucasian (14%), 2 were multiracial (11%), 1 was Asian (3%), and 2 did not identify as any race (6%). Neoplasms were located most commonly in the pancreatic tail (30%) and body (25%) in the operative group. NF-PNETs managed nonoperatively were found most frequently in the tail (33%) and head of the
pancreas (27%). Abdominal pain was the most common indication for abdominal imaging for patients in the operative group (40%) and the nonoperative group (40%). Treatment groups were comparable with respect to age, sex, ethnicity, location of neoplasm, indication for abdominal imaging, and NF-PNET baseline and end sizes (Table I). Operative group. Median size of tumors managed operatively was 2.3 cm based on imaging (range, 0.5–13.5). Seven patients were African American (35%), 6 were Hispanic (30%), 3 were Caucasian (15%), 2 were multiracial (10%), and 2 did not identify as any race (10%). Thirteen
ARTICLE IN PRESS 4 Rosenberg et al
Surgery j 2015
Table II. Outcomes for operative and nonoperative patients Log-Rank Nonoperative group
Operative group
n = 15 0 3 (20) 0 12 (80) N/A 10 0 0 0 10 (100) N/A 5 0 3 (60) 0 2 (40) N/A
n = 20 1 2 (10) 2 (10) 17 (85) 7 (35) 8 0 0 0 8 (100) 1 (13) 12 1 2 (17) 2 (17) 9 (75) 6 (50)
A. Outcomes for all neoplasms Tumor Growth/Recurrence, n (%) Metastases, n (%) Mortality, n (%) Progression/disease-free survivalz, n (%) Operative morbidity, n (%) B. Outcomes for neoplasms < 2 cm, n Tumor growth/recurrence, n (%) Metastases, n (%) Mortality, n (%) Progression/disease-free survivalz, n (%) Surgical morbidity, n (%) C. Outcomes for neoplasms $ 2 cm, n Tumor growth/recurrence, n (%) Metastases, n (%) Mortality, n (%) Progression/disease-free survivalz, n (%) Operative morbidity, n (%)
P-value* .4497 .3891 .4292 .5377 y y y y y y x x x x y
*P values were measured using Kaplan-Meier method; hazard ratios were measured using Cox-regression method for small samples. yInsufficient data for comparison. zProgression-free survival included tumor growth, metastases, and mortality as events for nonoperative patients; disease-free survival included tumor recurrence, metastases, and mortality as event for operative patients. xSample size is too small for statistical analysis. CI, Confidence interval; N/A, not applicable.
patients (65%) had a Ki-67 proliferative index of <2% and 6 patients (30%) had a Ki-67 index of 3– 20%. No Ki-67 data were available for 1 patient. In the operative cohort, 8 patients (40%) had NFPNETs < 2 cm, and 12 had NF-PNETs > 2 cm (60%). One resected tumor recurred, and 2 patients (10%) developed metastases, all of which were liver metastases. Additionally, an overall mortality rate of 10% and a DFS rate of 85% were observed for all NF-PNETs in the operative group (Table II). The operative complication rate for all NF-PNETs was 35% with pancreatic pseudocyst formation being the most common morbidity. Four patients (20%) developed grade I complications, which required no intervention. Three patients developed grade III complications, which required placement of drains. Distal pancreatectomy was the most commonly performed operation (Table III). Nonoperative group. NF-PNETs managed nonoperatively had a median baseline size of 1.4 cm (range, 0.8–11.0) on CT or MRI. No tumors grew in size, and no patients were moved to the resection group after date of diagnosis. Seven patients were African American (47%), 3 were Hispanic (20%), 2 were Caucasian (13%), 2 were multiracial (13%), and 1 was Asian. No patient had
a proliferative index of <2%, 4 patients (27%) had a Ki-67 index of 3–20%, and no Ki-67 data were available for 11 patients (73%). Ten patients (68%) had NF-PNETs < 2 cm, and 5 (33%) had NF-PNETs $ 2 cm (Table I). Five patients with tumors > 2 cm were not managed operatively owing to local invasion of the superior mesenteric artery. The median tumor size was 1.6 cm (range, 0.8– 12.2). No patients had tumor growth and 3 patients (20%) developed metastases, all in the liver. No patients in the observation group died. Patients in this treatment group had a PFS rate of 80% (Table II). Operative versus nonoperative. No difference in metastases was observed after 30 months (range, 12–143) of follow-up between the operative and nonoperative groups (P = .3891; Fig 1, A). Furthermore, however, no difference in overall survival rates between nonoperative (100%) and operative groups (90%; P = .4292; Fig 1, B). PFS rates in the nonoperative group (80%) and DFS rates in the operative group (85%) demonstrated no difference (P = .5337; Fig 1, C). In addition, no difference in tumor growth/recurrence was observed (P = .4497; Table II). The median tumor sizes of the operative and nonoperative groups were 2.3 and 1.4 cm, respectively. Because differences
ARTICLE IN PRESS Rosenberg et al 5
Surgery Volume j, Number j
Table III. Types of operation and related morbidities Operative group (n = 20) A. Type of operation, n (%) Distal pancreatectomy Pancreatoduodenectomy Distal pancreatectomy and splenectomy Subtotal Pancreatectomy Central Pancreatectomy Enucleation Unspecified B. Morbidity, n (%) Pancreatic pseudocyst Hematoma Pneumobilia Pancreatitis Any morbidity C. Clavien-Dindo postoperative complication,* n (%) Grade 1 Grade II Grade III Grade IV Grade V (death)
8 (40) 5 (25) 3 (15) 1 1 1 1 4 (25) 1 1 1 7 (35)
4 (20) 0 3 (15) 0 0
*Grade I: Any deviation from the normal postoperative course without the need for pharmacologic treatment or operative, endoscopic, and radiological interventions. Allowed therapeutic regimens are: drugs as antiemetics, antipyretics, analgesics, diuretics, electrolytes, and physiotherapy. This grade also includes wound infections opened at the bedside. Grade II: Requiring pharmacologic treatment with drugs other than such allowed for grade I complications. Blood transfusions and total parenteral nutrition are also included. Grade III: Requiring surgical, endoscopic or radiological intervention. Grade IV: Life-threatening complication requiring ICU management. Grade V: Death of a patient.
in baseline tumor size represent a confounding variable, both groups were divided into the following subgroups: small NF-PNETs (<2 cm) and large NF-PNETs ($2 cm). Comparison of outcomes was then made between similarly sized NFPNETs. Small versus large PNETs. Ten patients in the nonoperative group and 8 in the operative group presented with large NF-PNETs (<2 cm), representing 66% and 40% of their respective groups. There was no interval development of metastases, no tumor growth, and no mortality in patients with small NF-PNETs managed nonoperatively after 27 months of follow-up (range, 16–50). Likewise, in patients with small NF-PNETs managed operatively, no metastases, tumor recurrence, or mortality was observed during that time. (Table II) Five patients in the nonoperative group and 12 patients in the operative group presented with large NF-PNETs ($2 cm), representing 33% and
60% of each respective group (Table I). In patients with large NF-PNETs who were managed nonoperatively, no change was observed in the median primary tumor size over the course of 37 months of follow-up (range, 12–143). In contrast, 60% of nonoperatively managed patients (3 patients) had new metastases identified during that time. In comparison, in patients with large NF-PNETs that were managed operatively, 17% of the cohort (2 patients) had metastases, and 1 patient had a recurrence identified during that time. All metastases developed in the liver. In the subgroup of patients with large PNETs who were managed nonoperatively, there were no deaths recorded over the study period. In patients with large NFPNETs who were managed operatively, 2 deaths (17%) occurred. PFS was 40% and DFS was 75%, respectively, for nonoperatively and operatively managed large NF-PNET groups (Table II; Fig 2, A–C). DISCUSSION The optimal management of small, incidentally discovered NF-PNETs is currently a topic of debate in the literature.17,18 Our results demonstrate different outcomes for low-grade NF-PNETs <2 cm when compared with low-grade NF-PNETs $2 cm. Patients with small NF-PNETs managed nonoperatively have similar outcomes whether treated operatively by resection or nonoperatively. Small tumors (<2 cm) in both the nonoperative and operative subgroups did not develop metastases, did not grow or recur, and no deaths occurred over 28 months of follow-up. The only adverse outcome that patients with these small tumors encountered in their management was morbidity related to the operation. Small NF-PNETs in our population seem to be indolent, which is consistent with other studies.6 These data suggest that nonoperative management results in similar outcomes compared with those patients with small tumors managed operatively, while eliminating the risk of operatively related morbidity. The overall operative complication rate in this group was 35% (7/20), with 3 patients who required postoperative intervention. The rate of pseudocyst formation was greater than expected, which may be as a result of a subclinical pancreatic leak. These findings are similar to published averages of 40%. Patients may, therefore, avoid potential morbidity if the selective approach of observation of small NF-PNETs is adopted rather than operating on all patients with NF-PNETs, regardless of size. Therefore, we advocate an expectant approach, wherein small ($2 cm)
ARTICLE IN PRESS 6 Rosenberg et al
Surgery j 2015
Fig 1. (A) Time to metastasis for all patients, by treatment. (B) Overall survival for all patients, by treatment. (C) Progression/disease-free survival for all patients, by treatment. N, Number of patients at risk in in nonoperative group; S, number of patients at risk in the operative (surgical) group. *Progression-free survival included tumor growth, metastases, and mortality as events for nonoperative patients; disease-free survival included tumor recurrence, metastases, and mortality as event for operative patients.
NF-PNETs tumors are monitored to detect any change in growth to determine if and when operation becomes necessary. These data add to the growing discussion regarding management of small NF-PNETs. Results observed in our study were similar to those reported in both Lee et al5 and Gaujoux et al.10 These studies demonstrated comparable outcomes for patients managed nonoperatively and operatively with no signs of disease progression or mortality in patients with small NF-PNETs.5,10 These results support the argument that small NFPNETs can be managed safely with observation and serial imaging. In these studies, the only adverse outcome patients with small NF-PNETs experienced was as a result of operative resection.5,10 In comparison, a series of operatively managed patients demonstrated tumor-related mortality rates in well-differentiated tumors of 7.1% for tumors identified as benign and 14.0% for tumors identified as neither benign nor malignant.8 It is difficult to determine the clinical
importance of these results, because a nonoperative group was not used in the study. Patients managed nonoperatively could have demonstrated equally positive outcomes. Although operative management can be effective in managing small NF-PNETs, the outcome of an expectant approach may equivalent. Large NF-PNETs tended to recur, developed liver metastases, and mortality was greater. Although large tumors are typically managed by resection, 5 large NF-PNETs in this study were managed nonoperatively, because the primary tumor was technically unresectable owing to local invasion. These tumors were included in the nonoperative group, because they contribute to the understanding of the natural history of this heterogeneous cohort. Three of these 5 nonoperatively managed patients developed metastases, whereas 2 of the 12 operatively managed patients developed metastases, which is associated with disease-related mortality.19 These data suggest that operative management of NF-PNETs > 2 cm
ARTICLE IN PRESS Surgery Volume j, Number j
Rosenberg et al 7
Fig 2. (A) Time to metastasis for patients with nonfunctional pancreatic neuroendocrine tumors (NF-PNETs) $ 2 cm, by treatment. (B) Overall survival for patients with NF-PNETs $ 2 cm, by treatment. (C) Progression/disease-free survival for patients with NF-PNETs $ 2 cm, by treatment. N, Number of patients at risk in the nonoperative group; S, number of patients at risk in the operative (surgical) group. *Progression-free survival included tumor growth, metastases, and mortality as events for nonoperative patients; disease-free survival included tumor recurrence, metastases, and mortality as event for operative patients.
may be associated with improved outcomes. Therefore, NF-PNETs > 2 cm should be considered for resection if feasible. The 3 NF-PNETs managed operatively that progressed developed new metastatic lesions within 3 months of diagnosis (Fig 2, A). The baseline sizes of these 3 tumors were 2.8, 7.1, and 11 cm. Large primary tumor size is associated with increased disease progression.6,20 Therefore, the aggressive nature of NF-PNETs tumors may account for the short interval to the development of liver metastases; however, it is likely that those liver lesions were present at the time of diagnosis, but were below the sensitivity of detection on imaging studies. Different outcomes for small and large neoplasms may be observed, because NF-PNETs exist on a spectrum ranging from indolent neoplasms to those which behave more aggressively. Size may identify those tumors more likely to progress; small tumors have better outcomes.6,20 Small NF-PNETs may, therefore, represent a different biologic entity less likely to progress, metastasize, or result in
mortality. Accordingly, in our population, operative intervention did not yield improved outcomes in the smaller NF-PNETs. Conversely, patients with larger neoplasms may represent a more aggressive biology. NF-PNETs < 2 cm can be monitored for growth by CT or MRI at 6- to 12-month intervals. Any observation of growth suggests the neoplasm is of a more aggressive type and should be considered for resection. In addition, the Ki-67 proliferative index has been associated with disease progression in PNETs.21 Among the metastatic NF-PNETs in this study, 3 tumors had a Ki-67 index of 3–20%, 1 had a 0% Ki-67 index, and 1 did not have Ki-67 results. Of note, many stable tumors had a Ki-67 index of 3–20%. Ki-67 index was helpful in grading NF-PNETs to determine treatment options, but did not predict tumor aggressiveness in our sample. It is possible that the Ki-67 index did not correlate with tumor behavior because many patients in the nonoperative group had missing data and the population of studied patients is small.
ARTICLE IN PRESS 8 Rosenberg et al
There are limitations to this study. A sample size of 35 patients is small, limits the power of the study, and descriptive statistics are used to characterize the population. These data are retrospective and from a single institution, which may not be representative of all patient populations given its relative diversity. In addition, tumor grade could not be verified for several patients. Last, patients from a wide time frame were included in the study, and management has changed over time to our current algorithm where small NF-PNETs are observed and large NF-PNETs are managed operatively. Time frame analysis showed that prior to 2009, 50% of patients were managed operatively compared with 68% from 2009 to 2014. In conclusion, small low-grade NF-PNETs (<2 cm) can be managed nonoperatively without an increased risk of disease progression. No mortality, metastasis, or tumor growth/recurrence occurred in patients with small, low-grade NF PNETs in both the operative and nonoperative groups. An expectant approach may therefore, be used to treat patients with NF-PNETs < 2 cm. Patients with surgically managed NF-PNETs $ 2 cm may be less likely to develop metastases. Therefore, NF-PNETs measured initially as $ 2 cm or tumors that increase in size during follow-up should be considered for resection. Additional prospective studies are needed to determine the benefits and risks of operative versus nonoperative treatment of NF-PNETs. Therefore, a larger multicenter, randomized clinical trial is needed to address this controversy. REFERENCES 1. Dumlu EG, Karakoc D, Ozdemir A. Nonfunctional pancreatic neuroendocrine tumors: advances in diagnosis, management and controversies. Int Surg 2015;100:1089-97. 2. Yao JC, Hassan M, Phan A, Dagohoy C, Leary C, Mares JE, et al. One hundred years after ‘‘carcinoid’’: epidemiology of and prognostic factors for neuroendocrine tumors in 35,825 cases in the United States. J Clin Oncol 2008;26: 3063-72. 3. Metz DC, Jensen RT. Gastrointestinal neuroendocrine tumors: pancreatic endocrine tumors. Gastroenterology 2008;135:1469-92. 4. Oberg K, Eriksson B. Endocrine tumours of the pancreas. Best Pract Res Clin Gastroenterol 2005;19:753-81. 5. Lee LC, Grant CS, Salomao DR, Fletcher JG, Takahashi N, Fidler JL, et al. Small, nonfunctioning, asymptomatic pancreatic neuroendocrine tumors (PNETs): role for nonoperative management. Surgery 2012;152:965-74. 6. Kishi Y, Shimada K, Nara S, Esaki M, Hiraoka N, Kosuge T. Basing treatment strategy for non-functional pancreatic neuroendocrine tumors on tumor size. Ann Surg Oncol 2014;21:2882-8. 7. Kondo NI, Ikeda Y. Practical management and treatment of pancreatic neuroendocrine tumors. Gland Surg 2014; 3:276-83.
Surgery j 2015
8. Haynes AB, Deshpande V, Ingkakul T, Vagefi PA, Szymonifka J, Thayer SP, et al. Implications of incidentally discovered, nonfunctioning pancreatic endocrine tumors: short-term and long-term patient outcomes. Arch Surg 2011;146:534-8. 9. Sharpe SM, In H, Winchester DJ, Talamonti MS, Baker MS. Surgical resection provides an overall survival benefit for patients with small pancreatic neuroendocrine tumors. J Gastrointest Surg 2015;19:117-23; discussion 123. 10. Gaujoux S, Partelli S, Maire F, D’Onofrio M, Larroque B, Tamburrino D, et al. Observational study of natural history of small sporadic nonfunctioning pancreatic neuroendocrine tumors. J Clin Endocrinol Metab 2013;98:4784-9. 11. Birnbaum DJ, Turrini O, Vigano L, Russolillo N, Autret A, Moutardier V, et al. Surgical management of advanced pancreatic neuroendocrine tumors: short-term and longterm results from an international multi-institutional study. Ann Surg Oncol 2015;22:1000-7. 12. Kazanjian KK, Reber HA, Hines OJ. Resection of pancreatic neuroendocrine tumors: results of 70 cases. Arch Surg 2006; 141:765-9; discussion 769-70. 13. Akerstrom G, Hellman P. Surgery on neuroendocrine tumours. Best Pract Res Clin Endocrinol Metab 2007;21: 87-109. 14. Technology EHI. Clinical Looking Glass. Yonkers, NY 2009. p. A quality improvement health care surveillance software. 15. Eisenhauer EA, Therasse P, Bogaerts J, Schwartz LH, Sargent D, Ford R, et al. New response evaluation criteria in solid tumours: revised RECIST guideline (version 1.1). Eur J Cancer 2009;45:228-47. 16. Dindo D, Demartines N, Clavien PA. Classification of surgical complications: a new proposal with evaluation in a cohort of 6336 patients and results of a survey. Ann Surg 2004;240:205-13. 17. Libutti SK. Evolving paradigm for managing small nonfunctional incidentally discovered pancreatic neuroendocrine tumors. J Clin Endocrinol Metab 2013;98:4670-2. 18. Libutti SK, Inabnet WB 3rd. Force or stratagem? Surgery 2012;152:975-6. 19. Cherenfant J, Stocker SJ, Gage MK, Du H, Thurow TA, Odeleye M, et al. Predicting aggressive behavior in nonfunctioning pancreatic neuroendocrine tumors. Surgery 2013; 154:785-91; discussion 91-3. 20. Bettini R, Partelli S, Boninsegna L, Capelli P, Crippa S, Pederzoli P, et al. Tumor size correlates with malignancy in nonfunctioning pancreatic endocrine tumor. Surgery 2011;150:75-82. 21. Franchi G, Manzoni MF. Cytological Ki-67 in pancreatic endocrine tumors: a new ‘‘must’’? Gland Surg 2014;3: 219-21.
DISCUSSION Dr Marybeth Hughes (Bethesda, MD): Were these all sporadic and not Von Hippel-Lindau– type patients? Did you define tumor growth for when you decided to operate on these patients? Dr Alex M. Rosenberg (Bronx, NY): To address your first question, we only included sporadic tumors in our sample. Patients who were identified to have MEN1 or Von Hippel-Lindau were excluded from the sample. Your second question regarding what was considered significant tumor growth, we used the RESIST criteria, which would
ARTICLE IN PRESS Surgery Volume j, Number j
require 20% increase from baseline size in addition to a 5-mm increase in size from baseline. Dr Clive S. Grant (Colorado Springs, CO): As you know, we published something on this just a couple years ago and came to the same conclusion, that small tumors are reasonably observed. The question I have is, how is that to be done both before you decide not to operate and then the follow-up and the frequency of how you do that? What we had determined was that they needed to be #2 cm and with no invasion seen by the imaging. The third thing that we considered and has been since reported on is the ability to use endoscopic ultrasonography (EUS) for a core needle biopsy to determine ki-67 or a substitute mitotic count, then, of course, the follow-up. Would you recommend getting biopsies of these patients to learn ki-67 or mitotic count, or do you think that’s not critical? Dr Alex M. Rosenberg: Our current practice includes at the data first contact, we do crosssectional imaging of these tumors. We also do a fine-needle aspiration, and, when appropriate, we do an EUS to confirm the tumor size, location. You mentioned about interval follow-up. We do repeat follow-up at 6 months from baseline. If there is a change in size that is deemed significant, then we will consider resection. If there is an elevation in tumor markers, we will consider decreasing our interval follow-up time from 6 months to 3 months. If there is no significant change, we
Rosenberg et al 9
observe the following 6 months. If nothing changes, we then go to 1-year interval follow-up. Dr John A. Chabot (New York, NY): Nothing to disclose. Can you please clarify, I thought I heard you say you included both low- and moderategrade tumors in this series. Please tell us more about any moderate-grade tumors that are included here and what the outcome was there, because I think the vast majority of the world limits this to low-grade tumors. Dr Alex M. Rosenberg: We looked at low- and intermediate-grade tumors. For more information, there were 5 patients who experienced metastatic growth and disease progression. Three of those patients had intermediate grade, 1 had low grade, and 1 had data unavailable. Dr. John A. Chabot (New York, NY): Those are all in the resected group? Dr Alex M. Rosenberg: Across both groups. Dr Menno Vriens (Utrecht, The Netherlands): Nothing to disclose. Because those literature are there that says that somatostatin analogs have influence on growth, do you think there is a role for somatostatin analogs to use them in the sub2-centimeter lesions to remain a nonsurgical strategy? Dr Alex M. Rosenberg: Our current practice does not include using somatostatin analogs to treat these tumors. We simply will observe with serial imaging, or we will do surgery if they grow larger in size.
Resection versus expectant management of small incidentally discovered nonfunctional pancreatic neuroendocrine tumors Alex M. Rosenberg, BA,a Patricia Friedmann, MS,a Jaydira Del Rivero, MD,a,b Steven K. Libutti, MD,a,b and Amanda M. Laird, MD,a,b Bronx, NY
Background. Sporadic, nonfunctional pancreatic neuroendocrine tumors (NF-PNETs) are diagnosed with increasing frequency. We compared the risk of tumor growth, metastasis, and mortality between patients treated versus those treated expectantly. Method. A retrospective study of patients seen at our institution with sporadic NF-PNETs, with $12 months of follow-up. Kaplan-Meier analysis was performed. Results. Between 1999 and 2014, 35 patients with an incidentally discovered nonfunctional PNET were identified. Twenty underwent resection and 15 were followed with imaging. In the operative group, 8 had NF-PNETs < 2 cm, while 12 had NF-PNETs $ 2 cm. In the nonoperative expectant management by serial imaging group, 10 had NF-PNETs < 2 cm while 5 had NF-PNETs $ 2 cm. Small NF-PNETs (<2 cm) in either the operative or nonoperative groups demonstrated no evidence of progression or metastasis (median follow-up of 27.8 months). Morbidity in the operative group was 35% with pancreatic pseudocyst the most common. Conclusion. Incidentally discovered NF-PNETs < 2 cm in size can be observed safely with serial imaging. (Surgery 2015;j:j-j.) From the Albert Einstein College of Medicinea and Montefiore Medical Center,b Bronx, NY
PANCREATIC NEUROENDOCRINE TUMORS (PNETs) are rare neoplasms of islet cells in the pancreas. The reported incidence is 1–5/100,000.1,2 Detection is increasing owing to the widespread use of axial imaging.2 Presentation of these lesions depends on whether they are functional or nonfunctional (NF). Functional PNETs can secrete hormones such as insulin, glucagon, gastrin, vasoactive intestinal polypeptide, or somatostatin.3 In contrast, NF-PNETs do not cause hormonal imbalance and are typically asymptomatic. Diagnosis is most often incidental; the majority of NF-PNETs are discovered as a result of routine imaging or in the evaluation of nonspecific abdominal pain.3,4 Although
Presented at the 36th Annual Meeting of the American Association of Endocrine Surgeons, May 2015, Nashville, Tennessee. Accepted for publication October 14, 2015. Reprint requests: Dr Amanda M. Laird, MD, Montefiore Medical Center, 3400 Bainbridge Avenue, Bronx, NY 10461. E-mail: [email protected]. 0039-6060/$ - see front matter Ó 2015 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.surg.2015.10.013
resection is recommended generally for functional lesions, controversy regarding management of small NF-PNETs persists.5,6 Management of NFPNETs with resection has demonstrated a positive impact on mortality7; however, the benefit of resecting small NF-PNETs is limited, and management guidelines are mixed. Small tumors (<2 cm) have a more indolent course, and many studies suggest these NF-PNETs may be monitored safely through imaging.7 Management guidelines regarding NF-PNETs vary from institution to institution. Data from the Massachusetts General Hospital recommend resection of all NF-PNETs regardless of their size; a study by Haynes et al8 of 139 patients, all managed by surgical resection, demonstrated improved outcomes regardless of tumor size. These data suggested that NF-PNETs of all sizes should be resected to minimize disease progression; however, a nonoperatively managed control group was not used for comparison. In addition, data from the National Cancer Database demonstrated that patients small NF-PNETs managed by resection have an overall lesser mortality rate.9 Other data contradict these findings. Data from the Mayo Clinic SURGERY 1
ARTICLE IN PRESS 2 Rosenberg et al
support observation for small NF-PNETs (<2 cm), because these neoplasms have limited growth; observation had no impact on survival or development of metastases.5 Furthermore, these conclusions were based on outcomes from both an operative group and a nonoperative comparison group. Other groups have made similar findings.10 Patient selection is a critical issue in many of the published studies. The foregoing studies were retrospective, which resulted in selection bias in treatment groups. Randomizing intervention among patients with NF-PNETs would remove selection bias, but because NF-PNETs are rare, recruitment at a single institution for a randomized controlled trial is difficult. The morbidity rate for pancreatic resection of these lesions can be as great as 40%, which includes pancreatic leak and abscess formation.11-13 It is unclear whether all NF-PNETs need to be managed by resection, given the potential for postoperative complications. The present study sought to address the management of these lesions by evaluating outcomes with respect to distant metastases, tumor growth or recurrence, and overall rates of mortality, progression-free survival (PFS), and disease-free survival (DFS) in patients with NF-PNETs < 2 cm managed either operatively or nonoperatively by expectant management. Our secondary aim was to assess operative-related morbidity. We evaluated the extent to which smaller NF-PNETs (<2 cm) can be managed nonoperatively. We further hypothesized that larger NF-PNETs ($2 cm) are biologically distinct and that, as a result, outcomes may be improved with operative management. METHODS Patients who were diagnosed with pancreatic islet cell tumors at a single academic institution from January 1999 to November 2014 were identified from preexisting institutional registries using the Clinical Looking Glass, a health care, quality improvement software at Montefiore Medical Center.14 Patients were followed for $12 months. NFPNETs were selected based on pathology reports, computed tomography (CT), and MRI. Patients were divided into 2 cohorts based on whether they were managed operatively or observed. Patients deemed unresectable were included in the observational cohort. Within each cohort, patients were grouped based on a primary tumor size of >2 or <2 cm. Records were reviewed to catalog data including patient demographics, method of diagnosis, pathologic findings, Ki-67 index, imaging findings, operative interventions, follow-up imaging, tumor
Surgery j 2015
recurrence for those patients managed operatively, tumor growth for those patients managed nonoperatively, postoperative complications, and death. Tumor progression was determined using the Response Evaluation Criteria in Solid Tumors. Tumor growth was considered clinically relevant if the lesion demonstrated an increase in size of $20% from baseline and a 5-mm absolute increase in size from baseline.15 Postoperative complications were graded using the Clavien-Dindo classification system.16 Diagnosis of NF-PNET was based on classic and well-established CT or MRI findings or cytopathology; nonfunctionality was established by history of symptoms and laboratory analysis. Patients managed nonoperatively were followed in an outpatient setting. Every 6 months, these patients were evaluated using imaging studies including CT or MRI. Exclusion criteria included (1) pathologic signs of high grade features, including a Ki-67 of $20% index, (2) history of a familial syndrome associated with PNETs, including multiple endocrine neoplasia or Von Hippel Lindau, (3) existing metastases at the time of diagnosis, (4) patients #17 years, (5) or patients with <12 months of follow-up. Patient clinical characteristics, including survival and disease progression data, were cataloged from patient medical records. Survival in the nonoperative and operative groups was calculated from date of diagnosis to date of death or to date of last contact for patients who were still alive at the time of their last follow-up. Patients who were alive at the date of last contact were censored at this date. Outcomes including metastases, tumor growth/recurrence, overall mortality, and PFS for nonoperatively managed patients and DFS for operatively managed patients were compared. Time-to-event data were described using the Kaplan-Meier method. Survival curves were compared using the log-rank test. Categorical variables were analyzed using the Chi-square and Fisher’ exact tests. Continuous variables were analyzed using the non-parametric Wilcoxon rank-sum test. The Institutional Review Board at Albert Einstein College of Medicine approved this study. RESULTS Overall, 35 patients met the inclusion criteria. Of those, 20 underwent operative resection and 15 were managed nonoperatively. Median duration of follow-up for all patients was 30 months (range, 12–143), with 34 months (range, 12–143) for operatively managed patients and 28 months (range, 19–113) for nonoperatively managed
ARTICLE IN PRESS Rosenberg et al 3
Surgery Volume j, Number j
Table I. Patient and tumor characteristics Nonoperative group (n = 15) Median age, y (range) Sex, n (%) Male Female Ethnicity, n (%) African American Latino or Hispanic Caucasian Asian Multiracial Other/unknown Clinical status Neoplasm baseline size, cm (range) Neoplasms of baseline size < 2 cm, n (%) Neoplasms of baseline size $ 2 cm, n (%) Neoplasm end size, cm (range) Follow-up time, months (range) Tumor grade Ki-67 < 2%, n (%) Ki-67 3–20%, n (%) No Ki-67 data available, n (%) Location of neoplasm, n (%) Head Tail Neck Body Uncinate process Indeterminate Indication for abdominal imaging, n (%) Abdominal pain Other cancer surveillance Other disease surveillance Back Pain Other Unspecified
61 (31–70) 8 (53) 7 (47)
Operative group (n = 20) 61 (44–77)
P value .845* .42*
10 (50) 10 (50) .775y
7 3 2 1 2 0 1.4 10 5 1.6 28
(47) (20) (13) (7) (13) (0) (0.8–11.0) (67) (33) (0.8–12) (19–113)
0 (0) 4 (27) 11 (73)
7 6 3 0 2 2
(35) (30) (15) (0) (10) (10)
2.3 (0.5–13.5) 8 (40) 12 (60) N/A 34 (12–143)
.113* .118* .118* z .842*
13 (65) 6 (30) 1 (5) .428y
4 (27) 5 (33) 2 (13) 1 2 (13) 1
3 6 3 5 0 3
(15) (30) (15) (25) (0) (15)
6 (40) 3 (20) 3 (20) 0 1 2 (13.3)
8 (40) 1 5 (25) 1 1 4 (20)
.842y
*Chi-square test. yFisher’s exact test. zInsufficient data for comparison. N/A, Not applicable.
patients. Median age was 61 years (range, 31–70) in the nonoperative group and 61 years (range, 44–77) in the operative group. One-half of the operatively managed patients and 47% of nonoperatively managed patients were women. Our patient pool was diverse racially. Overall, 14 patients were African American (40%), 9 were Hispanic (26%), 5 were Caucasian (14%), 2 were multiracial (11%), 1 was Asian (3%), and 2 did not identify as any race (6%). Neoplasms were located most commonly in the pancreatic tail (30%) and body (25%) in the operative group. NF-PNETs managed nonoperatively were found most frequently in the tail (33%) and head of the
pancreas (27%). Abdominal pain was the most common indication for abdominal imaging for patients in the operative group (40%) and the nonoperative group (40%). Treatment groups were comparable with respect to age, sex, ethnicity, location of neoplasm, indication for abdominal imaging, and NF-PNET baseline and end sizes (Table I). Operative group. Median size of tumors managed operatively was 2.3 cm based on imaging (range, 0.5–13.5). Seven patients were African American (35%), 6 were Hispanic (30%), 3 were Caucasian (15%), 2 were multiracial (10%), and 2 did not identify as any race (10%). Thirteen
ARTICLE IN PRESS 4 Rosenberg et al
Surgery j 2015
Table II. Outcomes for operative and nonoperative patients Log-Rank Nonoperative group
Operative group
n = 15 0 3 (20) 0 12 (80) N/A 10 0 0 0 10 (100) N/A 5 0 3 (60) 0 2 (40) N/A
n = 20 1 2 (10) 2 (10) 17 (85) 7 (35) 8 0 0 0 8 (100) 1 (13) 12 1 2 (17) 2 (17) 9 (75) 6 (50)
A. Outcomes for all neoplasms Tumor Growth/Recurrence, n (%) Metastases, n (%) Mortality, n (%) Progression/disease-free survivalz, n (%) Operative morbidity, n (%) B. Outcomes for neoplasms < 2 cm, n Tumor growth/recurrence, n (%) Metastases, n (%) Mortality, n (%) Progression/disease-free survivalz, n (%) Surgical morbidity, n (%) C. Outcomes for neoplasms $ 2 cm, n Tumor growth/recurrence, n (%) Metastases, n (%) Mortality, n (%) Progression/disease-free survivalz, n (%) Operative morbidity, n (%)
P-value* .4497 .3891 .4292 .5377 y y y y y y x x x x y
*P values were measured using Kaplan-Meier method; hazard ratios were measured using Cox-regression method for small samples. yInsufficient data for comparison. zProgression-free survival included tumor growth, metastases, and mortality as events for nonoperative patients; disease-free survival included tumor recurrence, metastases, and mortality as event for operative patients. xSample size is too small for statistical analysis. CI, Confidence interval; N/A, not applicable.
patients (65%) had a Ki-67 proliferative index of <2% and 6 patients (30%) had a Ki-67 index of 3– 20%. No Ki-67 data were available for 1 patient. In the operative cohort, 8 patients (40%) had NFPNETs < 2 cm, and 12 had NF-PNETs > 2 cm (60%). One resected tumor recurred, and 2 patients (10%) developed metastases, all of which were liver metastases. Additionally, an overall mortality rate of 10% and a DFS rate of 85% were observed for all NF-PNETs in the operative group (Table II). The operative complication rate for all NF-PNETs was 35% with pancreatic pseudocyst formation being the most common morbidity. Four patients (20%) developed grade I complications, which required no intervention. Three patients developed grade III complications, which required placement of drains. Distal pancreatectomy was the most commonly performed operation (Table III). Nonoperative group. NF-PNETs managed nonoperatively had a median baseline size of 1.4 cm (range, 0.8–11.0) on CT or MRI. No tumors grew in size, and no patients were moved to the resection group after date of diagnosis. Seven patients were African American (47%), 3 were Hispanic (20%), 2 were Caucasian (13%), 2 were multiracial (13%), and 1 was Asian. No patient had
a proliferative index of <2%, 4 patients (27%) had a Ki-67 index of 3–20%, and no Ki-67 data were available for 11 patients (73%). Ten patients (68%) had NF-PNETs < 2 cm, and 5 (33%) had NF-PNETs $ 2 cm (Table I). Five patients with tumors > 2 cm were not managed operatively owing to local invasion of the superior mesenteric artery. The median tumor size was 1.6 cm (range, 0.8– 12.2). No patients had tumor growth and 3 patients (20%) developed metastases, all in the liver. No patients in the observation group died. Patients in this treatment group had a PFS rate of 80% (Table II). Operative versus nonoperative. No difference in metastases was observed after 30 months (range, 12–143) of follow-up between the operative and nonoperative groups (P = .3891; Fig 1, A). Furthermore, however, no difference in overall survival rates between nonoperative (100%) and operative groups (90%; P = .4292; Fig 1, B). PFS rates in the nonoperative group (80%) and DFS rates in the operative group (85%) demonstrated no difference (P = .5337; Fig 1, C). In addition, no difference in tumor growth/recurrence was observed (P = .4497; Table II). The median tumor sizes of the operative and nonoperative groups were 2.3 and 1.4 cm, respectively. Because differences
ARTICLE IN PRESS Rosenberg et al 5
Surgery Volume j, Number j
Table III. Types of operation and related morbidities Operative group (n = 20) A. Type of operation, n (%) Distal pancreatectomy Pancreatoduodenectomy Distal pancreatectomy and splenectomy Subtotal Pancreatectomy Central Pancreatectomy Enucleation Unspecified B. Morbidity, n (%) Pancreatic pseudocyst Hematoma Pneumobilia Pancreatitis Any morbidity C. Clavien-Dindo postoperative complication,* n (%) Grade 1 Grade II Grade III Grade IV Grade V (death)
8 (40) 5 (25) 3 (15) 1 1 1 1 4 (25) 1 1 1 7 (35)
4 (20) 0 3 (15) 0 0
*Grade I: Any deviation from the normal postoperative course without the need for pharmacologic treatment or operative, endoscopic, and radiological interventions. Allowed therapeutic regimens are: drugs as antiemetics, antipyretics, analgesics, diuretics, electrolytes, and physiotherapy. This grade also includes wound infections opened at the bedside. Grade II: Requiring pharmacologic treatment with drugs other than such allowed for grade I complications. Blood transfusions and total parenteral nutrition are also included. Grade III: Requiring surgical, endoscopic or radiological intervention. Grade IV: Life-threatening complication requiring ICU management. Grade V: Death of a patient.
in baseline tumor size represent a confounding variable, both groups were divided into the following subgroups: small NF-PNETs (<2 cm) and large NF-PNETs ($2 cm). Comparison of outcomes was then made between similarly sized NFPNETs. Small versus large PNETs. Ten patients in the nonoperative group and 8 in the operative group presented with large NF-PNETs (<2 cm), representing 66% and 40% of their respective groups. There was no interval development of metastases, no tumor growth, and no mortality in patients with small NF-PNETs managed nonoperatively after 27 months of follow-up (range, 16–50). Likewise, in patients with small NF-PNETs managed operatively, no metastases, tumor recurrence, or mortality was observed during that time. (Table II) Five patients in the nonoperative group and 12 patients in the operative group presented with large NF-PNETs ($2 cm), representing 33% and
60% of each respective group (Table I). In patients with large NF-PNETs who were managed nonoperatively, no change was observed in the median primary tumor size over the course of 37 months of follow-up (range, 12–143). In contrast, 60% of nonoperatively managed patients (3 patients) had new metastases identified during that time. In comparison, in patients with large NF-PNETs that were managed operatively, 17% of the cohort (2 patients) had metastases, and 1 patient had a recurrence identified during that time. All metastases developed in the liver. In the subgroup of patients with large PNETs who were managed nonoperatively, there were no deaths recorded over the study period. In patients with large NFPNETs who were managed operatively, 2 deaths (17%) occurred. PFS was 40% and DFS was 75%, respectively, for nonoperatively and operatively managed large NF-PNET groups (Table II; Fig 2, A–C). DISCUSSION The optimal management of small, incidentally discovered NF-PNETs is currently a topic of debate in the literature.17,18 Our results demonstrate different outcomes for low-grade NF-PNETs <2 cm when compared with low-grade NF-PNETs $2 cm. Patients with small NF-PNETs managed nonoperatively have similar outcomes whether treated operatively by resection or nonoperatively. Small tumors (<2 cm) in both the nonoperative and operative subgroups did not develop metastases, did not grow or recur, and no deaths occurred over 28 months of follow-up. The only adverse outcome that patients with these small tumors encountered in their management was morbidity related to the operation. Small NF-PNETs in our population seem to be indolent, which is consistent with other studies.6 These data suggest that nonoperative management results in similar outcomes compared with those patients with small tumors managed operatively, while eliminating the risk of operatively related morbidity. The overall operative complication rate in this group was 35% (7/20), with 3 patients who required postoperative intervention. The rate of pseudocyst formation was greater than expected, which may be as a result of a subclinical pancreatic leak. These findings are similar to published averages of 40%. Patients may, therefore, avoid potential morbidity if the selective approach of observation of small NF-PNETs is adopted rather than operating on all patients with NF-PNETs, regardless of size. Therefore, we advocate an expectant approach, wherein small ($2 cm)
ARTICLE IN PRESS 6 Rosenberg et al
Surgery j 2015
Fig 1. (A) Time to metastasis for all patients, by treatment. (B) Overall survival for all patients, by treatment. (C) Progression/disease-free survival for all patients, by treatment. N, Number of patients at risk in in nonoperative group; S, number of patients at risk in the operative (surgical) group. *Progression-free survival included tumor growth, metastases, and mortality as events for nonoperative patients; disease-free survival included tumor recurrence, metastases, and mortality as event for operative patients.
NF-PNETs tumors are monitored to detect any change in growth to determine if and when operation becomes necessary. These data add to the growing discussion regarding management of small NF-PNETs. Results observed in our study were similar to those reported in both Lee et al5 and Gaujoux et al.10 These studies demonstrated comparable outcomes for patients managed nonoperatively and operatively with no signs of disease progression or mortality in patients with small NF-PNETs.5,10 These results support the argument that small NFPNETs can be managed safely with observation and serial imaging. In these studies, the only adverse outcome patients with small NF-PNETs experienced was as a result of operative resection.5,10 In comparison, a series of operatively managed patients demonstrated tumor-related mortality rates in well-differentiated tumors of 7.1% for tumors identified as benign and 14.0% for tumors identified as neither benign nor malignant.8 It is difficult to determine the clinical
importance of these results, because a nonoperative group was not used in the study. Patients managed nonoperatively could have demonstrated equally positive outcomes. Although operative management can be effective in managing small NF-PNETs, the outcome of an expectant approach may equivalent. Large NF-PNETs tended to recur, developed liver metastases, and mortality was greater. Although large tumors are typically managed by resection, 5 large NF-PNETs in this study were managed nonoperatively, because the primary tumor was technically unresectable owing to local invasion. These tumors were included in the nonoperative group, because they contribute to the understanding of the natural history of this heterogeneous cohort. Three of these 5 nonoperatively managed patients developed metastases, whereas 2 of the 12 operatively managed patients developed metastases, which is associated with disease-related mortality.19 These data suggest that operative management of NF-PNETs > 2 cm
ARTICLE IN PRESS Surgery Volume j, Number j
Rosenberg et al 7
Fig 2. (A) Time to metastasis for patients with nonfunctional pancreatic neuroendocrine tumors (NF-PNETs) $ 2 cm, by treatment. (B) Overall survival for patients with NF-PNETs $ 2 cm, by treatment. (C) Progression/disease-free survival for patients with NF-PNETs $ 2 cm, by treatment. N, Number of patients at risk in the nonoperative group; S, number of patients at risk in the operative (surgical) group. *Progression-free survival included tumor growth, metastases, and mortality as events for nonoperative patients; disease-free survival included tumor recurrence, metastases, and mortality as event for operative patients.
may be associated with improved outcomes. Therefore, NF-PNETs > 2 cm should be considered for resection if feasible. The 3 NF-PNETs managed operatively that progressed developed new metastatic lesions within 3 months of diagnosis (Fig 2, A). The baseline sizes of these 3 tumors were 2.8, 7.1, and 11 cm. Large primary tumor size is associated with increased disease progression.6,20 Therefore, the aggressive nature of NF-PNETs tumors may account for the short interval to the development of liver metastases; however, it is likely that those liver lesions were present at the time of diagnosis, but were below the sensitivity of detection on imaging studies. Different outcomes for small and large neoplasms may be observed, because NF-PNETs exist on a spectrum ranging from indolent neoplasms to those which behave more aggressively. Size may identify those tumors more likely to progress; small tumors have better outcomes.6,20 Small NF-PNETs may, therefore, represent a different biologic entity less likely to progress, metastasize, or result in
mortality. Accordingly, in our population, operative intervention did not yield improved outcomes in the smaller NF-PNETs. Conversely, patients with larger neoplasms may represent a more aggressive biology. NF-PNETs < 2 cm can be monitored for growth by CT or MRI at 6- to 12-month intervals. Any observation of growth suggests the neoplasm is of a more aggressive type and should be considered for resection. In addition, the Ki-67 proliferative index has been associated with disease progression in PNETs.21 Among the metastatic NF-PNETs in this study, 3 tumors had a Ki-67 index of 3–20%, 1 had a 0% Ki-67 index, and 1 did not have Ki-67 results. Of note, many stable tumors had a Ki-67 index of 3–20%. Ki-67 index was helpful in grading NF-PNETs to determine treatment options, but did not predict tumor aggressiveness in our sample. It is possible that the Ki-67 index did not correlate with tumor behavior because many patients in the nonoperative group had missing data and the population of studied patients is small.
ARTICLE IN PRESS 8 Rosenberg et al
There are limitations to this study. A sample size of 35 patients is small, limits the power of the study, and descriptive statistics are used to characterize the population. These data are retrospective and from a single institution, which may not be representative of all patient populations given its relative diversity. In addition, tumor grade could not be verified for several patients. Last, patients from a wide time frame were included in the study, and management has changed over time to our current algorithm where small NF-PNETs are observed and large NF-PNETs are managed operatively. Time frame analysis showed that prior to 2009, 50% of patients were managed operatively compared with 68% from 2009 to 2014. In conclusion, small low-grade NF-PNETs (<2 cm) can be managed nonoperatively without an increased risk of disease progression. No mortality, metastasis, or tumor growth/recurrence occurred in patients with small, low-grade NF PNETs in both the operative and nonoperative groups. An expectant approach may therefore, be used to treat patients with NF-PNETs < 2 cm. Patients with surgically managed NF-PNETs $ 2 cm may be less likely to develop metastases. Therefore, NF-PNETs measured initially as $ 2 cm or tumors that increase in size during follow-up should be considered for resection. Additional prospective studies are needed to determine the benefits and risks of operative versus nonoperative treatment of NF-PNETs. Therefore, a larger multicenter, randomized clinical trial is needed to address this controversy. REFERENCES 1. Dumlu EG, Karakoc D, Ozdemir A. Nonfunctional pancreatic neuroendocrine tumors: advances in diagnosis, management and controversies. Int Surg 2015;100:1089-97. 2. Yao JC, Hassan M, Phan A, Dagohoy C, Leary C, Mares JE, et al. One hundred years after ‘‘carcinoid’’: epidemiology of and prognostic factors for neuroendocrine tumors in 35,825 cases in the United States. J Clin Oncol 2008;26: 3063-72. 3. Metz DC, Jensen RT. Gastrointestinal neuroendocrine tumors: pancreatic endocrine tumors. Gastroenterology 2008;135:1469-92. 4. Oberg K, Eriksson B. Endocrine tumours of the pancreas. Best Pract Res Clin Gastroenterol 2005;19:753-81. 5. Lee LC, Grant CS, Salomao DR, Fletcher JG, Takahashi N, Fidler JL, et al. Small, nonfunctioning, asymptomatic pancreatic neuroendocrine tumors (PNETs): role for nonoperative management. Surgery 2012;152:965-74. 6. Kishi Y, Shimada K, Nara S, Esaki M, Hiraoka N, Kosuge T. Basing treatment strategy for non-functional pancreatic neuroendocrine tumors on tumor size. Ann Surg Oncol 2014;21:2882-8. 7. Kondo NI, Ikeda Y. Practical management and treatment of pancreatic neuroendocrine tumors. Gland Surg 2014; 3:276-83.
Surgery j 2015
8. Haynes AB, Deshpande V, Ingkakul T, Vagefi PA, Szymonifka J, Thayer SP, et al. Implications of incidentally discovered, nonfunctioning pancreatic endocrine tumors: short-term and long-term patient outcomes. Arch Surg 2011;146:534-8. 9. Sharpe SM, In H, Winchester DJ, Talamonti MS, Baker MS. Surgical resection provides an overall survival benefit for patients with small pancreatic neuroendocrine tumors. J Gastrointest Surg 2015;19:117-23; discussion 123. 10. Gaujoux S, Partelli S, Maire F, D’Onofrio M, Larroque B, Tamburrino D, et al. Observational study of natural history of small sporadic nonfunctioning pancreatic neuroendocrine tumors. J Clin Endocrinol Metab 2013;98:4784-9. 11. Birnbaum DJ, Turrini O, Vigano L, Russolillo N, Autret A, Moutardier V, et al. Surgical management of advanced pancreatic neuroendocrine tumors: short-term and longterm results from an international multi-institutional study. Ann Surg Oncol 2015;22:1000-7. 12. Kazanjian KK, Reber HA, Hines OJ. Resection of pancreatic neuroendocrine tumors: results of 70 cases. Arch Surg 2006; 141:765-9; discussion 769-70. 13. Akerstrom G, Hellman P. Surgery on neuroendocrine tumours. Best Pract Res Clin Endocrinol Metab 2007;21: 87-109. 14. Technology EHI. Clinical Looking Glass. Yonkers, NY 2009. p. A quality improvement health care surveillance software. 15. Eisenhauer EA, Therasse P, Bogaerts J, Schwartz LH, Sargent D, Ford R, et al. New response evaluation criteria in solid tumours: revised RECIST guideline (version 1.1). Eur J Cancer 2009;45:228-47. 16. Dindo D, Demartines N, Clavien PA. Classification of surgical complications: a new proposal with evaluation in a cohort of 6336 patients and results of a survey. Ann Surg 2004;240:205-13. 17. Libutti SK. Evolving paradigm for managing small nonfunctional incidentally discovered pancreatic neuroendocrine tumors. J Clin Endocrinol Metab 2013;98:4670-2. 18. Libutti SK, Inabnet WB 3rd. Force or stratagem? Surgery 2012;152:975-6. 19. Cherenfant J, Stocker SJ, Gage MK, Du H, Thurow TA, Odeleye M, et al. Predicting aggressive behavior in nonfunctioning pancreatic neuroendocrine tumors. Surgery 2013; 154:785-91; discussion 91-3. 20. Bettini R, Partelli S, Boninsegna L, Capelli P, Crippa S, Pederzoli P, et al. Tumor size correlates with malignancy in nonfunctioning pancreatic endocrine tumor. Surgery 2011;150:75-82. 21. Franchi G, Manzoni MF. Cytological Ki-67 in pancreatic endocrine tumors: a new ‘‘must’’? Gland Surg 2014;3: 219-21.
DISCUSSION Dr Marybeth Hughes (Bethesda, MD): Were these all sporadic and not Von Hippel-Lindau– type patients? Did you define tumor growth for when you decided to operate on these patients? Dr Alex M. Rosenberg (Bronx, NY): To address your first question, we only included sporadic tumors in our sample. Patients who were identified to have MEN1 or Von Hippel-Lindau were excluded from the sample. Your second question regarding what was considered significant tumor growth, we used the RESIST criteria, which would
ARTICLE IN PRESS Surgery Volume j, Number j
require 20% increase from baseline size in addition to a 5-mm increase in size from baseline. Dr Clive S. Grant (Colorado Springs, CO): As you know, we published something on this just a couple years ago and came to the same conclusion, that small tumors are reasonably observed. The question I have is, how is that to be done both before you decide not to operate and then the follow-up and the frequency of how you do that? What we had determined was that they needed to be #2 cm and with no invasion seen by the imaging. The third thing that we considered and has been since reported on is the ability to use endoscopic ultrasonography (EUS) for a core needle biopsy to determine ki-67 or a substitute mitotic count, then, of course, the follow-up. Would you recommend getting biopsies of these patients to learn ki-67 or mitotic count, or do you think that’s not critical? Dr Alex M. Rosenberg: Our current practice includes at the data first contact, we do crosssectional imaging of these tumors. We also do a fine-needle aspiration, and, when appropriate, we do an EUS to confirm the tumor size, location. You mentioned about interval follow-up. We do repeat follow-up at 6 months from baseline. If there is a change in size that is deemed significant, then we will consider resection. If there is an elevation in tumor markers, we will consider decreasing our interval follow-up time from 6 months to 3 months. If there is no significant change, we
Rosenberg et al 9
observe the following 6 months. If nothing changes, we then go to 1-year interval follow-up. Dr John A. Chabot (New York, NY): Nothing to disclose. Can you please clarify, I thought I heard you say you included both low- and moderategrade tumors in this series. Please tell us more about any moderate-grade tumors that are included here and what the outcome was there, because I think the vast majority of the world limits this to low-grade tumors. Dr Alex M. Rosenberg: We looked at low- and intermediate-grade tumors. For more information, there were 5 patients who experienced metastatic growth and disease progression. Three of those patients had intermediate grade, 1 had low grade, and 1 had data unavailable. Dr. John A. Chabot (New York, NY): Those are all in the resected group? Dr Alex M. Rosenberg: Across both groups. Dr Menno Vriens (Utrecht, The Netherlands): Nothing to disclose. Because those literature are there that says that somatostatin analogs have influence on growth, do you think there is a role for somatostatin analogs to use them in the sub2-centimeter lesions to remain a nonsurgical strategy? Dr Alex M. Rosenberg: Our current practice does not include using somatostatin analogs to treat these tumors. We simply will observe with serial imaging, or we will do surgery if they grow larger in size.