Original article Annals of Oncology 6: 86 8, 25 doi:.93/annonc/mdi358 Published online 5 August 25 Endocrine pancreatic tumors: factors correlated with survival P. Tomassetti *, D. Campana, L. Piscitelli, R. Casadei 2, D. Santini 3, F. Nori, A. M. Morselli-Labate, R. Pezzilli & R. Corinaldesi Dipartimento di Medicina Interna e Gastroenterologia, 2 Dipartimento Clinico di Scienze Chirurgiche ed Anestesiologiche and 3 Dipartimento Clinico di Scienze Radiologiche ed Istopatologiche, Università di Bologna, Italy Received 29 April 25; revised 6 June 25; accepted 2 June 25 Introduction Background: The aim of this study was to evaluate the survival rate of patients with endocrine tumors of the pancreas, functioning or non-functioning, associated or not with MEN syndrome. Patients and methods: Eighty-three patients with pancreatic endocrine tumors diagnosed in our department from 978 to 23 were studied. Results: The study included 37 men (44.6%) and 46 women (55.4%). The median age of patients at diagnosis was 55 years (range 9 8 years). Fifty-two patients (62.7%) had non-functioning endocrine tumors, 6 (9.3%) had functioning endocrine tumors and 5 (8.%) had MEN disease with pancreatic involvement. Twenty-seven patients (32.5%) had liver metastases at the time of diagnosis, involvement of the lymph nodes was found in 47 out of 79 patients (59.5%). Forty patients (48.2%) had radical surgery, 2 (24.%) had palliative surgery and 53 were treated medically. The survival rate was significantly related to the presence of metastases, lymph node involvement, and the type of tumor and treatment. Conclusions: Tumor resection, the absence of liver and lymph node metastases, and the presence of MEN syndrome are related to a better survival rate. Radical surgery continues to have a central role in the therapeutic approach to endocrine tumors of the pancreas. Key words: endocrine pancreatic tumor, gastrinoma, multiple endocrine neoplasia, octreotide, surgical procedure Pancreatic endocrine tumors (PETs) are a heterogenous group of rare neoplasms, occurring in fewer than one in people per year [ 3]. Pancreatic adenocarcinomas are much more common than their endocrine counterparts, with a ratio of pancreatic adenocarcinomas to PETs of approximately 25: [4]. These tumors have attracted considerable attention in recent years, both because they are relatively easy to palliate and because they demonstrate the chronic effect of the particular hormone whose level is elevated. In about % 48% of cases, the tumor is not associated with obvious signs or symptoms of hormone hypersecretion; these are called non-functioning tumors [5 7]. Surgery is generally considered the treatment of choice and, for patients with localized disease, it can be curative. However, in patients with metastatic disease, the role of surgery is not clearly established. Moreover, the treatment is particularly controversial in patients affected by MEN- with pancreatic involvement; hence, other therapies should also be considered [8, 9]. Pharmacological therapy, e.g. with proton pump inhibitors (PPIs) and somatostatin analogs, has dramatically improved *Correspondence to: Prof P. Tomassetti, Department of Internal Medicine and Gastroenterology, University of Bologna, Policlinico S.Orsola-Malpighi, Via Massarenti, 9, 438 Bologna, Italy. Tel/Fax: +39-5-636486; E-mail: paola.tomassetti@unibo.it symptom control, and radiolabeled somatostatin analogs offer targeted therapy for metastatic or inoperable disease [9]. The aim of this study was to evaluate the survival rate of patients with endocrine tumors of the pancreas, functioning or non-functioning, as well as associated or not with MEN syndrome. We also evaluated which clinical, radiological and histological variables are associated with the survival rate of these patients. Finally, we evaluated which of the different treatment modalities are associated with the survival rate of patients with pancreatic endocrine tumors. Patients and methods This study included a total of 83 consecutive patients having endocrine tumors of the pancreas, admitted to our unit from 978 to 23. All patients underwent physical examination, laboratory and radiological investigations. To localize and stage the tumor, all patients underwent ultrasonography (US) and computed tomography (CT) and, in addition, some patients also underwent endoscopic ultrasonography (EUS) and magnetic resonance (MR). All had histopathologic confirmation of endocrine tumors. Ki-67 determination was performed only in the last 29 patients studied; it was not performed earlier because it was not available in our department. Endocrine tumors were classified as functioning on the basis of clinical symptoms of hormonal excess associated with increased serum peptide levels. Tumors without clinical symptoms were classified as non-functioning. In particular, in gastrinoma patients, we evaluated basal gastrin levels associated with basal acid output (BAO) and, when necessary to confirm the diagnosis, a secretin test was performed. In patients with insulinomas, the diagnosis was carried out Ó 25 European Society for Medical Oncology
87 using plasma determination of insulin, C-peptide eventually associated with fasting test. The patients enrolled in the study underwent a clinical check-up and abdominal US every 3 months. A CT scan was performed in all patients every 6 months during the first year after diagnosis and every 2 months thereafter. Surgical and medical procedures and survival rates were recorded in the follow-up period. Statistics Mean, median, standard deviation, range and frequencies were used as descriptive statistics. Survival curves were computed by using the Kaplan Meier method, whereas the Mantel Cox model was applied in order to identify factors significantly related to survival; the hazard ratios (HR), together with their respective 95% confidence intervals (CI), were also computed. The Wilcoxon signed ranks test was also applied. Statistical significance was defined as two-tailed P values less than.5. Statistical analyses were performed by means of the SPSS (Version 8. for Windows) statistical package. Results Table. Median size of the endocrine tumors of the pancreas at imaging techniques and at surgery. Data are reported as median and range (in brackets) Type of tumor Size at imaging (cm) Size at surgery (cm) P value Non-functioning 4. (.3.3) 4.5 (. 3.).2 Functioning 5.4 (.4 8.) 3.5 (.8 5.5).276 MEN 3.8 (..) 5.5 (.5.).599 All endocrine tumors 4. (..3) 5. (. 3.).43 There were 37 men (44.6%) and 46 women (55.4%). The median age of patients at diagnosis was 55 years (range 9 8 years) and the median age at their last observation was 6 years (range 29 82 years); the median follow-up period was 3 months (range 3 264 months). Fifty-two patients (62.7%) had non-functioning tumors, 6 patients (9.3%) had functioning tumors (seven gastrinomas, four insulinomas, two VIPomas, two somatostatinomas, one glucagonoma) and 5 patients (8.%) were affected by MEN with pancreatic involvement ( non-functioning tumors, three insulinomas, one gastrinoma). The tumor was localized in the head of the pancreas in 3 patients (37.3%), in the body in 24 patients (28.9%), in the tail in 2 patients (25.3%) and, in seven patients (8.4%), the tumor was diffuse throughout the pancreas. In Table, we report the size of tumors taken as a whole and also the size of each subtype. No significant differences in the size of the tumor were found between radiological and surgical exploration. Twentyseven patients (32.5%) had liver metastases at diagnosis and 43 patients (5.8%) developed metastases during the followup period. Involvement of lymph nodes was found in 47 out of 79 patients (59.5%); in four patients, it was not possible to obtain these data. The median Ki-67 evaluated on 29 histological specimens was 2.9% (range.% 84.%). Forty patients (48.2%) had radical surgery, 2 (24.%) had palliative surgery and 53 (63.8%) were treated medically. Of the latter, 53 underwent medical treatment with somatostatin analogues (63.8%), had chemotherapy (2%), nine chemoembolization (.8%) and six had treatment with alpha interferon (7.2%). None of the patients underwent neoadjuvant therapy. Forty-nine patients (59.%) were still alive at the time of the study; the median survival time was 9 months (95% CI 29 5 months) and the 5-year survival rate was 55.3% (Figure ). Table 2 reports the effect of various clinical, radiological and histological findings on the survival. The survival rate was significantly related to the presence of liver metastases at diagnosis (P <.) (Figure 2), the presence of lymph node involvement (P =.) (Figure 3), the type of tumor (overall P =.33; functioning versus MEN P =.; non-functioning versus 5 5 2 25 3 Figure. Cumulative survival (Kaplan Meier) among 83 patients with pancreatic endocrine tumors. Table 2. Effect of various clinical, radiological and histological findings on the survival in the 83 patients with endocrine tumors HR 95% CI P value Sex (male versus female).6.3.2.5 Patient age (decade).99.75.3.965 Tumor localization.646 Body versus head.5.67 3.38.37 Tail versus head.6.43 2.65.898 Diffuse versus head.72.6 3.25.667 Tumor size at imaging techniques (cm)..94.28.222 Tumor size at surgery (cm)..9.32.325 Presence of liver metastases at diagnosis 5.42 2.63.7 <. Presence of lymph node involvement 4.97.9 2.9. Type of tumor.33 Functioning versus MEN 7.63.64 35.5. Non-functioning versus MEN 4.65.7 2.3.4 Functioning versus non-functioning.228 Type of treatment <. Not operated versus radical surgery 5.2 2.2 2.7 <. Palliative surgery versus radical surgery 4.29.76.48. Not operated versus palliative surgery.626 Treatment with somatostatin analogues.7.56 2.46.672 HR, hazard ratio; 95% CI, 95% confidence interval.
88 MEN P =.4) (Figure 4) and the type of treatment (overall P <.; not operated versus radical surgery: P <.; palliative surgery versus radical surgery P =.) (Figure 5). On the other hand, the survival rate was not significantly related to sex (P =.5), the patient s age (P =.965), the localization of the tumor (P =.646), the size of the tumor both at imaging techniques (P =.222) and at surgery (P =.325), Ki-67 determination (P =.34) and treatment with somatostatin analogues (P =.672). Discussion Liver Metastases No Lymph Node Metastases Lymph Node Metastases No Liver Metastases 5 5 2 25 3 Figure 2. Cumulative survival (Kaplan Meier) among 83 patients with pancreatic endocrine tumors according to the presence or the absence of liver metastases at diagnosis (P <., Mandel Cox). 5 5 2 25 3 Figure 3. Cumulative survival (Kaplan Meier) among 83 patients with pancreatic endocrine tumors according to the presence or the absence of lymph node involvement (P =., Mandel Cox). The main aim of this study was to determine which clinical, radiological and histological variables are associated with survival by examining a large number of patients coming from a single center and having this disease. Regarding gender, we found a slight prevalence of women compared to men; this finding has been recently reported in a Multicenter Italian Study []. The average age at the time of diagnosis was 55 years, which is similar to that reported in previous studies [, ]. In our series, the percentage of patients with non-functioning tumors of the pancreas (62.7%) was higher than that referred to in the literature in which the percentage varies between % Nonfunctioning MEN Functioning 5 5 2 25 3 Figure 4. Cumulative survival (Kaplan Meier) among 83 patients with pancreatic endocrine tumors according to the type of the tumor (nonfunctioning, functioning, associated with MEN syndrome; overall P =.33, Mandel Cox). Radical Surgery No Surgery Palliative Surgery 5 5 2 25 3 Figure 5. Cumulative survival (Kaplan Meier) of 83 patients with pancreatic endocrine tumors following to radical surgery, palliative surgery or no surgery (overall P <., Mandel Cox). and 58% [7, 2, 3]. This difference may be explained by a more accurate diagnosis of pancreatic endocrine tumors according to the WHO classification [4]. The median dimension of the tumors at imaging (4. cm) and at surgery (5. cm) overlaps those of other authors [3, 5]. Specifically, in the present study, the median dimension of functioning pancreatic tumors at imaging (5.4 cm) was superior to that reported by other authors [5 7] but was in agreement with what was recently found by Chu et al. [3]. Furthermore, in our series, we did not find any significant differences in survival between patients with functioning and those with non-functioning endocrine pancreatic tumors. There is some controversy with regard to these data in the literature. Several series reported a lower 5-year survival rate in patients with non-functioning tumors than in those with functioning tumors [7, 5, 8]. On the contrary, other authors found no significant differences in the survival rate between these two groups [3, 6]. These data can be explained by the wide use of PPIs, which have been available to us since the 98s for compassionate use and which reduce the gastric acid hypersecretion due to ZES, thus delaying the diagnosis of a gastrinoma which is already metastatic when first observed [9]. It should always be kept in mind that multiple factors play a role in the evaluation
of the survival rate and, in agreement with recent reports, functioning and non-functioning tumors should be considered equally aggressive with similar prognoses [3, 6, 7]. Patients with MEN syndrome associated with pancreatic involvement showed a better survival rate than patients with sporadic pancreatic tumors. These data are not related to an earlier diagnosis, to the dimension of the tumors at diagnosis or to the type of neoplasia. In fact, according to the literature, in our series endocrine tumors of the pancreas associated with MEN were mainly non-functioning [2]. The higher survival rate in these patients may be explained by the indolent clinical course of this disease [8]. In our study, the 5-year survival rate was 55.3%, which is similar to that reported by other authors (between 35% and 6%) [, 2 23]. In our experience, sex, age of patients and localization of the tumor were not significantly related to survival [, 3]; size was not related, this finding being in contrast with other reports at least concerning non-functioning pancreatic tumors [, 24, 25]. One possible explanation is that we could not find an accurate definition of the tumor size that unequivocally affects patient s survival in the literature. As a rule, 3 4 cm diameter is considered the accepted limit, but there are controversial data [24, 26 28]. Moreover, in a recent study in patients with MEN-, Lowney et al., in agreement with our results, could not find a correlation between the size of the primary tumor and its metastatic potential [29, 3]. In our opinion, this discrepancy may be explained by the heterogeneity of the tumors grouped together as NETs. Due to the long period of time considered, it was not possible to separate well- and poorly-differentiated tumors according to the histopathological classification proposed by WHO in 2. For this reason a careful study of the natural history of each of these neoplasms is required in order to define better the relationship between the features of the tumor and patient survival. Moreover, in our patients, we could not find a significant relationship between survival and the Ki-67 proliferative index determination which is considered by several authors as a reliable prognostic factor [26, 27, 3]. We determined the Ki-67 index only in the last 29 patients studied. The limited number of determinations can affect the relevance of statistical significance. Regarding patients with metastases, survival was also significantly better in patients without metastases at diagnosis compared with those with metastases at diagnosis and this finding is in agreement with other recent reports. At present, metastases, and in particular liver metastases, are considered the factors that are more highly correlated with the survival rate [, 3, 24]. The overall survival rate in patients who underwent radical surgery was significantly better than in those not operated on and in patients who underwent palliative surgery; our data agree with those of other authors [, 7, 2, 24]. In addition, patients who underwent palliative surgery did not show a better survival rate with respect to patients not operated on. Moreover, recent reports suggest that conventional contraindications to surgical resection, such as nodal or distant metastases, the hallmarks of advanced malignancy, should be redefined, giving further support to surgical options [32]. Regarding medical treatment, available data did not show any statistical significance between patients treated with somatostatin analogues and those not treated. However, we have noticed a positive trend concerning survival in treated patients. It is important to stress that medical treatment with somatostatin analogues, available since the 99s, has produced a rapid disappearance of symptoms due to hormonal hypersecretion in patients with functioning endocrine tumors and an improvement in the quality of life in those with nonfunctioning ones. In conclusion, we can suggest that radical surgery continues to have a central role in the therapeutic approach to NETs of the pancreas. The crucial point is an early diagnosis that allows radical surgery when the tumor has not yet metastasized. Of course, our data cannot be proposed as a prognostic model for the treatment of endocrine tumors of the pancreas; we only report our experience on this topic even if we believe that a wider approach to surgical criteria can give better opportunities to patients. Medical treatment has a precise role in functioning endocrine tumors while, as far as non-functioning tumors are concerned, further studies on a larger number of patients are necessary. References 89. Eriksson B, Oberg K. Neuroendocrine tumours of the pancreas. Br J Surg 2; 87: 29 3. 2. Moldow RE, Connelly RR. Epidemiology of pancreatic cancer in Connecticut. Gastroenterology 968; 55: 677 686. 3. Lam KY, Lo CY. Pancreatic endocrine tumour: a 22-year clinicopathological experience with morphological, immunohistochemical observation and a review of the literature. Eur J Surg Oncol 997; 23: 36 42. 4. De Vita V, Hellman S, Rosemberg S. Cancer Principles and Practice of Oncology. Philadelphia: JB Lippincott 985. 5. Kloppel G, Heitz PU. Pancreatic endocrine tumors. Pathol Res Pract 988; 83: 55 68. 6. Maton PN, Gardner JD, Jensen RT. Diagnosis and management of Zollinger Ellison syndrome. Endocrinol Metab Clin North Am 989; 8: 59 543. 7. Phan GQ, Yeo CJ, Hruban RH et al. Surgical experience with pancreatic and peripancreatic neuroendocrine tumors: review of 25 patients. J Gastrointest Surg 998; 2: 472 482. 8. de Herder WW, Krenning EP, van Eijck CH, Lamberts SW. Considerations concerning a tailored, individualized therapeutic management of patients with (neuro)endocrine tumours of the gastrointestinal tract and pancreas. Endocr Relat Cancer 24; : 9 34. 9. Barakat MT, Meeran K, Bloom SR. Neuroendocrine tumours. Endocr Relat Cancer 24; : 8.. Gullo L, Migliori M, Falconi M et al. Nonfunctioning pancreatic endocrine tumors: a multicenter clinical study. Am J Gastroenterol 23; 98: 2435 2439.. Broder LE, Carter SK. Pancreatic islet cell carcinoma. I. Clinical features of 52 patients. Ann Intern Med 973; 79: 7. 2. Kloppel G, Heitz PU. Pancreatic endocrine tumors. Pathol Res Pract 988; 83: 55 68. 3. Chu QD, Hill HC, Douglass HO Jr et al. Predictive factors associated with long-term survival in patients with neuroendocrine tumors of the pancreas. Ann Surg Oncol 22; 9: 855 862.
8 4. Solcia E, Kloppel G, Sobin LH. Histological typing of endocrine tumors. Heidelberg: Springer-Verlag; 2. 5. Broughan TA, Leslie JD, Soto JM, Hermann RE. Pancreatic islet cell tumors. Surgery 986; 99: 67 678. 6. Venkatesh S, Ordonez NG, Ajani J et al. Islet cell carcinoma of the pancreas. A study of 98 patients. Cancer 99; 65: 354 357. 7. Mansour JC, Chen H. Pancreatic endocrine tumors. J Surg Res 24; 2: 39 6. 8. Eriksson B, Skogseid B, Lundqvist G et al. Medical treatment and long-term survival in a prospective study of 84 patients with endocrine pancreatic tumors. Cancer 99; 65: 883 89. 9. Corleto VD, Annibale B, Gibril F et al. Does the widespread use of proton pump inhibitors mask, complicate and/or delay the diagnosis of Zollinger-Ellison syndrome? Aliment Pharmacol Ther 2; 5: 555 56. 2. Doherty GM. Multiple endocrine neoplasia type. J Surg Oncol 25; 89: 43 5. 2. Evans DB, Skibber JM, Lee JE et al. Nonfunctioning islet cell carcinoma of the pancreas. Surgery 993; 4: 75 8. 22. Grama D, Eriksson B, Martensson H et al. Clinical characteristics, treatment and survival in patients with pancreatic tumors causing hormonal syndromes. World J Surg 992; 6: 632 639. 23. Thompson GB, van Heerden JA, Grant CS et al. Islet cell carcinomas of the pancreas: a twenty-year experience. Surgery 988; 4: 7. 24. Madeira I, Terris B, Voss M et al. Prognostic factors in patients with endocrine tumours of the duodenopancreatic area. Gut 998; 43: 422 427. 25. Wiedenmann B. Neuroendocrine tumors of the pancreas. Natural history and therapeutic options. Gastroenterol Clin Biol 23; 27 (3 Suppl): S26 S3. 26. La RS, Sessa F, Capella C et al. Prognostic criteria in nonfunctioning pancreatic endocrine tumours. Virchows Arch 996; 429: 323 333. 27. Rindi G, Candusso ME, Solcia E. Molecular aspects of the endocrine tumours of the pancreas and the gastrointestinal tract. Ital J Gastroenterol Hepatol 999; 3 (Suppl 2): S35 S38. 28. Weber HC, Venzon DJ, Lin JT et al. Determinants of metastatic rate and survival in patients with Zollinger-Ellison syndrome: a prospective long-term study. Gastroenterology 995; 8: 637 649. 29. Lowney JK, Frisella MM, Lairmore TC, Doherty GM. Pancreatic islet cell tumor metastasis in multiple endocrine neoplasia type : correlation with primary tumor size. Surgery 998; 24: 43 48. 3. Alexakis N, Connor S, Ghaneh P et al. Hereditary pancreatic endocrine tumours. Pancreatology 24; 4: 47 433. 3. Rigaud G, Missiaglia E, Moore PS et al. High resolution allelotype of nonfunctional pancreatic endocrine tumors: identification of two molecular subgroups with clinical implications. Cancer Res 2; 6: 285 292. 32. Norton JA, Warren RS, Kelly MG et al. Aggressive surgery for metastatic liver neuroendocrine tumors. Surgery 23; 34: 57 63.