A Woman With Metastatic Pancreatic Neuroendocrine Tumor

CASE REPORT

Dr. Shoushtari: A 47-year-old woman presented with 3 years of intermittent back pain radiating to the abdomen and 2 months of fatigue, abdominal distension, and mild weight loss. She had no flushing, diarrhea, or malabsorptive symptoms. Prior workup included an abdominal ultrasound notable for benign cysts, an upper gastroesophageal endoscopy, and a hepatobiliary iminodiacetic acid (HIDA) scan. She was referred for magnetic resonance imaging (MRI) of the abdomen.

Dr. Abou-Alfa: Dr. Haydar, can you comment on the images?

Dr. Haydar: This was an MRI with contrast with fat suppression (Figure 1a) showing a pancreatic mass with multiple lesions in the liver. On T2-weighted imaging (Figure 1b), some lesions were purely cystic, but some were mixed cystic and solid. The combination of a pancreatic mass and liver lesions suggested a primary pancreatic malignancy with metastatic disease.

Figure 1.

MRI of the abdomen at presentation, with contrast. Representative slices with (a) fat suppressed and (b) T2-weighted views. There were numerous hypodense metastatic infiltrates interspersed with hyperdense benign cysts.

Dr. O'Reilly: A differential diagnosis for a pancreatic mass with liver lesions in a young woman should include pancreatic adenocarcinoma, neuroendocrine tumors (NETs), and, more rarely, acinar cancer and solid pseudopapillary malignancy of the pancreas.^1,2 Well-defined, hypervascular lesions in general suggest neuroendocrine carcinoma over adenocarcinoma.³

Dr. Covey: Also, the pancreatic duct was not dilated, which argues against pancreatic adenocarcinoma.⁴

Dr. Abou-Alfa: Dr. Shoushtari, was there any family history of cancer?

Dr. Shoushtari: The patient had a history of multiple prior kidney stones and a parathyroid macroadenoma resected 15 years ago. Her family history was notable for a paternal grandmother with breast cancer diagnosed at age 65 and a maternal grandfather with colon cancer diagnosed at age 80, but no endocrine or autoimmune history. Over the course of her workup at an outside institution, the patient had tests to determine carcinoembryonic antigen (CEA), α-fetoprotein (AFP), CA 19-9, chromogranin-A, gastrin, insulin-like growth factor (IGF)-1, plasma serotonin, and C-peptide levels, all of which were within normal limits.

Dr. Kelsen: What familial syndromes come to mind for a young woman with an unusual tumor in her pancreas and a history of parathyroid macroadenoma?

Dr. O'Reilly: Multiple endocrine neoplasia (MEN)-1, which consists of parathyroid, pituitary, and pancreatic lesions that can include NETs.⁵ Other hereditary syndromes, such as von Hippel-Lindau, neurofibromatosis type 1, and tuberous sclerosis, are associated with pancreatic NETs (PNETs).⁶ Given the family history of breast cancer and colon cancer, it is worth noting that an evaluation for BRCA mutations and hereditary nonpolyposis colorectal cancer (HNPCC) might also be considered.

Dr. Abou-Alfa: Dr. Zaatari, can you comment on the pathology?

Dr. Zaatari: In the pancreatic biopsy (Figure 2a), the tumor cells were organized in trabeculae and sheets surrounded by desmoplastic stroma. Higher magnification (Figure 2b) revealed a somewhat uniform population of oval-to-polygonal cells, with a high nucleus-to-cytoplasm ratio and a moderate amount of granular cytoplasm, most consistent with a PNET.⁷ A mitotic figure was identified, which may be seen in these tumors but appears more frequently in intermediate and high-grade tumors. The observed morphologic features are unusual for ductal carcinoma of the pancreas; this tumor typically shows a much greater degree of pleomorphism and nuclear anaplasia. Moreover, the features are not suggestive of acinar carcinoma.

Figure 2.

Hematoxylin and eosin–stained core biopsies of (a) pancreas (×10), (b) pancreas (×100), and (c) liver (×40) demonstrating metastatic PNETs.

Dr. Shoushtari: The left hepatic lobe nodule was also biopsied.

Dr. Zaatari: Multiple fragments of liver tissue with parenchyma were shown, with evidence of infiltration by a metastatic NET. At a higher magnification (Figure 2c), there was more obvious cellular uniformity and an organoid appearance of the tumor cells; these features, with the metastasis to the liver, are suggestive of low- and intermediate-grade NETs, compatible with hepatic metastasis from a PNET.

Dr. Abou-Alfa: What do you use to determine grading?

Dr. Zaatari: Recently, the new World Health Organization (WHO) classification has dropped the term “carcinoid” and replaced it with the term “neuroendocrine tumor, not otherwise specified (NOS)” (grade 1). Although these tumors do not show worrisome cytologic features, some behave in a malignant fashion, and the clinical course is unpredictable. Tumors with obvious malignant cytologic features are classified as neuroendocrine carcinomas and are divided into intermediate grade (grade 2) and high grade (grade 3), depending on the degree of differentiation and mitotic activity. In the past, the distinction between these 2 groups was determined primarily by the number of mitotic figures per 10 high-power fields (HPFs); however, more recently, immunohistochemical staining for Ki-67, a marker for cellular proliferative activity, has been used for grading, and several studies have demonstrated good correlation between Ki-67 staining and prognosis. Tumors with Ki-67 staining less than 2% are classified as NET (grade 1), 2 to 20% is an intermediate grade malignancy (grade 2), and greater than 20% is a high-grade malignancy.⁸ There is greater reliance on Ki-67 in addition to mitotic counts in today's classification of NETs.

Dr. Naghy: Although the Ki-67 percentage has been shown to predict progression-free and overall survival,^9,10 there is controversy regarding evaluation of cases with grades near the Ki-67 cutoff points. Recent studies have shown that there is significant variability between pathologists in interpreting Ki-67 visually¹¹ and that Ki-67 can vary significantly, even within a single core biopsy.¹² Also, the number of mitotic figures per 10 HPFs can be variable between laboratories and even between individual microscopes.

Dr. Zaatari: The classification of NETs has always been challenging to pathologists, and I do not think we have all the criteria well defined and agreed upon. However, the grading and classification must include a combination of these features: morphology, mitotic index, and Ki-67. I agree with the problem of standardization of HPF counts, which are well known to vary and must be defined within each laboratory as part of their quality assurance process. I am fairly convinced that when the WHO reconvenes to review NET classification, there will be new criteria based on the new knowledge gained in this field.

Dr. O'Reilly: David Klimstra¹³ has pooled various international staging systems in an effort to reach consensus. High-grade tumors can have more heterogeneity. There are the anaplastic small-cell types considered high-grade, but some high-grade tumors may originate from a well-differentiated background.¹⁴ We may approach these 2 types differently, but this difference is not well captured in our current grading system.

Dr. Shoushtari: The final pathology showed metastatic, well-differentiated PNET. Synaptophysin and chromogranin-A were positive. There were 2 mitoses per 10 HPFs and Ki-67 was 15%, consistent with an intermediate grade.

Dr. Abou-Alfa: Dr. Shamseddine, do you still check 24-h urine for 5-hydroxyindoleacetic acid (5-HIAA) in these patients?

Dr. Shamseddine: This test is recommended in patients with secretory tumors, but we routinely check 5-HIAA in all patients.

Dr. O'Reilly: Most PNETs are nonfunctional, and this patient had a nonfunctional tumor without a secretory syndrome, characterized by flushing, diarrhea, or other symptoms of an insulinoma, gastrinoma, or vasoactive intestinal peptide secreting tumor (VIPoma) (Verner-Morrison syndrome). She had an extensive serologic workup at an outside institution, which we reviewed, but without a compelling history of a secretory syndrome, we at Memorial would not be likely to pursue an extensive list of hormonal tests, apart from measuring serum chromogranin-A and serotonin.

Dr. Abou-Alfa: I add that some PNETs present with unusual functional syndromes, including a series of concurrent Cushing syndrome and Zollinger-Ellison syndrome¹⁵ and a case report of adrenocorticotropic hormone (ACTH)-producing PNET.¹⁶ Also, serum chromogranin-A may be a biologic marker of tumor burden.¹⁷ One study suggested that it may be a prognostic marker, and higher levels of serum chromogranin may portend a worse outcome.¹⁸ It is worth noting that patients with renal and hepatic impairment may have elevated levels of chromogranin.¹⁹

Dr. Naghy: I concur with Dr. O'Reilly and Memorial's practice. The normal level of chromogranin-A can vary widely and is increased by commonly used medications, such as proton pump inhibitors,²⁰ and so it may not be a useful test for diagnostic purposes; however, when elevated, its level could be used to follow up responses to treatments.

Dr. Abou-Alfa: Dr. Kelsen, can you comment on prognosis?

Dr. Kelsen: The median duration of survival of unresectable PNET is approximately 24 months,²¹ although the range is varied. The survival is certainly better than that of adenocarcinoma, but it is still fairly poor.

Dr. Abou-Alfa: Dr. Mukherji, how would you approach treatment in this patient?

Dr. Mukherji: For this intermediate grade, unresectable PNET, there are now several treatment options. If this were a secreting tumor, the long-acting somatostatin analogues could be helpful in alleviating symptoms.²² In terms of systemic therapy, there are cytotoxic chemotherapy; targeted agents, with 2 classes showing benefit in phase III studies—mTOR and tyrosine kinase inhibitors; and local therapies, such as embolization of hepatic arteries and radioimmunotherapy.²³ Choosing among treatments, we should consider disease burden, performance status, access to treatment, and open clinical trials.

Dr. O'Reilly: This patient was somewhat symptomatic and understandably apprehensive, which informed much of the clinical decision-making process.

Dr. Kelsen: Dr. Haydar, what further imaging might you need to guide therapy?

Dr. Haydar: An indium-111 octreoscan is more sensitive for well-differentiated rather than poorly differentiated tumors.²⁴ For this patient, you could pursue an indium-111 octreoscan or a gallium-68 DOTA positron emission tomographic (PET) scan, which will bind somatostatin receptor-2.²⁵ For poorly differentiated neuroendocrine carcinomas, an [18F]-fluorodeoxyglucose (FDG) PET scan may be useful; FDG PET avidity has been linked to a poorer prognosis.²⁶

Dr. Abou-Alfa: Dr. Haydar, can you comment on the indium-111 octreotide scan in this patient?

Dr. Haydar: It showed multiple areas of uptake in the liver, pancreas, and a supraclavicular lymph node (Figure 3).

Figure 3.

Indium-111 octreotide scan obtained at diagnosis demonstrating hepatic and supraclavicular lymph node metastases.

Dr. Abou-Alfa: Are there any data that would support giving octreotide as an initial therapy in metastatic, nonfunctional NETs with a moderately avid octreotide scan?

Dr. O'Reilly: In the PROMID study, subjects with midgut tumors were enrolled, and 30 mg LAR Sandostatin was compared to placebo.²⁷ The results demonstrated a progression-free survival benefit over observation, but the trial was not designed to detect a survival outcome. However, an updated analysis of the results of this study has demonstrated a survival benefit for patients with low hepatic tumor load treated with initial long-acting somatostatin.²⁸ It is to be noted that the PROMID study did not require a confirmation of disease progression before enrollment of patients in the study.

Dr. O'Reilly: Dr. Covey, have you any comment regarding types of regional therapies and their suitability for our patient?

Dr. Covey: At Memorial we use mainly bland hepatic artery embolization, which is generally well tolerated, to treat patients who either have rapid progression, hormonal symptoms, or pain related to metastatic disease in the liver.^29,30 Other centers use radio-^31,32 or chemoembolization,^33,34 but there are no prospective data comparing these interventions, only small retrospective studies describing patients with various amounts of disease burden.^35,36 Given her lack of discrete, bulky liver metastases and significant extrahepatic disease, I would recommend against regional therapy at this time. Based on the number of hepatic lesions, the patient is not a good candidate for radiofrequency or cryoablation.

Dr. O'Reilly: Can you please comment on radioembolization?

Dr. Covey: This embolization method uses beads coated with a radionuclide such as yttrium-90 (Y-90).³⁷ It is more expensive and labor intensive than bland embolization and theoretically possesses a higher risk of hepatic insufficiency. As a result, we consider it up front only in intermediate or high-grade tumors that tend to respond more poorly to bland embolization or as a second-line treatment if bland embolization fails. There are no prospective data comparing the interventions head to head.

Dr. Abou-Alfa: Dr. Kelsen, can you describe peptide receptor radionucleotide therapy (PRRT)?

Dr. Kelsen: In contrast to Y-90-coated microspheres, which are regionally directed but not targeted, PRRT utilizes somatostatin analogues as a targeted delivery system to direct a radiopharmaceutical such as Y-90 or lutetium-177 (Lu-177).^38,39

Dr. O'Reilly: Until recently, PRRT has not been widely available in the United States. Typically, we send people to Europe. Our European colleagues have reported large, uncontrolled case series using various isotopes, but there are no randomized trials demonstrating the efficacy of this strategy. Several large case series reported tumor response in about a third of patients, but there are 2 significant toxicity concerns: myelosuppression and renal toxicity.^38,39 However, newer radioisotopes, improved dosimetry, and the infusion of an amino acids mixture may overcome some of these concerns.

Dr. Kelsen: The first randomized trial, which is international, is about to open in the United States for carcinoid tumors, to compare high-dose octreotide 60 mg with Lu-177 plus low-dose octreotide 30 mg.⁴⁰

Dr. Abou-Alfa: Dr. Shamseddine, what is your experience with PRRT?

Dr. Shamseddine: We have started using Lu-177 here in Lebanon. So far, we have treated 3 patients, with good response. The operational cost is much less than in Europe.

Dr. O'Reilly: I concur with Dr. Kelsen that the field needs a randomized study to define the efficacy, safety profile, and appropriate subset of patients who benefit most from this intervention. These data would help define where in the disease trajectory this therapy best fits.

Dr. Abou-Alfa: Dr. Shoushtari, how did the patient do?

Dr. Shoushtari: The patient was started on long-acting octreotide. After 2 months, she still had some pain, but her appetite improved and she stopped losing weight. In part because of patient preference, we restaged her with an MRI after 2 months of therapy.

Dr. Abou-Alfa: Dr. Haydar, can you comment on the MRI findings?

Dr. Haydar: With the limited views presented here (Figure 4), my impression is that the lesions may have mildly progressed.

Figure 4.

MRI of the abdomen with contrast after 2 months of octreotide treatment. (a) Fat-suppressed and (b) T2-weighted views.

Dr. O'Reilly: In most circumstances, after only 2 doses of long-acting octreotide, we would be likely to recommend continuing the same treatment, given the symptom improvement and only mild radiologic progression. After discussions with the patient, however, we felt it necessary to adjust therapy.

Dr. Abou-Alfa: Dr. Epstein, you are an expert in the psychosocial aspects of cancer care. How do you deal with a situation where a patient asks to switch therapies? Do you respond to the patient's request or say, “This is my treatment recommendation. Take it or leave it”?

Dr. Epstein: I am sure there were attempts to explore and respond to underlying emotions, and often it is useful to involve other disciplines (eg, psychiatry, palliative medicine, or social work) in situations punctuated by marked apprehension, as is described in this case. Through such measures, one attempts to strike a middle ground. One acknowledges and responds to the understandable degree of anxiety for an individual (or their loved ones) for whom serious illnesses like cancer often represent complete loss of control or a violation of their life's plans or dreams. I would present the medical situation at hand and the options available and then engage in decision-making with the patient in the hopes of meeting the patient's goals, given the situation. One is not obligated to order tests or to give therapies deemed medically inappropriate; however, the clinician must still attempt to meet the patient halfway, and sometimes that involves ordering tests or trialing certain therapies (especially in rarer disease such as this one), all to assist in decision-making and care for that individual. In the end, the best one can do is make recommendations based on the patient's personal components, and the medical facts.

Dr. O'Reilly: That is how we approached the situation. The patient has significant disease that is incurable and also some component of disease progression, and there is clearly a rationale for additional therapy at this juncture.

Dr. Shoushtari: We then considered cytotoxic chemotherapy. Many of the published studies of streptozocin, 5- fluorouracil, or temozolomide were performed in heterogeneous NET samples with very few PNET patients enrolled. In general, overall response rates (ORRs) have been low, ranging from 15 to 40%, with 2 notable exceptions. An older study reported a 69% ORR,⁴¹ but included serologic response criteria that are no longer valid. Temozolomide+capecitabine showed a nearly 70% ORR in a phase II study of 30 PNET patients.⁴² We contrasted this with data supporting everolimus. In RADIANT-3,⁴³ 410 patients were enrolled who had documented progression within 12 months. Less than half got concurrent octreotide, and the study design allowed cross-over, which confounded the overall survival data. The ORR was low, but stable disease was common, and progression-free survival was increased vs. placebo from 4.6 to 11 months. The sunitinib trial⁴⁴ was smaller because it was halted early by the data safety monitoring board for an excess of deaths and adverse events in the placebo group. Median duration of therapy in this trial was 4 months, and it was censored at progression, so it did show an overall survival benefit, but the confidence interval for the hazard ratio is quite wide.

Dr. Abou-Alfa: So, we have 2 therapies for the same disease published in the same edition of the New England Journal of Medicine.^43,44 Dr. Kelsen, can you comment on the approval process for sunitinib?

Dr. Kelsen: The approval process was carried out by the Oncologic Drugs Advisory Committee (ODAC) of the U.S. Food and Drug Administration (FDA). It was very controversial, although the ultimate decision was to approve the drug. The committee members included oncologists, a biostatistician, a patient representative, and an industry representative who could not vote. There was much controversy regarding the balance between the 2 arms of sunitinib vs. placebo, because the pathology on which they determined intermediate vs. low grade was from tissue obtained long before enrolment in the trial and frequently only by report, not central review. Although responses were uncommon and the survival data were controversial, it was felt that they provided enough benefit to recommend granting FDA approval in a split vote.⁴⁵ The point was made by a panel member that chemotherapy for patients with PNET should not be abandoned.

Dr. Abou-Alfa: Can you comment on the everolimus trial in regard to the independent radiology review of the objective responses?

Dr. Kelsen: In RADIANT-3, there was a dispute between the independent Response Review Committee and the investigator response that led to the convening of a third review committee. They withdrew the application for FDA approval of everolimus in well-differentiated NETs, and the study is being redone as RADIANT-4, which is already open for enrollment.⁴⁶

Dr. Abou-Alfa: Dr. Shoushtari, what did you elect to give the patient?

Dr. Shoushtari: After discussion with the patient, we elected to start everolimus 10 mg daily and to continue long-acting octreotide 20 mg monthly.

Dr. Abou-Alfa: It seems that many of us are leaning toward starting with everolimus as a targeted therapy in these cases.

Dr. O'Reilly: There are a couple things that influenced us to select everolimus over sunitinib. One of them was the larger size of the RADIANT-3 trial just mentioned. Also, the toxicity profile of everolimus appears to allow more tolerability for longer periods of time, although that is debatable. There is no consensus, however, on the optimal sequencing of these agents. If a patient has cardiac disease, one might avoid sunitinib; if the patient has difficult-to-control diabetes, one might not use everolimus as the first choice.

Dr. Abou-Alfa: Do our colleagues at the American University of Beirut and at the National Guard Hospital in Riyadh have any final comments?

Dr. Shamseddine: There are many unanswered questions in the field of NETs. For example, what do you do with a Ki67 level between 10 and 20%? Also, what is the role of primary chemotherapy in symptomatic patients with relatively high Ki67 with the combination of temozolomide and capecitabine?

Dr. Naghy: What is the optimal follow-up schedule for imaging with these patients? Are there different imaging modalities to use?

Dr. O'Reilly: I agree with both points. We have no consensus on which therapy should be up front or the optimal sequence for the available options. There is a range of treatment options, as we've outlined. Follow-up depends on several things: first, the biology of the disease in terms of the grade and differentiation; second, the symptom burden; and third, the patient. Two months is an early but not unreasonable timeframe for reimaging in a patient who has symptoms and reasonable disease bulk and whose pathology is on the higher end of the spectrum of intermediate grade.

Dr. Abou-Alfa: Thank you all. In summary, pancreatic NETs are a heterogeneous group of cancers with clinical behavior that can be better understood with evolving histopathologic criteria and new imaging techniques. Multiple new locoregional and systemic treatments of metastatic PNET have been developed over the past 20 years, but there is a critical need for prospective randomized data to better guide clinicians on how best to proceed with these various therapies. Treatment should be individualized by taking into account each patient's goals and preferences and offering a consensus opinion from a multidisciplinary panel of medical oncologists, surgeons, radiation specialists, and interventional radiologists.