Carcinoid Tumors: Past, Present, and Future

Abstract

Carcinoid tumors are being seen with increasing frequency by surgeons and have become the most common type of tumors of the small bowel. These tumors produce a variety of hormones, which leads to many unique characteristics in terms of symptoms and presentation. Our knowledge of the natural history and treatment of these tumors continues to evolve, and this article will summarize these advances.

History

Carcinoid tumors, now more specifically referred to as small bowel neuroendocrine tumors (SBNETs), were previously rare tumors which were recognized as distinct entities just over 100 years ago. They were likely to have first been described in 1888 when Lubarsch noted multiple small tumors in the ileum of 2 patients at autopsy [1]. Two years later, Ransom reported a patient who had diarrhea and shortness of breath who was found to have an ileal tumor and liver metastases [2]. Oberndorfer described 6 patients with these tumors in 1907, which resembled adenocarcinomas, but displayed more benign behavior and had not metastasized [3]. Later in his career, he conceded that these tumors actually do metastasize [4]. Masson demonstrated the enterochromaffin cells in the small bowel by silver staining and that carcinoid tumors arose in these cells in 1928 [5]. In 1953, Lembeck found that these tumors secrete serotonin [6], and a year later, Thorson et al. described a series of patients with these tumors and liver metastases who had diarrhea, flushing, asthma, and right-sided valvular disease [7].

The first large series of these patients was described by Moertel, who reported upon 203 patients in 1961, with an incidence of 0.65% on autopsy. He also noted that these tumors occurred more frequently moving distally along the small bowel, and that about 50% of tumors 1–2 cm in size metastasize [8]. His classic paper “An odyssey in the land of small tumors” from 1987 updated this series and described 183 patients who had had surgical resection [9]. Although 80% of patients were disease free at 5 years, in long-term follow-up, out to 25 years, only 23% had not recurred. Moertel also was an early adopter of somatostatin analogues for symptom relief and chemotherapy for metastatic disease.

Incidence

Our understanding of the biology and natural history of these tumors has gradually improved since these early days. These tumors were previously considered very rare, but their incidence has increased markedly between 1973 and 2012, from 2 cases to 12 cases per million [10]. There are several potential explanations for this trend. One is that in the early days of national databases (like the Surveillance, Epidemiology, and End Results [SEER] and the National Cancer Database [NCDB]), many SBNETs were considered to be benign and were not captured. Gradually, with the increasing recognition that these tumors were malignant and prone to metastasis, they advanced from being the second most common small bowel tumor in 1995 [11] to passing adenocarcinoma as the leading histology in the year 2000 [12]. Another major contributing factor to the growing incidence is the widespread use of CT scanning to evaluate patients with abdominal pain. CT scans may show liver metastases or suspiciously enlarged mesenteric nodes that lead to further work-up and diagnosis. In addition to these causes, environmental influences as well as and changes related to our modern lifestyles may be contributing to this increased incidence.

Diagnosis

History, Physical, and Biomarkers

These tumors secrete hormones, such as serotonin, histamine, and tachykinins, which are normally degraded within the liver [13]. When liver metastases are present, these compounds are not metabolized and pass directly into the systemic circulation [14]. This leads to carcinoid syndrome, where patients may manifest flushing, diarrhea, abdominal cramping, right-sided cardiac valvular disease, and wheezing [15]. Patients with earlier stage disease may be asymptomatic, but as tumors grow larger, they may cause narrowing of the bowel leading to abdominal pain and eventually small bowel obstruction. Elevation of serotonin leads to diarrhea, but perhaps not flushing [13]. The diagnosis can be made biochemically by the finding of elevated 5-HIAA in the urine [14], and chromogranin A, serotonin, and pancreastatin may have value in follow-up of patients [16–18].

Radiology

Many patients are diagnosed after finding enlarged mesenteric nodes or liver metastases on CT scan [19, 20]. Primary SBNETs may occasionally be seen on CT as a mass or possibly the site of obstruction, but in general, these tumors are small and not found. Enlarged mesenteric nodes, often with calcification, are the most frequent manifestation (Fig. 1a). MRI is better for characterizing liver lesions and is especially useful when cytoreductive surgery is planned. Octreoscan has been replaced by ⁶⁸Ga-DOTAPET/CT for functional imaging, which takes advantage of the abundance of somatostatin receptors on tumors [21–23]. This imaging is useful for determining that a mass seen on CT is a NET, identifying occult primaries, and determining the extent of metastases (Fig. 1b).

Fig. 1.

a CT scan showing a mesenteric mass in patient with a SBNET. b⁶⁸Ga-DOTAPET scan showing widespread metastases, with approximately 60% liver replacement by tumor, with uptake in a supraclavicular node, mesenteric nodes, and 3 small bowel lesions. c SBNET within Meckel’s diverticulum (white arrow) with adjacent calcified lymph nodes (black arrow). d MRI showing liver metastases in the same patient

Pathology

Biopsy of a liver metastasis or mesenteric mass can be helpful to make the diagnosis. Neuroendocrine tumors express synaptophysin and chromogranin by immunohistochemistry, and those of small bowel origin will also be positive for CDX2 and negative or weak for PAX6 and Islet1 (which are more specific for pancreatic NETs) [24, 25]. Determining the grade of the tumor is important using either the Ki-67 stain or mitotic rate, as this has important implications for prognosis [26].

Treatment

Exploration

Patients presenting with small bowel obstruction and suspicion of a SBNET by CT, or the characteristic mesenteric mass (often with calcifications), should be explored with or without biopsy proving the presence of a NET. Exploration is ideally carried out through a midline incision, where the abdominal cavity is surveyed for the presence of peritoneal nodules (most commonly found in the pelvis, under the diaphragm, or on the mesentery), liver lesions, enlarged mesenteric nodes, and ovarian masses. Next, the small bowel is run from the ligament of Treitz to the ileocecal valve, carefully palpating the entire length between thumb and forefinger [27]. Tumors can range from the size of a grain of rice to over 2 cm and are often subtle, firm masses within the wall of the bowel. The primary tumors may be fairly small, yet be accompanied by very large nodal and/or liver metastases (Fig. 1c and d). We measure the length of the bowel and record and mark the sites of each lesion found so as to determine the most sensible plan for resection. The average small bowel length is 5–600 cm, and resection of 100+ cm may be necessary to perform a single resection and anastomosis when multiple lesions (which occurs in 55% of patients) or large mesenteric nodes are present [28]. It is important to understand that if exploration is to be carried out laparoscopically, the incision that will be made for extracorporeal anastomosis should also be used to carefully palpate the entire length of the small bowel. Relying upon visualization of tumors through the camera or use of laparoscopic graspers for palpation will undoubtedly miss smaller tumors.

Lymphadenectomy

The length of bowel resection is also determined by the extent of nodal involvement, and the more proximally involved nodes extend up the mesentery, the greater likelihood of removing the vascular supply to a longer segment of bowel. In cases of distal ileal tumors, which are the most common site, ileocolic resection is recommended, removing the nodes up to the takeoff of the ileocolic artery and vein. For more proximal ileal (or the much less common jejunal) tumors, then removing the mesenteric nodes up to the takeoff of the segmental vessel to that portion of the bowel is recommended while preserving adjacent vessels wherever possible. If the nodes travel higher than this, then skeletonizing the SMA or SMV may be necessary; this may not be possible when nodes are heavily calcified and engulf these main trunks, and in these cases, it may be better to divide the nodes just distal to the segmental vessel takeoff. One can use an energy device or sequential clamping and tying to divide the mesentery, and staying along the edge of the mass as one comes to the proximal nodes is advised to avoid taking collateral vessels wherever possible. Beyond the most proximal nodes, it is best to try to identify and preserve the adjacent segmental vessels to maintain viability of the bowel that remains prior to division and suture ligation of the feeding segmental artery and vein. Other commonly involved nodal areas that may be seen on the preoperative CT scan are left periaortic, aortocaval, portocaval, and peripancreatic/duodenal nodes. The benefits or removing these nodes is unclear, and this adds further complexity to these procedures. We try to remove these nodes when we can, but the decision to proceed with this needs to be made within the context of what other areas of disease will be left behind (such as liver and bone metastases, more proximal mesenteric nodes, mediastinal or supraclavicular nodes), and how their removal might impact upon palliation and/or survival of the patient [29].

Peritoneal Disease

Tumors that grow through the serosa may seed the peritoneum with cells, which can end up anywhere, but most commonly in dependent areas. These tumor nodules may implant on the bowel wall, and lesions on the rectosigmoid colon may eventually lead to colonic obstruction. Other nodules may lead to kinking of the small bowel and obstruction, while others may not lead to problems. We recommend removing larger (> 5 mm) peritoneal lesions when possible to avoid future complications, and the use cautery or argon beam to treat smaller lesions. Heated intraperitoneal chemotherapy in these patients has not been shown to be of benefit [30].

Hepatic Cytoreduction

The liver is the most common site of distant metastasis from SBNETS. These tumors are often multiple, bilobar, and the extent of disease may not be appreciated unless a good-quality arterial and venous phase CT is performed, or an MRI with Eovist. If the lesions are small and diffuse, occupy more than 50% of the hepatic parenchyma, or number greater than 20–30, effective cytoreduction may not be possible [29, 31]. Otherwise, multiple lesions can be effectively treated by resection, including wedge resections and enucleations, as well as other parenchymal sparing techniques, such as microwave or radiofrequency ablation or electroporation. It used to be advocated that if 90% of the lesions could not be removed, then these procedures should not be done [32–34], but more recent studies have shown that achieving a threshold of 70% cytoreduction also likely provides a survival benefit [31, 35, 36]. It is important to remember that whatever one does in terms of cytoreduction, that recurrence is the rule rather than the exception. In one large series, 94% of patients had recurred within 5 years and 99% by 10 years [37], likely because there are many other smaller lesions present that cannot be appreciated at the time of cytoreduction [38].

Cholecystectomy

When patients have nodal or liver metastases at the time of exploration, one should carefully consider removing the gallbladder. There are several reasons for this, one of which is that patients who will be receiving long-term treatment with somatostatin analogues will eventually develop gallstones. Not all patients with gallstones will develop biliary colic or acute cholecystitis, but one study showed that up to 22% of NET patients needed cholecystectomy later when it was not performed at the initial operation [39]. Patients who undergo hepatic embolization may also be at risk for gallbladder necrosis following that procedure and cholecystectomy can avoid this problem [29].

Survival and Adjuvant Treatment

Surveys of national databases have shown a median survival of 170 months for patients with localized SBNETs, 145 months for regional disease, and 70 months for those with metastatic disease [10]. With aggressive hepatic cytoreduction of liver metastases, median survival may be improved to well over 100 months [35, 36]. Adjuvant therapy may not be necessary for patients with localized or low-volume regional disease, while it should be entertained in those with advanced disease or metastases. Somatostatin analogues have been shown to improve progression-free survival (PFS), but not overall survival [40, 41]. The mTOR inhibitor Everolimus has also been shown to modestly improve PFS (11 months vs. 4 months with placebo) [42]. Peptide receptor radiotherapy (PRRT) is also an option for those with advanced or metastatic disease. In the NETTER1 trial, the PFS of patients given ¹⁷⁷Lu-DOTATATE was not reached vs. 8.4 months in those receiving high-dose monthly Sandostatin [43]. After surgery, we generally place patients with metastatic disease on somatostatin analogues and use PRRT or everolimus when they develop progressive disease.

The Future

We have discussed the history and current management of SBNETS, but what does the future hold for continued progress for patients with these tumors? Radioguided surgery using ⁶⁸Ga-DOTATATE may allow for improved detection of primary and metastatic SBNETs [44, 45]. Genetic studies of SBNETs have shown frequent deletions in various chromosomal regions [46, 47], but few recurring mutations, except in the gene CDKN1B [48–50]. One family has been found with a germline mutation in the inositol polyphosphate multikinase gene, but this has not been found in other families [51], and therefore genetic testing for the predisposition to these tumors is not yet possible. Our current knowledge of genetics thus far has not yielded many new therapies, but progress has been made with taking advantage of gene expression profiles of SBNETs. These can be used to determine the site of unknown primaries [52, 53] and for follow-up and recurrence [54]. Knowledge of gene expression in primaries and metastases will also likely yield new targets for therapy [55], and the development of organoid and spheroid models of SBNETs will allow for more efficient drug testing [56]. Our approach to SBNETs has become increasingly sophisticated over the past century, but much work remains for us to translate our knowledge into improved therapies and survival for these patients.

Funding Information

Thanks to the NIH Iowa SPORE grant P50CA174521-01 for sponsoring our research.