Original Articles
Duodenal Carcinoid Tumors: How Aggressive Should We Be? NicholasJ. Zyromski, M.D., Michael L. Kendrick, M.D., David 211.Nagorney, M.D., Clive S. Grant, M.D., John H. Donohue, M.D., Michael B. Famzell, M.D., Geofiey B. Thompson, 111.D., David R. Farley, M.D., Michael G. Sam; M.D.
Duodenal carcinoid tumors are uncommon. It is not known whether they behave more like carcinoid tumors in the appendix (indolent course) or those in the ileum (often virulent)-crucial information for determining tbe need for radical resection. A retrospective review at our tertiary referral center (from 1976 to 1999) identified 27 patients with primary duodenal carcinoid lesions, excluding functional islet cell tumors. Endoscopic biopsy provided tbe diagnosis in 78% of patients. Treatment was by endoscopic excision (n = 1 l), transduodenal excision (n = S), pancreaticoduodenectomy (n = 3), segmental distal duodenectomy (n = 2), or palliative operation (n = 2). One patient did not undergo operation because of comorbidity. Eighteen of 19 patients with tumors smaller than 2 cm remained disease free after local (endoscopic or transduodenal) excision. The exception was a patient with a small periampullary carcinoid lesion. In contrast, all four patients with carcinoid tumors 2 cm or larger who were resected for cure developed a recurrence (2 to 9 years postoperatively). We conclude that duodenal carcinoid tumors smaller than 2 cm may be excised locally; to ensure complete resection we recommend open transduodenal excision for tumors between 1 and 2 cm. Endoscopic follow-up is indicated. It is unclear whether patients with larger tumors benefit from more aggressive locoregional resection. Ampullary/periampullary carcinoid tumors should be considered separately, as their behavior is unpredictable. (J GASTROINTEST SURG 2001;5:588-593.)
KEY WORDS:
Carcinoid
tumor, duodenum,
neuroendocrine
Carcinoid tumors are of neuroendocrine origin; they arise from enterochromafin cells and are found most commonly in the gastrointestinal tract. Primary duodenal carcinoid lesions account for less than 5% of all gastrointestinal carcinoid tumors,1-3 making it difficult to accrue a large enough experience to identify prognostic factors specific for the duodenum. Broader experience with carcinoid tumors in the remainder of the gastrointestinal tract has allowed more complete study and relatively clear therapeutic recommendations. Small appendiceal and rectal carcinoid tumors have an indolent course; tumors smaller than 2 cm in these locations without evidence of local invasion can be safely treated by conservative excision
tumor, endoscopy
without major concern about metastatic disease.4-7 Carcinoid tumors of the distal small bowel, on the other hand, are more aggressive. Even small tumors (especially in the ileum) are associated with mesenteric lymphatic metastases and distant metastases.@ Indeed, most ileal carcinoid tumors have already metastasized to regional lymph nodes by the time of diagnosis. Segmental resection with wide mesenteric
excision to include the draining lymph nodes has thus been considered minimal therapy for these tumors, regardless of size. In addition, some investigators have recently suggested that carcinoid tumors of the foregut may portend a worse prognosis than those arising elsewhere in the gastrointestinal tract, although
From the Gastroenterology Research Unit and Department of Surgery, Mayo Clinic, Rochester, Minn. Supported by a grant from the National Institutes of Health (NIH DK393 37 [Dr. Sarr]) and the Mayo Foundation. Presented in part at the Forty-Second Annual Meeting of The Society for Surgery of the Alimentary Tract, Atlanta, Ga., May 20-23,2001, and published as an abstract in Gaswoenterology 12O:A456, 2001. Reprint requests: Michael G. Sarr, M.D., Gastroenterology Research Unit @L-2-435), Mayo Clinic, 200 First Street SW, Rochester, MN 55905. 588
Vol. 5, No. 6 2001
these series included only a few duodenal carcinoid lesions.6,9 It remains unclear whether duodenal carcinoid tumors behave more like appendiceal and rectal tumors, with a more indolent course, or like ileal carcinoid lesions, with a greater likelihood of locoregional or distant hematogenous metastases-crucial information for determining the necessity of radical resection in this anatomic region where a formal segmental resection including regional lymph nodes may require a pancreaticoduodenectomy. This study was therefore undertaken to evaluate treatment and outcomes of patients with duodenal carcinoid tumors from our large tertiary referral practice. METHODS All consecutive patients evaluated for treatment of duodenal carcinoid tumors at the Mayo Clinic in Rochester, Minnesota, from January 1976 to December 1999, were identified by a retrospective chart review. We have defined “carcinoid tumor” according to the World Health Organization International Histological Classificationof Tumows. lo Those patients with neuroendocrine tumors that were functional at the time of diagnosis (e.g., gastrinoma, somatostatinoma), as well as those with tumors arising from structures surrounding the duodenum (pancreas, common bile duct), were specifically excluded from this study. Patient demographics, clinical features, diagnostic procedures, details of the operation, and pathologic findings were recorded. Because carcinoid tumors in other regions of the gut have been associated with multicentricity (up to 30%) and with other primary malignancies (up to 40%), we evaluated these parameters as well. Complete follow-up to date was obtained from patient charts, tumor registry, or by contacting patients or their primary physician by telephone or letter. The study protocol was approved by the Mayo Clinic Institutional Review Board. RESULTS Patient Characteristics Twenty-seven patients (15 men, 12 women) had duodenal carcinoid tumors. Median age at the time of diagnosis was 66 years (range 43 to 86 years). The most common presenting symptom was abdominal pain (48%). Other symptoms included vomiting, weight loss, gastrointestinal blood loss, pancreatitis, abdominal mass, and other nonspecific gastrointestinal complaints (Table I). Four patients were asymptomatic and diagnosed at the time of workup for an unrelated gastrointestinal condition. Diagnosis was secured in the vast majority of patients (78%) by endoscopic biopsy.
Duodenal Carcinoid Tumors
589
Table I. Symptoms of duodenal carcinoid tumors Symptoms
No. of patients (%)
Abdominal pain Vomiting Weight loss Gastrointestinal blood loss Pancreatitis Mass Nonspecific gastrointestinal complaints* Asymptomatic
13 (48) 405) 4 (15) 3 (11) 3 (11) 1 (4) 11 (41) 4 (15)
*Dyspepsia, nausea, bloating, heartburn, diarrhea.
Duodenal carcinoid tumors were found incidentally at operation for another unrelated condition in two patients. In two others, diagnosis was made by computed tomography-guided fine-needle aspiration. Pathologic Findings The majority of these tumors (16 of 27) were located in the first portion of the duodenum. Nine others were in the second portion, including two periampullary tumors, and only two were located in the distal duodenum. One patient with a periampullary tumor was noted to have several small (co.4 cm) localized tumors within the pathologic specimen; no other patient in this series had multicentric gastrointestinal carcinoid lesions. Nine patients (33%) had one or more other primary malignancies, including three patients each with adenocarcinoma of the colon and prostate, and one patient each with jejunal adenocarcinoma, lymphoma, small cell carcinoma of the lung, and renal cell carcinoma. Twelve tumors in this series underwent evaluation with special immunohistochemical stains. The majority (10 of 12) stained positive for the nonspecific neuroendocrine marker chromogranin. In addition, five tumors were positive for synaptophysin, two each for neuron-specific enolase and cytokeratin, and one each for gastrin, somatostatin, and the neuroendocrine marker nkhl . Treatment Enahscopic Excision. Eleven patients were treated by endoscopic excision, all with tumors 1.0 cm or smaller. All patients are alive and tumor free at a median follow-up of 4.2 years (range 1.5 to 8 years). One patient had an early local recurrence (noted at 3 months) treated successfully by endoscopic reexcision; he is alive with no further recurrence 7 years later.
590
Zyromski
Gastrointestinal
et al.
Journal of Surgery
Open Excision. Fifteen patients were managed operatively (Table II). Operative approaches included transduodenal local excision (n = 8), grossly curative pancreaticoduodenectomy (n = 3), grossly curative segmental excision of the third and fourth portions of the duodenum (n = 2), and palliative operation (n = 2; gastrojejunostomy with hepatic artery ligation in one and palliative pancreaticoduodenectomy in one). Thirteen of these patients thus underwent resection with curative intent (median follow-up 4.3 years, range 1.5 to 23 years). Five of the 13 patients resected for cure developed recurrent carcinoid tumor. The mean size of the primary tumor in patients with a recurrence was 3.6 cm (range 1 to 8 cm); the size of the primary tumor was greater than 2 cm in all patients except one who had a 1 cm periampullary carcinoid lesion. Four of the five patients with a recurrence had a primary tumor invading into the muscularis or through the serosa at the time of the initial operation (the exception being the one patient with a submucosal periampullary tumor). Only one patient with recurrent carcinoid tumor had metastatic lymph node involvement at the time of the initial operation. Two patients with recurrent tumor died of disease 7 and 12 years postoperatively, and three are alive with progressive disease 2,6, and 22 years postoperatively. Seven patients resected for cure are alive and tumor free (median follow-up 4.3 years, range 1.5 to 12.8 years), and one died of unrelated causes with no evidence of recurrence. All of these patients had primary carcinoid tumors smaller than 2 cm (mean size 1 cm, range 0.4 to 1.7 cm). The primary tumor was confined to the mucosa in seven of these patients and invaded the muscularis in one. The one patient managed by observation alone because of advanced age and significant comorbidity is alive 2.5 years after the initial diagnosis. Carcinoid Syndrome
3
ot-dwr.mmwbmo*br--m
wwr-lwwmwb~wwww*
Three patients in this series (all with advanced recurrent disease) developed symptoms of carcinoid syndrome (sweating, flushing, and diarrhea) after grossly curative resection. None of these patients had symptoms at the time of the initial diagnosis; all symptoms of carcinoid syndrome become evident after recurrence with multiple bulky hepatic metastases. The 24-hour urine 5-hydroxyindoleacetic acid (5’-HIAA) level was normal in one of these patients, and measurements were not performed in the other two. Another patient who developed metastatic disease with elevated urinary 5-HIAA levels did not have symptoms of carcinoid syndrome. No patient had elevated levels of 5’-HIAA at the time of initial presentation.
Vol. 5, No. 6 2001
DISCUSSION In 1907 Oberndorfer’ l coined the term “karzinoide” to describe a carcinoma-like tumor with less aggressive behavior, a term that remains appropriate today. Although carcinoid tumors are generally considered more indolent, especially when compared to other adenocarcinomas of the gastrointestinal tract, it is important to emphasize that the natural history of carcinoid tumors is variable depending on their location within the gastrointestinal tract. Small carcinoid tumors (~2 cm) arising in the appendix or rectum have a very low incidence of metastasis and are adequately treated by local excision (appendectomy or local excision with fulguration, respectively). In contrast, carcinoid tumors arising in the distal small bowel are more aggressive; even small tumors (~1 cm) here have a much greater propensity to metastasize.4Jjy12,13 Because of this aggressive behavior, segmental excision with wide local mesenteric excision is recommended for treatment of distal small bowel carcinoid tumors. Finding a carcinoid tumor in the duodenum presents a unique clinical dilemma because, unlike carcinoid lesions of the distal small bowel, appendix, or rectum, prognostic factors specific for carcinoid tumors arising in the duodenum are poorly characterized as a result of the rarity of this neoplasm. Proposed treatment of duodenal carcinoid tumors has spanned a broad spectrum including aggressive resection by pancreaticoduodenectomy, transduodenal local excision, full-thickness laparoscopic excision, and endoscopic excision by snare or strip biopsy.6J4-16 Optimal therapy for duodenal carcinoid tumors, however, remains unclear. The need to understand the specific natural history of carcinoid tumors in the duodenum is important for several reasons. First, radical duodenal resection is a more complex undertaking than wide segmental excision of the distal small bowel. Although major duodenal and pancreatic resections are being performed more commonly and safely in the current surgical era,“J* it is not clear whether a major resection provides any added therapeutic benefit over local excision for patients with duodenal carcinoid tumors. Second, the morbidity and mortality of carcinoid tumors are related to metastatic disease (liver replacement from metastases and carcinoid syndrome) and usually not from the primary tumor. A third important consideration is that duodenal carcinoid tumors will be identified more frequently with the increased use of upper gastrointestinal endoscopy. Indeed, almost 60% of patients in our series were diagnosed within the past 5 years, mostly by endoscopy. Most previous reports of treatment of duodenal carcinoid tumors have involved anecdotal experience, collective reviews of multiple small experiences, or re-
Duodenal
Carcinoid
Tumors
591
ports of unique technical approaches. There has been only one other large study of 99 duodenal carcinoid tumors, compiled by Burke et al.‘* at the Armed Forces Institute of Pathology. These pathologists identified size greater than 2.0 cm, presence of mitotic figures, and invasion of the muscularis as independent risk factors for metastasis. No patient in their series with a tumor smaller than 1.0 cm developed metastatic disease at minimum follow-up of 24 months. In our single-institution experience, although several different operative treatments were used, patients seem to be differentiated into two groups based on the size of the primary tumor and the presence of invasion of the muscularis, similar to the series from the Armed Forces Institute of Pathology.‘* Nineteen patients in our series had duodenal carcinoid tumors smaller than 2 cm; in only one of these patients had the tumor invaded past the submucosa into the muscularis. Of these 19 patients, 18 remain disease free after local resection (endoscopic excision in 11 and transduodenal excision in 7) with a mean follow-up of 4.7 years (range 1.5 to 13 years). The one exception was a patient with a 1 cm submucosal periampullary carcinoid lesion. Despite aggressive resection by pancreaticoduodenectomy, the patient had a recurrence 3 years postoperatively. She is alive with liver metastases 2.5 years after recurrence. This patient highlights the unpredictable behavior of even small ampullary/periampullary carcinoid tumors, which should probably stand alone when considering therapeutic options. The distinction between “ampullary” and “periampullary” carcinoid tumors in the literature is difficult to make; these tumors are quite rare, with only 89 cases reported. They do, however, appear to have a more aggressive course, with development of early metastatic disease despite very small size.19,*0 One other patient in our series had a periampullary carcinoid tumor. He underwent local excision of a 0.4 cm lesion and died of unrelated causes with no evidence of recurrence 3 years postoperatively. One of the 11 patients treated by endoscopic resection was found to have a local “recurrence” at 3-month endoscopic follow-up, which was most likely due to incomplete initial resection. This patient underwent endoscopic reexcision and remains free of disease 7 years later; this patient underscores the need for at least one endoscopic reexamination several months after any form of local excision (endoscopic or open). In contrast to patients with tumors smaller than 2 cm, all four patients in our series with duodenal carcinoid tumors 2 cm or larger who underwent curative resection (pancreaticoduodenectomy in 2 and segmental distal duodenectomy in 2) developed recurrent disease 2 to 9 years later. In ail four patients the primary tumor invaded through the submucosa into
592
Zyromski
et al.
the muscularis. It is also notable that only one of these patients had lymphatic nodal metastases at the time of the initial operation. Interestingly, two of these four patients remain alive 2 and 23 years postoperatively, and two have died 6 and 12 years postoperatively (4.5 and 7 years after metastatic recurrence). These latter patients further reinforce the less aggressive natural history of gastrointestinal carcinoid tumors, particularly when compared to gastrointestinal adenocarcinomas. From our single-institution experience and that of others,l* we suggest that duodenal carcinoid tumors smaller than 2 cm be excised locally. Endoscopic excision is appropriate for patients with tumors smaller than 1 cm, with at least one follow-up endoscopy 2 to 3 months later to ensure completeness of resection and rule out local recurrence. Tumors between 1 and 2 cm may best be treated by open transduodenal local excision, as it may be difficult to ensure complete endoscopic excision of these larger lesions. Appropriate management of duodenal carcinoid lesions 2 cm or larger is more problematic. All of our patients with tumors 2 cm or larger who were resected for cure developed recurrent disease despite segmental resection or pancreaticoduodenectomy. Only one of these patients had nodal metastases within the resected specimen. This experience questions the added benefit of an extended oncologic-type resection. Although a segmental resection may be needed to completely excise the primary tumor for local control, whether a more extended resection (pancreaticoduodenectomy) designed to include the draining nodal basin is of any added benefit remains unknown. Endoscopic ultrasonography may be helpful in determining the depth of invasion of the primary tumor; it is unclear whether an extended resection will benefit patients with small (<2 cm) tumors that invade into the muscularis. Although the collective experience with ampullary/periampullary carcinoid tumors is small, and it is difficult to draw conclusions from only two patients in this series, based on our collective experience and findings in the literature, these tumors appear to behave unpredictably and should probably be viewed as a distinct category of carcinoid tumor when considering treatment options. Even very small (< 1 cm) ampullary/periampullary carcinoid tumors display a distinctly different aggressive behavior and may metastasize early.19s20 Carcinoid tumors of the gastrointestinal tract have been associated with high rates of multicentricity (up to 30%)s~21~22 as well as with a 12% to 40% association with other primary malignancies.6~23 Although no patient in our series had distant multicentric carcinoid
Gastrointestinal
Journal of Surgery
tumors, nine patients had an associated second primary malignancy (either synchronous or metachronous), an important consideration during the workup or long-term management of patients with duodenal carcinoid tumors. In summary, our review of 27 patients with duodenal carcinoid lesions suggests that tumors smaller than 2 cm may be safely excised locally (endoscopically for tumors smaller than 1 cm or by open transduodenal local excision for tumors 1 to 2 cm). Close endoscopic follow-up of these patients is clearly warranted. Size greater than 2 cm, invasion of the tumor through the submucosa, or a periampullary location may predict a poorer prognosis. It is unclear whether patients with larger tumors will benefit from more aggressive locoregional resection; however, even patients with metastatic disease may enjoy a reasonably long survival. Ampullary and periampullary carcinoid tumors should be considered separately, as their behavior is more unpredictable.
We thank Deborah Frank for her expert assistancein preparing this manuwi~t. REFERENCES
1. Modlin 2.
3. 4.
5. 6.
7.
8.
9. 10.
11. 12.
IM, Sandor A. An analysis of 8305 cases of carcinoid tumors. Cancer 1997;79:813-829. Neugut AI, Jacobson JS, Suh S, Mukherjee R, Arber N. The epidemiology of cancer of the small bowel. Cancer Epidemiol Biomarkers Prev 1998;7:243-251. Godwin J. Carcinoid tumors: An analysis of 2837 cases. Cancer 1975;36:560-569. Moertel CG, Sauer W, Dockerty M, Baggenstoss AH. Life history of the carcinoid tumor of the small intestine. Cancer 1961;14:901-909. Moertel CG. An odyssey in the land of small tumors. J Clin Oncol 1987;5:1503-1522. Thompson GB, van Heerden JA, Martin JK Jr, Schutt AJ, Ilstrup DM, Carney JA. Carcinoid tumors of the gastrointestinal tract: Presentation, management, and prognosis. Surgery 1985;98:1054-1063. Naunheim KS, Zeitels J, Kaplan EL, Sugimoto J, Shen KL, Lee CH, Straus FH II. Rectal carcinoid tumors-treatment and prognosis. Surgery 1983;94:670-676. Makridis C, Oberg K, Juhlin C, Rastad J, Johansson H, Lorelius LE, Akerstrom G. Surgical treatment of mid-gut carcinoid tumors. World J Surg 1990;14:377-385. Akerstrom G. Management of carcinoid tumors of the stomach, duodenum, and pancreas. World J Surg 1996;20:173-182. Solcia E, KIoppel G, Sobin LH. Histological typing of endocrine tumours. In International Histological Classification of Tumours, 2nd ed. New York: Springer-Verlag, 2000, pp 61, 64-65. Oberndorfer S. Ueber die “kleinen Dunndarmcarcinome.” Verh Dtsch GesPathol 1907;11:113-116. Burke AP, Sobin LH, Federspiel BH, Shekita KM, Helwig EB. Carcinoid tumors of the duodenum. A clinicopathologic study of 99 cases. Arch Path01 Lab Med 1990;114:700-704.
Vol. 5, No. 6 2001
13. Strode1 WE, Talpos G, Eckhauser E Thompson N. Surgical therapy for small-bowel carcinoid tnmors. Arch Surg 1983; 118:391-397. 14. Toyonaga T, Nakamura K, Araki Y, Shimura H, Tanaka M. Laparoscopic treatment of duodenal carcinoid tumor. Wedge resection of the duodenal bulb under endoscopic control. Surg Endosc 1998;12:1085-1087. 15. Yamamoto C, Aoyagi K, Suekane H, Iida M, Hizawa K, Kuwano Y, Nakamura S, Fujishima M. Carcinoid tumors of the duodenum: Report of three cases treated by endoscopic resection. Endoscopy 1997;29:2 18-22 1. 16. Pemg CL, Lin HC, Wang K, Lai CR, Lee SD. Treatment of duodenal carcinoid by strip biopsy. J Clin Gastroenterol 1995;201:168-172. 17. Huang JJ, Yeo CJ, Sohn TA, Lillemoe KD, Sauter PK, Coleman J, Hruban RH, Cameron JL. Quality of life and outcomes after pancreaticoduodenectomy. Ann Surg 2000;23 1: 890-898. 18. Tsiotos GG, Farnell MB, Sarr MG. Are the results of pancreatectomy for pancreatic cancer improving? World J Surg 1999;23:913-919.
Duodenal Carcinoid Tumors
593
19. Makhlouf HR, Burke AP, Sobin LH. Carcinoid tumors of the ampulla of Vater: A comparison with duodenal carcinoid tumors. Cancer 1999;85:1241-1249. 20. Hatzitheoklitos E, Buchler M, Freiss H, Poch B, Ebert M, Mohr W, Imaizumi T, Beger HG. Carcinoid of the ampulla of Vater. Clinical characteristics and morphologic features. Cancer 1994;73:1580-1588. 2 1. Morgan JG, Marks C, Heam D. Carcinoid tumors of the gastrointestinal tract. Ann Surg 1974;180:720-727. 22. Shebani KO, Souba m, Finkelstein DM, Stark PC, Elgadi KM, Tanabe KK, Ott MJ. Prognosis and survival in patients with gastrointestinal tract carcinoid tumors. Ann Surg 1999; 229:815-823. 2 3. Marshall JB, Bodnarchuk G. Carcinoid tnmors of the gut. Our experience over three decades and review of the literature. J Clin Gastroenterol 1993;16:123-129.
