The evolving management of small bowel adenocarcinoma

Abstract

Background: Small bowel adenocarcinoma (SBA) is rare despite the fact that the small bowel represents the longest part and has the largest surface of all alimentary tract sections. Its incidence is 50-fold lower than that of colorectal carcinoma. It is often diagnosed at an advanced stage due to atypical and late symptoms, its low index of suspicion, difficult endoscopic access and poor detection by radiological imaging, resulting in impaired outcome. Due to its rarity and being molecularly a unique intestinal cancer, data regarding its optimal management are relatively sparse.

Material and methods: A PubMed search was performed to identify relevant manuscripts that were recently published. Emerging data regarding the pathogenesis, the diagnosis and the treatment of SBA that resulted from recent research are discussed in this comprehensive review.

Results: Genomic analysis has demonstrated that SBA is a molecularly unique intestinal cancer. Double balloon enteroscopy and capsule endoscopy are novel techniques which may result in earlier diagnosis and consequently in improvement of the generally poor prognosis. For clinically localized disease, the quality of surgery has recently been defined, with removal of at least 8–10 lymph nodes correlating with improved prognosis. Moreover, adjuvant chemotherapy seems to improve outcome of stage III disease. The combination of a fluoropyrimidine and oxaliplatin appears to be the most effective systemic chemotherapy for disseminated disease. Genomic profiling can identify potentially targetable genomic alterations in a significant proportion of SBA patients. The role of administration of targeted agents or immune checkpoint inhibitors is still unknown and subject of ongoing clinical trials. In the common case of peritoneal metastases, recent studies have shown that cytoreductive surgery and intraoperative hyperthermic intraperitoneal chemotherapy may be an attractive treatment option in selected patients.

Conclusions: SBA is a rare and unique malignancy, whose diagnostic approach and treatment are evolving, resulting in improved outcome.

Introduction

Approximately 5% of all gastrointestinal malignancies arise from the small bowel [1]. Small bowel adenocarcinoma (SBA) accounts for 30–45% of small bowel malignancies [2–6]. Although, adenocarcinomas have historically represented the most common histological subtype, recently the following histological distribution of small bowel malignancies was observed: carcinoid (44%), adenocarcinoma (33%), lymphoma (15%) and gastrointestinal stromal tumor/sarcoma (8%) [1,6]. Moreover, the distribution of histological subtypes varies across the small intestine, with adenocarcinoma representing the most common cancer of the duodenum [7].

According to the EUROCARE data, the estimated number of annual new cases of SBA in Europe is 3600, resulting in an estimated incidence rate of 5.7 new cases per million persons per year [8]. In the USA, the annual incidence of SBA has been estimated to be 3140 new cases [2]. The incidence of SBA seems to be increasing. In the USA, the estimated age-standardized incidence rose from 0.57 per 100,000 in 1973 to 0.73 per 100,000 in 2004 [6] and in the Netherlands from 0.5 per 100,000 in 1999 to 0.7 per 100,000 in 2013 [9]. The duodenum is the most frequently involved segment, being involved in 46–82% of the cases, followed by the jejunum (11–31%) and ileum (7–21%) [10–13].

One of the more interesting observations regarding SBA relates to its 50-fold lower incidence than the large bowel adenocarcinoma. This discrepancy occurs despite the fact that the small intestine represents approximately 75% of the length and 90% of the surface area of the alimentary tract [14].

The management of SBA is evolving. A PubMed search was performed to identify relevant manuscripts that were published recently, especially during the last decade. Emerging data regarding the pathogenesis, the diagnosis and the treatment of SBA that resulted from recent research are discussed in this comprehensive review.

Aetiology and pathogenesis

In contrast to colorectal cancer, studies on the pathogenesis of SBA are constrained by the rarity of the disease. Alcohol consumption and smoking have been associated with an increased risk of SBA. Other studies have reported an increased risk among high consumers of sugar, refined carbohydrates, red meat or smoked food, while a reduced risk was observed with higher intake of coffee, fish, fruit and vegetables [10].

The earlier noted apparent resistance of the small intestine to carcinogenesis may be explained by different hypotheses. However, limited experimental evidence exists to support any specific theory. Proposed hypotheses include: 1) rapid turnover of small intestine epithelium which precludes the accumulation of genetic damage, 2) increased lymphoid tissue in the small intestine, providing increased mucosal immune surveillance and 3) the inherent nature of the small intestine and its contents, which permits less exposure to carcinogenic agents in our diet as a result of rapid transit time, a dilute alkaline environment and lack of bacterial degradation activity [7]. Moreover, the epithelial cells of the small bowel are equipped with a wide range of microsomal enzymes, including benzopyrene hydroxylase, which may protect them against food-derived carcinogens [15].

The majority of cases are sporadic in nature, although a number of inherited cancer syndromes, such as hereditary non-polyposis colorectal cancer (HNPCC, Lynch syndrome), familial adenomatous polyposis and Peutz–Jeghers syndrome, are associated with an increased risk [11]. The two most common conditions linked to sporadic SBA, Crohn’s disease and celiac disease, are both associated with small bowel inflammation. SBA preferentially develops on sites of longstanding inflammation. A meta-analysis reported a relative risk for patients with Crohn’s disease 33.2 times greater than that of the general population [16]. The risk from Crohn’s disease reflects both the location of small bowel involvement, with 70% of cancers arising in the ileum, and the duration of disease, with an approximate risk of SBA of 2% after 25 years [17–19]. Other risk factors for the development of SBA in Crohn’s disease are a matter of debate, requiring further large cohort study analysis and include smoking, male gender and old age, proximal small bowel disease, use of 6-mercaptopurine, corticosteroids, azathioprine or TNF-alpha antibodies, young age at the time of diagnosis and surgically created non-functional small bowel loops [18].

A study comparing global DNA copy number alterations between small bowel, gastric and colorectal cancers suggested that SBAs are more similar to colorectal than to gastric cancers [20]. In addition, SBA demonstrates similar rates of mismatch repair system deficiency (MMR-D, 5–35%) as seen in colorectal cancer, which is characterized in the tumor by microsatellite instability [10,17,21,22]. Data from the literature suggest similar common carcinogenesis pathways in SBA and colorectal adenocarcinoma with molecular alterations, such as KRAS mutations, 18q loss and p53 loss [11,17,23,24]. However, striking difference exists in the rate of APC mutations, with reported percentages of 0–27% [11,22,25]. The lack of APC mutations in conjuncture with the infrequency of small bowel adenomas may suggest that the incidence difference between SBA and colorectal adenocarcinoma may reflect a difference in the early initiation phase of carcinogenesis. However, in a most recent large-scale comprehensive genomic analysis SBA appeared to be a molecularly unique intestinal cancer (Table 1) [22]. The frequency of genomic alterations seen SBA (n = 317) demonstrated distinct differences in comparison with either colorectal cancer (n = 6353; APC: 26.8 vs. 75.9%, p < .001; CDKN2A: 14.5 vs. 2.6%, p < .001) or gastric carcinoma (n = 889; KRAS: 53.6 vs. 14.2%, p < .001; APC: 26.8 vs. 7.8%, p < .001; SMAD4: 17.4 vs. 5.2%, p < .001). Although overall BRAF mutations were seen in similar percentages in colorectal adenocarcinoma and SBA (7.6 and 9.1%), V600E mutations were much less common in SBA, representing only 10.3% of BRAF-mutated cases. The ERBB2/HER2 point mutations (8.2%), microsatellite instability (7.6%) and high tumor mutational burden (9.5%) were all enriched in SBA. Significant differences were noted in the molecular profile of unspecified SBA compared with duodenal adenocarcinoma, as well as in inflammatory bowel disease-associated SBA. Targetable alterations in several additional genes, including PIK3CA and MEK1, and receptor tyrosine kinase fusions, were also identified in a significant number of patients.

Table 1. The significant differences in detection of the most common genomic alterations in small bowel adenocarcinoma when compared with colorectal cancer and gastric cancer demonstrate the molecular uniqueness of small bowel adenocarcinoma [22].

Mutations	SBA (%)(n = 317)	Colorectalcancer (%) (n = 6353)	Gastriccancer (%)(n = 889)
TP53	58	75	58
KRAS	54	52	14
APC	27	76	8
SMAD4	17	15	5
PIK3CA	16	18	13
CDKN2A	15	3	16
ARID1A	12	7	24
ERBB2/HER2	10	10	5
BRAF	9	8	2

The statistically significant differences in detection of each genomic alteration between small bowel adenocarcinoma (SBA) and colorectal or gastric cancer are marked by shading of the box.

Presentation and diagnosis

The median age at diagnosis is approximately 66 years, with over 85% of patients presenting after the age of 50 years [11,12,26–28]. There does not seem to be a clear gender preference [11,26,27]. The clinical presentation and diagnosis of SBA are usually delayed. While in approximately half of the patients with duodenal adenocarcinoma are asymptomatic, the majority of patients with jejunoileal adenocarcinoma is symptomatic [13]. The symptoms are initially rather nonspecific: abdominal pain, nausea, vomiting, weight loss and gastrointestinal bleeding [10,11,13,17,29]. Bowel obstruction is mainly observed in cases of jejunal or ileal tumor [27].

Nonspecific clinical symptoms coupled with the limited sensitivity of radiographic enteroclysis and conventional computed tomography for the detection of small bowel neoplasms led to the marked delay of the diagnosis. Multiphasic dynamic studies may have the potential to improve the diagnostic accuracy of multidetector computed tomography for small bowel neoplasms [30]. Only tumors in the proximal duodenum and the very distal ileum can be approached by conventional endoscopy. Newer investigation tools, such as computed tomography enteroclysis, magnetic resonance enteroclysis, wireless capsule endoscopy and double balloon enteroscopy now allow for an extensive exploration of the small bowel and should thus make early diagnosis possible [31–37]. Capsule endoscopy allows carrying out a complete small bowel exploration as an outpatient procedure. However, it should not be performed in the context of sub-occlusion. Double balloon enteroscopy can be used for the investigation of a wide range of small bowel pathologies [37]. Nevertheless, this procedure is less convenient than capsule endoscopy and should be used only if a biopsy or preoperative tattoo is required.

Final diagnosis is established after histological examination, usually after resection of the involved small bowel segment and less often after biopsies during endoscopy. Computed tomography of the chest and abdomen is recommended for staging of the disease. Upper and lower gastrointestinal endoscopy may be indicated to look for other tumors suggesting a predisposing genetic disease. A baseline plasma CEA and CA 19.9 assay should be performed, especially in cases of advanced disease, since the levels of these markers are of prognostic value [38]. In the context of a predisposing genetic disease or Crohn’s disease, a full small bowel exploration should be performed with computed tomography enteroclysis or with capsule endoscopy after small bowel stenosis has been excluded, to detect, if any, synchronous tumors [10]. Further, the patient may be tested for predisposing diseases like Crohn’s disease, Celiac disease and Lynch syndrome when indicated.

Prognosis

The 5-year overall survival for all SBA patients ranges from 14 to 33% [17]. The prognosis is mostly related to the disease stage. Disease staging is quite similar to that of colorectal cancer [39]. Stage I disease includes tumors which invade lamina propria, submucosa or muscularis propria in the absence of nodal or distant metastases. Stage II disease includes tumors with invasion of the subserosa, peritoneum or other organ or structures without apparent metastases. Stage III disease represents tumors with lymph node metastases, while stage IV disease includes all cases with distant metastases, such as liver, lung, peritoneal and distant, non-regional, lymph node metastases. Approximately, 6–12% of the patients have stage I disease at diagnosis, 27–37% stage II, 21–27% stage III and 32–37% stage IV [6,11,26,28]. This stage distribution contrasts with that of colon cancer, in which more patients (20%) present with stage I disease and less with stage IV disease (20%). This reflects most probably the delay in diagnosis. The 5-year overall survival is 50–60% for stage I, 40–55% for stage II, 10–40% for stage III and 3–5% for stage IV [2,3,10,12,26,28,29,40]. Moreover, among the stage III patients the number of lymph nodes involved is related to 5-year disease-specific survival, being 58% when less than three and 37% when three or more lymph nodes contain metastatic disease (p < .01) [12]. The prognosis of SBA appears to be intermediate between that of colon and gastric cancers.

In addition to stage, a number of additional factors have been associated with poor prognosis and include male gender, older age, impaired performance status, black ethnicity, symptomatic at diagnosis, duodenal location, low serum albumin, high CEA or CA 19.9, high lactate dehydrogenase, poor tumor differentiation and positive margins [6,13,27,38]. Approximately, 33–35% of the patients have poorly differentiated tumors, which is significantly higher than the 21% observed in colon cancer [11,26].

As seen with other tumor types, one of the most robust prognostic markers for resected cases is lymph node sampling. Recent data from the SEER database has demonstrated markedly improved outcomes for patients with increased nodal sampling, with one report finding eight lymph nodes as the optimal number of harvested lymph nodes [12]. For cases in which eight or more lymph nodes are assessed, 5-year cancer-specific survival improved significantly from 65.3 to 80.3% for stage I, from 55 to 69.9% for stage II and from 40 to 45.1% for stage III disease (p < .001) [12]. The fact that patients with duodenal adenocarcinoma appear to have a worse prognosis than patients with jejunal or ileal adenocarcinoma may be partly explained by understaging and incomplete lymph node harvest at the time of surgery [12]. When at least 8 lymph nodes were removed, there was no significant difference in survival between jejunoileal and duodenal adenocarcinoma for each disease stage, while with more limited lymph node dissection there were considerable differences between both tumor locations [12]. In another analysis of the SEER database which excluded duodenal adenocarcinoma [28], multivariate analysis demonstrated that removal of at least 10 lymph nodes was associated with significantly increased survival in patients with jejunoileal adenocarcinoma (p = .0009). This improvement was only seen in stage II disease and not in stage I and III disease. In a recent study [41], increasing lymph node assessment was associated with improved survival in patients with putative node negative (stage II) duodenal adenocarcinoma. Whether, the survival benefit after adequate lymph node dissection is due to stage migration or a real phenomenon remains to be answered. The fact that in the last two studies [28,41] survival improves with increasing lymph node harvest in stage II disease suggests stage migration as a cause, but on the other hand, in the third study [12] removal of more than seven lymph nodes was associated with significantly increased survival also in stage III disease (p < .001), suggesting a real benefit of adequate lymph node dissection. In one of the analyses of the SEER database [12], the positive to total lymph node ratio was an incrementally better predictor of survival than stratification by the number of positive lymph nodes among patients with stage III disease, also stressing the need for proper lymph node dissection. Adjuvant systemic chemotherapy could not compensate for inadequate lymph node dissection in patients with duodenal adenocarcinoma [41].

In a recent population-based comparison using the SEER database, even after accounting for lymph node sampling, SBA demonstrated worse cancer-specific survival stage for the stage than colon cancer [26]. Interestingly, even when comparing stage I cases with adequate lymph node sampling, 5-year cancer-specific survival was significantly lower for adenocarcinomas of the jejunum and ileum. Such data suggest that a fundamental biologic difference may exist between adenocarcinomas of the colon and those of the small bowel.

Treatment

Localized disease

Surgical resection is the mainstay of therapy for locoregional disease. As addressed above, data investigating the effect of lymph node assessment has clearly shown that SBA, and in particular duodenal adenocarcinoma, is markedly understaged and that an increased number of nodes for histological assessment is associated with improved outcome [12,28,41]. Such data support a more extensive surgical resection in order to harvest an adequate number of lymph nodes.

For duodenal tumors, a Whipple resection should be performed for a tumor located in the second segment of the duodenum or for an infiltrating tumor in the proximal or distal duodenum. Additionally, resection of the periduodenal, peripancreatic and hepatic lymph nodes should also be performed, as well as resection of the right side of the celiac and superior mesenteric arteries. A duodenal resection alone could be performed for a proximal duodenal tumor or a distal duodenal tumor with no infiltration of adjacent organs, despite the fact that this procedure is associated with poor prognosis [42]. An R0 resection is to be preferred, as R1 or R2 resections are strongly associated with poor prognosis [43]. For jejunal and ileal tumors, an R0 resection with lymph node resection and jejuno–jejunal or ileo–ileal anastomosis should be performed. If the last ileal loop or Bauhin’s valve is involved, an ileocecal resection or right hemicolectomy should be performed with ligation of the ileocolic artery so as to allow for adequate lymph node resection.

The relapse pattern for SBA is predominantly systemic, with one large retrospective study reporting distant and locoregional relapse accounting for 86 and 18% of all recurrences, respectively [29]. Although, a higher rate of local recurrence is seen with duodenal primaries, systemic relapse still predominates [43]. These data provide a potential role for adjuvant systemic treatment.

According to the National Cancer Database, the use of adjuvant chemotherapy has been significantly increasing, with rates of 8% of patients in 1985 to 24% in 2005 [6] and of 24.2% in 1998 to 43.4% in 2011 [27]. In part, this likely reflects the poor outcome of high-risk patients who undergo resection, the known activity of systemic fluoropyrimidine-based chemotherapy in the metastatic setting and extrapolation from the proven benefit of adjuvant treatment in colorectal cancer. At present, no randomized studies evaluating the benefit of adjuvant chemotherapy in SBA have been conducted. A number of single-institution retrospective studies have not demonstrated a clear benefit from adjuvant therapy, but these studies have all been limited by their small sample size, selection bias favoring the use of adjuvant therapy in higher-risk patients and the possibility that an inadequate chemotherapy regimen may have been administered [29,44]. In a most recent analysis of patients with resected SBA stage I-III who were identified in the US National Cancer Data Base for the period 1998–2011, 1674 from the 4746 patients had received adjuvant systemic chemotherapy and had a significantly decreased risk of death (hazard ratio 0.74, p < .001) [27]. A propensity score-matched analysis, used to minimize the effects of confounding by indication in different treatment groups, demonstrated a significant decrease in risk of death by the use of adjuvant systemic chemotherapy for stage III disease (median overall survival 42.4 vs. 26.1 months, p < .001). Adjuvant systemic chemotherapy was associated with a non-significant trend toward better overall survival for stage I disease, stage II disease, tumors infiltrating adjacent structures (T4) and involved surgical margins. Hence, the proven benefit of adjuvant chemotherapy for stage III disease and the potential absence of such an effect in stage I and II disease underlines the need for adequate staging by appropriate lymph node dissection. Recently, an international randomized trial (BALLAD (benefit of adjuvant chemotherapy for SBA) study, NCT02502370, Table 5) has been designed to evaluate the benefit of adjuvant chemotherapy after curative surgery for SBA. Patients will be allocated to observation or adjuvant treatment either with 5-fluorouracil and leucovorin or a combination of 5-fluorouracil, leucovorin and oxaliplatin (FOLFOX). For locally advanced and unresectable primary tumors initial treatment with chemotherapy or chemoradiotherapy should be considered [10,16].

In duodenal cancer, its retroperitoneal location and consequential higher risk for locoregional failure have led to the frequent use of adjuvant chemoradiotherapy. In a review of the US National Cancer Data Base, 11% of SBA patients received radiotherapy with or without chemotherapy, mainly for duodenal primary tumors [2]. Radiation was more frequently added to adjuvant chemotherapy when the surgical margins were involved (15.9 vs. 9.1%, p < .001) [45]. Support for this approach is limited. While in one retrospective study adjuvant chemoradiotherapy was not associated with better survival of duodenal adenocarcinomas resected with a curative intent [42], another study demonstrated a trend toward improved 5-year overall survival for those patients with an R0 resection who received adjuvant or neoadjuvant chemoradiotherapy compared with patients who underwent surgery alone (83 vs. 53%; p = .07) [46]. In a most recent analysis of patients with resected non-metastatic duodenal adenocarcinoma who were identified in the National Cancer Data Base for the period 1998–2012 [45], 694 patients had received adjuvant systemic chemotherapy, while 550 patients had received adjuvant systemic chemoradiotherapy. A propensity score-matched analysis did not demonstrate survival advantage for patients treated with adjuvant chemoradiotherapy compared with those treated with adjuvant chemotherapy (median overall survival 48.9 vs. 43.5 months, p = .67). Even in high-risk cases, such as those with involved surgical margins, infiltration of adjacent organs, inadequate lymph node staging, lymph node involvement and poorly differentiated histology, chemoradiotherapy was not associated with a significantly better overall survival. Of further note, neoadjuvant chemoradiotherapy for duodenal adenocarcinomas has been shown to be safe, with a number of studies reporting evidence of robust tumor downstaging in the pathologic specimens [46–48]. This approach deserves further investigation as cases of locally advanced unresectable disease have been converted to resectable disease following neoadjuvant therapy [47].

Systemic metastases

SBA is often diagnosed at an advanced stage due to its low index of suspicion, vagueness of symptoms, difficult endoscopic access and poor detection by radiological imaging. Approximately, one-third of the patients with SBA is presented with metastatic disease at the time of diagnosis [2,3,9,48]. Although no randomized studies have been performed to demonstrate a benefit of systemic chemotherapy in patients with advanced, i.e., unresectable or metastatic, disease, six retrospective comparative studies (Table 2) [29,39,44,48–52], including each between 37 and 165 patients with advanced SBA, all demonstrated a significant survival benefit for systemic chemotherapy. In these studies, overall response rates of 6–50% and disease control rates of 37–67% were observed. Systemic chemotherapy was associated with significantly higher median progression-free survival and overall survival (6 months vs. 1 month and 9–19 vs. 2–13 months, respectively). Various other series have demonstrated similar outcome for primary systemic chemotherapy [53–56].

Table 2. Retrospective studies on systemic chemotherapy versus no chemotherapy for unresectable or metastatic small bowel adenocarcinoma.

	N		Median overall survival (months)
Study	(chemotherapy vs. no)	Regimen	Chemotherapy	No chemotherapy
[29]	71 (68–32)	NA	12	2	p = .02
[44]	105 (44–61)	Various regimens	19	13*	p = .035*
[49]	37 (16–21)	5-FU based	16	8	NA
[40]	87 (24–53)	5-FU +/− cisplatin, gemcitabine based, other	9	4	p = .01
[48]	165 (NA)	NA	15.5	3.3	p < .01
[50]	91 (40–51)	5-FU, 5-FU + cisplatin, FOLFIRI	11.8	4.1	p < .01
[51]	59 (46–13)	Various combination chemotherapies	61%*	27%^a	p = .04
[52]	71 (56–15)	5-FU + cisplatin, FOLFOX, FOLFIRI, gemcitabine	13	2	p = .01

N: number of patients; NA: not available; 5-FU: 5-fluorouracil.

*In multivariate analysis.

^a1-year overall survival.

The next issue is to determine which chemotherapeutic regimen is the most effective. As already mentioned, there are no prospective comparative studies. In five retrospective series [38,52,57–59], various chemotherapeutic regimens were compared (Table 3). Although, retrospective comparative studies may be significantly biased and the number of patients per regimen was usually considerably small, it seems that the best results in means of response, survival and toxicity have been recorded for FOLFOX.

Table 3. Retrospective comparison of chemotherapeutic regimens for unresectable or metastatic small bowel adenocarcinoma.

		Systemic chemotherapy
Study	N	Regimen	N	Overall response (%)	p value	Disease control (%)	Median PFS (months)	p value	Median OS (months)	p value	Grade III/IV toxicity (%)
[57]	20	all	–	21	–	79	8	–	14	–	45
		5-FU + cisplatin	15	20	NA	–	–	–	–	–	53
		5-FU + carboplatin	2	0	–	–	–	–	–	–	0
		FOLFOX	3	33	–	–	–	–	–	–	33
[58]	80	all	–	25	–	–	5	–	13	–	–
		5-FU + platinum	29	41	.02		9	<.01	15	.10
		5-FU or other	51	16			4		12
[38]	93	all		26	.18	74	7	^–	15	^–	33
		5-FU	10	0		50	8	.16	14	.25	0
		5-FU + cisplatin	16	31	–	69	5	.02^a	9	.04^a	75
		FOLFIRI	19	9	–	73	6	–	11	–	39
		FOLFOX	48	34	–	79	7	–	18	–	46
[59]	132	all	–	–	–	–	–	^–	–	^–	–
		5-FU	45	20	NA		5	.03^b	14	.16^b	–
		5-FU + cisplatin	13	38	–	–	4	–	13	–	–
		FOLFIRI	8	25	–	–	6	–	9	–	–
		FOLFOX	17	42	–	–	8	–	22	–	–
		other	17	21	–	–	3	–	8	–	–
[52]	56	all	–	–	–	–	–	–	–	–	–
		5-FU + cisplatin	17	56	.75	–	9	NS	13	NS	–
		FOLFIRI	18	55	–	–	10	–	16	–	–
		FOLFOX	11	35	–	–	8	–	15	–	–
		gemcitabine	10	20	–	–	6	–	11	–	–

N: number of patients; PFS: progression-free survival; OS: overall survival; NA: not available; 5-FU: 5-fluorouracil; FOLFIRI: folinic acid +5-fluorouracil + irinotecan; FOLFOX: folinic acid +5-fluorouracil + oxaliplatin; NS: not significant.

^aFOLFOX vs. 5-FU + cisplatin.

^bFOLFOX vs. 5-FU monotherapy.

Six prospective phase II studies have been reported (Table 4). In first study [60], the combination of 5-fluorouracil, mitomycin C and doxorubicin was associated with a poor response and disease control rate (19 and 31%) as well as a high incidence of severe toxicity (68%). As shown in Table 4, the best results were recorded with a combination of a fluoropyrimidine and oxaliplatin. With either FOLFOX [61,62] or capecitabine and oxaliplatin (CAPOX) [63], a response rate of 45–50%, a disease control rate of 80–90% and a median overall survival of 15–20 months were observed. The addition of irinotecan to the combination of a fluoropyrimidine and oxaliplatin has resulted in higher efficacy but also severe hematological toxicity in colorectal cancer studies [64–66]. It appears that a common polymorphism in the number of TA repeats (7 vs. 6) of the uridine diphosphate glucuronosyltransferase enzyme 1A1 (UGT1A1) is responsible for decreased inactivation of irinotecan’s active metabolite SN-38 and hence increased hematological toxicity [67,68]. UGT1A1 genotype-guided drug dosage, i.e., standard irinotecan dosage for 6/6 UGT1A1 genotype and lower irinotecan dosage for heterozygous 6/7 UGT1A1 genotype and homozygous 7/7 UGT1A1 genotype may result in overall decreased systemic toxicity, while the exposure to SN-38 remains similar in all genotype groups [69,70]. In a most recent study [71], doses of capecitabine, irinotecan and oxaliplatin (CAPIRINOX) were administered according to the UGT1A1 genotype in 33 patients with advanced SBA. While, 17 patients had the common 6/6 UGT1A1 genotype, in 10 patients the 6/7 genotype was found and in 6 the 7/7 genotype. This genotype-guided dosed chemotherapy resulted in an objective response rate of 38%, a disease control rate of 81%, a median progression-free survival of 8.9 months and a median overall survival of 13.4 months in 33 patients with advanced SBA. Grade III/IV hematological toxicity was observed in 42% of the patients, which is considerably less frequent than in prior studies with unselected patients. With this genotype-guided dosed chemotherapy hematological toxicity as well as response rates did not significantly differ among the three genotype groups.

Table 4. Prospective phase II studies on systemic chemotherapy for unresectable or metastatic small bowel adenocarcinoma.

Study	N	Systemic chemotherapy regimen	Overallresponse (%)	Diseasecontrol (%)	Median PFS(months)	Median OS(months)	Grade III/IVtoxicity (%)
[60]	38	5-FU + mitomycin C + doxorubicin	19	31	5	8	68
[63]	30^a	CAPOX	50	90	11	20	–
[61]	33	FOLFOX	49	85	8	15	39
[62]	24	FOLFOX	45	80	5	17	>38
[71]	33	CAPIRINOX	38	81	9	13	79
[79]	30^b	CAPOX + bevacizumab	48	83	9	13	–

N: number of patients; PFS: progression-free survival; OS: overall survival; 5-FU: 5-fluorouracil; CAPOX: capecitabine + oxaliplatin; FOLFOX: folinic acid +5-fluorouracil + oxaliplatin; CAPIRINOX: capecitabine + irinotecan + oxaliplatin, dose based on UGT1A1 genotype.

^a18 patients with small bowel adenocarcinoma and 12 patients with ampullary adenocarcinoma.

^b23 patients with small bowel adenocarcinoma and seven patients with ampullary adenocarcinoma.

In conclusion, systemic chemotherapy seems to be beneficial in patients with advanced SBA, while the combination of a fluoropyrimidine and oxaliplatin (FOLFOX or CAPOX) appears to be the most effective front-line regimen. The addition of irinotecan is feasible, but its benefit remains yet unclear. FOLFIRI and the combination of 5-fluorouracil and cisplatin are alternative regimens. Further, systemic chemotherapy with FOLFIRI appeared to be effective as second-line treatment after failure on platinum-based chemotherapy in a small series of patients [72].

Genomic profiling can identify potentially targetable genomic alterations in a significant proportion of patients with SBA, whereas, the incidence of microsatellite instability and the higher tumor mutational burden suggest a potential role for immunotherapy [22]. The role of agents targeting vascular endothelial growth factor (VEGF), such as bevacizumab, regorafenib or epidermal growth factor receptor (EGFR) inhibitors such cetuximab or panitumumab, in the treatment of SBA has not been established yet, while ongoing studies are now underway (Table 5) [73]. The high percentage of tumors expressing both EGFR and VEGF suggests that patients with this rare malignancy may benefit from therapeutic strategies targeting EGFR and VEGF receptor [21], although in colorectal cancer no clear association has been found between EGFR and VEGF expression and the response to the EGFR and VEGF inhibitors [74–77]. Importantly, the rate of KRAS mutations in SBA is similar to that seen in colorectal cancer, and a handful of case reports have described responses to anti-EGFR therapy in patients with KRAS wild-type tumors [78]. At present, the results of only one such study have been reported. In a phase II study with CAPOX and bevacizumab for metastatic SBA and ampullary adenocarcinoma [79], in 48% of the 30 patients a response was observed, while median progression-free and overall survival were 8.7 and 12.9 months, respectively, and toxicity was acceptable. However, an exploratory analysis comparing the results of this study with those of the study with CAPOX alone [61], performed by the same institution, did not show significant difference in response rate and progression-free survival. It has to be noted that, although the study populations had similar characteristics, the number of patients enrolled was limited for powered statistical analysis. Treatment with immune checkpoint inhibitors, such the anti-PD1 monoclonal antibody pembrolizumab and the anti-PD-L1 monoclonal antibody avelumab has also been investigated. In a recent study [80], PD-1 and PD-L1 were highly expressed by most SBAs. Pembrolizumab has been proven highly effective in treatment-refractory MMR-D colorectal and non-colorectal tumors [81]. Phase II studies with pembrolizumab (NCT02949219) and avelumab (NCT03000179) are underway for patients with refractory SBAs, surprisingly regardless of the MMR status [73]. The results of ongoing studies with targeted agents, immune checkpoint inhibitors and novel drug formulations in the systemic treatment of SBA are eagerly awaited.

Peritoneal metastases

Peritoneal metastases affect 25–50% of the patients with metastatic disease from SBA [9,11,26,44,50,57–59,63,72], and are more common in tumors arising from the jejunum or ileum than in those arising from the duodenum [9]. There are no studies that specifically investigated the effect of systemic chemotherapy on peritoneal metastases of small bowel origin. In sub-analysis of a retrospective study regarding systemic chemotherapy in advanced disease [50], systemic chemotherapy also seems to be associated with better survival in patients with peritoneal metastases. However, the effect of systemic chemotherapy on peritoneal metastases from SBA may be less beneficial than for advanced disease in general. Clinical studies in colorectal cancer have demonstrated that peritoneal carcinomatosis is relatively resistant to systemic therapy compared to other metastatic sites, such as the liver [82–86]. An alternative treatment for isolated peritoneal metastases from SBA is cytoreductive surgery with hyperthermic intraperitoneal chemotherapy (HIPEC) [87]. This treatment modality has been increasingly used for peritoneal surface malignancies, such as pseudomyxoma peritonei, peritoneal mesothelioma and peritoneal metastases from colorectal, appendiceal, gastric and ovarian cancer [88]. Candidate patients for this treatment should have no or very limited liver metastases, have disease that can be macroscopically completely or almost completely (leaving very small deposits behind) resected and is in a general condition that allows for this major procedure [88,89]. The fact that cytoreductive surgery with HIPEC or another form of intraperitoneal chemotherapy may also be beneficial in peritoneal metastases from SBA is based on seven observational studies, including a total of 126 patients (Table 6) [90–96]. In these studies of cytoreductive surgery and intraperitoneal chemotherapy, median disease-free survival of 10–12 months, median overall survival of 16–47 months and grade III–V morbidity of 12–35% have been reported [90–96]. These data are quite comparable with those reported after cytoreductive surgery and HIPEC for peritoneal carcinomatosis of colorectal origin [88,97–99]. Given these promising survival results, efforts should be made to increase awareness regarding this treatment option. The relatively low number of patients in these studies demonstrates that awareness is currently lacking. In the Netherlands, 16 patients with peritoneal dissemination of SBA were treated with cytoreductive surgery and HIPEC over a 10 year time period [95], while the number of patients diagnosed with peritoneal dissemination of SBA was significantly higher in a Dutch epidemiological study, revealing 167 patients with the diagnosis of synchronous peritoneal metastases from SBA between 1999 and 2013 [9]. Approximately, 60% of these patients present with isolated peritoneal metastases without concurrent systemic (i.e., liver or lung) metastases (Dutch Cancer Registry) and are therefore potential candidates for this treatment modality. Since, data on metachronous peritoneal metastases from SBA are not available; the number of potential candidates is probably considerably higher [9].

Table 5. Ongoing clinical studies of systemic therapy for small bowel adenocarcinoma [73].

Drug regimen	Phase	Type of adenocarcinoma	Line of treatment	N	NCT identifier
5-FU vs. FOLFOX vs. observation	III	Small bowel	adjuvant	100	NCT02502370
Gemcatibine + oxaliplatin + erlotinib	I	Biliary, pancreatic, duodenal and ampullary	1st	28	NCT00987766
Panitumumab	II	Small bowel and ampullary^a	1st	27	NCT01202409
Nab-paclitaxel	II	Small bowel	≥2nd	10	NCT01730586
Pembrolizumab	II	Small bowel	≥2nd	40	NCT02949219
Avelumab	II	Small bowel	≥2nd	25	NCT03000179

5-FU: 5-fluorouracil; FOLFOX: folinic acid +5-fluorouracil + oxaliplatin; Nab-paclitaxel: nanoparticle albumin-bound paclitaxel; N: number of patients (estimated to be) enrolled.

^aKRAS wild-type only.

Table 6. Retrospective studies on cytoreductive surgery and perioperative intraperitoneal chemotherapy for small bowel adenocarcinoma with peritoneal metastases.

Study	N	Type of intraperitonealchemotherapy (N)	Median DFS (months)	Median OS (months)^a	Median OS (months)^b	Grade III/IV toxicity (%)
[90]	6	HIPEC + EPIC	–	16	–	–
[91]	6	HIPEC	–	30	–	–
[92]	31	HIPEC (21), EPIC (10)	–	47	–	35
[93]	17	HIPEC	–	18	37	12
[94]	16	HIPEC	10	31	–	25
[95]	31	HIPEC	–	36	51	26
[96]	16	HIPEC	11	25	–	25

N: number of patients; DFS: disease-free survival; OS: overall survival; HIPEC: intraoperative hyperthermic intraperitoneal chemotherapy; EPIC: early postoperative intraperitoneal chemotherapy.

^aFrom treatment with cytoreductive surgery and perioperative intraperitoneal chemotherapy.

^bFrom diagnosis of peritoneal metastases.

It has to be noted that the number of patients included in these studies is relatively small and there may be a significant bias in treatment results by patient selection and heterogeneity in treatment within and between the reported studies [90–96]. These variables include performance status, peritoneal tumor burden, completeness of cytoreductive surgery, intra-abdominal temperature, drug regimen, duration of intraperitoneal chemotherapy and the administration of (neo) adjuvant systemic chemotherapy. Limitations in available evidence stress the need for higher quality studies to confirm the promising results of cytoreductive surgery and HIPEC in selected patients with peritoneal metastases from SBA. Given the rarity of the disease, it is unlikely that this evidence will ever be provided by a prospective randomized study. A multi-institutional data registry on patients with peritoneal metastases from SBA will provide more insight in survival and morbidity outcomes in a larger cohort and prognostic patient, tumor and treatment-related variables that predict and influence survival and morbidity [87].

If a patient is no candidate for cytoreductive surgery and intraperitoneal chemotherapy due to extensive peritoneal disease, poor performance status and/or concurrent systemic disease, pressurized intraperitoneal aerosol chemotherapy (PIPAC) may be a future treatment option for symptomatic peritoneal disease [100]. This innovative locoregional treatment modality has shown promising preliminary results for peritoneal carcinomatosis of other origins with very limited toxicity [101–104].

Conclusions

SBA is a rare and unique malignancy, which is often diagnosed at an advanced stage due to atypical and late symptoms, its low index of suspicion, difficult endoscopic access and poor detection by radiological imaging. Its diagnostic approach and treatment are evolving. Double balloon enteroscopy and capsule endoscopy are novel techniques which may result in earlier diagnosis and consequently the improvement of the generally poor prognosis. Adequate lymph node dissection with an assessment of at least 8–10 lymph nodes correlates with improved prognosis. Adjuvant chemotherapy is increasingly being used to reduce the recurrence rate. The exact role of adjuvant chemotherapy and (neo) adjuvant radiotherapy has not been determined yet, but it seems that adjuvant chemotherapy is associated with significantly better outcome in stage III disease and probably not in stage I and II disease. The latter underlines once again the need for adequate staging with appropriate lymph node dissection. The combination of a fluoropyrimidine and oxaliplatin (FOLFOX or CAPOX) seems to be the most appropriate front-line systemic chemotherapy for disseminated disease. Genomic profiling can identify potentially targetable genomic alterations in the majority of patients with SBA, whereas the incidence of microsatellite instability and the higher tumor mutational burden suggest a potential role for immunotherapy. The role of administration of targeted agents or immune checkpoint inhibitors is still unknown and subject of various ongoing clinical trials. In the common case of peritoneal metastases, cytoreductive surgery and HIPEC may be an attractive alternative treatment option for selected patients.

Disclosure statement

The authors declare that they have no conflict of interest in any matter or financial affiliations to disclose.