Key points

Introduction

Although the small bowel represents 75% of the length and 90% of the overall mucosal surface of the gastrointestinal (GI) tract, it is a rare location for the development of neoplasms. Overall, small bowel tumors account for 3% to 6% of all GI neoplasms and 1% to 3% of all GI malignancies. Despite the identification of more than 40 different histologic types of small intestinal tumors, there are 4 major histologic subtypes: approximately 30% to 45% of small bowel tumors are adenocarcinomas, 20% to 40% neuroendocrine tumors, 10% to 20% lymphomas, and 10% to 15% sarcomas.

According to the US Surveillance, Epidemiology, and End Results (SEER) program, the rate of new cases, age-adjusted and based on 2009 to 2013 diagnoses, is 2.2 per 100,000 persons per year. Therefore, it is estimated that in 2016 approximately 10,000 people in the United States will be diagnosed with small intestinal cancer and approximately 1300 people will die of it. International data show that the incidence varies across countries, higher in Western counties and Oceania than in Asia, with an incidence ratio of 2 to 2.5 when US and Japanese populations are compared. A higher incidence rate is observed in the black US population compared with whites, and a high incidence rate is also reported among the Maori of New Zealand (approximately 4 cases per 100.000 per year) and Hawaiians. In most countries, the overall incidence of small bowel tumors is higher in men than in women; it starts increasing after the age of 40 to 45 years and tends to rise with age, until the age of 75. The median age at diagnosis is approximately 65 years.

Although some differences exist among studies, mostly concerning the incidence patterns of specific histologic subtypes, several studies consistently show an increasing overall incidence rate of small bowel neoplasms over time. The US SEER program reported that over the past 20 years (from 1992 to 2013) it increased from 1.5 to 2.2 cases per 100,000 inhabitants ( Fig. 1 ). It is, therefore, calculated that rates for new small bowel cancer cases have been rising on average 2.4% each year over the past 10 years. Similar time-trend figures were also reported in recent studies from European countries, such as the United Kingdom, Sweden, France, and Denmark. Currently, the reason for this increase remains largely unexplained. Taking into account the epidemiologic trends of predisposing diseases (ie, celiac disease and inflammatory bowel diseases), however, the behavioral and environmental risk factors associated with small bowel tumors (ie, obesity and cigarette smoking ) and the aging of populations, this trend is likely to continue, and small intestinal cancer is likely to become an increasing part of the gastroenterological workload in the next few years.

Age-adjusted incidence ratio of small bowel tumors between 1992 and 2013 in the Untied States based on SEER program.

( Modified from National Cancer Institute. SEER stat fact sheets: small intestine cancer. Available at: http://seer.cancer.gov/statfacts/html/smint.html . Accessed 25 March, 2016.)

Despite obvious differences among histologic subtypes, currently, the overall prognosis of small bowel tumors remains poor. Whereas cancers at other sites have increasing long-term survival rates over the past 2 decades, the US data from 1985 to 2000 showed no significant change in long-term survival rates for any of the 4 main histologic subtypes of small bowel tumors. Consistently, Lepage and colleagues showed no significant improvement in the survival rate over a 26-year period (1989–2001). Concerning small bowel tumors, earlier tumor stages at diagnosis (stage I and II), small tumor size, and curative resection have been identified as factors favoring overall survival. Conversely, poorly differentiated tumors, lymph node involvement, and distant metastases are the main factors predicting poorer prognosis. A majority of these factors largely depend on the timing of diagnosis, which is often delayed (delays of 3 years for benign tumors and 1.5 years for malignant tumors have been estimated). Some specific features of small bowel tumors, which grow slowly and extraluminally, remaining asymptomatic for years or presenting with nonspecific complaints, can contribute in delaying the diagnosis. Nevertheless, at least until recently, the technical difficulties in exploring the small bowel were one of the most important barriers in reaching a timely and definite diagnosis. Recent data suggest that the introduction of diagnostic techniques specifically dedicated to evaluation of the small bowel as well as new drugs for the treatment of specific histologic subtypes of tumors (ie, imatinib and sunitinib for GI stromal tumors [GISTs]) may change these figures. A recent study reported a shift in the diagnostic process of patients with small bowel tumors, with an increasing number of patients being diagnosed through CE. Consistently, a study from Portugal showed that in the time frame 2010 to 2014, more than 50% of patients diagnosed with small bowel tumors received the diagnosis by endoscopy or imaging techniques whereas only 24% of them received the final diagnosis at time of surgery. Moreover, Honda and colleagues underlined that in the time frame 2003 to 2011, when deep enteroscopy techniques were readily available, the median diagnostic delay was shorter than previously reported, 4.3 months for small bowel tumors overall and 4.9 months and 3.6 months for benign and malignant tumors, respectively. Consequently, an increased rate of patients received elective surgical intervention, instead of undergoing surgery for acute abdomen due to obstructive symptoms. Furthermore, the SEER program reported encouraging data: focusing on recent trends (2004–2013), despite the increasing incidence of small bowel tumors, the mortality rate has remained stable and an increasing overall 5-year survival rate has been observed. Similarly, according to EUROCARE-5 data, overall 5-year survival increased from 40.5% (1999–2001) to 48.7% (2005–2007).

The available evidence on the role of new tools in the diagnosis and management of small bowel tumors is reviewed, highlighting their strengths and limitations and how their combined use can possibly contribute in modifying both detection and prognosis of small bowel neoplasms.

Introduction

Although the small bowel represents 75% of the length and 90% of the overall mucosal surface of the gastrointestinal (GI) tract, it is a rare location for the development of neoplasms. Overall, small bowel tumors account for 3% to 6% of all GI neoplasms and 1% to 3% of all GI malignancies. Despite the identification of more than 40 different histologic types of small intestinal tumors, there are 4 major histologic subtypes: approximately 30% to 45% of small bowel tumors are adenocarcinomas, 20% to 40% neuroendocrine tumors, 10% to 20% lymphomas, and 10% to 15% sarcomas.

According to the US Surveillance, Epidemiology, and End Results (SEER) program, the rate of new cases, age-adjusted and based on 2009 to 2013 diagnoses, is 2.2 per 100,000 persons per year. Therefore, it is estimated that in 2016 approximately 10,000 people in the United States will be diagnosed with small intestinal cancer and approximately 1300 people will die of it. International data show that the incidence varies across countries, higher in Western counties and Oceania than in Asia, with an incidence ratio of 2 to 2.5 when US and Japanese populations are compared. A higher incidence rate is observed in the black US population compared with whites, and a high incidence rate is also reported among the Maori of New Zealand (approximately 4 cases per 100.000 per year) and Hawaiians. In most countries, the overall incidence of small bowel tumors is higher in men than in women; it starts increasing after the age of 40 to 45 years and tends to rise with age, until the age of 75. The median age at diagnosis is approximately 65 years.

Although some differences exist among studies, mostly concerning the incidence patterns of specific histologic subtypes, several studies consistently show an increasing overall incidence rate of small bowel neoplasms over time. The US SEER program reported that over the past 20 years (from 1992 to 2013) it increased from 1.5 to 2.2 cases per 100,000 inhabitants ( Fig. 1 ). It is, therefore, calculated that rates for new small bowel cancer cases have been rising on average 2.4% each year over the past 10 years. Similar time-trend figures were also reported in recent studies from European countries, such as the United Kingdom, Sweden, France, and Denmark. Currently, the reason for this increase remains largely unexplained. Taking into account the epidemiologic trends of predisposing diseases (ie, celiac disease and inflammatory bowel diseases), however, the behavioral and environmental risk factors associated with small bowel tumors (ie, obesity and cigarette smoking ) and the aging of populations, this trend is likely to continue, and small intestinal cancer is likely to become an increasing part of the gastroenterological workload in the next few years.

Age-adjusted incidence ratio of small bowel tumors between 1992 and 2013 in the Untied States based on SEER program.

( Modified from National Cancer Institute. SEER stat fact sheets: small intestine cancer. Available at: http://seer.cancer.gov/statfacts/html/smint.html . Accessed 25 March, 2016.)

Despite obvious differences among histologic subtypes, currently, the overall prognosis of small bowel tumors remains poor. Whereas cancers at other sites have increasing long-term survival rates over the past 2 decades, the US data from 1985 to 2000 showed no significant change in long-term survival rates for any of the 4 main histologic subtypes of small bowel tumors. Consistently, Lepage and colleagues showed no significant improvement in the survival rate over a 26-year period (1989–2001). Concerning small bowel tumors, earlier tumor stages at diagnosis (stage I and II), small tumor size, and curative resection have been identified as factors favoring overall survival. Conversely, poorly differentiated tumors, lymph node involvement, and distant metastases are the main factors predicting poorer prognosis. A majority of these factors largely depend on the timing of diagnosis, which is often delayed (delays of 3 years for benign tumors and 1.5 years for malignant tumors have been estimated). Some specific features of small bowel tumors, which grow slowly and extraluminally, remaining asymptomatic for years or presenting with nonspecific complaints, can contribute in delaying the diagnosis. Nevertheless, at least until recently, the technical difficulties in exploring the small bowel were one of the most important barriers in reaching a timely and definite diagnosis. Recent data suggest that the introduction of diagnostic techniques specifically dedicated to evaluation of the small bowel as well as new drugs for the treatment of specific histologic subtypes of tumors (ie, imatinib and sunitinib for GI stromal tumors [GISTs]) may change these figures. A recent study reported a shift in the diagnostic process of patients with small bowel tumors, with an increasing number of patients being diagnosed through CE. Consistently, a study from Portugal showed that in the time frame 2010 to 2014, more than 50% of patients diagnosed with small bowel tumors received the diagnosis by endoscopy or imaging techniques whereas only 24% of them received the final diagnosis at time of surgery. Moreover, Honda and colleagues underlined that in the time frame 2003 to 2011, when deep enteroscopy techniques were readily available, the median diagnostic delay was shorter than previously reported, 4.3 months for small bowel tumors overall and 4.9 months and 3.6 months for benign and malignant tumors, respectively. Consequently, an increased rate of patients received elective surgical intervention, instead of undergoing surgery for acute abdomen due to obstructive symptoms. Furthermore, the SEER program reported encouraging data: focusing on recent trends (2004–2013), despite the increasing incidence of small bowel tumors, the mortality rate has remained stable and an increasing overall 5-year survival rate has been observed. Similarly, according to EUROCARE-5 data, overall 5-year survival increased from 40.5% (1999–2001) to 48.7% (2005–2007).

The available evidence on the role of new tools in the diagnosis and management of small bowel tumors is reviewed, highlighting their strengths and limitations and how their combined use can possibly contribute in modifying both detection and prognosis of small bowel neoplasms.

Small bowel capsule endoscopy

The detection rate of small bowel tumors via CE ranges from 1.5% to 9% and from 3% to 5% in studies collecting more than 1000 patients. The rate of small bowel tumors has increased in patients undergoing CE for obscure GI bleeding. Among them, the detection of tumors is higher in those presenting with obscure-overt bleeding and in those under the age of 50. Although vascular lesions are the most common finding in older patients, in younger patients, small bowel CE often identifies inflammatory or neoplastic changes. These data contributed to raise awareness of the importance of prioritizing small bowel evaluation in young patients with obscure GI bleeding. Moreover, CE is perceived as a noninvasive test and it is better tolerated than other diagnostic methods for the study of the small bowel, mainly compared with DAE. Furthermore, CE offers the opportunity of a detailed and panoramic view of the entire small bowel. For all these reasons, CE often represents the first diagnostic test performed in patients with small bowel tumors.

Nevertheless, in a large retrospective study of more than 690 patients, Lewis and colleagues calculated that the CE miss rate for neoplasms was 18.9%. In the same study, however, the miss rates for vascular lesions and ulcers were 5.9% and 0.5%, respectively. Several studies have recently reported cases of small bowel tumors missed by CE and identified by means of DAE and/or dedicated small bowel cross-sectional imaging techniques. CE can miss neoplastic lesions because of limited visualization in the setting of poor bowel preparation, which occurs in approximately 10% to 15% of examinations, regardless of preparation regimen. In addition, the capsule is passively propelled by natural peristalsis, which can lead to quick capsule transit, mostly in the distal duodenum and proximal jejunum, where 25% to 30% of small bowel tumors are located. Small bowel motility studies have revealed that short bursts of a very fast movement (>15 cm/min) occur for approximately 45 minutes after pyloric passage, probably reflecting phase III of the migrating motor complex. Furthermore, it has been demonstrated that the sensitivity of CE in identifying glass beads sewn into the intestine of dogs is inferior to that of push enteroscopy (64% and 37%, respectively). Honda and colleagues found that the diagnostic yield of CE for small bowel tumors located in the proximal small bowel is significantly lower than that observed for small bowel tumors located in the distal small bowel (73% vs 90%; P = .040). Therefore, the visualization of the ampulla of Vater during the examination (which indicate an adequate proximal small bowel examination) should be considered a CE key quality indicator. In this regard, a recently developed capsule with a 360° lateral panoramic showed promising results.

Even when focusing the attention on the lower end of the small bowel (Hatzaras and colleagues reported that up to 30% of small bowel tumors are located in the distal small bowel), CE can miss neoplastic lesions mostly because the quality of small bowel preparation seems to decrease along the small bowel. Additionally, the capsule does not reach the ileocecal valve during recording time in approximately 15% to 20% of cases. Theoretically, prokinetic agents can be used to shorten the gastric transit time to improve completion rate. Therefore, various prokinetics (ie, erythromycin, mosapride, metoclopramide, and lubiprostone) have been investigated for bowel preparation of CE. Nevertheless, adding prokinetic agents to the bowel preparation does not seem to enhance either the completion rate or the diagnostic yield. Therefore, routine prokinetic administration is not recommended. Conversely, the targeted administration of a prokinetic, based on the systematic and repeated checking of capsule position through a real-time viewer, seems to significantly increase the rate of complete enteroscopy.

The passive uncontrollable capsule movement poses significant difficulty in differentiating a real submucosal mass from innocent mucosal bulges. Up to half of all small bowel malignancies, such as GISTs and neuroendocrine tumors ( Fig. 2 ), arise from the submucosal layer of the intestinal wall ; hence, they appear as bulges with certain surface morphologic characteristics. Girelli and colleagues attempted to devise and validate an index based on these characteristics (smooth, protruding lesion index on CE [SPICE]) to discriminate submucosal masses from innocent bulges. Although SPICE presents favorable diagnostic accuracy, it has not yet been incorporated into any proprietary reviewing software. Shyung and colleagues described a scoring system to interpret CE findings in small bowel tumors. The proposed tumor score comprises 5 components: bleeding, mucosal disruption, irregular surface, color, and white villi. These can be scored for probability of mass lesions seen at CE. To that end, it has also been shown that applying 3-D reconstruction to the standard 2-D video reading platform does not improve the performance of expert small bowel capsule endoscopy readers, although it significantly increases the performance of novices in distinguishing masses from bulges.

Small bowel carcinoid identified by CE ( arrows ).

Although capsule retention occurs in an estimated 2% of all patients undergoing this examination, small bowel tumors are one of the major diseases associated with this complication. Taking into account that the retention is generally asymptomatic and most patients with a small bowel tumor undergo surgical resection (with the possibility of easy retrieval of a capsule), however, in cases of suspected small bowel tumor, patency investigation before CE is still not recommended by current guidelines.

In light of the limitations of CE in the specific setting of small bowel tumors, several technical developments have been trialed to improve the capability of the device in identifying and classifying small bowel tumors. Recently, Fuji Intelligent Chromo Endoscopy (FICE), a spectral estimation technology based on arithmetical processing of ordinary images (Fujinon. Tokyo, Japan) that is integrated into proprietary software, has been tested. Unfortunately, it does not seem to improve the visualization of small bowel tumors. Augmented Live-Body Image Color-Spectrum Enhancement (ALICE), another spectral imaging technique for the MiroView system (IntroMedic, Seoul, Korea), has shown limited and equally disappointing data. Infrared fluorescence endoscopy, in conjunction with an infrared fluorescent-labeling contrast agent, is a well-known technique used for efficient early-stage cancer detection. There is currently a screening capsule prototype ( Fig. 3 ), which is able to detect infrared fluorescence emitted by indocyanine green (ICG) fluorophore dye. Rather than images, the capsule works as a high-sensitivity fluorometer that records fluorescence levels throughout the small intestine. Ex vivo experiments on ICG-impregnated swine intestine have shown that the prototype system is able to detect low concentrations of ICG in the nanomolar and micromolar regions. This could prove useful in detecting early cancer in the small bowel. Hyperspectral imaging is emerging as another promising technology for medical diagnosis. Hyperspectral images present large amounts of information from the mucosal surface, which are captured by sensors. Using this information and a set of complex classification algorithms, it is possible to determine the material or substance located at each pixel. This technique has been used by neurosurgeons in the process of brain tumor resection, avoiding the excessive extraction of normal tissue and unintentionally leaving small remnants of tumor. Such precise delineation of tumor boundaries improves the results of surgery. Preliminary results indicate that it is possible to discriminate between healthy and tumor tissues in the brain by exclusively processing the spectral information of the tissues.

Prototype of the infrared fluorescence-based cancer screening capsule for the small bowel.

Dedicated small bowel cross-sectional radiologic imaging

Small bowel barium follow-through studies were once the mainstay of small bowel imaging, but these have now largely been superseded by dedicated small bowel cross-sectional imaging techniques, namely magnetic resonance (MR) and CT. Because collapsed bowel loops can hide lesions or may mimic small bowel diseases, both techniques require luminal distension of the intestinal loops to identify small bowel lesions. This is achieved by administering luminal contrast agents.

Two types of oral contrast agents can be used for CT: positive and neutral. The former ( Fig. 4 ) provides an excellent background, facilitating detection of lesions protruding into the lumen and identification of mucosal details. It can hamper the detailed evaluation of mural features, however, when intravenous contrast is administered. Conversely, the latter (neutral or near-water attenuation contrast agent) ( Fig. 5 ) allows depiction of mucosal folds and assessment of mucosal enhancement and provides optimal bowel dilation. Oral contrast media for MR can be classified as positive, negative, or biphasic, according to the signal intensity of bowel lumen. With positive paramagnetic compounds (ie, water-based gadolinium solutions), the lumen signal is high (it appears as white), potentially masking the wall enhancement obtained by intravenous gadolinium administration. Negative contrast agents are superparamagnetic (ie, water-based solutions of iron oxide particles coated with silicone) and produce low-signal intensity (the lumen appears dark), allowing better visualization of the bowel wall and mesenteric fat. Finally, biphasic contrast agents (aqueous solutions of hyperosmotic compounds [ie, polyethylene glycol]) produce both the positive effect on T2-weighted sequences and the dark lumen appearance on T1-weighted sequences; therefore, they are the most versatile agents, which are frequently used when small bowel tumors are suspected.

Coronal CT enterography image with positive oral contrast demonstrates a short segment of proximal ileal thickening ( red arrows ) due to small bowel lymphoma, with low-volume but size-significant mesenteric nodes ( blue arrows ).

CT enterography axial ( A ) and coronal ( B ) images demonstrate a carcinoid mass within the small bowel mesentery ( red arrows ) with flecks of calcification; white arrow shows pelvic ascites.

For both CT and MR, luminal contrast agents can be administered either orally (enterography) or through a nasojejunal tube (enteroclysis). Enteroclysis may offer better bowel distension (the contrast is infused by means of a pressure-controlled electric pump at a rate of 100–160 mL/min), but it requires nasojejunal tube placement beyond the ligament of Treitz. On the other hand, enterography is better tolerated, but adequate loop distension strictly depends on patient compliance, because patients are asked to drink a large volume of liquids (eg, 1000–2000 mL) in a short time frame (approximately 1.5–2 h), according to a predefined schedule, which varies across centers. The choice between the 2 approaches remains controversial and depends on several factors, such as expected disease location, local radiology practice, and the diagnostic algorithms of different centers. Nevertheless, enterography is frequently used in everyday clinical practice, given its better patient tolerance and acceptability.

In conjunction with luminal contrast agents, intravenous iodinated contrast agents are also administered to optimize the assessment of small bowel during CT; although gadolinium-based contrast agents are not always required during MR studies, they are generally used when assessing inflammatory or potentially malignant small bowel processes ( Fig. 6 ). Antispasmodic agents (ie, hyoscine butylbromide or glucagon) are also routinely administered intravenously at time of examination to reduce artifact due to peristaltic contractions.

Jejunal small bowel GISTs. MR enterography: coronal ( A ) and axial ( B ) views of jejunal GIST (marked by the red arrows in panels A and B ). Postgadolinium MRIs (coronal [ C ] and axial [ D ]) views demonstrate enhancement of the jejunal GIST (marked by the red arrows in panels C and D ).

MRI has been shown more sensitive than CT for detecting mucosal lesions of the small bowel, and it seems to facilitate the identification of subtle mucosal changes. These findings may be due to the better soft tissue contrast that can be achieved with MRI; diffusion-weighted imaging is now a standard sequence while performing MR small bowel imaging because it improves the detection of small bowel tumors and inflammatory small bowel pathology. Recent data seem to suggest that for the specific subset of small bowel tumors, MR has higher diagnostic performance compared with CT. Moreover, because of the exposure to ionizing radiation with CT, the latter can be performed at only a few time points. Therefore, repeated dynamic imaging, and hence assessment of small bowel peristaltic activity, is not usually possible with CT, and an intermittent spasm or peristaltic contraction during the CT examination can also be misdiagnosed as a small bowel neoplasm. Nevertheless, the use of the latest generation of multidetector CT scanners and specific acquisition protocols can significantly reduce the radiation dose. It has also been shown that CT produced higher-quality images, mostly due to rapid acquisition, compared with MR and it seems to have better performance than MR in the assessment of potential perforation or complete bowel obstruction. CT is more frequently performed in patients who have difficulty in holding their breath or in those presenting with acute abdomen, whereas MR is preferred in young patients or in those with known renal impairment, in whom the administration of intravenous iodinated contrast agent can worsen renal function. In everyday clinical practice, however, MRI is less readily available and more time consuming. Taking into account the pros and cons of CT and MR techniques in the setting of suspected small bowel tumors, the choice between them often relies on opportunistic reasons (availability, local expertise, costs, and radiologist preference), rather than on clinical or technical issues.

The use of dedicated small bowel cross-sectional imaging techniques allows accurate detection of small bowel tumors with sensitivities and specificities of 85 to 94 and 95% to 97%, respectively. Furthermore, some studies report excellent interobserver agreement. Although the overall sensitivity of radiologic imaging tests can be influenced by the small number of patients included in single studies and by strict patient selection (in a majority of studies, only patients with suspected small bowel tumors were included), both CT and MR seem excellent diagnostic techniques for detecting small bowel tumors. As for CE, however, both CT and MR radiologic imaging techniques lack specificity to differentiate between different subtypes of tumors. Although different small bowel tumors have characteristic imaging features, no feature is absolutely specific. The identification of small bowel tumors by radiologic imaging warrants further examinations, aimed at obtaining tissue samples.

Cases of false-positive (ie, patients with adhesions or inflammatory bowel thickening) and false-negative examinations (in which the final diagnosis has been reached by other techniques) have been reported. Despite the highlighted limitations, however, dedicated small bowel cross-sectional imaging techniques have an obvious advantage over endoscopic tests for the study of the small bowel: they allow detailed panoramic extraluminal evaluation, providing diagnosis and staging at the same time. For this reason, when they are not used as first-line examination (eg, in patients with occlusion), radiologic examinations are almost universally performed in the diagnostic work-up of small bowel tumors, as a confirmatory or preoperative test.

There are other advanced radiologic techniques (ie, octreoscan and PET with gallium 68), not discussed in this article, which are essential diagnostic tests as far as specific tumors (ie, neuroendocrine) of the small bowel are suspected.