Diagnostic imaging of gastrointestinal neuroendocrine neoplasms with a focus on ultrasound

Transabdominal ultrasound is usually a first-line diagnostic imaging procedure for GEP-NEN. The use of various ultrasound techniques: harmonic and Doppler imaging, computer processing and compound imaging (CI) makes ultrasound a commonly used method for the diagnosis and monitoring of NEN. It is also possible to perform ultrasound-guided biopsy of tumours located in the liver, pancreas and retroperitoneal space.

The sensitivity of this method depends on the location of abnormalities, anatomical conditions, experience of the physician performing the procedure and technical aspects of the ultrasound device⁽¹⁾. The sonographic image of NEN, both the primary lesion and liver metastases is non-characteristic; it depends on the type of NEN, proliferation index, tumour size and location and on the presence of regressive changes⁽¹⁾.

A pancreatic neuroendocrine neoplasm is often well-defined, hypoechoic (less commonly hyper- or normoechoic), sometimes with a hyperechoic halo, areas of necrosis and calcifications^(1,4).

Transabdominal ultrasound has limited utility for detecting a primary lesion in the small intestine. It is rarely possible to detect a lesion in the intestinal wall using high-frequency probes⁽¹²⁾. A case of small intestinal tumour diagnosis with abdominal ultrasound despite negative results of other imaging studies has been reported in the literature⁽¹⁰⁾. Secondary lesions in the mesentery (enlarged, hypoechoic lymph nodes, mesenteric fibrosis) may be an indirect indication of a tumour in the intestinal wall⁽¹²⁾.

Hepatic metastases of a primary NEN tumour located in the lungs, stomach or colon are usually hypo- or isoechoic compared to the hepatic parenchyma. In tumours arising from the small intestine and pancreas the echogenicity of secondary lesions is various and non-characteristic (hypo-, iso-, hyperechoic or sometimes mixed lesions). In insulinoma tumours, liver metastases usually have decreased echogenicity^(13–15). Areas of necrosis are usually found in large tumours (>3 cm) with a Ki-67 proliferation index of >2% and in individuals who have undergone treatment (interferon, embolisation, chemotherapy, radiofrequency ablation). No necrotic changes were observed in patients treated with cold or hot somatostatin analogues^(13–15).

On Doppler ultrasound, primary NEN foci and liver and lymph node metastases have a rich network of blood vessels. In tumours with Ki-67 >2% the vascular network may be irregular^(13–15).

The sensitivity of transabdominal ultrasound in the diagnosis of metastatic lesions in the liver is 82–88% and its specificity is 92–95%. The sensitivity for pancreatic tumours is lower: 39%⁽¹⁾.

Endoscopic ultrasound is an invasive imaging method which involves the insertion of a flexible endoscope with a high-frequency ultrasound probe (7.5–12 MHz), which is positioned near the pancreas, stomach, duodenum or rectum. The minimisation of the distance between the probe and the abnormality to be evaluated makes it possible to thoroughly assess anatomical relationships, depth of tumour invasion through the gastrointestinal wall and the presence of metastases in local lymph nodes⁽¹⁾.

Endoscopic ultrasound is a fundamental diagnostic tool for the detection of NEN in the pancreas and the distal part of the colon. Combined with simultaneous fine-needle biopsy, it is the most sensitive method of pancreatic NEN diagnosis. It is a method of choice for the detection of small lesions (even 1–2 mm), particularly those hormonally active and a screening method for patients with genetic abnormalities (e.g. with MEN1 syndrome)⁽⁴⁾. In a multi-centre retrospective study investigating the size of pancreatic NEN in patients with MEN1 syndrome using EUS more than 90% of small lesions of less than 7 mm were visualised⁽¹⁶⁾. In other studies the sensitivity of EUS depended on the location of the focal lesion and was 77–100% for tumours located in the head and body of the pancreas and 75–80% for tumours located in the distal part of the pancreas^(1,17). Recent studies report that the sensitivity of EUS may be higher than that of computed tomography (91.7% vs 63.3%), particularly for insulinoma (84.2% vs 31.6%)^(1,2,4). The sensitivity of EUS in the detection of doudenal lesions is 45–60%^(1,2). Comparative studies on conventional examination methods (transabdominal ultrasound, CT and MRI) demonstrated their lower sensitivity compared to EUS⁽¹⁶⁾.

Hormonally active pancreatic neuroendocrine tumours present as single homogeneous, hypoechoic solid tumours (less often normo- and hyperechoic ones) which are characterised by intensive enhancement following contrast administration⁽¹⁶⁾. A cystoid form of the tumour may occur in approximately 10–20% of patients⁽¹⁸⁾. Calcifications in hormonally active and inactive NEN of the pancreas may be found in approximately 25% of cases, while being rarely observed in adenocarcinomas. In larger tumours, the pancreatic duct may become dilated and the pancreatic parenchyma may undergo atrophy^(1,18). EUS plays an important role in disease staging, assessment of blood vessel invasion and differential diagnosis of pancreatic tumours of unclear nature (distinguishing between NEN and adenocarcinomas, cystoid lesions and tumours with atypical, poor vasculature). It may be stated that EUS combined with Doppler ultrasound, contrast, elastography and EUS-guided biopsy makes it possible to detect all pancreatic NEN^(1,18).

In diagnostic imaging of rectal tumours endoscopic ultrasound is also the most sensitive method for preoperative local staging. A neuroendocrine neoplasm in the form of a polyp presents as a well-defined, iso- or hypoechoic, homogeneous lesion on EUS⁽¹⁹⁾. The sensitivity of this method in the assessment of the tumour and the depth of its invasion in the rectal wall is 76–93%. For the assessment of metastases in the nearby lymph nodes the sensitivity is 61–88%^(1,2,4,19).

The assessment of lesions located in the distal part of the ileum is possible on EUS using high-frequency miniprobes inserted through the biopsy channel of a colonoscope^(4,19).

An advantage of EUS is also the fact that it is possible to insert a radiofrequency ablation (RFA) probe in the echoendoscope channel, which causes thermal coagulation necrosis of the tumour. It has been demonstrated that this technique is safe and effective in patients with hormonally active NEN of the pancreas who were denied surgery due to comorbidities⁽¹³⁾. The disadvantages of EUS include the fact that the result depends on the skills and experience of the examiner and that the field of vision is limited, which may reduce its efficacy in detecting lesions beyond the area of interest⁽⁴⁾.

It is possible to achieve an even better visualisation of pancreatic NEN in the immediate vicinity of the pancreatic duct and abnormalities in its lumen by inserting a miniprobe to the pancreatic or bile duct through the biopsy channel of an endoscope (IDUS)⁽¹⁾. The sensitivity of the procedure is around 94% and approximates 100% if the abnormalities in the pancreatic duct are larger than 3 mm⁽²⁰⁾.

Third-generation contrast agents used for ultrasound diagnosis are gas-filled microvesicles coated with phospholipids with a long half-life and perfusion-dependent greyscale image enhancement⁽¹⁾. On CEUS, focal lesions are assessed based on characteristic vasculature patterns indicating benign or malignant abnormality⁽¹⁵⁾. Primary pancreatic NEN, even small ones, have a good network of blood vessels. Following contrast administration, pancreatic NEN show intensive enhancement in the arterial phase⁽¹⁶⁾. The majority of NEN metastases in the liver (78–86%) show enhancement in the arterial phase following contrast administration with a quick contrast clearance in the venous phase^(1,16,17). Wen-Tao et al. demonstrated that metastases with a rich vasculature demonstrated a longer contrast clearance time compared to those with a poorer vasculature. The assessment of both primary and metastatic tumours using CEUS may replace other contrast-enhanced imaging techniques in the majority of cases, particularly if there are contraindications to those procedures⁽¹⁶⁾. CEUS is useful in the detection of small tumours of less than 2 cm, with a comparable sensitivity to EUS (95%)⁽¹⁾.

Ultrasound imaging using elastography is an increasingly available technique allowing one to assess tumour hardness. This method may be used to assess focal lesions in the pancreas on EUS and focal lesions of the liver on transabdominal ultrasound. In a study by Iglesias-Garcia et al., the sensitivity of quantitative EUS elastography in differentiating between pancreatic adenocarcinoma and NET was 100% and its specificity was 88% for the strain ratio (SR) cut-off value of 26.6^(21,22). In general, the sensitivity and specificity of differentiating between benign and malignant lesions were 93% and 86%, respectively, while the SR limit value was 6.0. In a metanalysis on qualitative EUS elastography in the diagnosis of pancreatic tumours a high sensitivity (95%) and an acceptable low specificity (67%) were observed^(21,22). Qualitative and quantitative EUS elastography may be a valuable supplementary method used to differentiate between benign and malignant pancreatic lesions⁽²¹⁾. According to the recommendations of the European Federation of Societies for Ultrasound in Medicine and Biology (EFSUMB), EUS elastography may be helpful in making therapeutic decisions in cases where fine-needle biopsy performed during EUS has no diagnostic value. However, elastography alone cannot replace cytological assessment^(21,22). Elastography of focal hepatic lesions allows one to make a preliminary distinction between MEN metastases and primary benign focal hepatic lesions. However, the results may be doubtful in obese patients or in the case of lesions located near the ribs^(21,22).

Intraoperative ultrasound (IOUS) is used mainly to diagnose focal lesions in the liver and pancreas, particularly in patients with multiple endocrine neoplasia (MEN) syndromes. The procedure also makes it possible to detect multifocal tumours and metastatic lesions in the liver and assess the distance of a pancreatic tumour, particularly a small one, from the pancreatic duct⁽¹⁾. The sensitivity of IOUS in the detection of tumours is 90%⁽¹⁾.