Intravenous ultrasound contrast agents are microbubbles consisting of gas covered with a lipid or phospholipid membrane [41]. A certain range of acoustic power induces microbubble oscillation or breakage [42, 43]. Contrast-enhanced harmonic ultrasonography selectively depicts the signals produced by microbubble oscillation or breakage. The first-generation ultrasound contrast agent was Levovist (Schering AG, Berlin, Germany), which is composed of microbubbles of room air covered with a palmitic acid membrane. Levovist requires high acoustic power to oscillate or break its microbubbles [41, 42]. The second-generation ultrasound contrast agents include SonoVue (Bracco Imaging, Milan, Italy), Sonazoid (Daiichi-Sankyo, Tokyo, Japan; GE Health care Milwaukee, WI, USA), and Definity (Lantheus Medical Imaging, North Billerica MA, USA) and are composed of gases (not room air). They can be oscillated or broken by lower acoustic powers [41, 43], and thus are more suitable for the EUS, because the small transducer of EUS produces limited acoustic power [44]. Immediately before performing contrast enhancement, the ultrasound contrast agents are reconstituted with sterile water. After images of the ideal scanning plane are displayed on the specific mode for contrast enhancement, a bolus injection of the ultrasound contrast agent is administered through a 22-gauge cannula placed in the antecubital vein.

Intravenous ultrasound contrast agents enhance EUS images by depicting the vessels. Contrast-enhanced EUS includes contrast-enhanced doppler EUS and contrast-enhanced harmonic EUS. The former modality is based on the principle that the phase shift of the signals received from ultrasound contrast agents produces pseudo-doppler signals [42, 43]. On contrast-enhanced doppler EUS, these pseudo-doppler signals increase the sensitivity with which color and power doppler imaging depict the doppler signals from vessels [45–51]. However, this modality remains limited in terms of real time vessel imaging, because of artefacts such as blooming.

The other contrast-enhanced EUS modality, namely, contrast-enhanced harmonic EUS, selectively depicts harmonic components that are integer multiples of the fundamental frequency [9, 10, 42–44]. When microbubbles are oscillated or broken after receiving a certain range of acoustic power, harmonic components are produced. The harmonic component derived from microbubbles is higher than that from tissues; thus, contrast harmonic imaging more intensively depicts signals from the microbubbles than those from the tissue by selectively detecting the harmonic components [9, 10, 42–44]. Contrast-enhanced harmonic EUS can be performed with a wideband transducer equipped with EUS and the second-generation ultrasound contrast agents which are oscillated or broken by low acoustic powers [9, 10, 44].

On contrast-enhanced EUS, the solid pancreatic lesions can be characterized on the basis of their enhancement patterns relative to their surrounding tissue, namely, hypo-enhancement, iso-enhancement, or hyper-enhancement [52–55]. The contrast-enhanced harmonic EUS depicts pancreatic ductal carcinomas as nodules with hypo-enhancement that mostly have irregular network-like vessels (Fig. 4a) [52–55]. By contrast, most inflammatory pseudotumors exhibit iso-enhancement (Fig. 4b) and most neuroendocrine tumors exhibit hyper-enhancement (Fig. 4c) [52–55]. When Sonazoid which is the most sensitive ultrasound contrast is used for contrast-enhanced harmonic EUS, all pancreatic carcinomas possess a certain enhancement although the most are with hypo-enhancement, while benign necrotic tissues exhibit non-enhancement [54]. A recently published meta-analysis showed that when pancreatic adenocarcinomas are diagnosed on the basis of hypo-enhancement in contrast-enhanced harmonic EUS, the pooled diagnostic sensitivity and specificity are 94 and 89 %, respectively [56]. Moreover, when contrast-enhanced harmonic EUS was compared to conventional EUS, the former detected pancreatic adenocarcinomas (defined as hypo-enhanced lesions) with better sensitivity and specificity (96 and 64 %, respectively) than conventional EUS, where pancreatic adenocarcinomas were defined as hypoechoic lesions (86 and 18 %, respectively) [52]. Contrast-enhanced harmonic EUS also improves the depiction of the outline of ductal carcinomas with uncertain conventional EUS findings [52, 54, 57].

The contrast-enhanced harmonic EUS and contrast-enhanced CT are comparable in terms of differentiating ductal carcinomas from other masses, although contrast-enhanced harmonic EUS (91 % sensitivity and 94 % specificity) is superior to contrast-enhanced CT (71 % sensitivity and 92 % specificity) for small (≤ 2 cm) carcinomas [55]. In particular, the contrast-enhanced harmonic EUS is useful for characterizing small neoplasms that contrast-enhanced CT cannot identify [55].

Since contrast-enhanced harmonic EUS can identify vessels in the echogenic structure of cystic lesions, it discriminates mural nodules from mucous clots more sensitively than conventional EUS (Fig. 5) [58]. Mural nodules can also be further classified by contrast-enhanced EUS into four classes; low papillary, polypoid, papillary, and invasive nodules. Of these, the papillary and invasive nodules associate frequently with malignancy [58].

Compared to conventional EUS, contrast-enhanced harmonic EUS improves the preoperative T-staging of pancreatobiliary tumors [59]. In particular, contrast-enhanced harmonic EUS is superior in diagnosing portal invasion by pancreatobiliary adenocarcinomas. With respect to lymph node metastases, these exhibit enhancement defects on contrast-enhanced doppler EUS [51]. When this property is used, contrast-enhanced doppler EUS distinguishes benign from malignant lymph nodes with significantly higher sensitivity (100 %) and specificity (86 %) than conventional EUS variables (88 and 77 %, respectively) [51]. The contrast-enhanced harmonic EUS is also useful for diagnosing intra-abdominal lesions of undetermined origin. When it is used, 96 % of malignant lesions exhibit heterogeneous enhancement, whereas 75 % of benign lesions exhibit homogeneous enhancement [60].

When contrast-enhanced harmonic EUS was compared to conventional EUS in terms of the differential diagnosis of gallbladder wall thickening, the former had better diagnostic accuracy [61]. On contrast-enhanced harmonic EUS, inhomogeneous enhancement is strongly predictive of malignant gallbladder wall thickening (Fig. 6) [61]. With respect to gallbladder polyps, the presence of irregular intratumoral vessels or perfusion defects seen on the contrast-enhanced harmonic EUS are sensitive and accurate predictors of malignant gallbladder polyps [62].

The contrast-enhanced harmonic EUS studies of subepithelial tumors in the upper gastrointestinal tract revealed that gastrointestinal stromal tumors (GISTs) had significantly higher echo intensity than benign tumors such as lipomas [63]. In addition, the contrast-enhanced harmonic EUS allows the vessels flowing from the periphery to the center of GISTs to be visualized (Fig. 7) [64]. By contrast, the contrast-enhanced CT cannot identify most of these vessels. All high-grade malignancy GISTs possess these contrast-enhanced harmonic EUS-depicted irregular vessels. When using higher echo intensity and detection of irregular vessels to diagnose high-grade malignancy GISTs, the contrast-enhanced harmonic EUS is more sensitive than conventional EUS findings (namely, a large size, a lobular border, and a heterogeneous structure) [64]. These results suggest that the contrast-enhanced harmonic EUS can be used to identify GISTs and estimate their malignant potential.

The classification of pancreatic lesions on the basis of their enhancement patterns on the contrast-enhanced harmonic EUS is a convenient way to characterize the conventional EUS-detected lesions. However, the classification system depends on subjective assessment, which means that different readers can differ in their interpretations. This problem may be eliminated by quantitative analyses using a time-intensity curve with the contrast-enhanced harmonic EUS. Several time-intensity curve variables have been shown to be useful for the differential diagnosis of pancreatic masses. In particular, a low ratio of the uptake inside the mass to the uptake of the surrounding parenchyma [65], a low median intensity [66], a low maximum intensity [66], a long time to peak [67], a high area under the curve [67], and a high echo intensity reduction rate [68] have been found to be predictive of ductal adenocarcinomas.

It is unclear whether combining the contrast-enhanced EUS and EUS elastography yields a better diagnostic ability than each individual method. When Sãftoiu et al. assessed the diagnostic accuracy of the contrast-enhanced EUS and/or EUS elastography for solid masses in the pancreas, they found that the contrast-enhanced doppler EUS and EUS elastography had comparable sensitivity (91 and 85 %, respectively), but their specificities were less than 70 % [29]. By contrast, when hypo-enhancement on the contrast-enhanced doppler EUS and hard elasticity on EUS elastography were used to diagnose these masses, this combination was highly specific (95 %); thus, combining the two methods may be useful for reducing the number of false-positive cases [29]. However, Hocke et al. reported that the combination of fundamental B-mode, elastography and the contrast-enhanced doppler imaging (90 % sensitivity and 64 % specificity) did not improve on the result of the contrast-enhanced doppler EUS alone (90 % sensitivity and 92 % specificity) [26]. Further studies are needed to establish the contrast-enhanced harmonic EUS and EUS elastography criteria that can be used to diagnose pancreatic tumors.

The contrast-enhanced harmonic EUS can complement EUS-FNA in terms of identifying pancreatic ductal carcinomas that have false-negative EUS-FNA findings [53–55]. When pancreatic lesions with hypo-enhancement on contrast-enhanced harmonic EUS were regarded as ductal carcinomas in patients with negative EUS-FNA findings, the sensitivity of ductal carcinoma diagnosis increased from 92 % (EUS-FNA alone) to 100 % (both EUS-FNA and the contrast-enhanced harmonic EUS) [54]. Moreover, a French multicenter study showed that five false-negative EUS-FNA cases were correctly classified by the contrast-enhanced harmonic EUS [55].

Since the contrast-enhanced harmonic EUS clearly depicts subtle lesions that conventional EUS cannot identify, it can also be used to identify the target of EUS-FNA (Fig. 8) [52, 54, 57]. Moreover, it can be used to locate a specific site within a lesion that would be more suitable for EUS-FNA than other sites. Identifying and avoiding the avascular sites in a lesion may help avoid sampling necrotic areas and improve the sensitivity of pancreatic tumor diagnosis by EUS-FNA [69]. The contrast-enhanced EUS may also be helpful in terms of assessing lymph nodes that cannot be accessed by EUS-FNA, because of an intervening tumor or vessels; it can also help eliminate the time and risk associated with performing EUS-FNA at a second site [70].

The endoscopic biliary drainage is a common first-choice approach for obstructive jaundice. In standard endoscopic biliary drainage, a stent is inserted into the bile duct through the duodenal papilla by using a side-viewing endoscope; however, deep cannulation of the bile duct is difficult in some patients. Moreover, in some patients, the reconstruction after resection of the digestive tract or the duodenal infiltration of the tumor makes it impossible to reach the duodenal papilla. In such patients, endoscopic transpapillary treatment cannot be selected and percutaneous transhepatic biliary drainage (PTBD) and surgery had to be considered. In 2001, Giovannini et al. reported EUS-guided biliary drainage [12]. Since then, various EUS-guided biliary drainage techniques have been reported as an alternative treatment of obstructive jaundice after unsuccessful endoscopic transpapillary biliary drainage [12, 13, 71–108, 110–112].

The EUS-guided biliary drainage procedure is based on the EUS-guided pancreatic cyst drainage procedure. It consists of four steps, as follows: selection of the puncture site/route, puncture, dilation, and stenting [12, 13, 71–112]. As this procedure is performed in an organ with a specific shape, viz., the biliary tract, various puncture sites/routes and stenting methods can be used. This procedural variability is a feature of differing from EUS-guided pancreatic cyst drainage. There are two puncture routes, as follows: either from the gastric body (or small intestine after total gastrectomy with reconstruction of the digestive tract) to the intrahepatic bile duct, or from the duodenal bulb to the extrahepatic bile duct [71–112]. Another route from the duodenal bulb (or antrum) to the gallbladder may also be an option [113–121]. Stenting is classified into three methods: transmural drainage in which a stent is inserted into the site of puncture [12,13, 80, 81, 82, 90, 94, 96, 103, 104, 105, 109, 111], rendezvous drainage in which EUS-guided puncture is performed as an auxiliary procedure for transpapillary treatment [78, 88, 99, 102, 106], and antegrade stenting in which a stent is antegradely inserted/placed from the site of puncture into a stenotic site of the bile duct along a guide-wire after puncture [89, 101, 107, 110]. Because these different puncture sites/routes or stenting methods have not been systematically compared, it remains to be clarified which is the most effective method. The treatment choice depends on whether the site of biliary obstruction and papilla can be reached. For rendezvous drainage, an endoscope is needed to reach the papilla [78, 88, 99, 102, 106]. Furthermore, the EUS-guided transduodenal treatment is indicated for obstruction of the lower bile duct [81, 82, 83, 90, 94, 96, 103, 105, 109]. It is important to review which treatment method is appropriate for individual patients prior to the start of treatment and to plan a therapeutic strategy, so that a second choice can be selected in case the first-choice route of treatment or stenting method is difficult.