Endoscopy 2014; 46(05): 380-381
DOI: 10.1055/s-0034-1365425
© Georg Thieme Verlag KG Stuttgart · New York

Diagnosis of pancreatic cancer by contrast-harmonic endoscopic ultrasound (EUS): complementary and not competitive with EUS-guided fine-needle aspiration

Pietro Fusaroli
American University of Beirut Medical Center, Beirut, Lebanon
Mohamad A. Eloubeidi
American University of Beirut Medical Center, Beirut, Lebanon
› Author Affiliations
Further Information

Publication History

Publication Date:
16 April 2014 (online)

Contrast-harmonic endoscopic ultrasound (CH-EUS) is a relatively new technique, based on observation of a specific tissue harmonic after intravenous injection of an ultrasound contrast agent [1]. SonoVue (sulfur hexafluoride microbubbles; Bracco Imaging, Milan, Italy), the most widely adopted contrast agent, acts as a pure intravascular enhancer; intraprocedural observation of the microvasculature of the lesion of interest is achieved and compared with that of the surrounding parenchyma and vessels. Researchers have identified different patterns of SonoVue uptake in pancreatic tumors, and classified them into three broad categories: hypoenhanced, isoenhanced, and hyperenhanced. Further classification can be derived by analyzing the network of vessels (homogeneous or inhomogeneous) and the timing of contrast washout (fast washout during the arterial phase or slow washout during the venous phase).

The classical CH-EUS features of pancreatic adenocarcinoma are a hypoenhanced pattern, inhomogeneous network, and fast washout. Conversely, a pancreatic neuroendocrine tumor typically looks hyperenhanced, with homogeneous or inhomogeneous network according to its differentiation, and slow washout. Additional patterns have been described in other pancreatic and extrapancreatic lesions. However, the above-mentioned patterns have been reported by single centers that pioneered the technique of CH-EUS and which published their observations as either retrospective or prospective studies [2] [3] [4] [5]. Skeptics of CH-EUS have often questioned the reproducibility of these observations, which are based merely on qualitative analysis of contrast agent uptake. Moreover, CH-EUS has been criticized because it is unable to avoid the need for EUS-guided fine-needle aspiration (FNA).

The study by Gincul et al. [6] is a milestone in the effort of scientific accreditation of CH-EUS. A total of 100 patients with a pancreatic mass of unknown origin were enrolled prospectively at three major endoscopy centers in France. After standard EUS investigation with a radial echoendoscope, CH-EUS and EUS-FNA were carried out using a linear echoendoscope (GF-UCT 180; Olympus Medical Systems, Tokyo, Japan), the Prosound Alpha 10 processor (Aloka, Tokyo, Japan), and SonoVue. Five experienced endosonographers performed all examinations, paying particular care to standardize every step of the procedure in order to obtain comparable data from each case. In particular, all CH-EUS examinations were recorded and subsequently reviewed anonymously during a 2-day consensus session among the same endosonographers plus two beginners. Thus, both intraobserver and interobserver agreement were evaluated in a highly selected cohort of patients with solid pancreatic tumors. The diagnostic gold standard was represented either by pathology (EUS-FNA and/or surgery) or by long term follow-up.

The results reported by Gincul et al. are sound. The intraobserver agreement was good for the beginners and excellent for the experienced endosonographers, and the overall interobserver agreement was good. Interestingly, out of 69 cases of pancreatic adenocarcinoma, 66 were hypoenhanced at CH-EUS. In addition, 29 of the remaining 31 lesions that were not adenocarcinomas (including neuroendocrine tumors, focal pancreatitis, and renal cancer metastases) were iso- or hyperenhanced. Compared with the gold standard, the hypoenhancement pattern at CH-EUS showed 95 % accuracy, 96 % sensitivity, and 94 % specificity for the diagnosis of pancreatic adenocarcinoma.

The French study is also a very good benchmark for EUS-FNA, as tissue sampling was obtained in 100 % of the patients enrolled. EUS-FNA showed 95 % accuracy, 93 % sensitivity, and 100 % specificity for the diagnosis of adenocarcinoma. Interestingly, five false-negative cases by EUS-FNA were correctly classified as pancreatic adenocarcinoma by CH-EUS.

We believe that CH-EUS and EUS-FNA are complementary and should always be performed during the same investigation to obtain as much information as possible. For instance, in cases with a hypoenhanced lesion with a negative EUS-FNA, surgery should be considered anyway, particularly if there is a high clinical suspicion of malignancy. Alternatively, in cases with an iso- or hyperenhanced lesion with a negative EUS-FNA, pancreatic cancer could be ruled out, and either follow-up or repeat EUS-FNA should be indicated based on clinical grounds ([Fig. 1]). However, performing CH-EUS during a second examination only in cases of negative EUS-FNA would be less cost effective. We believe this algorithm deserves further study, and should be adjusted as more data emerge in this growing field.

Zoom Image
Fig. 1 Suggested algorithm based on contrast-enhanced endoscopic ultrasound (CH-EUS). FNA, fine-needle aspiration.

As suggested by Gincul et al. [6], CH-EUS may improve EUS-FNA by targeting the best area for sampling and avoiding nonhypoenhanced areas of the tumor, which may correspond to inflammatory peritumoral reaction and could potentially lead to false-negative results. Further studies are warranted to assess whether EUS-FNA accuracy can be improved by CH-EUS.

The outstanding results of CH-EUS in this French study were obtained by consensus among seven endosonographers who reviewed all of the videos on a different day from the study procedure. Postprocedure review of videos is current practice among those performing examination with contrast agents, as many details may have gone unnoticed during intraprocedural evaluation. Nevertheless, individual accuracy of CH-EUS in the diagnosis of pancreatic adenocarcinoma was also very good and was never below 81 %. The good intraobserver and interobserver agreement reported by the authors suggest that only a short learning curve is required before CH-EUS can be included in clinical routine with pancreatic EUS. Another study investigating interobserver agreement among experienced and less-experienced endosonographers from Italy reported only moderate overall agreement [7]. However, the latter study included solid as well as cystic pancreatic lesions and also submucosal tumors. It may be inferred that while the use of CH-EUS for the diagnosis of pancreatic adenocarcinoma is quite straightforward, more experience and training are required to investigate more subtle changes in other fields of application of CH-EUS.

Time will tell whether ultrasound contrast agents will become as indispensable for EUS as radiological contrast agents already are for computed tomography. However, CH-EUS has several advantages: the cost of a vial of SonoVue is relatively cheap and significantly less than a needle for EUS-FNA; second, CH-EUS is very quick, taking only 2 – 3 minutes; third, the published results so far are encouraging and are supported by strong levels of evidence. Based on all of these premises, it appears that CH-EUS will not be washed out very fast from our practice and is here to stay. More data are needed, however, in order to confirm these findings so that CH-EUS is fully integrated into routine EUS performance, to sharpen and enhance endosonographers’ skills, to better diagnose pancreatic cancer, and to minimize false-negative diagnosis by EUS-FNA.