Endoscopy 2015; 47(01): 4-5
DOI: 10.1055/s-0034-1391079
© Georg Thieme Verlag KG Stuttgart · New York

Novel diagnostic approach to pancreatic cysts: is confocal laser endomicroscopy bridging the gap?

Marco J. Bruno
Department of Gastroenterology and Hepatology, Erasmus Medical Center, Rotterdam, The Netherlands.
› Author Affiliations
Further Information

Publication History

Publication Date:
22 December 2014 (online)

With the advent and growing use of cross-sectional imaging, an increasing number of individuals are being diagnosed with a cyst in the pancreas. In the majority of cases, scanning is done for reasons other than the suspicion of pancreatic disease, and the finding of a pancreatic cyst is therefore coincidental. The reported prevalence rates vary, but even the most conservative estimates indicate that a sizable number of individuals are affected, all of whom are potential candidates for diagnostic and/or follow-up investigations [1].

Detection and surveillance of an asymptomatic pancreatic cyst offers the opportunity for timely detection of advanced neoplasia or a small cancer, and early detection is likely to increase the survival of this otherwise deadly disease. However, there is also a flip side to the coin: there is a danger of medicalization, posing burden to the individuals being investigated, both physically and psychologically, and using valuable healthcare resources with possibly little or even no gain. More importantly, there is a potential risk of doing harm if patients are wrongly selected to undergo a pancreatic resection, and are exposed to associated risks of morbidity and mortality.

To date, the management of asymptomatic pancreatic cysts is largely driven by consensus guidelines with low grades of scientific evidence available for the majority of recommendations [2] [3]. We lack good-quality data pertaining to evidence-based management. We need to understand more about the behavior and natural progression of pancreatic cysts in different individuals, as well as morphologic and biomolecular cyst characteristics. We also need to establish the (cost-)effectiveness of investigational tools to timely detect the progression of neoplastic cysts during surveillance.

First and foremost, we need a method to reliably differentiate between the various pancreatic cyst types. Rather than keeping individuals with a pancreatic cyst of uncertain origin under surveillance for indefinite periods of time, there would be great value in a test that enables the accurate determination of its origin, in particular to ascertain the risk of malignant progression in cases of mucinous cystadenoma or intrapapillary mucinous neoplasm. Individuals with truly benign lesions, such as pseudocysts and serous cystadenoma, could be discharged from life-long follow-up and thus avoid the anxiety of repeated investigations. Moreover, the burden and risk of surgery would be avoided in these individuals. In a recent large, retrospective, multinational survey in 598 patients with a serous cystadenoma, 63 % had undergone a surgical resection because of uncertainty about the preoperative diagnosis [4].

In this issue of Endoscopy, Napoléon et al. take up the challenge of pancreatic cyst differentiation and tested the ability of endoscopic ultrasound (EUS)-guided needle-based confocal laser endomicroscopy (nCLE) to diagnose pancreatic serous cystadenomas [5]. CLE involves in vivo microscopic imaging of the mucosal epithelium during endoscopy. It is also known as confocal fluorescent endomicroscopy or optical endomicroscopy. The term “confocal” denotes that both illumination with a low-power laser and collection of reflected light are oriented in the same focal plane. The fluorescent dye most frequently used is fluorescein. CLE is proposed as a technique to guide “smart” targeted biopsies or to replace traditional histology. It has been utilized in colorectal adenomas and cancer, inflammatory bowel disease, and Barrett’s esophagus [6]. Recently, the use of CLE has expanded into the field of hepatopancreaticobiliary disorders, with the probe being introduced into the biliary tract and pancreatic duct [7].

In 2013, Konda et al. published the results of a multicenter study on EUS-guided nCLE in 66 patients. The authors defined and applied specific criteria to identify pancreatic cystic neoplasms, including intraductal papillary mucinous neoplasms, mucinous cystic adenoma, or adenocarcinoma in a post-procedural, blinded, consensus review [8]. The presence of epithelial villous structures based on nCLE was associated with pancreatic cystic neoplasms, and provided a sensitivity of 59 %, specificity of 100 %, positive predictive value of 100 %, and negative predictive value of 50 %. These data suggest that nCLE has a high specificity in the detection of pancreatic cystic neoplasms, but may be limited by a low sensitivity.

Napoléon et al. add to these data by testing the ability of EUS-guided nCLE to diagnose pancreatic serous cystadenomas [5]. Based on the visualization of a dense network of subepithelial capillaries shown to be typical and highly characteristic for serous cystadenoma, the authors were able to differentiate serous cysts from other types of cysts with an accuracy of 87 %, a sensitivity of 69 %, a specificity of 100 %, a positive predictive value of 100 %, and a negative predictive value of 82 %. These are good results, particularly for specificity and positive predictive value, and match or outweigh other available diagnostic modalities, including morphology (computed tomography, magnetic resonance imaging – magnetic resonance cholangiopancreatography, EUS), cyst fluid analysis, and EUS-guided pathology [9]. An additional advantage of EUS-guided nCLE is that the procedure is quick (mean duration 7 minutes, range 3 – 10 minutes [5]), and that it provides instant results.

The concept of EUS-guided nCLE warrants some considerations, in particular pertaining to needle size. The presently available confocal miniprobe (Cellvizio; Mauna Kea Technologies, Paris, France) fits a 19-G needle. It is preloaded into the needle, after which the cyst is punctured and the miniprobe is advanced outside the needle sheath. Manipulation of a 19-G needle may be difficult or impossible when a lesion is located in the pancreatic head. This proved not to be an obstacle in the series by Napoléon et al., with 15 out of 31 cysts being located in the pancreatic head, although the results could be different in less-experienced hands. Another incentive for a smaller needle size and matching confocal miniprobe is the notion that examinations were limited to cysts bigger than 2 cm. In clinical practice, the majority of cysts that pose a differential diagnostic dilemma, whether to offer surveillance or to discharge a person from further follow-up, are smaller than 2 cm, and puncture of a smaller cyst is easier with a smaller needle. A final argument for using a smaller needle concerns the complication rate and avoidance of post-procedural pancreatitis, which in both series occurred in about 3 % of cases.

Of note, CLE might even have potential beyond cyst differentiation. In a prospective, double-blind, controlled study in a rat model, Li et al. evaluated the accuracy of in vivo CLE to differentiate between the various neoplastic stages using cathepsin E as a probe [10]. Cathepsin E (CTSE) is highly and specifically expressed in pancreatic ductal adenocarcinoma and pancreatic intraepithelial neoplasia, and its expression gradually increases with disease progression [11]. The accuracy of CLE-CTSE to diagnose all grades of these two lesion types was greater than 82.7 %.

Two reports have now reported favorably on the use of EUS-guided CLE for pancreatic cyst differentiation. Given the diagnostic challenge in a growing number of individuals with an asymptomatic pancreatic cyst, the lack of other sensitive and specific diagnostic modalities, and the dilemma of whether or not to institute long-term follow-up with associated burden and costs, EUS-guided nCLE clearly bears promise. Whether nCLE will ultimately bridge the gap where other diagnostic modalities have so far failed, remains to be proven. Before drawing such conclusion, larger series of patients are needed, applying the criteria as proposed by Konda et al. [8] and Napoléon et al. [5], and with pathological confirmation of cysts that are resected and long-term follow-up of individuals who are not operated upon.