Kourosh F. Ghassemi

Variation in learning curves and competence for ERCP among advanced endoscopy trainees by using cumulative sum analysis

BACKGROUND AND AIMS There are limited data on learning curves and competence in ERCP. By using a standardized data collection tool, we aimed to prospectively define learning curves and measure competence among advanced endoscopy trainees (AETs) by using cumulative sum (CUSUM) analysis. METHODS AETs were evaluated by attending endoscopists starting with the 26th hands-on ERCP examination and then every ERCP examination during the 12-month training period. A standardized ERCP competency assessment tool (using a 4-point scoring system) was used to grade the examination. CUSUM analysis was applied to produce learning curves for individual technical and cognitive components of ERCP performance (success defined as a score of 1, acceptable and unacceptable failures [p1] of 10% and 20%, respectively). Sensitivity analyses varying p1 and by using a less-stringent definition of success were performed. RESULTS Five AETs were included with a total of 1049 graded ERCPs (mean ± SD, 209.8 ± 91.6/AET). The majority of cases were performed for a biliary indication (80%). The overall and native papilla allowed cannulation times were 3.1 ± 3.6 and 5.7 ± 4, respectively. Overall learning curves demonstrated substantial variability for individual technical and cognitive endpoints. Although nearly all AETs achieved competence in overall cannulation, none achieved competence for cannulation in cases with a native papilla. Sensitivity analyses increased the proportion of AETs who achieved competence. CONCLUSION This study demonstrates that there is substantial variability in ERCP learning curves among AETs. A specific case volume does not ensure competence, especially for native papilla cannulation.

Suboptimal accuracy of carcinoembryonic antigen in differentiation of mucinous and nonmucinous pancreatic cysts: results of a large multicenter study.

BACKGROUND AND AIMS The exact cutoff value at which pancreatic cyst fluid carcinoembryonic antigen (CEA) level distinguishes pancreatic mucinous cystic neoplasms (MCNs) from pancreatic nonmucinous cystic neoplasms (NMCNs) is unclear. The aim of this multicenter retrospective study was to evaluate the diagnostic accuracy of cyst fluid CEA levels in differentiating between MCNs and NMCNs. METHODS Consecutive patients who underwent EUS with FNA at 3 tertiary care centers were identified. Patients with histologic confirmation of cyst type based on surgical specimens served as the criterion standard for this analysis. Demographic characteristics, EUS morphology, FNA fluid, and cytology results were recorded. Multivariate logistic regression analysis to identify predictors of MCNs was performed. Receiver-operating characteristic (ROC) curves were generated for CEA levels. RESULTS A total of 226 patients underwent surgery (mean age, 61 years, 96% white patients, 39% female patients) of whom 88% underwent Whipples procedure or distal pancreatectomy. Based on surgical histopathology, there were 150 MCNs and 76 NMCNs cases. The median CEA level was 165 ng/mL. The area under the ROC curve for CEA levels in differentiating between MCNs and NMCNs was 0.77 (95% confidence interval, 0.71-0.84, P < .01) with a cutoff of 105 ng/mL, demonstrating a sensitivity and specificity of 70% and 63%, respectively. The cutoff value of 192 ng/mL yielded a sensitivity of 61% and a specificity of 77% and would misdiagnose 39% of MCN cases. CONCLUSIONS Cyst fluid CEA levels have a clinically suboptimal accuracy level in differentiating MCNs from NMCNs. Future studies should focus on novel cyst fluid markers to improve risk stratification of pancreatic cystic neoplasms.

Predictors for Surgical Referral in Patients with Pancreatic Cystic Lesions Undergoing Endoscopic Ultrasound: Results from a Large Multicenter Cohort Study

Objective Endoscopic ultrasound (EUS) plays an integral role in the evaluation of pancreatic cysts lesions (PCLs). The aim of the study was to determine predictors of surgical referral in patients with PCLs undergoing EUS. Methods We performed a multicenter retrospective study of patients undergoing EUS for evaluation of PCLs. Demographics, EUS characteristics, and fine-needle aspiration results were recorded. Patients were categorized into surgery or surveillance groups on the basis of post-EUS recommendations. Univariate and multivariate analyses were performed to identify predictors of surgical referral. Results 1804 patients were included. 1301 patients were recommended to undergo surveillance and 503 patients were referred for surgical evaluation, of which 360 patients underwent surgery. Multivariate analysis revealed the following 5 independent predictors of surgical referral: symptoms of weight loss on presentation (odds ratio [OR], 2.69; 95% confidence interval [CI], 1.44–5.03), EUS findings of associated solid mass (OR, 7.34; 95% CI, 3.81–14.16), main duct communication (OR, 4.13; 95% CI, 1.71–9.98), multilocular macrocystic morphology (OR, 2.79; 95% CI, 1.78–4.38), and fine-needle aspiration findings of mucin on cytology (OR, 3.06; 95% CI, 1.94–4.82). Conclusions This study identifies factors associated with surgical referral in patients with PCLs undergoing EUS. Future studies should focus on creation of risk stratification models to determine the need for surgery or enrollment in surveillance programs.

Tu1919 Diagnostic Accuracy of Carcinoembryonic Antigen (CEA) in Cyst Fluid Analysis in Histologically Confirmed Pancreatic Cysts: Results From a Large, Multicenter Cohort Study

Background: The phosphatidylinositol 3-kinase (PI3K)/Akt/mTORC1/S6K pathway plays a pivotal role in the proliferation and survival of pancreatic ductal adenocarcinoma (PDAC) cells and is aberrantly activated in pancreatic cancer tissues. In addition to growth-promoting signaling, mTORC1/S6K also mediates negative feedback loops that restrain upstream signaling via insulin/IGF receptor and other tyrosine kinase receptors. Suppression of these feedback loops by selective mTORC1 inhibitors, e.g. by rapamycin and its analogs, unleashes overactivation of the PI3K/Akt pathway that potentially oppose the anti-proliferative effects of mTOR inhibitors. This prompted the development of active-site mTORC1/2 kinase inhibitors (TOR-KIs) and dual PI3K and mTOR inhibitors (PI3K/TOR-KIs). Recently, we reported that TOR-KIs induce an unexpected increase in the activity of the ERK pathway in PDAC cells through a PI3K-independent pathway. Here, we examined whether PI3K/TOR-KIs also induce ERK pathway over-activation in PDAC cells. Results: To determine the effect of dual PI3K/TOR inhibition on ERK activation, we treated serum-starved cultures of PDAC cells (MiaPaca-2 and PANC-1) with increasing concentrations of the dual PI3K/TOR-KI NPVBEZ235 for 2 h followed by stimulation with insulin and neurotensin, a potent mitogenic combination for these cells. As expected, prior exposure to NPV-BEZ235 potently blocked mTORC1 activation (scored by S6 phosphorylation at Ser-240/244) and mTORC2-mediated Akt phosphorylation at Ser-473, in a concentration-dependent manner. Strikingly, we also demonstrate, for the first time that exposure to NPV-BEZ235 markedly enhanced the increase in the phosphorylation of ERK at Thr-202 and Tyr-204. Maximal ERK over-activation coincided with complete inhibition of Akt phosphorylation at Ser-473 (produced at 100500 ηM NPV-BEZ235). ERK over-activation was also seen when PDAC cells were stimulated with 2% fetal bovine serum instead of insulin and neurotensin and when a different PI3K/ TOR-KI (PKI-587) was used instead of NPV-BEZ235. Treatment with the MEK inhibitors U126 or PD0325901 (1-5 μM) prevented ERK over-activation induced by PI3K/TOR-KIs. In order to examine whether the over-activation of the ERK pathway counterbalances the growth-suppressive effect of PI3K/TOR-KIs, PDAC cells were treated with NPV-BEZ235, PD0325901 or a combination of NPV-BEZ235 and PD0325901. The combination of these drugs caused a more pronounced inhibition of cell growth than that produced by each inhibitor added individually. Conclusion: Collectively, our results highlight the importance of discovering novel signaling crosstalk to anticipate mechanisms of tumor resistance to new drugs. The capability of predicting drug resistance can assist in developing rational and effective strategies for developing combination therapies in PDAC.

The Clinical Utility of Evaluating the Luminal Upper Gastrointestinal Tract During Linear Endoscopic Ultrasonography.

Background: The clinical utility of performing esophagogastroduodenoscopy (EGD) before linear endoscopic ultrasonography (L-EUS) to evaluate the luminal upper gastrointestinal (GI) tract is not well established. Goals: The study was aimed to determine the prevalence of clinically meaningful luminal abnormalities (any luminal finding requiring further evaluation with mucosal biopsy or initiation of treatment) in patients undergoing L-EUS. The study also sought to compare the ability of the gastroscope and the linear echoendoscope in identifying these lesions. Study: A prospective, multicenter cohort study enrolled patients undergoing L-EUS for nonluminal indications. All patients underwent EGD followed by L-EUS by 2 different endoscopists. The second endoscopist was blinded to the results of the initial EGD. The identification of clinically meaningful luminal lesions and quality of endoscopic visualization of the upper GI tract were measured. Results: In the cohort of 175 patients, 52 (29.7%) patients had clinically meaningful luminal findings seen in the upper GI tract. There was no significant difference in the number of clinically meaningful lesions identified on EGD and L-EUS (25.1% vs. 22.9%, P=0.39). No significant difference was found in the miss rate of clinically meaningful lesions between the 2 modalities (EGD: 4.5% vs. EUS: 6.9%, P=0.39). Conclusions: A substantial minority of patients undergoing L-EUS for nonluminal indications will have clinically meaningful luminal findings. The endoscopic evaluation of the luminal upper GI tract can be adequately achieved using the linear echoendoscope.

Su1499 Creation of a Prediction Tool (M-PACT) to Accurately Identify Premalignant and Malignant Cysts in Patients Undergoing Endoscopic Ultrasound (EUS) for Evaluation of Pancreatic Cystic Lesions: Results From a Large Multicenter Cohort

Introduction: The nature of pancreatic cysts often remains uncertain, despite several tests, resulting in unoptimal management for patients. nCLE is an imaging technique, enabling microscopic observation of solid organs, in vivo and in real-time, during an EUS-FNA procedure. This technique could potentially provide useful information to establish a differential diagnosis between mucinous and non mucinous neoplasms. A recent study, INSPECT, described nCLE criteria for the characterization of intraductal papillary mucinous neoplasms (IPMN). Further descriptions were needed to identify nCLE criteria for the characterization of mucinous cystadenomas neoplasms (MCN) and non mucinous lesions. A prospective multicenter French study (CONTACT) aims at assessing the yield of nCLE for the diagnosis of pancreatic cystic tumors. Interim results presented criteria specific of serous cystadenomas (SCA). However, nCLE criteria for the characterization of MCN are still unknown. Materials and methods: Over 10 months, 31 patients without chronic pancreatitis and with a lonely pancreatic cyst > 2 cm large were enrolled. Following EUS examination, the AQ-Flex 19 miniprobe was introduced in a 19G needle and real-time video sequences of the cyst wall were recorded. Intracystic fluid obtained was analyzed. Final diagnosis of MCN (n=6) was based on surgery. The other lesions were considered as IPMN (n=5), pseudocysts (n=7) and SCA (n=13) on pathological consideration or expert consensus. Four gastroenterologists and two pathologists, unblinded to the final diagnosis studied the 31 nCLE videos records and compared their findings to the pathological specimen. Results: A particular pattern was furtively and recurrently seen in sequences acquired in three of the MCN lesions: an epithelial border lined the cyst wall, with or without deep blood vessels, and without papillary organization. These features were only observed in MCN sequences. These criteria correlate with the histological structure of those tumors that are characterized by a tall columnar mucin-producing epithelium underlined by a thick fibrous tissue with pseudo-ovarian stroma and vessels. Conclusions: The mucinous columnar epithelial border of MCN can be highlighted by nCLE. The validation of this criterion by an external group is undergoing and results will be available by the beginning of 2014. If proven to be accurate, this new sign might be useful for the diagnosis of mucinous cystadenoma neoplasms.