*T1591 The Negative Predictive Value (NPV) of Endoscopic Ultrasound (EUS) in a Large Series of Patients with Suspected Pancreatic Cancer Jason B. Klapman, Kenneth J. Chang, Phuong T. Nguyen BACKGROUND: EUS has been shown to be highly accurate in the diagnosis and staging of pancreatic cancer. Many studies report high sensitivity rates for EUS in the detection of pancreatic cancer, particularly smaller tumors. However, the accuracy of EUS in predicting the absence of pancreatic cancer in a large series of patients with a clinical suspicion of pancreatic cancer is not well documented. Our aim was to determine the NPV of EUS in patients with a suspicion of pancreatic cancer. METHODS: We retrospectively reviewed, from our EUS data base, 693 patients who were suspected of having pancreatic cancer and presented for EUS exams between 1/99-3/03. EUS exams were performed with the radial echoendoscope and/or the linear echoendoscope if the pancreas was not clearly visualized with 1 instrument. Patients with a pancreatic lesion or a focal pancreatic abnormality were excluded. 155 patients were found by EUS to have a normal pancreas. The mean age of the patients was 61 years (range 31-86) with a slight female predominance (85F/70M). Indications for EUS in these patients included: patient symptoms of weight loss/abdominal pain; and/or pancreatic enlargement/ fullness on computed tomography (CT) and/or; bile duct/pancreatic duct narrowing on ERCP; and/or an elevated CA19-9. Follow-up information was obtained from patient phone calls and/or physician visits and/or CT scan. Followup data was obtained in 135 (87%) patients with mean follow-up time of 25 months (range 8-58 mo.) RESULTS: No patients developed pancreatic cancer or required surgery during the follow-up period. Following the EUS examination, no further diagnostic tests were required in 119/135 (88%) of patients. Further diagnostic evaluation with follow-up CT ($6 months) was performed in 16 patients, none of which showed a pancreatic mass. The survival rate of patients with follow-up data was 132/135 (98%). All 3 patients died of causes unrelated to pancreatic disease. The negative predictive value of EUS in excluding pancreatic cancer in those patients with follow-up was 100%. CONCLUSION: EUS is highly specific in the diagnosis of pancreatic cancer with a NPV of 100% and obviates the need for further diagnostic testing in the vast majority of patients. In patients with a clinical suspicion of pancreatic cancer, EUS should be considered as the initial diagnostic modality.
*T1592 New Endoscopic Ultrasound (EUS) Criteria for Diagnosis of Chronic Pancreatitis (CP) Enrique Vazquez-Sequeiros, Daniel Boixeda, Victor Moreira, Antonio Gracia Plaza Background: The presence > 3 EUS diagnostic criteria for CP (EDCCP), allows one to establish a diagnosis of CP with a sensitivity > 90%. Unfortunately, specificity of EDCCP is low, as EDCCP may also be found in patients (PT) with no CP. Aims: 1.) To determine the pattern of pancreatic EUS normality and analyze which factors are associated with EUS changes suggestive of CP. 2.) To design new EDCCP that count on those factors associated with presence of EDCCP. Methods: 08/2002-05/2003. Prospective EUS exam in 155 asymptomatic patients (Inclusion Criteria: 1.) PT > 18 years candidate for EUS and no clinical suspicion for CP. Exclusion Criteria: 1.) History of pancreatobilliary disease, stones or colecystectomy, 2.) PT refuses to participate), in whom a prospective blinded assessment of EDCCP (Hyperechoic foci, linear strands, lobulation, calcifications, cysts, lithiasis, hyperechoic duct wall, irregular or dilated Wirsung) was performed. Factors associated with presence of EDCCP in this cohort of 155 PT was evaluated. In a different cohort of 75 PT (50 asymptomatic PT, 25 PT with CP), performance characteristics of conventional and modified EDCCP (accounting for confounding factors identified) were compared (Fisher, McNemar test. Logistic regression). Results: 155 consecutive PT included: : M/F (57%/43%), age (60.3 + 14.4 y.o. (61)), smoker (yes: 57%; cigs/d: 14 + 15.9 (10)), alcohol (yes: 64%; amount: 26.3 + 29.7 gr/d (30 gr/d), (BMI) (27.1 + 5.1 Kg/m2 (27.3)), diabetes mellitus (12%), EPOC (12%), cardiopathy (11%). Number of EDCCP +: 1.4 + 1.2 (median 1). Hyperechogenic foci (69% of PT), hyperechoic strands (32%), lobulation (21%). Uni and multivariate analysis of factors: alcohol intake > 30 gr/d and a PT age > 61 years were statistically associated with presence of EDCCP + (p=0.0001, p=0.0002) (BMI, sex, smoker p>0.05). In a different cohort of 75 PT, conventional EDCCP (CP = > 3 EDCCP +) and modified by age and alcohol EDCCP (CP = a.) 3 or more EDCCP + in PT < 60 y.o y < 30 grs/d alcohol; b.) 4 or more EDCCP + in PT > 60 y.o or > 30 grs/d alcohol; c.) 5 or more EDCCP + in PT > 60 y.o. and > 30 grs/d alcohol) were compared. EDCCP (conventional vs modified): sensitivity (100% vs 100%, p = 1), specificity (76% vs 88%, p = 0.04), diagnostic accuracy (84% vs 92%, p = 0.04). Conclusions: 1.) An alcohol intake > 30 gr/d and a PT age > 60 years are both associated with EUS changes suggestive of CP. 2.) Modified EDCCP provide a higher specificity and accuracy compared with conventional criteria, with no reduction in sensitivity.
VOLUME 59, NO. 5, 2004
*T1593 Detection of K-ras Point Mutation Increases the Efficacy of EUS-FNA for the Diagnosis of Ductal Pancreatic Adenocarcinoma Fauze Maluf-Filho, Marcia S. Kubrusly, Jose Eduardo M. da Cunha, Paulo Sakai, Jose Jukemura, Andre Montagnini, Marcel C. C. Machado, Shinichi Ishioka Background: In larger series, EUS-FNA sensitivity and accuracy for the diagnosis of pancreatic adenocarcinoma (PA)is around 85% and 89%, respectively. Mutational activation of the K-ras at codon 12 has been demonstrated in 70100% of the cases of PA. Aim: to evaluate the efficacy of EUS-FNA biopsy combined with K-ras point mutational analysis for the diagnosis of PA. Patients and Methods: From May/02 to Oct/03, ductal pancreatic adenocarcinoma was confirmed in 48 patients. Control group consisted of patients with neuroendocrine tumor (n=4), chronic pancreatitis (n=4), intraductal papillary mucinous tumor (n=2), mucinous cystadenoma(n=1)and breast cancer metastasis to the pancreas (n=1).EUS-FNA was performed under conscious sedation, using linear echoendoscope GFUC160P (Olympus Inc, Tokyo, Japan)and EUSN-3 needle (Wilson-Cook, WS, USA). A minimum of three passes were intended during each procedure. Residual material was rinsed and used for detection of codon 12 K-ras mutation by restriction fragment lenght polymorphism-PCR analysis. Efficacy of EUS-FNA biopsy and K-ras mutational analysis either alone or in combination was determined using a combination of surgical findings and/or clinical follow-up as ‘‘gold standards’’ for definitive diagnosis. Results: A total of 193 samples were obtained (3.2 passes/patient; range: 2-5. A cytopathologist was present in the EUS suite in 38% of the cases. The samples were considered adequate in all cases. The molecular study was possible in 52 from 60 patients (86,7%). EUS-FNA biopsy sensitivity, specificity and accuracy for diagnosis of PA were 81.3%, 100%, 84.5%, respectively. K-ras mutational analysis achieved a 73.2% sensitivity, a 100% specificity and a 78,1% accuracy for diagnosis of PA. When both techniques were combined the sensitivity reached 91.7% and accuracy, 93,1%. Conclusion: EUSFNA biopsy combined with K-ras mutational analysis is a strategy of higher sensitivity and accuracy for the diagnosis of ductal pancreatic adenocarcinoma than each test alone.
*T1594 Assessment of Telomerase Activity in Samples Obtained from Pancreatic Lesions by Endoscopic Ultrasound Guided Fine Needle Aspiration (EUS-FNA) Is Clinically Useful Girish Mishra, Doug Case, Mary Zhao, David Bridgers, Ben Pineau, John Sweeney, Kim Geisinger, Simon Bergman, James Cappellari IV, Coty Ho, Susan Melin, Russel Howerton, Ed Levine, Perry Shen, Jamal Ibdah Background and Aims: Telomerase activity is virtually absent from normal human somatic cells but is upregulated in germ line cells and in approximately 85% of solid tumors including pancreatic cancer. The objective of this study was to determine if telomerase activity obtained by EUS-FNA cytology can aid in the diagnosis of pancreatic cancer and distinguish benign cysts and inflammatory masses. Methods: Patients referred for EUS-FNA of solid and cystic pancreatic lesions were enrolled consecutively between January 2002 and June 2003. The final aspirate was inoculated into 10mL of RPMI media, centrifuged, and the pellet was frozen at -808F. Telomerase activity was assessed using a photometric enzyme immunoassay utilizing the TRAP ELISA kit (Roche Molecular Biochemicals). Clinical data and patient outcomes including age, gender, race, symptom onset, size and location of lesion, stage, nodal involvement, treatment, cytology (positive/ negative), serum CA19-9, and telomerase levels were collected prospectively. Chisquared tests, log-rank tests, and Cox’s proportional hazards regression were used to assess associations between telomerase levels and patient outcomes such as the detection of cancer and survival. Results: 70 patients were enrolled; median age 69 (range 37-94). Telomerase was significantly associated with a cancer diagnosis (p<.0001). Moreover, telomerase was positive in 5 of 30 (17%) patients who were cytology negative by EUS and later confirmed to have pancreatic adenocarcinoma. The sensitivity, specificity, PPV, and NPV for telomerase was 80%, 100%, 100%, and 73% respectively. Telomerase was negative in all cysts (n=18) proven to be benign by either surgery or clinical follow-up. Telomerase was negative in 4 patients with chronic pancreatitis and accurately detected cancer in this setting in 1 patient. The Kaplan-Meier estimated median survival for telomerase positive patients with pancreatic cancer is 11 months (.66-22.1), whereas the median survival for telomerase negative patients has not been reached yet (p=.0001). Even after controlling for cytology, maximal tumor diameter, and surgery, telomerase was still borderline significantly associated with survival. Conclusions: Determination of telomerase activity in specimens obtained at EUS-FNA is feasible and accurately predicts malignancy. Cytology negative lesions with a positive telomerase should be pursued aggressively to exclude malignancy.
GASTROINTESTINAL ENDOSCOPY
P223