Mo1518An Evaluation of Risk Factors for Inadequate Cytology UsingContrast-Enhanced Harmonic Endoscopic Ultrasound inEndoscopic Ultrasound-Guided Fine Needle Aspiration ofSuspected Pancreatic MalignancyMasahiro Itonaga*, Kazuki Ueda, Hiroko Terada, Takashi Tamura,Yasunobu Yamashita, Hiroki Maeda, Takao Maekita, Mikitaka Iguchi,Hideyuki Tamai, Jun Kato, Masao IchinoseSecond Department of Internal Medicine, Wakayama MedicalUniversity, Wakayama City, Japan[OBJECTIVES] Contrast-enhanced harmonic endoscopic ultrasound(CH-EUS) isuseful for differential diagnosis of pancreatic mass lesions, but risk factors forinadequate cytology using this modality in endoscopic ultrasound-guided fineneedle aspiration(EUS-FNA) remain unclear. The present study aimed to evaluatethe role of CH-EUS in EUS-FNA. [METHODS] Generally, CH-EUS is evaluated inthe perfusion phase (60-90 s after contrast agent infusion), but we performedevaluations in the early phase (10-30s after contrast agent infusion) andcategorized into two patterns: Pattern A, homogeneous enhancement; andPattern B, heterogeneous enhancement with partial hypoenhancement. Weselected typical Pattern A and Pattern B cases of pancreatic ductaladenocarcinoma that had undergone surgery and performed histopathologicalexaminations with hematoxylin and eosin staining and immunohistochemicalstaining. Compared with Pattern A, the Pattern B case showed necrotic or fibrouslesions and fewer vessels. The Pattern A case showed tumor cellshomogeneously, but the Pattern B case showed few tumor cells in the necroticor fibrous regions and tumor cells were distributed heterogeneously. We thushypothesized that accuracy of EUS-FNA would differ between Patterns A and B.From January 2009 to July 2012, a total of 61 patients with suspected pancreaticmalignancy were included. After CH-EUS was performed, we performed EUS-FNA at the center of each mass. Risk factors for inadequate cytology wereevaluated retrospectively. [RESULTS] Male:female ratio was 36:25, mean(�standard deviation) age was 68.8�8.96 years, mean lesion size was 37.6�15.3cm, head:body or tail ratio was 28:33; 19G/22G ratio was 37/24; number ofpasses was 2.77�1.28 and Pattern A:B ratio was 23:38. FNA was successfullyperformed in all cases and adequate samples for histological examination wereobtained in 53 of the 61 cases (86.9%). Only Pattern B was identified as asignificant risk factor for inadequate cytology (p�0.02) using the �2-test andStudent’s t-test. [CONCLUSION] As expected, accuracy in EUS-FNA differedsignificantly between Patterns A and B. We expect CH-EUS to improve thediagnostic value of EUS-FNA if regions of hypoenhancement on CH-EUS areavoided.

Mo1519Tumor Vessel Depiction With Contrast-Enhanced EndoscopicUltrasonography Predicts Efficacy of Chemotherapy inPancreatic CancerYasunobu Yamashita*, Kazuki Ueda, Takashi Tamura,Masahiro Itonaga, Takeichi Yoshida, Hiroki Maeda, Takao Maekita,Mikitaka Iguchi, Hideyuki Tamai, Jun Kato, Masao IchinoseSecond Department of Internal Medicine, Wakayama MedicalUniversity, Wakayama, JapanObjectives: Contrast-enhanced endoscopic ultrasonography (CE-EUS) is a newimaging modality for pancreatic lesions. The aim of this study is to evaluate ifCE-EUS is useful for predicting treatment efficacy before pancreatic cancerchemotherapy by assessing intratumoral vessel flow. Methods: Thirty-ninepatients with unresectable advanced pancreatic cancer underwent CE-EUS beforechemotherapy. Patients were divided into two groups according to theintratumoral vessel flow observed with CE-EUS: vessel sign positive and vesselsign negative groups. Patient prognosis was investigated according to presenceor absence of the vessel sign. Results: Two patients were excluded due to poorvisualization of CE-EUS images, so 37 patients were analyzed. CE-EUS revealedpositive vessel sign in 20 patients, while negative vessel sign in 17 patients. Bothprogression free survival and overall survival were significantly longer in thepositive versus negative vessel sign groups (P � 0.037, and P � 0.027,respectively). Multivariate analysis demonstrated that the positive vessel sign wasan independent factor associated with longer overall survival (hazard ratio �0.22, 95% confidence interval: 0.08-0.53). Conclusions: Evaluation of intratumoralvessel flow by CE-EUS could be useful for predicting efficacy of chemotherapyin pancreatic cancer patients. CE-EUS could be used prior to chemotherapy forinoperable pancreatic cancer.

Mo1520EUS Fine Needle Aspiration and Biopsy Sample Quality Analysisfor Hent1 Immunohistochemical Staining in Pancreatic CancerManagementPuleo Francesco*1, RaphaëL MaréChal1, Pieter Demetter2,Laurine Verset2, Pierre Eisendrath1, Annabelle Calomme1,Emmanuel Toussaint4, Jean-Baptiste Bachet3, Marianna Arvanitakis1,Jacques M. Deviere1, Van Laethem Jean-Luc1

1Gastroenterology, Erasme Hospital, Brussels, Belgium; 2Pathology,Erasme Hospital, Brussels, Belgium; 3Gastroenterology, Hôpital Pitié-Salpêtrière, Brussels, France; 4Gastroenterology, Bordet Institute,Brussels, BelgiumIntroduction: In the era of biomarker studies and biomarker-driven protocols,tissue acquisition is the limiting factor. In pancreatic cancer (PAC) thisrepresents a challenge. Data from several retrospective studies suggest humanEquilibrative Nucleotide Transporter-1 (hENT1) as a biomarker to selectgemcitabine responders in the adjuvant setting. No studies have beenpublished on the impact of hENT1 status in neoadjuvant, locally advancedand metastatic settings, the limiting factors being PAC tissue availability inpatients who will not undergo surgery. Endoscopic ultrasound (EUS) fineneedle aspiration (FNA) and biopsy (FNB) are the gold standards for tissueacquisition. We aimed to evaluate the feasibility of immunohistochemical(lHC) staining for hENT1 on EUS FNA/FNB and provide information onpredictive factors of sample quality. Materials and Methods: In our institutionaldatabase of PAC patients we identified 192 patients who underwent EUS-FNA(1/2006-5/2011) (Cohort 1). We prospectively collected tissues from 38patients who underwent EUS FNA or FNB for pancreatic masses with apathologically confirmed diagnosis of PAC from 6/2011 to 11/2012 (Cohort2). We compared the two cohorts in terms of type of needle, cellblockpreparation, presence of neoplastic cells in cellblock and yield to performIHC staining for hENT1. We evaluated factors correlated to the success orfailure to perform IHC. hENT1 staining was evaluated semi-quantitatively bytwo pathologists. Results: The two cohorts were comparable in terms of age,sex, localisation of the lesion, presence of chronic pancreatitis (CP), rapidon-site cytopathological examination (ROSE), tumor stage and number ofpasses. For 53 patients in cohort 1 and 4 patients in cohort 2, no cellblockswere available (p�0.017). No neoplastic cells were found in 71 and 10cellblocks in cohort 1 and 2 respectively. hENT1 IHC was possible in 68 of192 EUS-FNA in cohort 1 and 24 of 34 EUS-FNA/FNB in cohort 2 (36,98% vs63,16%, p�0.015). Factors associated with the possibility to perform IHCwere: prospective collection of sample (p�0.015), absence of CP (p�0.024),�3 passes (p�0.0001). Size, type of needle, localisation, ROSE and tumorstage were not predictive of the ability to perform IHC. Overall, hENT1staining was negative, positive and inconclusive in 71.4%, 18,2% and 10,4%of patients. Conclusions: PAC tissue availability still remains a challenge,especially for inoperable and locally advanced disease. In the prospectivearm of this study, we obtained enough tissue to perform hENT1 staining in63% of patients. Predictive factors include a higher number of passes and theabsence of CP. We should take this into account in future tissue biomarker-driven protocols and efforts should be focused on increasing the yield oftissue availability, not only for diagnostic but also for therapy-guidedpurpose.

Mo1521Potential Risks and Benefits of Preoperative EndosonographicEvaluation of Resectable Pancreatic MassesVasco Eguia*1, Austin L. Chiang1, Theodore P. Doukides1,2,Amrita Sethi1,2, John M. Poneros1,2, John D. Allendorf2,John A. Chabot2, Charles J. Lightdale1, Tamas A. Gonda1,2

1Division of Digestive and Liver Diseases, Columbia University MedicalCenter, New York, NY; 2The Pancreas Center, Columbia UniversityMedical Center - NYPH, New York, NYIntroduction: EUS-FNA is the diagnostic test of choice when evaluatingpancreatic masses. In cases when imaging is suggestive of malignancy andthe lesion appears surgically resectable it may be reasonable to proceed todefinitive therapy without further sampling. Our goal in this study was toassess the benefits of preoperative EUS-FNA in such cases by evaluating theability of EUS -FNA to (1) identify patients with a non-malignant disease andthus avoid potentially unnecessary surgery; (2) assess the accuracy of EUS tostage malignancy. We also examined the potential long term risk of a pre-operative EUS. Methods: In a prospectively maintained surgical andendoscopic database, we identified a cohort of 985 patients who had apancreatic mass and underwent surgical resection without receiving pre-operative therapy between 2006-2012. We collected data on pre-operativeand final pathologic diagnosis and stage and post operative survival.Categorical values were expressed as medians, ranges and percentages.Comparisons were made using Chi-square test and performed univariate andmultivariate Cox proportional hazards analysis. Results: There were 713 (72%)

Abstracts

AB412 GASTROINTESTINAL ENDOSCOPY Volume 77, No. 5S : 2013 www.giejournal.org

malignancies, 194 (20%) cystic neoplasms and 78 (8%) benign or non-neoplastic resections. 272 (27%) of cases had preoperative evalution by EUSFNA. The overall accuracy of EUS-FNA in diagnosing malignancy was 84%(sensitivity of 82.4 %). 1.8% of resections were benign in the EUS group and5.3% in the non-EUS group. There were significantly fewer benign lesionsresected in the pre-operative EUS group than in the surgery only group(p�0.024). With regards to accuracy of staging, overall, EUS was able todifferentiate T1/T2 from locally advanced disease (T3/T4) with an accuracy of91.2%, sensitivity of 90.5% and a specificity of 91.5%. The accuracy of N-stageby EUS was 64.7%, sensitivity was 94.7%, and specificity was 53.1%. Byunivariate survival analysis there was no difference in overall survival amongpatients undergoing resection for pancreatic cancer with and withoutpreoperative EUS FNA (Log rank p�0.1628). In a multivariate Coxproportional hazards model, undergoing preoperative EUS-FNA was not asignificant predictor of time to recurrence (p�0.16). Conclusions: Our studysuggests that preoperative EUS-FNA of suspicious but resectable appearingpancreatic masses may reduce the number of resections of benign pancreaticlesions. EUS had a high diagnostic and T staging accuracy but a relativelylow N stage accuracy among patients with confirmed malignancy. We foundno association between EUS-FNA and recurrence or survival in patients withresectable pancreatic cancer.

Mo1522Yield of Endoscopic Ultrasound in Screening for PancreaticCancer At a Tertiary Cancer CenterAmanpal Singh*1, Nathaniel H. Kwak1, Harshad S. Ladha2,Abhik Bhattacharya1, Somashekar G. Krishna2, William A. Ross1,Manoop S. Bhutani1, Eric P. Tamm3, Priya Bhosale3,Gottumukkala S. Raju1, Jason B. Fleming4, Jeffrey H. Lee1

1Division of Gastroenterology, Hepatology & Nutrition, University ofTexas MD Anderson Cancer Center, Houtson, TX; 2Division ofGastroenterology, Hepatology & Nutrition, Ohio State UniversityMedical Center, Columbus, OH; 3Department of Radiology,University of Texas MD Anderson Cancer Center, Houston, TX;4Department of Surgical Oncology, University of Texas MDAnderson Cancer Center, Houston, TXBackground: Family history has been noted to be a significant risk factor forpancreatic cancer. Limited data is available on the yield of endoscopicultrasound (EUS) for screening patients with family history of pancreaticcancer. The aim of this study is to report the outcomes among patients whounderwent screening for pancreatic cancer at a tertiary cancer center.Methods: Our endoscopy database was searched to find patients whounderwent EUS for screening for pancreatic cancer. Data was extracted inregards to the demographics, findings of the EUS, CT, and MRI and numberof first- and second-degree relatives with pancreatic cancer. Follow up datawas collected on patients with positive findings. Results: During 2003-2012,EUS was performed for screening of pancreatic cancer in 52 patients (meanage 56 � 7.8 years, males 42%). At least 2 first-degree relatives (FDR) hadhistory of pancreatic cancer among 19 patients, 1 FDR and 1 second-degreerelative (SDR) in 23 patients and 1 FDR only in 7 patients, 1 SDR and Lynchsyndrome in 1 patient, 1 SDR and BRCA2 mutation in 1 patients and BRCA2mutation only in 1 patient. Among 52 patients, BRCA2 mutations were notedamong 6 patients and BRCA1 mutation in 1 patient. EUS found anabnormality in 15 (28.8%) patients. Cysts were noted on EUS in 11 (21.2%)patients. The median cyst size was 4.8mm (range 2.6-10.9mm). Differentialsof cysts were IPMN in 5 patients and non-specific cysts in 6 patients. Fineneedle aspiration (FNA) was not performed in any patient with cyst. EUSfound a pancreatic mass lesion in 4 patients. FNA was performed in 3patients and found benign tissue. In 1 patient with mass lesion withechogenicity similar to spleen, FNA could not be performed due to a bloodvessel crossing the path of the needle. CT and MRI found lesions in 8/52(13.5%) and 5/52 (7.7%) patients, which were confirmed on EUS in 3/52(5.8%) and 4/52 (7.7%) patients, respectively. None of the patients underwentpancreatic resection. Table 1 shows the follow up information on patientswith cystic and mass lesions noted on the EUS. The median follow upduration was 12 months (rang 0-42 months) Conclusion: EUS screening forpancreatic cancer among high-risk patients has a low yield of significantabnormalities that need further management. Cystic lesions were mostcommonly found among the abnormalities. Although CT scan found moreabnormalities than MRI, more findings of MRI could be confirmed on EUSthan those of CT scan.

Follow up of Patients with Positive Findings on the Screening EUS

Patient Risk Initial Findings

Length ofFollow Up

(mo.) Follow Up Conclusions

CysticLesions

65/M 2 FDR 9.2 mm SB-IPMN on EUS 42 Unchanged on EUS at 6 months.Normal pancreas on CT at 42months

66/F 2 FDR Benign 4.7 mm cyst onEUS

22 Unchanged on EUS at 10months. Unremarkable pancreason CT at 22 months

46/F 1 FDR,2 SDR

10.9 mm SB-IPMN onEUS

32 Unchanged on EUS at 18months. Stable 7 mm IPMN onCT at 32 months.

64/M 1 FDR,3 SDR

2.9 mm SB-IPMN on EUS 0

62/F 2 FDR,1 SDR

2 benign cysts (2.8, 3.5mm) on EUS

29 Unremarkable pancreason MRIat 29 months

53/F 1 FDR 2.6 mm SB-IPMN on EUS 39 SB- IPMN increased to 7.3 mmon EUS at 13 months. SB-IPMN 7mm on MRI at 39 months

60/M 2 FDR 4 mm SB-IPMN or cysticneoplasmon EUS

10 Cyst no longer present on EUSat 10 months

55/F 1 FDR,1 SDR,1 TDR

Benign 4.8 mm cyst onEUS

31 Unremarkable pancreason MRIat 31 months

58/F BRCA2

5 benign cysts (2.5-6.6mm) on EUS

0

78/M 1 FDR 4 mm cyst, 5.2 mmdilated side branch onEUS

14 Unremarkable pancreason MRIat 14 months

62/F 1 FDR,1 SDR,1 TDR

3 benign cysts (2-5 mm)on EUS

6 �3 mm cyst on MRI at 6 months

Mass Lesions53/M 1 FDR,

1 SDR6.3 mm mass lesion onEUS , suspectedsplenule

12 Unchanged on EUS at 4 monthsand 12 months

73/M 2 FDR 2 nodules (5.2, 6.6 mm)s/p FNA, cytology - non-malignant

11 Nodules no longer present onEUS at 11 months

49/M 2 FDR 9 mm hypoechoic areas/p FNA, cytology - non-malignant

3 Unchanged on EUS at 3 months

54/F 1 FDR;BRCA2

8 mm hypoechoic areas/p FNA, cytology - non-malignant

3 Hypoechoic area decreased to4.8 mm on EUS at 3 months s/pFNA, cytology - non-malignant

�SB-side branch

Mo1523Does FISH Increase the Positive Yield of EUS-Guided FNA ofPancreatic Masses?. Results of an Ongoing TrialJose-Guillermo De La Mora-Levy*, Antonio Hidalgo,Miguel A. RAMíRez-RAMíRez, Sergio R. Sobrino-Cossio,Juan Octavio Alonso-Larraga, Julio Sanchez Del Monte,Angelica Hernandez-GuerreroGastroenterology, Instituto Nacional de Cancerología, Estado DeMexico, MexicoBackground: EUS-guided FNA cytology is currently the suggested gold standardfor diagnosis of patients with pancreatic masses. A number of ancillarytechniques have been proposed to increase the yield of FNA; such as thepreparation of cell blocks or the addition of Digital Imaging Analysis (DIA) orFluorescence-Indiced in Situ Hybrization (FISH). Our aim was to find if theaddition of FISH to FNA-obtained material would increase positive yield ofcytology and cell block analysis, Material and Methods: All patients undergoingEUS-guided FNA of a pancreatic mass and suspicion of adenocarcinoma, wereincluded. An electronic linear echoendoscope (Olympus GF-UCT160) was usedby two experienced endosonographers. Either EZ-shot or Echotip needles(Cook-medical), were used (22 or 25 gauge). No cytologist or technician waspresent and at most 3 passes were taken per patient. With the material obtained,air-dried slides were prepared, as well as cell blocks from the material sent in aspecial fixative (Carbo-wax). FISH was performed from two slides, using acommercially available kit (UroVysion/Abbott Molecular) that detects aneuploidyin chromosomes 3, 7,17 as well as loss of region 9p21, which has been validatedin previous studies. The gold standard was either a compatible clinical scenario

Abstracts

www.giejournal.org Volume 77, No. 5S : 2013 GASTROINTESTINAL ENDOSCOPY AB413