W1332

EUS-FNA for Biliary Strictures Following Non-Diagnostic

ERCP CytologyDarren A. Pavey, Josh George, Fletcher D. Srygley, Shalini Balapurdi,Carlos G. Micames, M. Stanley Branch, Paul S. Jowell, Frank G. GressBackground: Accurate pre-operative diagnosis of biliary strictures is important toguide clinical management. The sensitivity of ERCP brush cytology is reported to bebetween 33-58%. EUS-FNA offers an alternative approach when ERCP brushcytology is non-diagnostic. Aim: To report the diagnostic yield of EUS-FNA for biliarystrictures following non-diagnostic ERCP brush cytology. Methods: All patients whounderwent both ERCP and EUS between January 2000 and July 2005 were identifiedfrom a prospectively collected database. Patients who underwent ERCP for a biliarystricture and had non-diagnostic or unsuccessful brush cytology were included.Results: 108 patients with non-diagnostic ERCP brush cytology were identified. EUSrevealed a mass lesion in 81/108 (75%) including 27 (25%) in whom previousimaging failed to detect a mass lesion. EUS-FNA was performed in 102/108 (94%)patients resulting in the following cytological diagnoses: adenocarcinoma 46 (45%),atypical 8 (8%), suspicious 9 (9%), negative 36 (35%), and other malignancies 3(3%). Surgical histopathology was available for 41 (38%) patients and confirmedmalignancy in 25/41 (61%): adenocarcinoma (22), other cancers (3); and a benignprocess in 16/41 (39%) patients. EUS-FNA diagnosed malignancy in 2/8 (25%)upper, 4/10 (40%) middle and 39/85 (46%) lower third bile duct strictures. Inpatients confirmed to have malignancy at surgery, EUS-FNA was non-diagnostic formalignancy in 9/25 (36%) including 3 with negative and 6 with atypical/suspiciouscytology. There were no false positive EUS-FNA cytology findings. The sensitivity,specificity, positive predictive value, negative predictive value and accuracy of EUS-FNA for the diagnosis of malignancy were: 67%, 100%, 100%, 59% and 78%respectively. EUS-FNA altered management in 68/102 (67%) patients. Two patientshad complications following ERCP: mild pancreatitis (1) and cholangitis (1). Therewere no complications associated with EUS-FNA. Conclusions: EUS-FNA isa sensitive method for the diagnosis of biliary strictures following non-diagnosticERCP brush cytology and alters management in the majority of these patients. Inaddition, EUS may detect a mass where previous imaging studies have beennegative.

W1333

Endoscopic Ultrasound-Guided Fine Needle Aspiration Cytology

of Lesions Within the Liver and the Porta Hepatis (Liver Hilum).

A Single Center Report of 67 Lesions in 59 PatientsIoannis Karoumpalis, Thomas Anastasiou, Charitini Salla,Athanasios Chatzinikolaou, Anastasios Konstantinidis, Irini Doumani,Vassilios Delis, Vassilios Balatsos, Katerina Karadima, Nikolaos SkandalisObjectives: Endoscopic ultrasound-guided fine needle aspiration (EUS-FNA) isincreasingly being used in gastrointestinal oncology. The liver and the porta hepatisconstitute a common site of primary tumors or metastasis, sometimes difficult to benon-surgically accessible. Therefore EUS-FNA of liver and hilar lesions is ofparamount importance. The aim of the study was to report the indications,cytological diagnosis, diagnostic effectiveness, safety, and clinical impact of EUS-FNA of liver (LL) and hilar (HL) lesions in our tertiary facility center. Methods:Between May, 2003, and October, 2005, a total of 59 patients with 67 different LLand HL were enrolled in this study. Demographic data of patients as well asindications, complications and findings of EUS-FNA were collected and reviewed.Results: EUS-FNA was performed on 43 LL and 24 HL (in total 67) in 59 patients (43male and 16 female with a mean age of 61 G 12 years). Of these 59 patients, 35 hadonly LL, 18 only HL and 6 had both LL and HL. The main indication for theprocedure was previous imaging of a liver mass in 14 out of 59 patients (23,7%),a pancreatic mass in 13 (22%), or abnormal pancreatobiliary tree findings in 10patients (16,9%). On average 1,5 G 0,6 needle passes were performed througheach of the 67 lesions. No complications occured. Of these 67 lesions, 45 (67,2%)were cytologically diagnosed as malignant (36 LL and 9 HL), 18 (26,9%) benign (4 LLand 14 HL) and 4 (5,9%) were nondiagnostic (3 LL and 1 HL). A total of 18 out of 22nonmalignant lesions were confirmed as negative by follow up. Regardingmalignancy, EUS-FNA had a sensitivity of 91,8%, specificity of 100%, yieldinga negative predictive value of 81,8% and a positive predictive value of 100%. Amongpatients with malignancy (40 out of 59) the most common cytological diagnosis wasmetastatic adenocarcinoma from the pancreas in 16 of them (40%), followed bycholangiocarcinoma in 8 patients (20%). Of the 41 patients with liver lesions, atleast 17 (41,5%) had normal previous noninvasive imaging, in 14 of whom (34,1%)malignancy was diagnosed. EUS-FNA had a positive impact on patient managementin 82,4% of subjects with cytology positive for malignancy. Conclusions: EUS-FNA isa safe and effective procedure for determining the malignant potential of lesions inthe liver and the porta hepatis. It can have major impact on clinical managementrevealing undiagnosed lesions, upstaging tumors, obtaining tissue for diagnosis inpatients with inoperable disease and avoiding surgery. Our experience suggests thathaving a pathologist present in the endoscopy unit provides optimal amount oftissue for correct cytological diagnosis.

W1334

Diagnosis of Intraabdominal and Mediastinal Sarcoidosis Using

Endoscopic Ultrasound Guided Fine Needle AspirationHazar Michael, Sammy Ho, Bonnie Pollack, Mala Gupta, Frank GressBackground: In the presence of a compatible clinical picture, the diagnosis ofsarcoidosis requires pathologic confirmation of non-caseating epithelioidgranuloma in affected tissues. The standard procedure of choice for most patientsis bronchoscopy with transbronchial biopsy (TBB), which has a diagnostic yieldranging from 40-90%. The lowest yield with TBB is encountered in cases withpredominant mediastinal or intraabdominal lymphadenopathy (LN) and minimalparenchymal lung involvement. In these settings, endoscopic ultrasound (EUS)guided fine needle aspiration (FNA) may provide an alternative diagnostic modality.Methods: Retrospective analysis of 21 patients (10 male and 11 female) with clinicalsuspicion of sarcoidosis and mediastinal and/or intraabdominal LN or masses on CTwho underwent EUS-guided FNA. Results: EUS-guided FNA diagnosed sarcoidosisin 18 out of 21 (86%) patients. In three patient EUS-guided FNA was inconclusiveand patients underwent mediastinoscopy with lymphadenectomy whichestablished the diagnosis of sarcoidosis. Five of the 21 patients had intraabdominalLN/masses, and EUS-guided FNA of the intraabdominal pathology was diagnostic ofsarcoidosis in three out the five patients (60%). One of the five patients wasdiagnosed via EUS-guided FNA of mediastinal LN and the other patient viamediastinoscopy. Three of the 21 patients had prior history of malignancy and EUS-guided FNA had ruled out the recurrence of malignancy in this subset of patients.Conclusion: EUS-guided FNA offers a practical, minimally invasive technique for thediagnosis of sarcoidosis in patients presenting with predominant mediastinal orintraabdominal LN or masses.

W1335

The Incremental Yield of EUS Over Endobronchial US and CT

for Detection of Extrathoracic Metastases in Lung CancerJason Conway, Gerard Silvestri, Carolyn Reed, Joseph Romagnuolo,Robert Hawes, James Ravenel, Brenda HoffmanBackground: Endoscopic Ultrasound (EUS), especially when combined with fineneedle aspiration (FNA), is an established modality for mediastinal staging of non-small cell lung cancer. EUS-FNA can also detect and biopsy suspected extrathoracicmetastases in the celiac area, left adrenal, and liver. Endobronchial US (EBUS) hasthe ability of reviewing thoracic stations not accessible by EUS, howeverextrathoracic areas are not accessible by EBUS. Aim: To describe the frequency ofceliac area, left adrenal, and liver metastases in patients referred for EUSmediastinal staging of lung cancer. Methods: The EUS database at our institutionwas queried to identify all the mediastinal staging cases for known or suspectedlung cancer from 1994 to 2005. Review of the EUS report as well as the electronicmedical revealed data on visualization of the celiac area, left adrenal and liver andFNA cytology results. Data on cross sectional imaging were also collected. Binomialconfidence intervals (CI) were calculated for the incremental yield of EUS inextrathoracic disease. Results: The database search identified 299 patients. Meanage was 65, 68% were male, and 82% were white. The celiac area, left adrenal, andliver were commented on in 100%, 89%, and 75% of reports, respectively.Table 1 shows the number of suspicious lesions seen on CT and/or EUS and theEUS-FNA result. Overall, 40 (13.4%) patients had suspicious celiac area, left adrenal,or liver lesions seen at EUS; only 47% of these lesions were also seen on CT. 6.7%(95% CI 4.1-10%) of all patients had confirmatory cytology by EUS-FNA. CT notedsuspicious extrathoracic lesions in 6.3% of all patients, missing 5 involved celiacnodes, 3 adrenal lesions, and 3 liver lesions. Conclusion: In a small but importantproportion of patients undergoing staging for lung cancer EUS-FNA can confirmextra-thoracic metastases, many of which are missed by CT and would be missed byEBUS. Routine review of the celiac area, left adrenal, and liver is critical in EUSstaging of lung cancer.

Table 1. Lung cancer extrathoracic metastases on CT and EUS-FNA.

Celiacn (%)

Left adrenaln (%)

Livern (%)

Lesions noted on CT 1 (0.3%) 14 (4.7%) 4 (1.3%)Lesions noted at EUS 10 (3.3%) 23 (7.7%) 7 (2.3%)EUS-FNA cytology malignant or atypical 5 (1.5%) 11 (3.7%) 4 (1.3%)

Abstracts

AB270 GASTROINTESTINAL ENDOSCOPY Volume 63, No. 5 : 2006 www.giejournal.org