Chemosensitivity Testing to Predict Chemosensitivity to Gemcitabine Using EUS-FNA Specimens from Unresectable Pancreatic Carcinoma

Abstracts

quality and yield of samples may increase with the number of sites sampled (‘‘4x5’’ appears better than the other techniques).

Characteristics medications

W1282 Prospective Comparison of Endoscopic Ultrasound-Guided Fine-Needle Aspiration Processed As Cell-Block Using High Suction Versus Regular Suction Rami El Abiad, Leana Guerin, Chris S. Jensen, Henning Gerke Background: Endoscopic Ultrasound (EUS)-guided fine needle aspiration (FNA) is an established technique for tissue acquisition from lesions within proximity to the gastrointestinal tract. Tissue in aspirates can be accumulated by centrifugation and processed as cell-blocks. The aim of this study was to investigate if the specimen quality and diagnostic yield of cell-blocks can be increased by using higher suction during aspiration (HS-FNA) compared to regular suction (RS-FNA). Patients and Methods: Fifty-two patients with 53 solid lesions requiring EUS-guided tissue sampling were prospectively included. HS-FNA for cell-block preparation was performed with a 22 gauge needle and 35cc of suction obtained with a 60cc syringe. Additional needle passes with RS-FNA (5-10cc of 10cc syringe) were obtained for cell-block and for smears as guided by on-site cytology assessment. Sample quality between HS-FNA and RS-FNA cell-blocks was compared using a scoring system. Diagnostic accuracies of cell-blocks and FNA-cytology were assessed. Biopsy sites included pancreatic masses (nZ18), mediastinal lymph nodes (nZ6), abdominal lymph nodes (nZ8), lung masses (nZ3), mediastinal masses (nZ 2), abdominal masses (nZ2), liver lesions (nZ3), adrenal gland masses (nZ3), gastric subepithelial tumors (nZ5), duodenal subepithelial tumors (nZ2) and one ampullary lesion. The final diagnosis was malignant in 42 cases and benign in 11 cases. Results: Tissue quality was better with HS-FNA in 34.0% of cases (18/53), with RS-FNA in 18.9% of the cases (10/53) and equal in the remainder (25/53). The accuracy of the combined cytologic diagnosis was 90.6% (48/53). Diagnostic accuracy was 52.8% (28/53) for HS-FNA and 30.2% (16/53) for RS-FNA cytology (pZ0.03). Conclusion: Specimen quality and diagnostic yield of cell-blocks is higher when performed with HS-FNA compared to RS-FNA. This method enables a diagnosis in 53% of cases with a single needle stick, which may be of particular value if an on-site cytopathology service is not available or if accessory stains are required. However, it does not replace the preparation of multiple smears during FNA.

W1283 Endoscopic Ultrasound-Guided Celiac Plexus Neurolysis (EUS-CPN) in Patients with Locally Advanced and Unresectable Pancreatic Adenocarcinoma: A Randomized Prospective Study of 10 mL Versus 20 mL of Alcohol Julia K. Leblanc, Michelle Symms, Susan M. Rawl, Cynthia S. Johnson, John M. DeWitt, Lee McHenry, Stuart Sherman, Kathleen A. McGreevy, Mohammad A. Al-Haddad, Thomas J. Howard, Keith D. Lillemoe, Thomas F. Imperiale Introduction: The accepted dose of alcohol used for neurolysis during EUS-CPN is 10 ml. The optimal dose to achieve pain relief in patients with unresectable pancreatic adenocarcinoma is unknown. Aim: To examine the effect of two doses of 98% alcohol (10 ml versus 20 ml) in EUS-CPN and its impact on pain relief and quality of life in subjects with unresectable pancreatic adenocarcinoma. Methods: 20 subjects with clinical stage 3 or 4 pancreatic adenocarcinoma and pain were randomized to receive 10 ml or 20 ml of 98% alcohol during EUS-CPN. Subjects received 20 ml of 0.75% bupivacaine into the celiac region followed by 98% alcohol with a 22 gauge needle. The celiac ganglia were directly injected when identified. If the celiac ganglia were not identified the medication was injected in the region of the celiac ganglia. Baseline pain was assessed using a numeric rating scale (NRS) from 0-10 and validated brief pain inventory (BPI). Subjects were contacted by phone 24 hours post EUS-CPN and weekly thereafter to assess pain relief by an interviewer who was blinded to the randomization. Pain was measured using NRS and BPI. Pain relief was defined as a 30% or greater decrease in pain scores without an increase in medications used to control pain. Quality of life surveys (QLQ-C30 & QLQ-PAN26) were administered at baseline, week 2, week 4, and every 4 weeks thereafter. Duration of pain relief, onset of pain relief, pain medications, and quality of life were measured. Patients were followed until their pain returned to baseline. Results: None of the subjects had a prior EUS-CPN. There were no differences in pain relief between groups with respect to palliative chemotherapy (pZ0.77) or radiation (pZ0.36). In subjects who did not have their ganglia directly injected (3 in 10 ml group, 6 in 20 ml group) the median (range) of pain relief was 21 (1-35) and 28 (7-112) days for the 10 ml and 20 ml alcohol groups respectively (pZ0.19). Median (range) onset of pain relief for this subgroup was 1 (1-14) days. Quality of life scores directly correlated with pain relief on multiple measures. There were no complications. Conclusions: 20 ml of 98% alcohol may be more efficacious in providing pain relief but does not affect onset of pain relief. Quality of life is directly related to degree of pain relief.

www.giejournal.org

Group A (NZ10)

20 ml bupivacaine, 10 ml 98% alcohol mean age (SD) 66 (14) ganglia injected 5 overall effectiveness (%) 40 median (range) onset of pain 4 (1-14) relief (days) mean (95%I) duration of 21 (5.7, 11.4) effectiveness (days)

Group B (NZ10) 20 ml bupvicaine, 20 ml alchol 62 (10) 5 90 1 (1-14) 51.3 (2.8, 99.9)

0.25 1.0 0.029 0.73 0.59

W1284 Chemosensitivity Testing to Predict Chemosensitivity to Gemcitabine Using EUS-FNA Specimens from Unresectable Pancreatic Carcinoma Takeru Wakatsuki, Atsushi Irisawa, Goro Shibukawa, Tadayuki Takagi, Hidemichi Imamura, Yuta Takahashi, AI Sato, Masaki Sato, Tsunehiko Ikeda, Takuto Hikichi, Katsutoshi Obara, Hiromasa Ohira Background: Although gemcitabine has been associated with both clinical benefit and prolongation of survival in patients with advanced pancreatic carcinoma, the treatment results have never been satisfactory. We had studied chemosensitivity to gemcitabine using EUS-FNA biopsy specimens from unresectable pancreatic carcinoma, and had reported fairly good results (DDW 2008 in San Diego). The aim of this study is to confirm whether this chemosensitivity testing is useful to predict the clinical response to gemcitabine for unresectable pancreatic carcinoma. Patients and method: We enrolled seventeen consecutive patients with unresectable pancreatic cancer, who underwent EUS-FNA in Fukushima Medical University Hospital between Oct. 2006 and Aug. 2008. None of the patients had received any chemotherapy or radiotherapy before this study. Chemosensitivity testing by ATP assay was performed with fresh EUS-FNA speciemens from primary tumor. Chemosensitivity (treated/control ratio: T/C) was calculated as the percentage of quantity of ATP of a tumor treated with gemcitabine and of control. All patients were treated with gemcitabine (1,000 mg/m2, 30-minutes iv infusion) administered on days 1, 8, and 15, which was repeated every 4-weeks, and prospectively followed. Tumor response was evaluated by computed tomography scan every 2 cycles according to RECIST. We assessed correlation between T/C ratio and progression free survival (PFS) for confirmation of clinical usefulness. Results: There was a tendency toward correlation between T/C ratio and PFS (rsZ-0.5011, pZ0.0708). When T/C was divided into two groups at 70%, 12 patients were T/C ZO 70% and 5 patients were T/C ! 70%. One patient was PR, 5 patients were SD, and 6 patients were PD in T/C ZO 70%, whereas all patients were SD in T/C ! 70%. Median of PFS was 60 days in T/C ZO 70%, whereas 246 days in T/C ! 70%, there was a tendency toward difference between them (pZ0.0586). Conclusion: From the result of this prospective study, it might be said that this chemosensitivity testing is useful to predict the clinical response to chemotherapy leading to designing tailor-made treatments for the patients with unresectable pancreatic carcinoma.

W1285 Adequacy of EUS Guided True Cut Liver Biopsy Ruben D. Acosta, William Twadell, Peter E. Darwin Background: Liver biopsies performed via the percutaneous route are quick and well tolerated. In some patients, biopsies are obtained via the transjugular approach to minimize risk. While EUS guided liver core biopsies have been reported, there is no data to guide the number of passes required to assess for routine liver histology. Our objective was to determine if the Quick Core ultrasound biopsy needle could obtain adequate liver biopsy samples on cadaveric tissue as compared to a standard percutaneous biopsy method. Methods: Liver biopsies were obtained using the Wilson Cook Quick Core (Cook Medical, Bloomington, IN) 19 gauge 2 cm ultrasound biopsy needle (Q-C) on two separate explanted cadaveric livers under direct visualization. Five passes were made on each cadaveric liver with the Q-C needle and then each was placed in a separate container. For controls, 2 passes were also made using a standard 16 gauge Monopty (Conmed Corp/Bard, Covington, GA) Liver Biopsy Gun (MBG) on each liver. A GI pathologist assessed the adequacy of biopsy samples between the two methods in terms of the length of the specimen and number of portal tracts. Based upon standard pathologic criteria, samples were deemed adequate if they contained at least 5 portal tracts. An additional surrogate marker utilized was a length of at least 1 cm. Results: In cases where the sample was fragmented, the total length of the core was recorded as the sum of the fragments. Length and portal tract calculations were averaged for both cadaver livers. The mean length of the 4 passes with the MBG was 1.8 cm with a range of 1.2-2.2 cm. The mean length of the 10 passes with the Q-C was 0.8 cm with a range of 0.5-1.3 cm. The mean number of portal tracts per pass for the MBG

Volume 69, No. 5 : 2009 GASTROINTESTINAL ENDOSCOPY AB331