Inclusion Criteria:

1. Malignant-appearing pancreatic mass, or proven pancreatic cancer involving the pancreatic genu, body, or tail. 2. Any level of abdominal or back pain considered to be potentially related to the mass: 1. New onset pain (<3 months) 2. Constant. 3. Centrally located. 4. With or without irradiation to the back. 5. No obvious other source of pain based on history and physical examination by the attending endosonographer. 3. No possibility of surgical management. 4. Signed, informed consent. 5. Celiac axis accessible for bilateral neurolysis at EUS.

Exclusion Criteria:

1. Allergy to bupivacaine

Pancreatic malignancies are the second most frequent gastrointestinal malignancy in Canada. From cancer mortality statistics in 2014, there were 4,700 new cases of pancreatic malignancies second only to colorectal cancer, representing 2.4% of all cancers. Even with chemotherapy, the median survival for patients with pancreatic adenocarcinoma is 6 to 10 months. Few patients are diagnosed at a resectable stage (12%-20%) so many patients are candidates for palliation only. In this context, one of the most important symptoms is pain because it often affects both quality of life and survival. 70 to 80 % of patients with pancreatic cancer have abdominal pain at the time of diagnosis. Adequate pain control is therefore an essential component of care in these patients. In the initial phase, the pain is visceral, but with disease progression, somatic pain may occur, especially due to the peri-pancreatic invasion of neural structures, such as the celiac plexus. Standard analgesics such as acetaminophen are usually ineffective and the use and effectiveness of opioids is frequently limited by side effects such as nausea, constipation, somnolence, confusion or respiratory depression. The celiac plexus is immediately adjacent to the gastric wall. Endoscopic ultrasound (EUS) allows EUS-guided trans gastric injection of neurolytic agents around the celiac plexus (celiac plexus neurolysis [CPN]). Under conscious sedation, the echoendoscope is advanced into the stomach, just distal to the gastro-esophageal junction. The region of the celiac plexus is identified around the takeoff of the celiac artery from the aorta. Then, under real-time ultrasound guidance, a 19g needle is used to inject a neurolytic agent such as absolute alcohol around the base of the celiac artery. The entire procedure takes approximately 5 minutes. Absolute alcohol causes the immediate destruction of the celiac plexus neurons, by precipitation of endoneural lipoproteins and mucoproteins. The effectiveness of CPN, is well established. It is safe, produces significant pain reduction, significantly reduces narcotic requirements, and may even increase survival. The investigators were the first to publish a randomized, sham-controlled trial demonstrating the efficacy of EUS-CPN for pain due to pancreatic cancer, and authored the most recent published guidelines on the use of EUS-CPN. Based on our experience in over 1000 neurolysis procedures, patients undergoing EUS-guided CPN may experience pain, acutely during alcohol injection, and sometimes post-procedure, for up to a few hours. (Unpublished observations) It is possible that the presence of pain during injection of alcohol indicates that the celiac plexus has been accurately targeted and may therefore portend better long-term pain control. Currently, during the neurolysis procedure, it is standard procedure to inject bupivacaine immediately before injecting absolute alcohol, to theoretically prevent pain during and after the procedure. The true value of bupivacaine during neurolysis has never been studied. There are no data showing whether bupivacaine injection has any real influence on intra-procedural, immediate post-procedural, or long-term pain control. The injection of bupivacaine before the alcohol may have no effect, a synergistic effect, or an antagonistic effect, by diluting the alcohol, and reducing its neurolytic capacity. Inadvertent intravascular injection of bupivacaine may also cause irreversible cardiac arrhythmias and death. In other words, in the worst case scenario, the injection of bupivacaine may increase procedural risk, without any associated benefit in terms of pain reduction. The EUS team at the CHUM stopped using bupivacaine during neurolysis approximately 2 years ago and has noticed no obvious difference in pain during the procedure or in the immediate post-procedure recovery period, no increase in complications, and a possible reduction in requests for repeat neurolysis

• - suggesting that neurolysis without bupivacaine may be more effective.

(Unpublished observations) The investigators therefore propose a randomized clinical trial to determine whether the exclusion of bupivacaine during EUS-guided CPN improves outcomes, or not.

Arms

Active Comparator: EUS-CPN with bupivacaine

Endoscopic ultrasound guided celiac plexus neurolysis with absolute alcohol 20 mL preceded by injection of 10 ml of bupivacaine 0.5%.

Experimental: EUS-CPN without bupivacaine

Endoscopic ultrasound guided celiac plexus neurolysis with absolute alcohol 20 mL only.

Interventions

Procedure: - EUS-CPN without bupivacaine

Under conscious sedation, an echoendoscope is advanced into the stomach, just distal to the gastro-esophageal junction. The region of the celiac plexus is identified around the takeoff of the celiac artery from the aorta. Then, under real-time ultrasound guidance, a 19g needle is used to inject a neurolytic agent such as absolute alcohol around the base of the celiac artery.

Procedure: - EUS-CPN with bupivacaine

Under conscious sedation, an echoendoscope is advanced into the stomach, just distal to the gastro-esophageal junction. The region of the celiac plexus is identified around the takeoff of the celiac artery from the aorta. Then, under real-time ultrasound guidance, a 19g needle is used to firstly inject bupivacaine and secondly a neurolytic agent such as absolute alcohol around the base of the celiac artery.