SSAT Abstracts 1031 Natural History of Patients Followed Radiographically With Mucinous Cystic Neoplasms of the Pancreas Linda Ma, Michael D'Angelica, Ronald P. DeMatteo, Peter Kingham, Vinod Balachandran, William R. Jarnagin, Peter Allen OBJECTIVE: The aim of this study was to evaluate the outcome of patients presumed to have IPMN (elevated cyst fluid carcinoembryonic antigen CEA ≥192 ng/mL), who were initially selected for radiographic surveillance. METHODS: All patients who had a cystic lesion of the pancreas and had a cyst fluid CEA ≥192 ng/mL and evaluated at our institution over the past 15 years were included. Patients with less than 6 month follow-up were excluded. Patients were stratified by treatment strategy: those who underwent initial surgical resection (within six months of first radiographic diagnosis), and those who were recom- mended for radiographic surveillance. The natural history and disease progression of these two groups was examined. RESULTS: During the 15 year study period, 227 patients were identified who had a documented cyst fluid CEA ≥192 ng/mL (median 961, interquartile range 448-3474 ng/mL) and met inclusion criteria. The majority of these 227 patients had branch-duct IPMN (190/227, 83%), 26 were main-duct (11%), and 11 were mixed-type (5%). The minority of patients underwent immediate resection (63/227, 28%). Within the group of 164 patients who were selected for initial radiographic surveillance, 87% were branch-duct lesions, median cyst diameter was 1.9 cm (range 0.4-10.8 cm), median cyst fluid CEA was 813 ng/ml (interquartile range 414-2823 ng/mL), and 87% met consensus criteria for radiographic surveillance (142 patients met consensus guidelines, 22 were outside of consensus guidelines). With a median follow-up of 56 months, 48 of the 164 patients (29%) have undergone resection. There were three cases (2%) of high-grade dysplasia, and two cases of invasive carcinoma (1%) found on pathology. Three of these five cases of high- risk disease were in the 22 patients who were followed outside of guidelines. CONCLUSIONS: Appropriately selected patients with small branch-duct IPMNs can be safely followed with serial surveillance with a low risk of malignant progression. 1032 Improved Survival Following Pancreatic Cancer Among Patients Receiving Metformin Marcelo Cerullo, Faiz Gani, Joseph K. Canner, Timothy M. Pawlik Background: Preclinical evidence has demonstrated anti-tumorigenic effects of metformin, associating metformin use with improved survival and decreased recurrence following resec- tion of endometrial and gastric cancer. The effects of metformin following pancreatic cancer remain undefined. We sought to assess the association between metformin use and survival using a large, nationally representative sample of patients undergoing surgery for pancreatic cancer. Methods: Patients undergoing a pancreatic resection between January 01, 2010 and December 31, 2012 were identified using the Truven Health MarketScan database. Clinical data, including history of metformin use, as well as operative details and information on long- term outcomes were collected. Multivariable Cox proportional hazards regression analysis was performed to assess the effect of metformin use on overall survival (OS). Results:A total of 3,401 patients were identified. The mean age of patients was 54.2 years (SD=9.1 years). Roughly one-half of patients were female (n=1,739, 51.2%); 49.1% (n=1,669) presented with a Charlson co-morbidity index of 3 or greater (CCI ≥3) and 18.7% (n=636) had diabetes. At the time of surgery, 60.1% (n=2,044) of patients underwent a pancreaticoduodenectomy, 35.5% (n=1,207) a partial/distal pancreatectomy, while 4.4% (n=150) had a total pancreatec- tomy. On pathology, 1,059 (31.1%) had lymph node metastasis. Metformin use was identified in 284 patients (8.4%) and was more commonly administered among patients without locally advanced disease (6.80% vs. 9.06%, p=0.028). While OS was comparable between patients within the first year of surgery (OS at 1 year: 65.4% [95%CI 63.5%-67.3%] vs. 71.2% [95%CI 64.8%-76.6%]), patients who received metformin demonstrated an improved OS beginning at 18 months following surgery. On multivariable analysis adjusting for patient and clinicopathologic characteristics, risk factors associated with a higher risk of death included age (HR 1.03, 95%CI 1.02-1.04, p<0.001) and CCI ≥3 (HR 1.64, 95%CI 1.33- 2.02, p=0.02). In contrast, metformin use was independently associated with a decreased risk of mortality (HR 0.71, 95% CI: 0.58-0.88, p<0.001). Conclusions: Metformin use was associated with an improved overall survival among patients undergoing pancreatic surgery for pancreatic cancer. Further work is necessary to better understand metformin's role in modifying cancer specific and overall health outcomes. S-1202 SSAT Abstracts 1033 Extent of Resection Does Not Affect the National Failure to Treat Localized Pancreas Cancer with Post-Resection Adjuvant Chemotherapy John R. Bergquist, Christopher R. Shubert, Tommy Ivanics, Rory Smoot, Michael L. Kendrick, David M. Nagorney, Michael B. Farnell, Mark J. Truty Introduction: Survival benefit of post-resection adjuvant chemotherapy has been well dem- onstrated for patients with localized pancreas adenocarcinoma (PDAC). Despite this, up to 40% of patients do not receive adjuvant therapy after curative intent resection. Due to higher expected perioperative morbidity for pancreatoduodenectomy (PD) compared to distal partial pancreatectomy (DPP), we hypothesized that patients undergoing the less morbid DPP would be more likely to receive adjuvant therapy and have associated survival benefit compared to those undergoing PD. Methods: The National Cancer Data Base (2004-2012) was reviewed for patients with localized (AJCC Stage I/II) PDAC who underwent DPP and PD. Patients receiving neoadjuvant therapy were excluded. Univariate and multivariable analysis were used to identify factors associated with therapy receipt. Unadjusted Kaplan-Meier analysis and adjusted Cox proportional hazards modeling of overall survival (OS) were performed. Results: 9503 patients underwent pancreatectomy for localized PDAC: 1645 (17.3%) DPP and 7858 (82.7%) PD. Despite less extensive resection, an equal proportion of patients did not receive adjuvant chemotherapy between groups (DPP 36.2% vs PD 36.0%, p=0.904). Prognostic characteristics were similar between groups except DPP patients had greater pre- operative comorbidity (Charlson Deyo 2+ 9.4% vs. 6.6%, p<0.001), fewer N1 lesions (50.2% vs. 70.0%, p<0.001), and a lower rate of positive margin (20.1% vs. 23.9%, p=0.004). Of patients with data on surgical approach, DPP were more likely to be completed in minimally invasive fashion (p<0.001). DPP had shorter length of stay (median 7 vs 10 days, p<0.001), but 30-day readmission rates were equivalent (DPP 8.7% vs. PD 8.2%, p=0.448). There was no difference in type of treatment center (academic vs. community, p=0.23) or in receipt of radiation therapy (35.4% vs 37.6%, p=0.248). Multivariable analysis confirmed that type of procedure was not independently associated with receipt of adjuvant chemotherapy (p= 0.385). Median OS was improved for patients receiving adjuvant chemotherapy compared to surgery alone (PD 22.2 vs. 15.1 months, p<0.001; DPP 25.4 vs. 17.4 months, p<0.001, Figure). This result was confirmed on multivariable analysis controlling for significant demo- graphic and pathologic factors (adjuvant therapy/surgery alone HR 0.56 p<0.001; PD/DPP HR 0.98, p=0.74). Conclusions: Patients with localized PDAC who undergo DPP are equally unlikely to receive post-resection adjuvant chemotherapy as are patients undergoing PD. Patients with localized PDAC who receive post-resection adjuvant therapy have improved survival regardless of surgical procedure performed. These data suggest that factors unrelated to extent of resection and associated morbidity are driving the nationwide failure to provide adjuvant therapy in patients with localized PDAC. Unadjusted Kaplan-Meier estimates of overall survival for PDAC patients undergoing pancrea- toduodenectomy (left panel) and distal partial pancreatectomy (right panel) with surgery alone vs. surgery plus post-resection adjuvant chemotherapy. 1034 Response to Neoadjuvant Therapy and Its Impact on Survival in Resected Pancreatic Cancer Katelin A. Mirkin, Erin K. Greenleaf, Christopher S. Hollenbeak, Joyce Wong Background: Surgical resection and systemic therapy offers the only hope for long-term survival in pancreatic cancer. Neoadjuvant therapy (NAT) has been increasingly employed to optimize outcomes; however, little is known about its impact on pathologic response. This study evaluates the response to NAT and its impact on survival in resected pancreatic cancer. Methods: The National Cancer Data Base (2003-2011) was retrospectively reviewed for patients with clinical stages 1-3 pancreatic carcinoma who underwent NAT and surgery vs. surgery alone. Response to NAT, determined by comparison of clinical with pathologic staging, was classified as downstaged, no change, or upstaged. Univariate, Kaplan-meier, and multivariate analyses using Weibull model were performed. Results: 11,989 patients who underwent NAT and surgery and 6,338 patients who underwent surgery alone were included. With NAT, 362 (3%) were downstaged, 7,007 (58%) had no change, and 4,620 (39%) progressed to a higher stage. Patients who were downstaged tended to be younger, male, have less comorbidities, and higher clinical stage disease, as compared to those who had no change in stage, or progressed despite NAT. Of the NAT cohort, 9,328 (78%) had evaluable therapy information: 4,267 (46%) underwent neoadjuvant chemotherapy (NAC) and 5,061(54%) underwent neoadjuvant chemotherapy and radiation (NACR). Neoadjuvant chemotherapy and NACR seemed to afford similar median survival in clinical stage I and II disease. In clinical stage III disease, NACR extended median survival over NAC (16.4 vs. 11.9 months in those downstaged, 16.3 vs. 13.7 months in no change, and 14.4 vs. 11.1 months in those upstaged, respectively). On multivariate analysis, utilization of NAT was associated with a survival benefit (0.82 HR with NAC and 0.73 HR with NACR), as compared to surgical resection alone. Downstaging of disease afforded a 13% lower hazard of mortality up to 5 years, as compared to no change, while upstaging demonstrated a 22% higher