Search

TPS620 Background: Gallbladder carcinoma (GBC) is a rare disease with a poor prognosis, with an overall estimated 5-year survival rate of 5-13%. Approximately 70% of gallbladder cancers in the US are found incidentally on pathologic analysis after elective cholecystectomy for presumed benign disease. Current management guidelines for incidental gallbladder cancer (IGBC) recommend re-resection for T1b, T2, and T3 lesions, which entails a partial hepatectomy and lymph node dissection. Up to 75% of patients have residual locoregional disease and 20% have disseminated disease at the time of re-resection, both factors strongly associated with poor prognosis and inoperability, respectively. For those who undergo re-resection, the recurrence rate at one year can be as high as 40%. Therefore we designed a Phase II/III trial to assess the benefit of a perioperative chemotherapy approach around the re-resection procedure compared to standard of care adjuvant therapy alone. Methods: We have enrolled 18 of 186 planned patients (Phase III design) on study NCT 04559139/EA2197. Current eligibility criteria include: Incidentally diagnosed T2 / T3 gallbladder cancer with no evidence of metastatic disease or inoperable locoregional disease. At randomization, patients must be within 12 weeks of their index cholecystectomy. Patients are randomized 2:1 to receive perioperative chemotherapy with gemcitabine and cisplatin vs adjuvant therapy alone. The primary endpoint is overall survival. Secondary endpoints include resectability, presence of residual disease, and progression-free survival. Clinical trial information: 04559139 .

TPS776 Background: Neoadjuvant chemotherapy (NAC) and/or chemo-RT may confer benefit to patients with localized PDAC, by better tolerability, tumor down-staging, and increased R0 resections. mFFX or GA are the current NAC backbones; however, a lack of robust predictive biomarker(s) hampers identification of patients most likely to benefit from mFFX or GA. Further, desmoplastic stroma/poor vascularity compromise NAC efficacy, but angiotensin II receptor inhibitor, losartan, might remodel vascular perfusion to enhance chemotherapy activity. We designed the NeoOPTIMIZE trial for patients with newly diagnosed localized PDAC to provide a flexible clinical platform to: 1) evaluate the feasibility and efficacy of early switching of mFFX to GA, and 2) establish a robust biomarker/imaging discovery platform to optimize the NAC backbone. Methods: NeoOPTIMIZE (NCT04539808) is an open-label, non-randomized, phase II trial assessing the preliminary efficacy of an adaptive treatment strategy that allows for early switching of NAC in patients with localized PDAC. Sixty patients (n = 40 resectable/BRCP; n = 20 locally advanced unresectable [uLAPC]) will be enrolled to receive 2 months of preoperative mFFX (oxaliplatin, 85 mg/m2; folinic acid, 400 mg/m2; irinotecan, 150 mg/m2; 5-FU, 2400 mg/m2), then restaging by a multidisciplinary tumor board (multiD-TB). Absent progression (by panc protocol CT and CA19-9 decline/increase < 30% from baseline), patients continue mFFX (4 cycles). If progression (by panc protocol CT; CA19-9 increase > 30%), patients switch to GA (nab-paclitaxel, 125 mg/m2; gemcitabine, 1000 mg/m2) for 2 months. After 4 months of mFFX or mFFX/GA, another restaging multiD-TB will decide to proceed with: a) RT (if vascular involvement) then resection, b) resection, or c) continued chemo (if unresectable). Losartan (50 mg PO QD) is given throughout NAC and RT regimens. The primary endpoint estimates the proportion of resectable/BRPC patients with R0 resection. Assuming that the proportion of R0 is 60%, a sample size of 32 will provide a 95% CI (0.41, 076). To account for a 20% dropout, 40 patients will be enrolled towards primary endpoint. A separate exploratory cohort of 20 uLAPC patients will be enrolled. Secondary endpoints include DFS, PFS, OS, and AEs. Exploratory objectives include correlating clinical outcomes data with changes in blood-based biomarkers (CA19-9, ctDNA, circulating tumor cells etc.) and research DCE-MRI. To date, the trial has enrolled 19 patients: 8 resectable, 7 BRCP, and 4 uLAPC. Clinical trial information: NCT04539808 .

4097 Background: Given a high recurrence rate, surgical resection for localized intrahepatic cholangiocarcinoma (IHCC) is curative in only 30-35% of patients. BILCAP has established a standard of care for adjuvant therapy with monotherapy capecitabine. Gemcitabine, cisplatin and nab-paclitaxel combination chemotherapy (GAP) was associated with a response rate of 45% in biliary cancer and 20% of patients who were previously inoperable underwent margin negative resection. Based on these data, we conducted a neoadjuvant study of GAP for resectable but high-risk IHCC. Methods: A multi-institutional prospective single-arm phase II trial was conducted for patients with resectable high-risk disease, as defined by tumor size > 5cm, multiple tumors, presence of radiographic major vascular invasion, and lymph node involvement. Patients were administered 4 cycles (3 months) of preoperative GAP (gemcitabine 800 mg/m2, cisplatin 25 mg/m2 and nab-paclitaxel 100 mg/m2 on days 1 and 8 of a 21-day cycle) prior to an attempt at curative-intent surgical resection. The primary endpoint was completion of all therapy, including both preoperative chemotherapy and resection. Secondary endpoints were toxicity, radiologic response according to Response Evaluation Criteria in Solid Tumor (RECIST) criteria, RFS, and OS. Thirty evaluable patients provided 73% power to reject a null therapy completion rate of 50%, with a target completion rate of 70% using a one-sided exact test with a Type I error of 0.05. Results: Thirty-seven patients were screened and 30 evaluable patients were enrolled. The trial was sequentially activated at each of the 4 sites from 09/18-02/21 and the final patient was enrolled in 09/21. Median age was 60.5 years and 40% were female. Twenty-three patients (77%, 90% CI: 60.6-88.5%; p = 0.0026) completed all preoperative chemotherapy and underwent surgical resection. Ten patients (33%) experienced grade 3-4 treatment-related adverse events, the most common being neutropenia and diarrhea; 47% required at least one dose reduction. Partial response rate was 23% and disease control rate was 90% (PD: 10%, PR: 23%, SD: 67%). Of the 23 patients who successfully underwent surgical resection, 2 (9%) had minor postoperative complications. Median size of largest tumor was 5.5cm, median number of tumors was 3, and 39% were lymph node positive. Median length of hospital stay was 4 days. There was zero treatment related mortality. Conclusions: This study met its primary endpoint and demonstrated that neoadjuvant gemcitabine/cisplatin/nab-paclitaxel is feasible and safe prior to resection of intrahepatic cholangiocarcinoma and does not adversely impact perioperative outcomes. Continued follow-up for RFS and OS with this treatment strategy is underway and larger validation studies are planned. Clinical trial information: NCT03579771.

TPS500 Background: The majority of patients with intrahepatic cholangiocarcinoma (ICC) present with advanced liver dominant disease rendering them unresectable. The current standard of care supports palliative treatment of unresectable ICC with gemcitabine plus cisplatin. Alternatively, continuous liver-directed therapy using a surgically implanted hepatic arterial infusion (HAI) pump allows for maximal treatment of liver tumors with limited systemic side-effects and can facilitate conversion to a resectable status. No prospective clinical trial has utilized FOLFIRINOX in combination with HAI floxuridine-dexamethasone for ICC. Methods: HELIX-ICC is a first-line, single-center, single-arm, phase II clinical trial enrolling patients with liver-dominant unresectable or multifocal ICC. Only patients with microsatellite stable cancer and no prior liver radiotherapy will be included. Patients are administered dose-modified FOLFIRINOX systemically for 4 cycles over 8 weeks. Those with evidence of disease control on restaging imaging and laparoscopy will proceed to HAI pump placement. Treatment continues with two 28-day cycles of combined HAI floxuridine (1.08 mg/kg) with dexamethasone for 14 days and systemic dose-modified FOLFIRI starting on day 15. The primary objectives are to evaluate the safety and efficacy (disease control rate [DCR] at 6 months) of this novel treatment approach that utilizes our institutional dose-reduced HAI floxuridine protocol in combination with modified systemic regimens to facilitate control of liver disease and maximize patient quality of life (QoL). HELIX-ICC includes many exploratory analyses through longitudinal collection of blood samples, liver biopsies (including an end of trial research biopsy), and QoL metrics (Table). The study is open to n = 30 with an initial safety run-in of 6 patients, of which 4 have enrolled at the time of submission. This study is designed to allow for drop out of 9 patients, with total accrual of 21 patients to protocol completion achieving 80.2% power at 0.05 significance to detect a 25% increase in DCR at 6 months. Clinical trial information: NCT04251715. [Table: see text]

4104 Background: Biliary tract cancers are rare tumors with a median overall survival (OS) of 16 months for node-positive (N+) and 37 months for node-negative (N0) disease despite resection. Lymph node status is a known strong prognostic factor for local recurrence with an average estimated 2-year disease-free survival (DFS): 65.5% for N0 and 29.7% for N+ tumors. The Phase II Intergroup Trial S0809 showed that adjuvant capecitabine and gemcitabine followed by radiotherapy and concurrent capecitabine improved OS in patients with extrahepatic cholangiocarcinoma (EHCC) and gallbladder cancer (GBC) compared to historical controls. We hypothesized that nodal status is a prognostic factor for local recurrence in this patient population who received adjuvant therapy. Methods: This analysis included patients with stage pT2-4, N+ or positive margin EHCC or GBC. Treatment included four cycles of gemcitabine (1,000 mg/m2 intravenously on days 1 and 8) and capecitabine (1,500 mg/m2 per day on days 1 to 14) every 21 days followed by concurrent capecitabine (1,330 mg/m2 per day) and radiotherapy (45 Gy to regional lymphatics; 52.5 to 59.4 Gy to tumor bed). S0809 patients who did not receive radiotherapy were excluded from analysis. Correlations between nodal status, resection margin, and other clinicopathological factors, patterns of recurrence and survival were analyzed, and Cox regression models were used to estimate the prognostic significance of nodal status. A Z-test was used to compare DFS rates between these patients and historical data. Results: A total of 69 patients [EHCC n = 46 (66%); GBCA n = 23 (33%)] were evaluated with a median age of 61.7 (26.1-80.6). The majority of N0 patients were female (17/24, 70.8%), whereas most N+ patients were male (25/45, 55.6%; p < 0.04). Distribution of R0 (66.7%) and R1 (33.3%) resections was similar in the N0 and N+ groups. Thirty-four patients with EHCC had N+ disease (73.9%) compared with 11 patients with GBCA (47.8%, p = 0.03). Nodal status did not significantly impact OS (HR = 2.03, 95% CI 0.92-4.49, p = 0.08) or DFS (HR = 1.75, 95% CI 0.85-3.59, p = 0.13). Two-year OS was 70.6% for N0 and 60.9% for N+ disease (p = 0.11). Nodal status was not significantly associated with 2-year DFS: 62.5% for N0 and 49.8% for N+ (p = 0.20). N+ vs N0 tumors showed higher rates of distant failure (51.1% vs 25.0%, p < 0.04), but similar local recurrence (17.8% vs 12.5%, p = 0.88). The observed 2-year DFS in patients with N+ tumors was significantly longer compared to the historical rate of 29.7% (p = 0.004). Conclusions: This combination adjuvant treatment regimen following curative resection for EHCC and GBCA provides favorable outcomes regardless of nodal status. These data suggest that adjuvant chemoradiation may positively impact local control in N+ patients. These findings need to be validated in future clinical trials. Clinical trial information: NCT00789958.

771 Background: As neoadjuvant Rx for resected PDAC often includes chemoradiation, the PV of PR includes its impact. We began analysis of the impact of NC alone in this setting. Methods: Patients (pts) were identified from the Virginia Mason Pancreaticobiliary Cancer Database. Inclusion criteria: 1) Dx 1/2010 - 3/2019; 2) Path dx PDAC stage I-III; 3) NC ( any type) as sole neoadjuvant Rx; 4) complete surg path data; 5) longitudinal OS known. Exclusion criteria: 1) neoadjuvant chemoradiation; 2) unknown NC (outside providers only). Histologic response was scored as follows: ( 0=complete response, 1 ≥95% response, 2=50-95% response, 3

430 Background: Neuroendocrine tumors (NET) or carcinoids of the ampulla are exceedingly rare in comparison to duodenal NET. Surgical management is widely accepted as the treatment of choice, but for patients that refuse surgery or are poor operative candidates, endoscopic resection may be option. We present a consecutive case series at a tertiary care center describing our experience with endoscopic resection of ampullary NET. Methods: This is a restrospective review with a long-term follow-up of patients with ampullary NET that were endoscopically resected. Outcomes were analyzed based on the histopathologic classification system proposed by the World Health Organization in 2000. Results: Twelve patients (9 male, 3 female), ranging in age from 41 to 86 (mean 59) underwent endoscopic ampullectomy for ampullary NET, with a mean follow-up time of 5 years. Patients had refused surgery or were poor surgical candidates. All, but one incidentally found case, were symptomatic on presentation, with gastrointestinal bleeding being the main symptom. No patients had a hormonal syndrome. The mean size of the lesions was 21 mm (6 mm to 35 cm). Six (50%) patients had a well-differentiated, benign lesion, 6 (50%) patients had a well-differentiated NET with unknown malignant potential (gangliocytic paragangliomas). Eight (67%) were completely excised during the initial endoscopy with 4 requiring re-excision. Only 2 patients developed recurrence, after 2.5 and 10 years and this necessitated a pancreaticoduodenectomy. Five patients had complications (2 for bleeding and 3 for post-ERCP pancreatitis), with zero deaths. Conclusions: Unlike duodenal carcinoids, ampullary NET are rare. Pancreaticoduodenectomy can be offered to fit patients except for gangliocytic paragangliomas that do not require an aggressive operation. However, for those that refuse surgery or are poor candidates, endoscopic ampullectomy can be an option with acceptable short and long-term outcomes.

379 Background: Somatostatin analogue functional imaging with Ga-68 Dotatate PET/CT has demonstrated superiority in lesion detection in patients with NETs. The effect of this novel imaging modality on US clinical practice and its usefulness in different types of NETs is not well described. We describe the impact of initial NETSPOT imaging on diagnosis and management in NET patients at a large urban medical center. Methods: Consecutive patients diagnosed with NETs and referred to our institution who received an initial Ga-68 Dotatate PET/CT between 07/2017-09/2018 were included. Imaging was reviewed and compared to prior available CT, MRI, and/or In-111 Pentetreotide scans. Results: Among 101 patients, 51/50 were female/male, tumor origins were gastroenteropancreatic (GEP) (75%), Unknown Primary (UP) (13%), lung (8%), thymic (2%), and other (2%). All tumors were histologically well/moderately differentiated and 37/51/3/10 were G1/G2/G3/Unknown, respectively. Initial imaging with Ga-68 Dotatate PET/CT revealed additional metastatic disease in 37 of 77(48%) patients with prior evidence of metastatic disease. Most common sites were distant lymph nodes (18), bone (15) and liver (9), peritoneal/pleural (4). A previously UP tumor was identified in 3 patients. No patients with metastatic lung carcinoids (6 atypical, 2 typical) or thymic NETs (2 atypical/G2) NETs had evidence of Ga-68 Dota PET/CT uptake above reference liver SUV levels. Results of imaging altered patient management as follows: 14 initiated systemic therapy due to documentation of progression, in 6 surgical therapy was altered, in 4 biopsy/other management was changed. In 11 patients with no tumor Ga-68 Dotatate uptake, decisions about use of PRRT and somatostatin analogues was altered. Conclusions: In this series, Ga-68 Dotatate PET/CT altered diagnosis and management in 35/101 NET patients. Among GEP and UP NETs, Dotatate imaging diagnosed primarily new nodal, bone, liver and pleural/peritoneal metastases missed by other imaging modalities. Results support the routine use of Ga-68 Dotatate PET/CT in the care of patients with advanced and early stage NETS.

376 Background: The prognostic value of pathologic variables is not consistent for gastroenteropancreatic neuroendocrine tumors (GEP-NETs). We previously demonstrated a limited prognostic role of lymph node (LN) positivity in small bowel NETs (SBNET) compared to pancreatic NETs (panNET). Although minority race is often associated with worse cancer outcomes, the interaction of race with pathologic and oncologic outcomes of pts with GEP-NETS is not known. Methods: Pts with GEP-NETs who underwent curative intent resection at eight institutions of the US NET Study Group from 2000-16 were included. Given few pts of other races, only Black and White race pts were analyzed. Results: Of 2,182 pts, 1,143 met inclusion criteria. Median age was 58 yrs, median follow up was 3 yrs, 48% were male, 14% (n = 157) were Black, and 86% (n = 986) were White. Black pts were more likely uninsured (7 vs 2%, p = 0.005), had symptomatic bleeding (13 vs 7%, p = 0.006), required emergency surgery (7 vs 3%, p = 0.003), and had LN positive disease (47 vs 36%, p = 0.016). Despite this, Black pts had improved 5 yr recurrence free survival (RFS) compared to White pts (90 vs 80%, p = 0.008). The quality of care received was comparable between both groups, demonstrated by similar LN yield at surgery, neg margin resection rate, post-op complications, and need for reoperation or readmission (all p > 0.05). Black pts were more likely to have SBNET (22 vs 13%) and less likely to have panNET (43 vs 68%) compared to White pts (p < 0.001). Consistent with prior data, pts with LN pos panNET had decreased 5yr RFS (67 vs 83%, p = 0.001); however, for SBNET, LN involvement was not prognostic (77 vs 96%, p = 0.08). The prognostic value of LN pos disease was similar between Black and White pts in both SBNET (p = 0.34) and panNET (p = 0.95). Conclusions: Black pts with GEP-NET present with more advanced disease, including higher LN positivity. Despite this, Black pts have improved RFS compared to White pts. Although there may be delays in seeking or reaching care, Black pts received similar quality of care compared to White pts. The improved RFS seen in Black pts may be attributed to the epidemiologic differences in the site of presentation of GEP-NETs and variable prognostic value of LN pos disease.

212 Background: Preoperative factors that reliably predict lymph node (LN) metastases in pancreatic neuroendocrine tumors (PanNETs) are unclear. The number of LNs needed to accurately stage PanNETs has not been defined. Methods: Patients who underwent curative-intent resection of primary non-functional PanNETs at 8 institutions from 2000-2016 were analyzed. Tumors with poor differentiation and Ki-67 > 20% were excluded. Preoperative factors associated with LN metastases were identified. A procedure specific target for LN retrieval to accurately stage patients was determined. Results: Of 2182 pts with GI NETs, 695 underwent resection of PanNETs. 33% of tumors were proximal (head/uncinate), and 67% were distal (neck/body/tail). 26% of pts (n = 158) had LN+ disease, which was associated with worse 5-yr recurrence-free survival (RFS) (60% vs 86%; p < 0.001). Increasing number of +LNs was not associated with worse RFS. Preoperative factors associated with +LNs included tumor size ≥2 cm (OR 6.6; p < 0.001), proximal location (OR 2.5; p < 0.001), moderate vs well differentiation (OR 2.1; p = 0.006), and Ki-67≥3% (OR 3.1; p < 0.001). LN metastases were also present in tumors without these risk factors: < 2cm (9%), distal location (19%), well differentiated (23%), and Ki-67 < 3% (16%). Median LN retrieval was 13 for pancreatoduodenectomy (PD), but only 9 for distal pancreatectomy (DP). Given that PD routinely includes a complete regional lymphadenectomy, a minimum number of LNs to accurately stage pts was not identified. For DP, however, removal of < 7 LNs failed to discriminate 5-yr RFS between LN (+) and (-) pts ( < 7 LNs: 72% vs 83%, p = 0.198; ≥7 LNs: 67% vs 86%, p = 0.002). Conclusions: Tumor size ≥2 cm, proximal location, moderate differentiation, and Ki-67≥3% are preoperative factors that predict LN positivity in resected non-functional PanNETs. Given the 9-23% incidence of LN metastases in patients without such risk factors, routine regional lymphadenectomy should be considered. Pancreatoduodenectomy inherently includes sufficient LN retrieval, while distal pancreatectomy should aim to remove ≥7 LNs for accurate staging.

371 Background: Insurance status predicts access to medical care in the United States. Previous studies show uninsured and government insured patients have worse outcomes than those with private insurance. However, the impact of insurance status on survival in patients with Gastrointestinal Neuroendocrine Tumors (GI-NETs) is unclear. We evaluate the association between insurance status and survival in patients with GI-NETs. Methods: Our analysis includes 2022 patients who had surgical resection of GI-NETs at 8 institutions in the U.S. Neuroendocrine Study Group. Patients were categorized based on insurance as private (PI), government (GovI) or uninsured (UI). Factors associated with insurance status were assessed by uni- and multi-variate analysis. Primary endpoint was overall survival. Results: Patient demographics between the insurance categories were similar in ECOG performance status and tumor size at presentation. GovI patients had a higher median age than PI or UI (66 vs. 54 vs. 56 years respectively; p

139 Background: Gastric carcinoids are rare neuroendocrine tumors of the GI tract. They are typically managed according to their etiology. However, there is little known about the impact of surgical strategy on the long-term outcomes of these patients. Methods: All patients who underwent resection of gastric carcinoids at 8 institutions from 2000-2016 were analyzed retrospectively. Tumors were stratified according to subtype (I, II, III, IV) and resection type (local resection LR or formal gastrectomy FG). Clinicopathological parameters, recurrence-free (RFS), and overall survival (OS) were compared between groups. Results: Of 79 patients identified with gastric carcinoids, 34 had type I lesions associated with atrophic gastritis, 4 had type II lesions associated with a gastrinoma, 37 had type III sporadic lesions, and 4 had type IV poorly-differentiated lesions. The mean age of presentation was 56 years in predominantly Caucasian (77%) and female (63%) patients. Mean tumor size was 2.4 cm and multifocal tumors were found in 24 (30%) of patients with the majority occurring in those with type I tumors. Lymph node positive tumors were seen in 15 (19%) patients and 7 (8%) had M1 disease; both most often in type IV followed by type III tumors. R0 resection was achieved in 56 (71%) patients while 15 (19%) had R1 resections and 6 (8%) had R2 resections. Patients with type I and III tumors were equally likely to have a LR (50% and 43% respectively) compared to FG while those with type II and IV all had FG with one exception. Type IV tumors had the poorest RFS and OS while Type II tumors had the most favorable RFS and OS (p < 0.04 and p < 0.0004, respectively). While there was no difference in RFS in those patients undergoing FG versus LR, OS was worse in the FG group (p < 0.017). This trend persisted when type II and type IV groups were excluded (p < 0.045). Conclusions: Gastric carcinoid treatment should be tailored to tumor type, as biologic behavior rather than resection technique is the more important factor contributing to long-term outcomes.

265 Background: Small bowel neuroendocrine tumors (SB-NETs) frequently involve regional lymph nodes (LNs). The prognostic value of LN positivity on recurrence of disease is not well defined. The number of LNs needed to accurately stage patients is unknown. Methods: All patients with primary SB-NETs who underwent curative-intent resection at 8 institutions in the US Neuroendocrine Tumor Study Group between 2000 and 2016 were identified. Patients with distant metastatic disease were excluded. The association of LN positivity with recurrence of disease and the extent of lymphadenectomy required were analyzed. Results: Of 2182 patients with resected NETs, 203 had SB-NETs. Median age was 60 yrs, 56% were male, and median follow-up was 39 months. 83.5% of patients (n = 157) had LN positive disease. There was no difference in 3-year recurrence free survival (3-yr RFS) among patients with 1 or 2 positive LNs compared to patients with negative LNs (p = 0.63). Patients who had 3 or more positive LNs had a worse 3-yr RFS compared to those with 0, 1, or 2 positive LNs (n = 92 vs n = 73; 3-yr RFS 82% vs 92%; p < 0.001). Retrieval of 8 or more LNs was associated with a higher positive LN count compared to less than 8 LNs (4.6 vs. 1.6; p = 0.002). However, an increasing LN ratio was not associated with 3-yr RFS. When examining patients who had less than 8 lymph nodes retrieved, there was no difference in 3-yr RFS in those patients with 3 or more positive LNs compared to those with 0, 1, or 2 positive LNs (3-yr RFS: 100% vs 91%; p = 0.37). Retrieval of more than 8 lymph nodes, however, accurately discriminated patients with 3 or more positive LNs compared to those with 0, 1, or 2 positive LNs (3-yr RFS: 79.7% vs 93.5%; p = 0.005). Conclusions: For patients undergoing curative-intent resection of small bowel NETs, accurate lymph node staging requires a minimum of 8 lymph nodes for examination. 3 or more positive LNs is associated with decreased 3-yr RFS compared to 0, 1, or 2 positive lymph nodes. A thorough regional lymphadenectomy is critical for accurate staging and management of patients with small bowel neuroendocrine tumors.

216 Background: Neutrophil, monocyte and lymphocyte counts are important biomarkers of the tumor-induced systemic inflammatory response. A lower monocyte-to-lymphocyte ratio (MLR) and neutrophil-to-lymphocyte ratio (NLR) is associated with a favorable prognosis for certain tumors. However, this association has not yet been demonstrated in resected pancreatic neuroendocrine tumors (PNETs). The aim of this study was to investigate the prognostic significance of MLR and NLR in patients with resectable PNETs with regards to overall survival (OS). Methods: Eligible patients undergoing surgery for PNETs between 2000 and 2016 were identified using a national multi-center cohort dataset. Their pre-operative neutrophil, monocyte and lymphocyte counts were imported and NLR and MLR were calculated. Receiver operating characteristic (ROC) curve analysis was used to determine the optimal cut offs for NLR and MLR. Univariate analyses were used to compare patient factors and tumor characteristics in different groups. The difference in OS between high and low groups was explored with the use of Kaplan Meier curves and compared using log-rank tests. Results: A total of 635 patients were analyzed. Univariate analysis revealed that higher lymphocyte, low monocyte and low neutrophil counts were associated with improved OS (P = 0.0002, 0.014 and 0.011 respectively). The mean NLR and MLR level was 4.11 ± 0.219 and 0.4295 ± 0.0196, respectively. Lower NLR ( < 4) and lower MLR ( < 0.25) were significantly associated with improved OS (p < 0.001 & p < 0.0001). Median survival was 141 months in the low NLR group (n = 474) and 119 months in the high NLR group (n = 161). The median survival in the high MLR group (n = 464) was 120 months and has not been reached in the low MLR group (n = 171). On multivariate analysis, low MLR was independent predictor of improved OS when controlling for age, race and comorbidities (p < 0.001). Conclusions: For resectable PNETs, low pre-operative NLR and MLR levels are significantly associated with improved survival. NLRs and MLRs may serve as valuable markers to stratify PNET patients for subsequent therapies and clinical trial enrollment.

694 Background: Intestinal neuroendocrine tumors (NETs) can often be indolent, with survival commonly exceeding 10 years. However, a subset can be more aggressive. The search for a prognostic test based on preoperative lab results would be of benefit. An elevated NLR has been found to be predictive of diminished overall survival (OS) in several cancers, and we hypothesized that this same trend would hold true for intestinal NETs. Methods: Utilizing the 8 member institutions of the retrospective U.S. NET Study Group, patients undergoing resection of intestinal NETs were included. After excluding patients undergoing emergent surgery, those with preoperative sepsis or distant mets, and those with missing neutrophil/lymphocyte lab values, 281 patients were included for analysis. The cutoff values for NLR were determined using receiver operator curve (ROC) analysis. Univariate analysis was performed based on these groups. A Kaplan-Meier survival curve was created based off optimal NLR from ROC analysis. OS was calculated as date of surgery to date of death/ last follow-up. SPSS 23.0 was used for analysis. Results: Of the 281 patients identified, 9% were appendiceal, 9% colonic, 21% duodenal, 5% rectal, and 56% jejunal/ileal. Using a NLR of >2.47, the ROC area under the curve was 0.609 ( p=0.002). Median survival for the 130 patients in the 3% in 30% vs. 26%, respectively, p=0.57), and post-op complications (36 vs. 42%, respectively, p=0.310) were similar between groups. A significantly greater number of patients in the NLR high group had positive lymph nodes (60 vs. 47%, p =0.002). Conclusions: For patients undergoing curative resection of intestinal NETs with no distant mets, a preoperative NLR of ≥2.47 was significantly predictive of decreased OS, with those in the elevated NLR group having an OS 3.7 years less than those in the

e15771 Background: Optimum therapy (Rx) for BRPC is unknown. Since 2008, we have used neoadjuvant Rx with extended course chemotherapy (CT) but not routine neoadjuvant chemoradiation (CRT). Initial results were presented in 2013 (Rose et.al. J Clin Oncolabstr 4043). We present updated findings here. Methods: Patients (pts) were prospectively identified in our institutional PC database. Inclusion criteria: 1) bx-proven PC; 2) radiographic staging per AHPBA/NCCN criteria; 3) no prior Rx; 4) negative staging laparoscopy whenever feasible; 5) all neoadjuvant Rx at our institution, 6) followup ≥ 24 weeks (wks) from initial Rx . Unless disease progression or Rx intolerance noted, pts received gemcitabine/docetaxel (G/D) as neoadjuvant CT x 24 weeks . At that time, all pts felt likely to achieve R0 resection offered surgery; other pts offered 5FU –based CRT if medically fit. Results evaluated by "intent to treat". Results: Among 129 pts, characteristics include median age 66 yrs (range 33-88 yrs), ECOG PS 0/1/2+ 86/36/7, 97%/43% venous /arterial involvement. 78% (101/129) pts completed ≥80% intended CT; 38% (6/16) pts age > 80 and /or ECOG 2+ (p < 0.01) 51% (66/129) pts were resected (44 R0, 22 R1 ≤ 1 mm margin ). 49% (63/129) pts were not (24 disease progression, 17 surgeon decision (anatomy/safety), 11 Rx toxicity/ comorbidity, 6 unresectable at surgery, 5 pt withdrawal). 45% (31/66) and 23% (16/66) resected pts received postop CRT/CT respectively. Median f/u is 47 months (mo). 58% (38/66) resected pts recurred: 18% (12/66) local, 41% (27/66) systemic. For resected pts, median PFS is 25.0 (95% CI: 16.2-32.7) mo, median OS is 37.6 (95% CI: 27.4-58.5) mo. 5-yr OS is 28% ( 95% CI :14-44%). Median OS for non-resected pts is 13.1 (95% CI 10.9-16.9) mo .Median OS for all pts is 22.1 (95% CI: 18.7-28.4) mo. Conclusions: 1.Our series is distinctive with respect to size, use of laparoscopic staging and neoadjuvant G/D CT as standards,and median f/u 2. Pt selection is key; e.g.approach suboptimal for pts > 80 yrs/ ECOG PS 2+ 3. Reported OS compares favorably with other BRPC and de novo resectable PC cooperative group results 4. Detailed analysis of our study (e.g. prognostic factors) will aid future clinical research in BRPC.

373 Background: AT with chemotherapy (CT) + radiation (RT) has been shown to improve PDAC survival over surgery alone. Although race and socioeconomic status can affect outcomes in PDAC, the impact of rural or remote residence on the delivery and effect of AT has not been studied. Methods: Patients undergoing pancreatectomy for PDAC were identified from the National Cancer Data Base between 2006 and 2013. Individuals were classified as living in a metro area, urban/rural adjacent to metro area (URA), and urban/rural remote area (URR). Patients with less than 6 months follow-up were excluded. Logistic regression was performed to assess residence as a predictor of receiving AT. Overall survival (OS) as a function of inhabitance was estimated by the method of Kaplan and Meier and prognostic factors were identified by Cox regression. Results: A total of 32,521 individuals underwent pancreatectomy for PDAC. The majority of AT was delivered in academic research facilities in 56% of patients while only 29% of patients received both CT and RT. Univariate analysis demonstrated individuals in URR were less likely to receive CT (55% vs 58%, p < 0.01) but not RT (30% vs 31%, p < 0.261) and had a longer interval to AT (82 vs 75 days, p < 0.009) than those in metro areas. However on multivariate analysis URR inhabitance was no longer predictive of any form of AT (OR = 0.892, 95% CI: 0.792-1.006, p = 0.062). Hispanic ethnicity, Medicaid insurance, uninsured status, and lower education were all predictive of decreased likelihood of receiving AT. Median OS was inferior for URR dwellers with pathologic T2 and T3 tumors compared to those in metro areas (19.8 vs. 24.4 months, p = 0.044 and 17.5 vs. 19.4 months, p < 0.001). Cox regression revealed URR location remained independently associated with poorer OS (HR 1.076, 95% CI: 1.008-1.149, p < 0.029). Conclusions: While living in a URR does not lead to reduced access to AT, it is associated with a worse OS in resected PDAC. This may be due to inadequate AT or other socioeconomic factors present in URR patients. Attention must be focused on improving oncologic care for groups susceptible to treatment disparities.

392 Background: Gemcitabine-taxane combination chemotherapy has demonstrated a survival benefit in metastatic pancreatic cancer (PC). We present our experience with gemcitabine/docetaxel (gem/tax) based adjuvant treatment (Rx) following curative intent surgery. Methods: Patients with upfront resectable PC from January 1, 2010 to December 31, 2015 were identified from our institutional database and tumor registry. We included resected patients who received gem/tax as initial Rx administered exclusively at our institution with or without chemoradiation (CRTx). Patients were excluded if they died or recurred prior to Rx or received neoadjuvant Rx. Survival was estimated by Kaplan-Meier methods and prognostic factors identified by Cox regression. Results: Of 185 eligible patients, 58 met study criteria. Characteristics included: median age of diagnosis 64.9 years, female gender 56.9%, Whipple procedure 69% and R1 resection in 55.2%%. Tumour characteristics included: median size 28.0mm, poor differentiation 53.5% and AJCC stage 2A/2B 27.6%/67.2%. Patients completing ≥ 80% of 24 week Rx was 89.7% (n = 52). Of those patients, 71.2% received post gem/tax CRTx. With a 51.2 month median follow-up (95%CI: 37.1-55.7), median disease-free survival (DFS) and overall survival (OS) were 35 months (95%CI: 20.7-NR) and 52.3 months (95%CI: 27.4-NR), respectively. Five year OS was 49.5% (95%CI: 33.7-63.4). Patients receiving CRTx (n = 37) had a superior OS compared to patients who only completed gem/tax Rx (n = 15)(61.5 vs. 26.7 months, p = 0.04). Prognostic variables impacting OS on multivariate analysis (HR, 95%CI) included: margin status (4.55, 1.46-14.2, p = 0.01), AJCC stage (10.9, 2.46-47.9, p = 0.002) and administration of CRTx (0.08, 0.03-0.24, p < 0.000). Variables impacting DFS included: AJCC stage (3.29, 1.15-9.41, p = 0.03), tumour size (2.59, 1.11-6.04, p = 0.03) and administration of CRTx (0.23, 0.09-0.56, p = 0.001). Conclusions: In this select cohort of resected PC patients with adverse pathologic features, adjuvant gem/tax with or without CRTx is feasible with favorable DFS and OS. These findings support further prospective studies of gem/taxane-based adjuvant Rx in PC.

365 Background: Pancreatic ductal adenocarcinoma (PDAC) is characterized by dense stroma and connective tissue growth factor (CTGF) overexpression. Although prolonged overall survival (OS) can be achieved through resection, only approximately 15% to 20% of patients are treated with surgery. Previously, pamrevlumab (P), a fully human monoclonal antibody blocking CTGF, was combined with gemcitabine/erlotinib (G/E) in a study of advanced PDAC. P plasma levels of > 150 μg/mL and low baseline CTGF levels resulted in prolonged OS. We hypothesize that neoadjuvant (NA) treatment of LAPC with P + G/nab-paclitaxel (NP) may alter LAPC’s stroma, increase R0 resectability, and improve OS similar to that of patients resectable at presentation. Methods: LAPC subjects, confirmed by NCCN criteria and staging laparoscopy, were randomized 2:1 to G/NP ± P for 6-cycle/24-week treatment. Safety and efficacy (resection rate, OS, progression-free survival, tumor response) were assessed. Resection was performed in subjects who met protocol-defined criteria with no contraindications. Results: As of 26 Sep 2016, 25 subjects were enrolled. Of 14 subjects in Arm A (G/NP + P), 7 completed, 2 discontinued. Of 11 subjects in Arm B (G/NP), 4 completed, 4 discontinued. 6 subjects (43%) in Arm A and 4 (36%) in Arm B experienced SAEs; no P-related SAEs were reported. Out of 7 eligible Arm A subjects, 3 were not resected. Successful resection was achieved in 4 Arm A subjects (3 R0, 1 R1) and 1 Arm B subject (R0) (Table). Conclusions: The combination of G/NP + P is feasible and well-tolerated with no incremental safety signals in this study. The addition of P suggests a trend toward increased resectability among LAPC subjects. Clinical trial information: NCT02210559. [Table: see text]

e15693Background: Adjuvant rx using single-agent gem in resected PC typically has produced median / 5-year (yr) survivals of 18-22 months (mo)/ 10%, respectively. As taxanes added to gem improved O...

321 Background: As neoadjuvant therapy of locally advanced, borderline resectable pancreatic cancer (BRPC) is becoming more widely utilized; better indicators of progression are needed to help guide therapeutic decisions. The aim of this study is to determine if CA19-9 response during treatment predicts disease progression. Methods: A retrospective review was performed on all patients with BRPC (by AHPBA/SSO consensus criteria) between 2008-2015 who received 24 weeks of neoadjuvant gemcitabine and docetaxel. Patients with medical comorbidities limiting treatment completion were excluded. Serum CA19-9 levels were checked at baseline and every 3 weeks while on therapy. A normal CA19-9 level was defined as < 37.5 units/mL and levels with concomitant biliary obstruction were censored. CA19-9 response was analyzed as a predictor of disease progression, recurrence, and survival. Results: Eighty patients were included with a mean of 11 CA19-9 levels checked per patient during treatment. Thirty-two (40%) progressed on treatment (18 local and 14 distant) and 48 (60%) were resected (79% R0). CA19-9 responses were categorized into 5 groups: 1) Always normal [n = 13]; 2) Increasing [n = 3]; 3) Slow decline [n = 7]; 4) Rapid decline with plateau [n = 41]: and 5)Rapid decline with late rise [n = 16]. Univariate logistic regression analysis found that a final CA19-9 decline > 50% of baseline (OR 0.06, p =

362 Background: Duodenal and ampullary adenocarcinomas are rare gastrointestinal cancers that share similar anatomic location and treatment strategy. We report a single-institution experience regarding the association between clinicopathologic features, treatment, and survival outcomes. Methods: A retrospective review of all patients resected with curative intent for duodenal adenocarcinoma (DUO) between 2005-2015 and ampullary adenocarcinoma (AMP) between 2011-2015 at VMMC was performed. For AMP, histologic subtyping into intestinal (IT) and pancreatobiliary (PB) phenotypes was determined. Demographic and clinicopathologic parameters were compared between DUO and AMP patients using Chi-square test. Overall survival was calculated using Kaplan-Meier analysis and prognostic factors were identified by univariate Cox regression. Results: Patients with DUO (n = 44) presented at higher T-stage (p = 0.002) and with larger tumors (4.35cm vs 2.33cm, p < 0.001) than AMP patients (n = 46). DUO patients had a higher rate of surgical complications (68% vs 41%, p = 0.01) with a trend for more pancreatic fistulas (36% vs 20%, p = 0.08). There was no difference in median overall survival between groups. Factors positively influencing survival included Caucasian race (p = 0.02) and normal CA19-9 level at diagnosis (p = 0.01). Tumor factors negatively influencing survival included positive lymph nodes (p = 0.04), lymphovascular invasion (p = 0.001), and perineural invasion (p = 0.02). Within AMP, the PB subtype presented at higher T-stage (p = 0.01) and with more positive lymph nodes (p = 0.03) than IT. There was no difference in survival between subtypes. Majority received adjuvant chemotherapy (88.8% in AMP, 76.3% in DUO), fewer received adjuvant radiotherapy (23.3% in AMP, 30% in DUO), but no survival difference was seen. Conclusions: DUO presents with larger tumors and higher T-stage than AMP and is associated with more surgical complications. The PB phenotype has more advanced pathologic features than IT. No survival difference was seen between anatomic locations or subtypes. Better surgical and chemotherapeutic strategies may be needed to overcome high risk features. Longer follow-up with more patients is needed to confirm these findings.

457 Background: Curative therapy for PDAC is currently achievable through resection only, but ≤ 15% of patients are candidates for surgery. FG-3019, a recombinant human mAb to connective tissue growth factor (CTGF), plus gemcitabine (G) significantly reduced tumor growth, increased apoptosis and improved survival vs G alone in the KPC mouse model. Higher FG-3019 exposure and lower baseline (BL) CTGF were independently and significantly associated with improved overall median and 1-year survival (OS) in a Ph2a trial with G and erlotinib in stage 3 or 4 PDAC. Neoadjuvant treatment of LAPC with FG-3019 combined with multidrug therapy may change the fibrotic character of LAPC and thus may make it amenable to surgical resection and potentially improve OS. Methods: Patients with unresectable LAPC (confirmed by laparoscopy) are randomized to G/NP ±FG-3019 for 6-cycle/24-week treatment. Safety, efficacy, PK and changes in CT, PET, & CA19.9 from BL through end of treatment (EOT) are assessed. Up to 4 endoscopic ultrasound (EUS) guided core biopsies are collected at BL. Protein microarrays of pre- and post-treatment biopsies are prepared from tumor and stromal cells isolated by laser capture microdissection. Surgical exploration is performed in subjects who achieved the protocol-defined criteria for possible resection. Results: To date, 11 subjects were enrolled with 5 randomized to +FG-3019 arm. There were no complications from laparoscopy or EUS core biopsies that were clinically significant or delayed dosing. Per investigator assessment, no FG-3019-related SAEs were reported, nor any identified toxicity suspected from combination of FG-3019 with chemotherapy. Overall, CA19.9 mean change from BL was −62.7% at Week 8 (n = 8), and target lesions were reduced by 20.6% (n = 6) at Week 16. Three out of 4 subjects had PET SUVmax reductions of ≥ 30% at EOT. Additional data from protein microarrays will be presented. Conclusions: Based on the limited data collected so far, this trial design is feasible and the G/NP +FG-3019 combination is well-tolerated and appears to be active. Summary results on surgical exploration of subjects will be provided later. Clinical trial information: NCT02210559.

TPS500 Background: Only 15% of patients with pancreatic ductal adenocarcinoma (PDAC) are eligible for resection. Patients with locally advanced PDAC may become candidates for surgery following donwstaging with neoadjuvant therapy. PDAC often exhibits a high degree of desmoplasia, characterized by extensive connective tissue stroma. FG-3019 is a recombinant human monoclonal antibody that binds to connective tissue growth factor and significantly reduces tumor growth, neovascularization, and metastases in mouse models of PDAC. Therefore, FG-3019 may change the fibrotic character of PDAC to one more amenable to resection. In a phase I trial of stage III/IV patients, FG-3019 did not add toxicity when combined with gemcitabine and erlotinib. Methods: Phase 1/2, randomized, open-label trial to evaluate safety and tolerability of gemcitabine plus Nab-paclitaxel with FG-3019 or gemcitabine plus Nab-paclitaxel in subjects with locally advanced, unresectable PDAC (clinicaltrials.gov #NCT02210559). Primary endpoint is treatment-related adverse events. Secondary endpoints include R0 resection rate, surgical complications, tumor response, and survival. Patients with biopsy-proven PDAC are included following radiographic and laparoscopic confirmation of irresectability and no metastases. Sample size of 40 is based on clinical judgment for initial exploration of safety, tolerability, and treatment effect. Each cycle is 28 days and subjects receive up to six cycles of treatment. Those who complete treatment undergo resection based on the following criteria: reduction in CA 19-9 level by more than 50%, FDG-PET SUVmax decrease by >30%, or radiological tumor response per RECIST to meet the definition of resectable or borderline resectable. Safety will be assessed through adverse event monitoring, clinical laboratory testing, ECG, ECOG performance status, vital signs, and physical exam until 28 days after the last FG-3019 infusion. All resected subjects will be followed for 30 days after discharge and undergo safety assessment per standard of care. Since the trial opened in 8/2014, we have enrolled 1/40 patients. Clinical trial information: NCT02210559.

4043 Background: The optimal surgical (S) approach to BRPC is unknown. We evaluated an approach to BPRC using extended course chemotherapy (CT) without routine neoadjuvant chemoradiation (CRT). Clinical outcomes were evaluated on an "intent to treat" basis. Methods: Patients (pts) were identified from a prospectively-maintained database started in 2008. Pts required 1) Dx BRPC with non-tail primary per radiographic staging using AHPBA/NCCN guidelines 2) No prior therapy (Rx) 3) Negative staging/exploratory laparoscopy prior to S 4) All cancer Rx at VMMC prior to S. Pts received gemcitabine/docetaxel (G/D) as initial CT. Pts with systemic progression/comorbid complication prior to 24 wks were not offered local Rx; pts with localized cancer at 24 wks judged likely to achieve R0 resection were offered S; all other pts were offered fluoropyrimidine-based CRT. Results: Of 76 identified pts (median age 66), 12 (16%) are on initial CT and not fully evaluable. G/D as sole CT achieved 24 wk disease control in 46/64 (72%) pts and >50% decline in baseline CA 19.9 in 39/55 pts (71%). 53/64 fully evaluable pts (83%) completed 24 wks CT, 11/64 pts (17%) did not (4 systemic progression, 3 Rx- related, 4 intercurrent illness). 50/53 pts attempted local Rx (40 S, 10 CRT, 2 unfit, 1 refused). 30/40 pts (75%) had successful S (all R0), 10/40 pts (25%) had inoperable disease (4 local-subsequently received CRT, 6 systemic). 23/30 pts (77%) have received some form Rx post R0 resection. With median f/o of 20 mo, 23/44 (52%), 17/30 (57%), and 6/14 (43%) receiving any local Rx, S, and CRT-only remain progression free. 13/30 S pts recurred, 3 local (10%) and 10 systemic (33%). 1-yr, 2-yr, 3-yr, and median overall survivals (OS) for fully evaluable pts respectively are 82%, 50%, 36%, and 27 mo. Median OS for pts receiving CT-only (20 pts), CT+CRT (14 pts), and R0 pts are 12 mo, 17 mo, and >20 mo, respectively. 25/30 R0 pts remain alive (range 6-54 mo). Conclusions: 1) Extended neoadjuvant CT without routine neoadjuvant CRT is a feasible approach to BRPC. 2) G/D has significant activity in BRPC. 3) Pretreatment surgical staging combined with the above Rx yields a high probability of R0 resection. 4) OS for both resected and non-resected pt was superior to usual literature comparators.

177 Background: Distinguishing bile duct carcinoma from other diagnoses is often difficult using endoscopic or percutaneous techniques. The cell surface protein CEACAM6 is over-expressed in many gastrointestinal cancers and may be selectively elevated in biliary adenocarcinoma. The aim of the present study is to determine if CEACAM6 can be detected in the bile of patients with biliary cancer and can serve as a diagnostic biomarker for cholangiocarcinoma. Methods: Bile duct and gallbladder bile from patients with benign biliary disease and cholangiocarcinoma (hilar, intrahepatic and distal) was collected at the time of index operation. The concentration of CEACAM6 was quantified by sandwich enzyme-linked immunosorbent assay (ELISA) and correlated to pathologic diagnosis. Diagnostic capability of CEACAM6 was evaluated by Wilcoxon rank-sum, logistic regression, and receiver operating characterisitic (ROC) curve analysis. Results: Bile from 73 patients was analyzed: 40 with benign disease and 33 with cholangiocarcinoma. Patients in the benign cohort were younger, predominantly female, and had lower mean biliary CEACAM6 levels than patients in the malignant cohort (102 ng/ml vs. 239 ng/ml; p=0.0006). Logistic regression analysis determined CEACAM6 to be a positive predictor of extrahepatic but not intrahepatic cholangiocarcinoma (p=0.0015). ROC curve with a CEACAM6 level >22.1 ng/ml was associated with 85% sensitivity, 73% specificity, and a likelihood ratio of 3.15 (AUC 0.82) for extrahepatic cholangiocarcinoma. Conclusions: CEACAM6 levels in bile can predict patients with extrahepatic cholangiocarcinoma with acceptable accuracy. Further investigation is warranted in a larger cohort of patients prior to clinical application.

236 Background: The optimum approach to maximize R0 resection rate in BRPC is unknown. We evaluated an approach in BRPC using extended duration neoadjuvant CT, typically without neoadjuvant CRT to assess R0 resection rate on an "intent to treat"basis. Methods: Patients (pts) were identified from a prospectively-maintained database from 2008-2012. Pts required (1) BRPC per radiographic staging using NCCN/AHPBA criteria, (2) no prior PC therapy (rx) PC, (3) negative staging laparoscopy prior to rx, and (4) all cancer rx at VMMC prior to decision on "downstaging" surgery (S). Pts were initially treated with gemcitabine/docetaxel CT until (a) disease progression, or (b) 24 weeks (wks), before assessment for local rx. Pts with systemic progression and/or inability to complete 24 wks of CT were excluded from local rx; pts with localized cancer at 24 wks judged unlikely to achieve R0 resection were offered 5FU-based CRT then reassessed; all other pts were offered S. Results: Of 70 identified pts, 12/70 (17%) are on initial CT; 58 are fully evaluable. 51/58 pts (88%) completed 24 wks of CT. Response rates to CT were (a) serologic 42/58 (72%) > 50% decline in baseline CA19.9 (median 85%), and (b) radiographic of 47 pts evaluable by RECIST, PR/SD/PD were 60%/32%/8% respectively. 14/58 pts (24%) did not receive local rx (7 intercurrent illness, 7 disease progression). 44/58 pts (76%) were offered local rx- 12/58 (21%) CRT and 32/58 (55%) S. 2 pts receiving CRT later underwent S. CRT. 29 pts (50% total evaluable pts, 85% pts undergoing S) achieved R0 resection;10/29 pts (35%) via en bloc venous resections. 5 pts were inoperable (3 local - subsequently given CRT, 2 systemic). At median f/u of 16 mo, 21/29 (72%) R0 resection pts remain disease free (range 7+-47+mo) along with 4/12 pts (33%) receiving CRT w/o R0 resection (range 12+ -30 mo). 25/58 (43%) fully evaluable pts remain progression free on an "intent to treat" basis. Median OS for all pts is 27 mo (95% CI 15-39 mo); 25/29 R0 resected pts remain alive ( median OS> 20 mo). Conclusions: (1) extended neaoadjuvant CT is a feasible approach in BRPC. (2) Gemcitabine/docetaxel has significant activity in localized PC. (3) Neoadjuvant CRT may not be essential to R0 resection in BRPC.

207 Background: The entire biliary tree is at risk for malignant change, but little is known about differences in molecular pathogenesis with respect to anatomic site. CEACAM6 is a membrane protein involved in cell adhesion and signaling that is overexpressed in pancreatic adenocarcinoma and associated with poor prognosis. This study examines CEACAM6 expression in the entire spectrum of biliary carcinomas and its relationship to outcome. Methods: Tissue microarrays containing triplicate cores of paraffin-embedded surgical specimens from patients with bile duct carcinoma (hilar, intrahepatic, distal) and control tissue were stained for CEACAM6 by immunohistochemistry. Clinical, pathologic and survival data were analyzed and correlated with CEACAM6 expression. Survival was estimated using the Kaplan-Meier method and compared with log rank test. Results: One hundred twenty cases of bile duct carcinomas from 1992-2007 were assembled in the tissue microarrays. Strong CEACAM6 signal was present in 30/60 (50%) of hilar tumors, 7/45 (16%) of intrahepatic tumors, 7/15 (47%) of distal tumors, and none of the control tissues. Overall median survival and follow-up were 36.4 months and 98.3 months, respectively. CEACAM6 staining did not correlate with sex, grade, positive lymph nodes, vascular invasion or metastases but was associated with age > 65 (p Conclusions: CEACAM6 may serve as a marker of poor outcome in patients with intrahepatic cholangiocarcinoma and should be further evaluated as a means of selecting patients for adjuvant therapy after resection.

201 Background: Intraductal papillary neoplasm of the bile duct (IPNB) is a type of cholangiocarcinoma characterized by intraductal growth, mucin production and a better outcome compared to the more common nodular-sclerosing type. IPNB may be analogous to IPMN of the pancreas and may be a precursor of invasive cholangiocarcinoma, but its pathogenesis and natural history are ill-defined. This study examines the incidence, clinicopathologic features and outcome of IPNB in a single center. Methods: A consecutive cohort of patients with bile duct cancer (hilar, intrahepatic or distal) was reviewed and those with papillary features identified. Histopathologic morphology and immunohistochemical staining for cytokeratin and mucin proteins were utilized to classify IPNB into subtypes. Survival data were analyzed and correlated with clinicopathological parameters. Results: Between 1993 and 2008, 40 IPNBs were identified in hilar (24/144), intrahepatic (4/86) and distal (12/113) bile duct cancer specimens (11.7%). Histopathologic examination revealed 27 pancreatobiliary, 4 gastric, 3 intestinal, and 6 oncocytic subtypes; cytokeratin and mucin staining was similar to that of IPMNs of the pancreas. Invasive carcinoma was seen in 29/40 (72%) IPNBs. Overall median survival was 59 months and was not different between IPNB locations or subtypes. Factors associated with a worse median survival included depth of invasion (39 months for > 5mm, 128 months for < 5mm, and 144 months for none, p Conclusions: IPNBs are an uncommon variant of bile duct cancer, representing approximately 10% of all cases, occur throughout the biliary tract and share histologic and clinical features with IPMNs of the pancreas. These lesions may represent an alternative carcinogenesis pathway. Given their significant malignant potential, they should be treated aggressively with margin-negative resection. No significant financial relationships to disclose.

e21040 Background: SPARC is an albumin-binding protein associated with poor prognosis that may mediate intratumoral accumulation of nab-P. SPARC is associated with both tumor cells and stromal comp...