Search

Background Prolonged survival of patients after pancreaticoduodenectomy can be associated with late complications due to altered gastrointestinal anatomy. The incidence of gastric cancer is increasingly reported. We set out to examine our experience with gastric cancer as a late complication after pancreaticoduodenectomy with a focus on incidence, risk factors, and outcomes. Methods We queried our prospectively collected institutional database for patients that developed gastric cancer after pancreaticoduodenectomy and conducted a systematic review of the literature. Results Our database revealed 6 patients who developed gastric cancer following pancreaticoduodenectomy, presenting with a mean age of 62.2 years and an even sex distribution. All of those patients underwent pancreaticoduodenectomy for malignant indications with an average time to development of metachronous gastric cancer of 8.3 years. Four patients complained of gastrointestinal discomfort prior to diagnosis of secondary malignancy. All of these cancers were poorly differentiated and were discovered at an advanced T stage (≥ 3). Only half developed at the gastrointestinal anastomosis. Four underwent surgery with a curative intent, and 2 patients are currently alive (mean postgastrectomy survival = 25.5 months). In accordance with previous literature, biliopancreatic reflux from pancreaticoduodenectomy reconstruction, underlying genetic susceptibility, and adjuvant therapy may play a causative role in later development of gastric cancer. Conclusion Long-term survivors after pancreaticoduodenectomy who develop nonspecific gastrointestinal complaints should be evaluated carefully for complications including gastric malignancy. This may serve as an opportunity to intervene on tumors that typically present at an advanced stage and with aggressive histology.

Background Pancreatic ductal adenocarcinoma (PDAC) does not respond to immune checkpoint inhibitors (ICI) therapy as single agent treatments including anti-PD-1 antibody. One of the mechanisms for the resistance of PDAC to ICI is now attributed to the immunosuppressive microenvironment (TME) in PDAC. Myeloid cells are thought to be the predominant immunosuppressive cells in the TME. Human interleukin-8 (IL-8) is a pro-inflammatory chemokine in the CXC family and has the capability of recruiting myeloid cells into the TME to promote tumor progression and immune escape. Therefore, several anti-IL-8 blockade antibodies were developed including HuMax-IL8 and B108-IL8, which both are fully human IgG1 kappa monoclonal antibodies. We therefore tested whether anti-IL-8 antibodies can potentiate anti-tumor activity of anti-PD-1 antibody in a humanized model of PDAC. Materials and Methods We reconstituted the immune system of the NGS mice with ex vivo activated human T cells and a combination of CD14+ and CD16+ myeloid cells after the mice were orthotopically implanted with human PDAC cells. 10x single nuclei RNA-Seq data processing was further performed to analyze differentially expressed genes among certain cell clusters. Results Our results showed that anti-PD-1 antibody alone had a minimal anti-tumor activity when mice was reconstituted with ex vivo activated T cells. Interestingly, the infusion of the combination of CD14+ and CD16+ myeloid cells together with anti-PD-1 antibody resulted in a modest anti-tumor activity. Adding either HuMAX-IL8 or B108-IL8 led to a significantly enhanced anti-tumor activity. Both CD14+ and CD16+ myeloid cells appeared to be needed for the full anti-tumor activity of IL-8 blockade because mice infused with only CD14+ myeloid cells did not respond to IL-8 blockade and mice infused with only CD16+ myeloid cells responded partially to IL-8 blockade. This result suggested that the target of IL-8 is mainly present in CD16+ myeloid cells and is likely to be granulocytes. Tumor infiltrating immune cells were isolated and demonstrated that IL-8 blockade increases CD45+CD11b+CD15+CD14- myeloid cells, which is known to comprise neutrophils and granulocytic myeloid derived suppressive cells (G-MDSC), in the tumors. Reconstitution of the mice with myeloid cells led to a decrease of CD8+ T cells in the tumors; however, IL-8 blockade brought the CD8+ T cell number back to the baseline. Consistent with an effect of IL-8 blockade on the increase of CD15+CD14- myeloid cells, single nuclear RNA sequencing analysis of the tumor tissues showed that the innate immune response and cytokine response pathways in the myeloid cell cluster were activated by IL-8 blockade. Conclusions This result suggested that IL-8 blockade did not simply inhibit myeloid cells as previously anticipated, but potentiated myeloid cells for the innate immune response and concomitant production of type I cytokines. Such immune responses may subsequently activate the effector T cells as the single nuclear RNA sequencing analysis demonstrated enhanced activation signals in the T cell cluster from the tumors treated by anti-IL-8 antibodies. Taken together, this study supports further testing of anti-IL-8 antibodies including B108-IL8 and HuMax-IL8 in combination with anti-PD-1 antibodies for PDAC treatment. Disclosure Information P. Li: None. N. Rozich: None. J. Wang: None. J. Gai: None. J. Wang: None. Y. Xu: None. B. Herbst: None. R. Yu: A. Employment (full or part-time); Significant; NovaRock. S. Muth: None. N. Niu: None. K. Li: None. V. Fune: None. A. Osipov: None. C. Wolfgang: None. M. Lei: A. Employment (full or part-time); Significant; NovaRock. T. Liang: None. L. Zheng: B. Research Grant (principal investigator, collaborator or consultant and pending grants as well as grants already received); Significant; Bristol-Meyer Squibb, Merck, iTeos, Amgen, NovaRock, Inxmed, Halozyme. E. Ownership Interest (stock, stock options, patent or other intellectual property); Significant; Alphamab, Mingruzhiyao. F. Consultant/Advisory Board; Significant; Biosion, Alphamab, NovaRock, Akrevia/Xilio, Ambrx, Novagenesis, Datarevive, Snow Lake Capitals, Mingruzhiyao. Other; Significant; Aduro Biotech.

OBJECTIVE. Pancreatic ductal adenocarcinoma (PDAC) is often a lethal malignancy with limited preoperative predictors of long-term survival. The purpose of this study was to evaluate the prognostic utility of preoperative CT radiomics features in predicting postoperative survival of patients with PDAC. MATERIALS AND METHODS. A total of 153 patients with surgically resected PDAC who underwent preoperative CT between 2011 and 2017 were retrospectively identified. Demographic, clinical, and survival information was collected from the medical records. Survival time after the surgical resection was used to stratify patients into a low-risk group (survival time > 3 years) and a high-risk group (survival time < 1 year). The 3D volume of the whole pancreatic tumor and background pancreas were manually segmented. A total of 478 radiomics features were extracted from tumors and 11 extra features were computed from pancreas boundaries. The 10 most relevant features were selected by feature reduction. Survival analysis was performed on the basis of clinical parameters both with and without the addition of the selected features. Survival status and time were estimated by a random survival forest algorithm. Concordance index (C-index) was used to evaluate performance of the survival prediction model. RESULTS. The mean age of patients with PDAC was 67 ± 11 (SD) years. The mean tumor size was 3.31 ± 2.55 cm. The 10 most relevant radiomics features showed 82.2% accuracy in the classification of high-risk versus low-risk groups. The C-index of survival prediction with clinical parameters alone was 0.6785. The addition of CT radiomics features improved the C-index to 0.7414. CONCLUSION. Addition of CT radiomics features to standard clinical factors improves survival prediction in patients with PDAC.

The aim of this study was to evaluate outcomes of patients with resectable pancreatic adenocarcinoma (PDAC) who underwent neoadjuvant chemotherapy. The MEDLINE and PubMed databases were searched to identify relevant original articles investigating neoadjuvant therapy in resectable PDAC. Qualitative analyses were performed to investigate patient selection, disease stage, impact on perioperative outcomes, and cost-effectiveness. Forty-three studies met inclusion criteria for this review. Neoadjuvant chemotherapy for upfront resectable PDAC is cost-effective, safe, may result in lower stage disease and has potential survival advantages. With proper patient selection, neoadjuvant chemotherapy is an appropriate approach for upfront resectable PDAC. Nevertheless, the risk for disease progression and losing a curative surgical window highlights the need for appropriate patient identification, further discovery of superior biomarkers or molecular profiles representative of positive treatment response, and additional prospective comparative study.

OBJECTIVE The aim of this study was to critically analyze the surgical experience of managing autoimmune pancreatitis (AIP) in an era of modern diagnostics and compare these patients with those who were managed conservatively. METHODS Two prospectively maintained databases were used to retrospectively identify patients with AIP who were either managed conservatively or underwent pancreatectomy. RESULTS Eighty-eight patients were included in the study, of which 56 (63.6%) underwent resection and 32 (36.4%) were managed conservatively. Patients who underwent resection were more likely to present with jaundice (64.3% vs 18.1%, P < 0.001) and weight loss (53.6% vs 15.6%, P = 0.005). The cohort who underwent resection had a significantly higher median carbohydrate antigen 19-9 (40.0 vs 18.6 U/mL, P = 0.034) and was less likely to have elevated immunoglobulin G4 (26.1% vs 50.0%, P < 0.001). The most frequent initial diagnosis in the cohort who underwent resection was ductal adenocarcinoma (82.1%). Nine patients (28.1%) in the conservatively managed cohort experienced AIP relapse compared with 6 patients (10.7%) in the cohort who underwent resection. CONCLUSIONS The most frequent reason for surgical resection of AIP is concern for malignancy. Carbohydrate antigen 19-9 elevations were more common than immunoglobulin G4 in our cohort, suggesting that this laboratory profile is suboptimal for this population.

Background Pancreatectomy has a significant rate of procedure-specific morbidity which can result in readmission. Readmission has been proposed as a measure of quality. The goal of this study is to determine what factors are associated with readmission after pancreatectomy and whether readmission can be prevented. Methods A retrospective review of a single institution’s pancreatectomies between January 2011 and April 2015 was performed. Demographic, perioperative, and outpatient data were collected from the medical record. Primary outcome was 90-day readmission. Univariate and multivariable analyses were performed to determine which factors were associated with increased risk for readmission. Results A total of 257 patients met inclusion criteria; the 90-day readmission rate was 32.7%. The median time to readmission was 13 days. Readmitted patients were more likely to have a postoperative pancreatic fistula (POPF) on univariate analysis. Surgical site infections were more common in readmitted patients (18% vs 6.4%, P = .0138). Upon multivariable adjustment, only POPF ( P = .0005) remained significant. A positive dose-response relationship was noted between POPF grade and the odds of readmission with odds ratios (ORs) ranging from 1.6 (95% Confidence Interval (CI): .6-4.1) for grade A to 16.7 (95% CI: 1.8-156.8) for grade C, albeit with limited precision. Conclusions Readmission after pancreatectomy is a common occurrence despite the many advancements in perioperative care. Our data suggest that POPF is a risk factor for readmission after pancreatectomy. Presently, this factor is not clearly preventable. This suggests that readmission may not be the best measure of quality to utilize in the evaluation of pancreatic surgery.

Background We hypothesize that women undergoing cytoreductive surgery (CRS) and hyperthermic intraperitoneal chemotherapy (HIPEC) for peritoneal carcinomatosis from appendiceal cancer will have a survival advantage compared to men. Methods The National Cancer Database (NCDB) public user file (2004–2014) was used to select patients with PC undergoing CRS and HIPEC from appendiceal cancer. Univariate and multivariable analyses were performed. Results 1,190 patients with PC from appendiceal cancer underwent HIPEC and CRS. OS was significantly longer for women than for men, with mean and median OS being 73.8 months and 98.2 months for women vs 58.7 months and 82.5 months for men, respectively (p = 0.0032). On multivariable analysis, male sex (HR: 1.444, 95% CI: 1.141–1.827, p = 0.0022) and increasing age (HR: 1.017, 95% CI: 1.006–1.027, p = 0.0017) were both found to be independent risk factors for worse OS. Conclusion Women undergoing CRS and HIPEC for PC from appendiceal origin live longer than men undergoing the same treatment. Increasing age was also found to be independent risk factors for worse survival

OBJECTIVE: Patients with pancreatic cancer (PDAC) who undergo surgical resection and receive effective chemotherapy have the best chance of long-term survival. Unfortunately, we lack predictive biomarkers to guide optimal systemic treatment. Ex-vivo generation of patient-derived organoids (PDO) for pharmacotyping may serve as predictive biomarkers in PDAC. The goal of the current study was to demonstrate the clinical feasibility of a PDO-guided precision medicine framework of care. METHODS: PDO cultures were established from surgical specimens and endoscopic biopsies, expanded in Matrigel, and used for high-throughput drug testing (pharmacotyping). Efficacy of standard-of-care chemotherapeutics was assessed by measuring cell viability after drug exposure. RESULTS: A framework for rapid pharmacotyping of PDOs was established across a multi-institutional consortium of academic medical centers. Specimens obtained remotely and shipped to a central biorepository maintain viability and allowed generation of PDOs with 77% success. Early cultures maintain the clonal heterogeneity seen in PDAC with similar phenotypes (cystic–solid). Late cultures exhibit a dominant clone with a pharmacotyping profile similar to early passages. The biomass required for accurate pharmacotyping can be minimized by leveraging a high-throughput technology. Twenty-nine cultures were pharmacotyped to derive a population distribution of chemotherapeutic sensitivity at our center. Pharmacotyping rapidly-expanded PDOs was completed in a median of 48 (range 18–102) days. CONCLUSIONS: Rapid development of PDOs from patients undergoing surgery for PDAC is eminently feasible within the perioperative recovery period, enabling the potential for pharmacotyping to guide post-operative adjuvant chemotherapeutic selection. Studies validating PDOs as a promising predictive biomarker are ongoing.

Objective The aim of this study was to characterize the patterns and treatment of disease recurrence in patients achieving a pathological complete response (pCR) following neoadjuvant chemoradiation for advanced pancreatic ductal adenocarcinoma (PDAC). Summary of background data A pCR is an independent predictor for improved survival in PDAC. However, disease recurrence is still observed in these patients. Methods Patients with advanced PDAC who were treated with neoadjuvant therapy and had a pCR were identified between 2009 and 2017. Overall survival (OS) was determined from the initiation of neoadjuvant, disease-free survival (DFS) from the date of surgery, and post-recurrence survival (PRS) from the date of recurrence. Factors associated with recurrence were analyzed using a Cox-regression model. Results Of 331 patients with borderline resectable or locally advanced PDAC, 30 achieved a pCR following neoadjuvant treatment and pancreatectomy. The median DFS for pCR patients was 29 months and OS 76 months. Recurrence was observed in 14 patients. No clinicopathologic or treatment characteristics were associated with survival. The median PRS following recurrence was 25 months. Treatment following recurrence included chemotherapy, radiation or ablation, and surgical resection. Hepatectomy or completion pancreatectomy was accomplished in 2 patients that remain alive 13 and 62 months, respectively, following metastasectomy. Conclusions A pCR following neoadjuvant therapy in patients with advanced PDAC is associated with remarkable survival, although recurrence occurs in about half of patients. Nevertheless, patients with pCR and recurrence respond well to treatment and survival remains encouraging. Advanced molecular characterization and longitudinal liquid biopsy may offer additional assistance with understanding tumor biologic behavior after achieving a pCR.

An area of growing interest in precision medicine is the utilization of in vitro and in vivo models of diseases that are reliable and accurate representations. These model systems not only provide valuable insight into disease mechanism and pathophysiology, but also provide a platform with which to test and develop targeted therapeutic agents. The ability to tailor treatment for individual patients is truly reliant on the establishment of effective and efficient model systems. Thus, models that accurately recapitulate disease processes and reliably predict therapeutic efficacy are invaluable to moving the field of precision medicine forward.

BACKGROUND Postoperative pancreatic fistula (POPF) remains a significant source of morbidity following distal pancreatectomy (DP). There is a lack of information regarding the impact of trauma on POPF rates when compared with elective resection. We hypothesize that trauma will be a significant risk factor for the development of POPF following DP. METHODS A retrospective, single-institution review of all patients undergoing DP from 1999 to 2017 was performed. Outcomes were compared between patients undergoing DP for traumatic injury to those undergoing elective resection. Univariate and multivariable analyses were performed using SAS (version 9.4). RESULTS Of the 372 patients who underwent DP during the study period, 298 met inclusion criteria: 38 DPs for trauma (TDP), 260 elective DPs (EDP). Clinically significant grade B or C POPFs occurred in 17 (44.7%) of 38 TDPs compared with 41 (15.8%) of 260 EDPs (p < 0.0001). On multivariable analysis, traumatic injury was found to be independently predictive of developing a grade B or C POPF (odds ratio, 4.3; 95% confidence interval, 2.10-8.89). Age, sex, and wound infection were highly correlated with traumatic etiology and therefore were not retained in the multivariable model. When analyzing risk factors for each group (trauma vs. elective) separately, we found that TDP patients who developed POPFs had less sutured closure of their duct, higher infectious complications, and longer hospital stays, while EDP patients that suffered POPFs were more likely to be male, younger in age, and at a greater risk for infectious complications. Lastly, in a subgroup analysis involving only patients with drains left postoperatively, trauma was an independent predictor of any grade of fistula (A, B, or C) compared with elective DP (odds ratio, 8.6; 95% confidence interval, 3.09-24.15), suggesting that traumatic injury is risk factor for pancreatic stump closure disruption following DP. CONCLUSION To our knowledge, this study represents the largest cohort of patients comparing pancreatic leak rates in traumatic versus elective DP, and demonstrates that traumatic injury is an independent risk factor for developing an ISGPF grade B or C pancreatic fistula following DP. LEVEL OF EVIDENCE Prognostic study, Therapeutic, level III.

Background Mismatch repair proficient (MMRp) colorectal cancer (CRC) has been refractory to single‐agent programmed cell death protein 1 (PD1) inhibitor therapy. Colon GVAX is an allogeneic, whole‐cell, granulocyte‐macrophage colony‐stimulating factor ‐secreting cellular immunotherapy that induces T‐cell immunity against tumor‐associated antigens and has previously been studied in combination with low‐dose cyclophosphamide (Cy) to inhibit regulatory T cells. Methods We conducted a single‐arm study of GVAX/Cy in combination with the PD1 inhibitor pembrolizumab in patients with advanced MMRp CRC. Patients received pembrolizumab plus Cy on day 1, GVAX on day 2, of a 21‐day cycle. The primary endpoint was the objective response rate by Response Evaluation Criteria in Solid Tumors (RECIST) version 1.1. Secondary objectives included safety, overall survival, progression‐free survival, changes in carcinoembryonic antigen (CEA) levels, and immune‐related correlates. Results Seventeen patients were enrolled. There were no objective responses, and the disease control rate was 18% by RECIST 1.1. The median progression‐free survival was 82 days (95% confidence interval [CI], 48‐97 days) and the median overall survival was 213 days (95% CI 179‐441 days). Biochemical responses (≥30% decline in CEA) were observed in 7/17 (41%) of patients. Grade ≥ 3 treatment‐related adverse events were observed in two patients (hemolytic anemia and corneal transplant rejection). Paired pre‐ and on‐treatment biopsy specimens showed increases in programmed death‐ligand 1 expression and tumor necrosis in a subset of patients. Conclusions GVAX/Cy plus pembrolizumab failed to meet its primary objective in MMRp CRC. Biochemical responses were observed in a subset of patients and have not previously been observed with pembrolizumab monotherapy in MMRp CRC, indicating that GVAX may modulate the antitumor immune response., Mismatch repair proficient colorectal cancer has been refractory to single‐agent programmed cell death protein 1 inhibitor therapy. Colon GVAX is an allogeneic, whole‐cell, granulocyte‐macrophage colony‐stimulating factor‐secreting cellular immunotherapy that induces T‐cell immunity against tumor‐associated antigens and has previously been studied in combination with low‐dose cyclophosphamide (Cy) to inhibit regulatory T cells. In this single‐arm study, we find that GVAX/Cy plus pembrolizumab results in biochemical responses but not radiographic responses.

Background: Systemic disease control is the principle driver of prognosis in patients with pancreatic cancer (PDAC). Currently, there are no predictive biomarkers of clinical response to guide selection of an optimal treatment approach. Our work explores the logistics and feasibility of generating patient tumor-derived models to serve as biomarkers. Methods: Patients with borderline resectable or locally advanced PDAC enrolled to participate in a multi-institutional prospective, randomized, trial were eligible. Pretreatment core needle biopsies from diagnostic endoscopy were shipped via commercial post to a centralized organoid laboratory. Patient-specific organoids (PDO) were established as previously described by our group. Putative predictive biomarkers of clinical treatment response, including whole-exome sequencing (WES), RNA-sequencing, and pharmacotyping (chemotherapeutic sensitivity testing) were performed prospectively. Putative predictive biomarkers will be compared to clinical outcomes when approved by the trial's Data Safety Monitoring Board. Results: Approximately 40% (42/98) of patients participated in tissue acquisition for organoid development. Culture maturation can be described in three relevant phases: establishment, expansion, and characterization. The establishment phase describes the emergence of viable ductal organoid development in the setting of a culture being cleared of other components of the tumor microenvironment. During these first 3-5 passages, there are numerous potential pitfalls, including technical and logistical challenges. Out of 42, 26 (62%) PDOs were successfully established. Notably, the establishment phase appears to be an area of the greatest improvement, as the rates of successful establishment improve while the collaborative team builds experience. Expansion involves biomass accumulation in clean culture. Of those that succeeded in establishment, 77% (20/26) entered in expansion phase. We can currently report 15 PDOs having reached the characterization phase (WES, RNA-seq with subtyping, and/or pharmacotyping). Based on RNA expression profile, 10/14 were classified as classical and 4 as basal-like subtype. Mean time to pharmacotyping was 179 days. The PDOs showed a variable population distribution of sensitivity to standard-of-care chemotherapeutics (gemcitabine, paclitaxel, irinotecan, 5-FU, oxaliplatin). Conclusions: Developing a living organoid biobank from PDAC biopsies in a multi-center trial setting is a tractable approach in precision medicine initiatives. Amongst a collaborative team of physician scientists, the establishment phase is critical. PDOs from pre-treatment biopsies can be individually characterized within a clinically relevant time frame and unique tumor-specific ex vivo-derived therapeutic sensitivities can be assessed. Citation Format: Toni T. Seppälä, Jacquelyn W. Zimmerman, Noah Rozich, Alex Blair, Ammar Javed, John L. Cameron, William R. Burns, Jin He, David Tuveson, Christopher L. Wolfgang, David P. Ryan, Alec Kimmelman, Joseph M. Herman, Wells Messersmith, Theodore S. Hong, David T. Ting, Richard Andrew Burkhart. Patient-derived organoids may facilitate precision medicine in pancreatic cancer: Demonstrating feasibility in the context of a multi-center clinical trial [abstract]. In: Proceedings of the Annual Meeting of the American Association for Cancer Research 2020; 2020 Apr 27-28 and Jun 22-24. Philadelphia (PA): AACR; Cancer Res 2020;80(16 Suppl):Abstract nr LB-011.

Intestinal anastomoses are commonly performed in both elective and emergent operations. Even so, anastomotic leaks are a highly feared complications of colonic surgeries and can occur in up to 26% of surgical anastomoses, with mortality being up to 39% for patients with such a leak. Currently, there remains a paucity of data detailing the cellular mechanisms of anastomotic healing. Devising preventative strategies and treatment modalities for anastomotic leak could be greatly potentiated by a better understanding of appropriate anastomotic healing. A murine model is ideal as previous studies have shown that the murine anastomosis is the most clinically similar to the human case as compared with other animal models. We offer an easily reproducible murine model of colonic anastomosis in mice that will allow for further illustration of anastomotic healing.

176 Background: CRC patients with PM have a significantly worse median overall survival (OS) when compared to those with liver and lung metastases. We studied MSLN protein expression using immunohistochemistry (IHC) on primary resected CRC (PRC), its correlation with future development of PM and its effect on OS. Our hypothesis is that PRC tumors from patients who later developed PM will have an increased proportion of positive MSLN scores based on IHC. Methods: We performed a retrospective case control study of all localized CRC (Stage I, II, III) patients who underwent curative resection from 2000- 2017 at the University of Oklahoma. The target population consisted of PRC patients who had distant recurrence during surveillance. Control group comprised of patients who never recurred after a minimum of 5 years of surveillance. MSLN score was calculated by the sum of staining percentage and intensity scores on each MSLN stained slide. Kaplan-Meier method was used to estimate OS and differences in survival were estimated using log-rank test. Results: Among 100 patients with recurrent CRC, 19 patients had PM, 51 had solid organ only metastases (SOM) and 66 are in the control group. MSLN staining results by IHC were satisfactory in 87 patients (Table). 91%, 46% and 21% of the patients in PM, SOM, and control groups had a positive MSLN score on PRC respectively. A significant association between MSLN score and future development of PM was observed (p [Table: see text]

Background The effect of cigarette smoking on postoperative morbidity following pancreaticoduodenectomy (PD) for cancer is unclear. We hypothesize that smoking is associated with higher morbidity following PD. Methods A retrospective review of patients undergoing PD for cancer from 2010 to 2016 at a single institution was performed. Patients who had never smoked were compared to current or past-smokers with at least 1 pack-year history. Univariate and multivariable analyses were performed. Results Two hundred fifty-two patients met inclusion criteria. On univariate analysis, there was a significant difference between smokers and never-smokers in age at diagnosis (65.5 versus 68.6 y, P = 0.013) and fistula rate (28.5% versus 16.2%, P = 0.024). Male sex was significantly associated with fistula rate compared with female sex (15.5% versus 7.1%, P = 0.023). Comparing males and females separately, smoking correlated with higher fistula development only in the male cohort (22.5% versus 5.8%, P = 0.016 in men and 7.3% versus 9.1%, P = 1.00 in women). On multivariable analysis, current and past smoking was independently predictive of developing a fistula: odds ratio of 2.038 (P = 0.030). For current and past-smokers, male sex was an independent risk factor for developing a fistula: odds ratio 2.817 (P = 0.022). There were no other significant differences between groups in rates of postoperative complications. Conclusions Smoking status is independently predictive of postoperative pancreatic fistula following PD for cancer. Among smokers, male sex is an independent risk factor for fistula. Further studies are needed to determine if smoking cessation before surgery decreases this risk, and if so, the optimal duration of cessation.

Bispecific T cell engager (BiTE®) antibody constructs are designed to redirect T cells to induce lysis of tumor cells through simultaneous binding to CD3 on T cells and to a tumor associated antigen. BiTE® antibody constructs have previously been shown to be effective at depleting blood and tissue targets in B cell malignancies, suggesting a therapeutic potential for BiTE® antibody constructs in solid tumors. Pancreatic ductal adenocarcinoma (PDA) is a devastating malignant disease with a dismal prognosis even with currently available therapies. The tumor differentiation antigen mesothelin (MSLN) is highly expressed in over 80% of pancreatic tumors. Expression of MSLN is restricted in normal tissue implicating MSLN as an attractive target for BiTE® targeted therapy in PDA. In this study, we report the preclinical characterization of an anti-MSLN/CD3 BiTE® antibody construct with half-life extension (MSLN HLE BiTE®) in xenograft models of PDA. Previously the MSLN HLE BiTE® was found to bind CD3 and MSLN with low nanomolar affinity and have low picomolar cytotoxic activity against MSLN-positive cells in vitro. Moreover, the MSLN HLE BiTE® can mediate redirected lysis of cancer cell lines resistant to chemotherapy. Our study examines the efficacy of the highly potent and specific MSLN HLE BiTE® antibody construct in vivo by utilizing the orthotopic implantation of human AsPC-1 luciferase expressing PDA cells in immunocompromised NOD-Prkdcscid IL2rgnull (NSG) mice. We found that mice receiving the MSLN HLE BiTE® antibody construct survived significantly longer and had significantly reduced tumor burden comparing to control mice. Moreover, tumor bearing mice were given fluorescently labeled human T cells to measure T cell localization. Using live animal fluorescent imaging, we found mice treated with the MLSN HLE BiTE® antibody construct had significantly increased T cell localization and retention to the PDA tumor area compared to control. Mice receiving orthotopic implantation of MSLN knock out human AsPC-1 tumors treated with the MSLN HLE BiTE® did not show increased survival, reduction of tumor burden or increased T cell localization, confirming the BiTE® specificity for the tumor target MSLN. Furthermore, using the KPC mouse model of PDA, which develops cancer spontaneously upon conditional pancreatic expression of KRASG12D and TP53R172H mutations, we tested if a mouse-specific surrogate MSLN HLE BiTE® antibody construct could localize to a naturally formed dense tumor microenvironment. We found that mice treated with a murine MSLN HLE BiTE® labeled with CW800 demonstrated strong localization of the MSLN HLE BiTE® through the dense microenvironment to the PDA tumor. Taken together, these results suggest that a MSLN HLE BiTE® may be a unique targeted immunotherapy for PDA patients. Citation Format: Noelle R. Jurcak, MacKenzie Zarecki, Fei Lee, Noah Rozich, Stephen Muth, Elizabeth Jaffee, Sabine Stienen, Julie Bailis, Lei Zheng. Evaluation of mesothelin BiTE® antibody constructs in models of pancreatic ductal adenocarcinoma [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2019; 2019 Mar 29-Apr 3; Atlanta, GA. Philadelphia (PA): AACR; Cancer Res 2019;79(13 Suppl):Abstract nr 1561.

563 Background: Mismatch repair proficient (MMRp) colorectal cancer (CRC) is refractory to single-agent programmed cell death protein 1 (PD1) inhibitors. Cancer vaccines may prime the tumor microenvironment for anti-PD1 therapy. Colon GVAX is an allogeneic, whole-cell, GM-CSF-secreting vaccine that induces T-cell immunity against tumor-associated antigens. GVAX has previously been studied in combination with low-dose cyclophosphamide (Cy) to inhibit regulatory T cells. Methods: We conducted an open label, single-arm, phase 2 study of GVAX/Cy in combination with the PD1 inhibitor pembrolizumab in patients with MMRp CRC who had received at least two prior lines of therapy in the metastatic setting. Patients received pembrolizumab plus Cy on day 1, GVAX on day 2, of a 21 day cycle through 4 cycles, and were then continued on a maintenance regimen of pembolizumab every 3 weeks with cy/GVAX given every 12 weeks. Results: Seventeen patients were enrolled. There were no objective responses, and the disease control rate was 18% by RECIST and 29% by irRC. The median progression free survival was 2.7 months and the median overall survival was 7.0 months. Biochemical responses (≥30% decline in CEA) were observed in 7/17 (41%) of patients. Two patients (12%) had grade 3/4 adverse events that were attributed to study therapy. To test the hypothesis that the induction of a humoral response against CEA protein had resulted in the observed biochemical responses, we measured patient titers of anti-CEA antibodies. Anti-CEA antibody titers increased 13 of 13 patients (100%) with available paired pre- and on-treatment research blood samples, but the change in anti-CEA antibody titers were similar among CEA responders and non-responders. Multiplex immunohistochemistry performed on pre- and post- biopsy specimens will be reported at the conference. Conclusions: GVAX/Cy plus pembrolizumab is well tolerated in advanced MMRp CRC and resulted in CEA responses but not radiographic responses. PFS and OS compare favorably to historical controls in this small cohort. CEA responses were not observed with PD1 monotherapy in MMRp CRC, suggesting GVAX can modulate the antitumor immune response. Clinical trial information: NCT02981524.