Transcriptomic profiles of neoantigen-reactive T cells in human gastrointestinal cancers.

Tumor-infiltrating neoantigen-reactive T cells can mediate regression of metastatic gastrointestinal cancers yet remain poorly characterized. We performed immunological screening against personalized neoantigens in combination with single-cell RNA sequencing on tumor-infiltrating lymphocytes from bile duct and pancreatic cancer patients to characterize the transcriptomic landscape of neoantigen-reactive T cells. We found that most neoantigen-reactive CD8+ T cells displayed an exhausted state with significant CXCL13 and GZMA co-expression compared with non-neoantigen-reactive bystander cells. Most neoantigen-reactive CD4+ T cells from a patient with bile duct cancer also exhibited an exhausted phenotype but with overexpression of HOPX or ADGRG1 while lacking IL7R expression. Thus, neoantigen-reactive T cells infiltrating gastrointestinal cancers harbor distinct transcriptomic signatures, which may provide new opportunities for harnessing these cells for therapy.

Evaluation of Indocyanine Green Fluorescence Imaging for Intraoperative Identification of Liver Malignancy.

INTRODUCTION: There is early evidence that indocyanine green (ICG) fluorescence imaging has the ability to detect metastatic and primary malignancies in the liver that are too small to be identified by other methods. However, the rate of false positives and false negatives remains unknown. MATERIALS AND METHODS: This is a single institution prospective single-arm study. Patients with suspected hepatic or pancreatic malignancies were intravenously injected with ICG one to three days prior to their scheduled surgical therapy. At the beginning of the procedure, the liver was assessed with fluorescence imaging and all identified lesions were biopsied and evaluated. RESULTS: Twenty-three patients were enrolled from April 2015 through February 2016. Fifteen patients with confirmed malignancy had adequate fluorescence imaging evaluation of the liver; 10 with pancreatic primary malignancies and five with hepatic primaries. Fluorescence imaging was the only modality that identified nine concerning hepatic lesions, all of which were benign on pathology examination. Out of 11 malignant hepatic masses, six were visible on fluorescence imaging. Out of nine benign hepatic lesions, five were visible. No side effects or complications of the fluorescence imaging were encountered. The sensitivity for ICG fluorescence was 45.5%, the specificity 21.2%, the positive predictive value 25%, and the negative predictive value 40%. CONCLUSION: Intraoperative hepatic assessment with ICG fluorescence imaging to identify malignancy in the liver is feasible and safe. However, in this study the significant number of false positives limit the utility of the technique. Our preliminary data do not support its routine use for detection of malignancies in the liver.

Role of yttrium-90 selective internal radiation therapy in the treatment of liver-dominant metastatic colorectal cancer: an evidence-based expert consensus algorithm.

Surgical resection of colorectal liver metastases is associated with greater survival compared with non-surgical treatment, and a meaningful possibility of cure. However, the majority of patients are not eligible for resection and may require other non-surgical interventions, such as liver-directed therapies, to be converted to surgical eligibility. Given the number of available therapies, a general framework is needed that outlines the specific roles of chemotherapy, surgery, and locoregional treatments [including selective internal radiation therapy (SIRT) with Y-90 microspheres]. Using a data-driven, modified Delphi process, an expert panel of surgical oncologists, transplant surgeons, and hepatopancreatobiliary (HPB) surgeons convened to create a comprehensive, evidence-based treatment algorithm that includes appropriate treatment options for patients stratified by their eligibility for surgical treatment. The group coined a novel, more inclusive phrase for targeted locoregional tumor treatment (a blanket term for resection, ablation, and other emerging locoregional treatments): local parenchymal tumor destruction therapy. The expert panel proposed new nomenclature for 3 distinct disease categories of liver-dominant metastatic colorectal cancer that is consistent with other tumor types: (I) surgically treatable (resectable); (II) surgically untreatable (borderline resectable); (III) advanced surgically untreatable (unresectable) disease. Patients may present at any point in the algorithm and move between categories depending on their response to therapy. The broad intent of therapy is to transition patients toward individualized treatments where possible, given the survival advantage that resection offers in the context of a comprehensive treatment plan. This article reviews what is known about the role of SIRT with Y-90 as neoadjuvant, definitive, or palliative therapy in these different clinical situations and provides insight into when treatment with SIRT with Y-90 may be appropriate and useful, organized into distinct treatment algorithm steps.

Preoperative imaging characteristics predict poor survival and inadequate resection for left-sided pancreatic adenocarcinoma: a multi-institutional analysis.

BACKGROUND: Optimal treatment of pancreatic ductal adenocarcinoma of the neck, body and tail (PDAC-NBT) necessitates R0 surgical resection. Preoperative radiographic identification of patients likely to achieve successful oncologic resection remains difficult. This study seeks to identify preoperative imaging characteristics predictive of non-R0 resections or impaired survival for PDAC-NBT. METHODS: Patients at five high-volume centers who underwent resection for PDAC-NBT were retrospectively analyzed. The most immediate preoperative cross-sectional scan was assessed along with outcome measures of overall survival and margin status. RESULTS: 330 patients were treated between 2001 and 2016. Margin status included 247 R0 (78.2%), 67 R1 (21.2%), and 2 R2 (0.6%). A non-R0 resection predicted worse survival (p = 0.0002). On preoperative imaging, patients with tumors greater than 20 mm, tumor attenuation greater than 70 Hounsfield units, or who demonstrated pancreatic atrophy and/or calcifications also had worse survival (p = 0.010, p = 0.036, p = 0.025 respectively). Patients with tumors interfacing with the splenic artery or vein or extending posteriorly achieved fewer R0 resections (p = 0.0006, p = 0.0004, p = 0.001, respectively). CONCLUSION: Preoperative cross-sectional imaging can identify tumor characteristics associated with poor survival and non-R0 resection. Further investigation is needed to identify the appropriate surgical and treatment modifications necessary to clinically benefit this subset of patients.

The Miami International Evidence-based Guidelines on Minimally Invasive Pancreas Resection.

OBJECTIVE: The aim of this study was to develop and externally validate the first evidence-based guidelines on minimally invasive pancreas resection (MIPR) before and during the International Evidence-based Guidelines on Minimally Invasive Pancreas Resection (IG-MIPR) meeting in Miami (March 2019). SUMMARY BACKGROUND DATA: MIPR has seen rapid development in the past decade. Promising outcomes have been reported by early adopters from high-volume centers. Subsequently, multicenter series as well as randomized controlled trials were reported; however, guidelines for clinical practice were lacking. METHODS: The Scottisch Intercollegiate Guidelines Network (SIGN) methodology was used, incorporating these 4 items: systematic reviews using PubMed, Embase, and Cochrane databases to answer clinical questions, whenever possible in PICO style, the GRADE approach for assessment of the quality of evidence, the Delphi method for establishing consensus on the developed recommendations, and the AGREE-II instrument for the assessment of guideline quality and external validation. The current guidelines are cosponsored by the International Hepato-Pancreato-Biliary Association, the Americas Hepato-Pancreato-Biliary Association, the Asian-Pacific Hepato-Pancreato-Biliary Association, the European-African Hepato-Pancreato-Biliary Association, the European Association for Endoscopic Surgery, Pancreas Club, the Society of American Gastrointestinal and Endoscopic Surgery, the Society for Surgery of the Alimentary Tract, and the Society of Surgical Oncology. RESULTS: After screening 16,069 titles, 694 studies were reviewed, and 291 were included. The final 28 recommendations covered 6 topics; laparoscopic and robotic distal pancreatectomy, central pancreatectomy, pancreatoduodenectomy, as well as patient selection, training, learning curve, and minimal annual center volume required to obtain optimal outcomes and patient safety. CONCLUSION: The IG-MIPR using SIGN methodology give guidance to surgeons, hospital administrators, patients, and medical societies on the use and outcome of MIPR as well as the approach to be taken regarding this challenging type of surgery.

Can high-volume teams of anesthesiologists and surgeons decrease perioperative costs for pancreatic surgery?

BACKGROUND: Pancreatic surgery outcomes are associated with surgeon and center experience. Anesthesiologists as potential value drivers for pancreatic surgery have not been explored. We sought to evaluate whether anesthesiologists impact perioperative costs for pancreatic surgery. METHODS: Within an integrated health care system, 796 pancreatic surgeries (526 PDs and 270 DPs) were performed from January 2014 to June 2017. Mean direct operative and anesthesia costs driven by anesthesiologists (operating room (OR) time, anesthesia billing and anesthesia procedures) were determined for each case. The volumes of pancreatic cases per anesthesiologist were calculated, and those above the 75th percentile for volume (4 cases) were considered high-volume. A multivariable analysis of OR/anesthesia costs was performed. RESULTS: Mean OR and anesthesia costs for PD were $7064 for low-volume anesthesiologists (LVA), higher than $5968 for high-volume anesthesiologists (HVA) (p < 0.001). By multivariable analysis, HVA were associated with decreased costs of $2278 (p < 0.001). Teams of HVA and high-volume surgeons (HVS) were also associated with decreased mean costs of $1790 (p = 0.04). CONCLUSION: These data suggest that anesthesiologists experienced in the management of complex pancreatic operations such as PDs may contribute to improved efficiencies in care by reducing perioperative costs.

What Are the Most Significant Cost and Value Drivers for Pancreatic Resection in an Integrated Healthcare System?

BACKGROUND: An initiative was established to improve value-based care for pancreatic surgery in a large nonprofit health system. Cost data were presented bimonthly to a hepatobiliary clinical performance group via videoconference. STUDY DESIGN: The direct costs were calculated for all patients undergoing distal pancreatectomy (DP) and pancreaticoduodenectomy (PD) between January 2014 and July 2017. Median length of stay, 30-day and 90-day mortality rates, readmission rate, and costs were stratified by surgeon volume using 2 published criteria: "volume pledge" criteria (≥5 PDs/year) and Leapfrog criteria (≥11 PDs/year). RESULTS: There were 270 DPs and 526 PDs performed in 14 hospitals spanning 4 states. Median PD costs were lower for high-volume surgeons (≥5 PDs/year), $21,026 vs $24,706 (p = 0.005). High-volume surgeons had a shorter length of stay (9 days vs 11 days; p < 0.001) for PD and DP (6 days vs 7 days; p = 0.001). Increased costs for low-volume surgeons included operative/anesthesia costs ($7,321 vs $6,325; p = 0.03), room and board ($5,828 vs $4,580; p = 0.01), and intensive care costs ($4,464 vs $3,113; p = 0.04). Operating time was increased for high-volume surgeons for DP and PD (p < 0.001). There was no difference in 30-day or 90-day mortality rates or readmissions for DP or PD when stratified by volume pledge criteria. There was no difference in total costs for DP or PD when stratified by Leapfrog criteria. CONCLUSIONS: There was a significant cost reduction for PD but not DP when the threshold of 5 PDs was used as a definition of high volume. The sharing of detailed financial data with HPB surgeons on a regular basis provides an opportunity to evaluate practice patterns and thereby reduce direct costs.

Regional Metastatic Behavior of Nonfunctional Pancreatic Neuroendocrine Tumors: Impact of Lymph Node Positivity on Survival.

OBJECTIVES: Literature addressing the significance of lymph node positivity in the management of nonfunctional pancreatic neuroendocrine tumors (PNETs) is conflicting. METHODS: The National Cancer Data Base was queried for patients who underwent surgical resection of nonfunctional PNETs between 1998 and 2011. Clinical data and overall survival were analyzed using χ and Cox proportional hazards regression. Multiple imputation was used as a comparative analysis because of the high number of patients missing data on tumor grade. RESULTS: Two thousand seven hundred thirty-five patients were identified. The overall incidence of lymph node metastasis was 51%. In the subset of patients with grade 1 tumors less than 1 cm, 24% had positive lymph nodes. Overall median survival for patients with negative lymph nodes was 11 years compared with 8 years for lymph node-positive patients (P < 0.001). On multivariate survival analysis, tumor grade, distant metastases, regional lymph node involvement, positive surgical margins, male sex, and older age were predictive of decreased overall survival. CONCLUSIONS: Lymph node positivity was associated with decreased overall survival. The incidence of lymph node involvement in resected low-grade tumors less than 1 cm is higher than previously reported. Patients selected for resection of PNETs should be offered lymphadenectomy for staging.

Novel Simulation Device for Targeting Tumors in Laparoscopic Ablation: A Learning Curve Study.

INTRODUCTION: A novel 3-dimensional (3D) guidance system was developed to aid accurate needle placement during ablation. METHODS: Five novices and 5 experienced hepatobiliary surgeons were recruited. Using an agar block with analog tumor, participants targeted under 4 conditions: in-line with the ultrasound plane using ultrasound, in-line using 3D guidance, 45° off-axis using ultrasound, and off-axis using 3D guidance. Time to target the tumor, number of withdrawals, and the National Aeronautics and Space Administration Task Load Index were collected. Initial and final parameters for each of the conditions were compared using a within-subjects paired t test. RESULTS: A significant reduction was seen in the number of required withdrawals in all situations when using the 3D guidance (0.75 vs 3.65 in-line and 0.25 vs 3.6 for off-axis). Mental workload was significantly lower when using 3D guidance compared with ultrasound both for novices (29.85 vs 41.03) and experts (31.98 vs 44.57), P < .001 for both. The only difference in targeting time between first and last attempt was in the novice group during off-axis targeting using 3D guidance (115 vs 32.6 seconds, P = .03). CONCLUSION: Though 3D guidance appeared to decrease time to target, this was not statistically significant likely as a result of lack of power in our trial. Three-dimensional guidance did reduce the number of required withdrawals, potentially decreasing complications, as well as mental workload after proficiency was achieved. Furthermore, novices without experience in ultrasound were able to learn targeting with the 3D guidance system at a faster pace than targeting with ultrasound alone.

Standardizing terminology for minimally invasive pancreatic resection.

BACKGROUND: There is a growing body of literature pertaining to minimally invasive pancreatic resection (MIPR). Heterogeneity in MIPR terminology, leads to confusion and inconsistency. The Organizing Committee of the State of the Art Conference on MIPR collaborated to standardize MIPR terminology. METHODS: After formal literature review for "minimally invasive pancreatic surgery" term, key terminology elements were identified. A questionnaire was created assessing the type of resection, the approach, completion, and conversion. Delphi process was used to identify the level of agreement among the experts. RESULTS: A systematic terminology template was developed based on combining the approach and resection taking into account the completion. For a solitary approach the term should combine "approach + resection" (e.g. "laparoscopic pancreatoduodenectomy); for combined approaches the term must combine "first approach + resection" with "second approach + reconstruction" (e.g. "laparoscopic central pancreatectomy" with "open pancreaticojejunostomy") and where conversion has resulted the recommended term is "first approach" + "converted to" + "second approach" + "resection" (e.g. "robot-assisted" "converted to open" "pancreatoduodenectomy") CONCLUSIONS: The guidelines presented are geared towards standardizing terminology for MIPR, establishing a basis for comparative analyses and registries and allow incorporating future surgical and technological advances in MIPR.

Analysis of 340 Patients with Solid Pseudopapillary Tumors of the Pancreas: A Closer Look at Patients with Metastatic Disease.

INTRODUCTION: Current literature addressing the treatment of solid pseudopapillary neoplasms (SPNs) of the pancreas is limited, particularly for patients with distant metastases. We aimed to define predictive indicators of survival in a large series of patients and assess the outcome of patients with distant metastases. METHODS: The National Cancer Database was queried for patients diagnosed with SPNs of the pancreas between 1998 and 2011. Single predictor univariate analyses were performed on variables including demographics, tumor characteristics, and surgery outcomes, and multivariate Cox proportional hazards survival analysis was then completed with backward elimination. RESULTS: Overall, 340 patients were identified: 82% were female, median age was 39 years, and 84% had no comorbidities. Patients undergoing any type of surgical resection experienced long-term survival (85% 8-year survival). Patients undergoing surgical resection (n = 296) had superior survival (hazard ratio [HR] 21 for no surgery, p < 0.0001), as did patients treated at academic centers and those with private insurance (HR 3.9, p = 0.009; HR 4.9, p = 0.007). Sex, age, tumor size, presence of lymph node metastases, positive surgical margins, and presence of distant metastases were not significant predictors of survival in multivariate analysis. Of 24 patients with distant metastases, seven were treated surgically and experienced long-term survival similar to that of patients without metastases treated surgically (HR 2, p = 0.48). CONCLUSION: SPNs of the pancreas are rare neoplasms with excellent overall survival; however, in a low number of patients they metastasize. Of the few patients with metastatic disease selected for resection, most experienced long-term survival.

Worldwide survey on opinions and use of minimally invasive pancreatic resection.

BACKGROUND: The introduction of minimally invasive pancreatic resection (MIPR) into surgical practice has been slow. The worldwide utilization of MIPR and attitude towards future perspectives of MIPR remains unknown. METHODS: An anonymous survey on MIPR was sent to the members of six international associations of Hepato-Pancreato-Biliary (HPB) surgery. RESULTS: The survey was completed by 435 surgeons from 50 countries, with each surgeon performing a median of 22 (IQR 12-40) pancreatic resections annually. Minimally invasive distal pancreatectomy (MIDP) was performed by 345 (79%) surgeons and minimally invasive pancreatoduodenectomy (MIPD) by 124 (29%). The median total personal experience was 20 (IQR 10-50) MIDPs and 12 (IQR 4-40) MIPDs. Current superiority for MIDP was claimed by 304 (70%) and for MIPD by 44 (10%) surgeons. The most frequently mentioned reason for not performing MIDP (54/90 (60%)) and MIPD (193/311 (62%)) was lack of specific training. Most surgeons (394/435 (90%)) would consider participating in an international registry on MIPR. DISCUSSION: This worldwide survey showed that most participating HPB surgeons value MIPR as a useful development, especially for MIDP, but the role and implementation of MIPD requires further assessment. Most HPB surgeons would welcome specific training in MIPR and the establishment of an international registry.

Validation of Fistula Risk Score calculator in diverse North American HPB practices.

BACKGROUND: Fistula Risk Score (FRS) is a previously developed tool to assess the risk of clinically relevant postoperative pancreatic fistula (CR-POPF) following pancreatoduodenectomy (PD). METHODS: Prospectively collected databases from 4 university affiliated and non-affiliated HPB centers in United States and Canada were used. The influence of individual baseline characteristics, FRS and FRS group on CR-POPF was assessed in univariate and multivariate analyses. FRS calculator performance was assessed using a C-statistic. RESULTS: 444 patients were identified. Pathology, soft pancreas texture and pancreatic duct size were associated with CR-POPF rates (p < 0.001 for each); EBL was not (p = 0.067). The negligible risk group consisted of 50 (11.3%) patients, low risk of 118 (26.6%), moderate 234 (52.7%) and high risk group of 42 (9.5%) patients. The overall rate of CR-POPF was 20%. Of the patients in the negligible risk group, 2% developed CR-POPF, 13.6% of the low risk, 23.1% moderate and 42.9% in the high risk group (p < 0.001). Overall C-statistic was 0.719. CONCLUSION: FRS is robust and able to stratify the risk of developing CR-POPF following PD in diverse North American academic and non-academic institutions. The FRS should be used in research and to guide clinical management of patients post PD in these institutions.

Proceedings of the first international state-of-the-art conference on minimally-invasive pancreatic resection (MIPR).

The application of minimally-invasive techniques to major pancreatic resection (MIPR) has occurred steadily, but slowly, over the last two decades. Questions linger regarding its safety, efficacy, and broad applicability. On April 20th, 2016, the first International State-of-the-Art Conference on Minimally Invasive Pancreatic Resection convened in Sao Paulo, Brazil in conjunction with the International Hepato-Pancreato-Biliary Association's (IHPBA) 10th World Congress. This report describes the genesis, preparation, execution and output from this seminal event. Major themes explored include: (i) scrutiny of best-level evidence outcomes of both MIPR Distal Pancreatectomy (DP) and pancreatoduodenectomy (PD), (ii) Cost/Value/Quality of Life assessment of MIPR, (iii) topics in training, education and credentialing, and (iv) development of best approaches to analyze results of MIPR - including clinical trial design and registry development. Results of a worldwide survey of over 400 surgeons on the practice of MIPR were presented. The proceedings of this event serve as a platform for understanding the role of MIPR in pancreatic resection. Data and concepts presented at this meeting form the basis for further study, application and dissemination of MIPR.

Characterization of Pancreaticojejunal Anastomotic Healing in a Porcine Survival Model.

INTRODUCTION: Anastomotic leak after pancreaticoduodenectomy is the most important cause of postoperative morbidity and mortality. Histological studies of bowel anastomoses have provided valuable insights regarding causes of anastomotic failure. However, this crucial information is lacking for pancreatico-enteric anastomoses. METHODS: Pancreaticoduodenectomy was performed in a porcine model. Animals were survived up to 10 days and then the pancreatico-enteral anastomosis specimen was resected en bloc. Anastomotic bursting pressure was measured and histological sections of the anastomoses were examined. RESULTS: Six out of 8 animals had excellent healing of the anastomoses. One animal developed a clinically significant leak at the pancreaticoduodenal anastomosis (12.5%) and one animal had a subclinical duodeno-duodenal leak discovered on necropsy (12.5%). Both anastomoses that failed had a collagen-to-tissue ratio less than 40%. In contrast, none of the anastomoses with a ratio greater than 40% showed any evidence of disruption. CONCLUSION: Our results indicate that quantitative measurement of collagen deposition at the pancreatic anastomosis provides objective assessment of healing of the pancreatic anastomosis. A survival porcine model of pancreaticoduodenectomy results in a similar leak rate to published data on pancreaticoduodenectomy in humans and will be useful for future studies assessing novel pharmacologic or technical interventions aimed at improving outcomes.

The influence of fellowship training on the practice of pancreatoduodenectomy.

BACKGROUND: There has been a proliferation of gastrointestinal surgical fellowships; however, little is known regarding their association with surgical volume and management approaches. METHODS: Surveys were distributed to members of GI surgical societies. Responses were evaluated to define relationships between fellowship training and surgical practice with pancreatoduodenectomy (PD). RESULTS: Surveys were completed by 889 surgeons, 84.1% of whom had completed fellowship training. Fellowship completion was associated with a primarily HPB or surgical oncology-focused practice (p < 0.001), and greater median annual PD volume (p = 0.030). Transplant and HPB fellowship-trained respondents were more likely to have high-volume (≥20) annual practice (p = 0.005 and 0.029, respectively). Regarding putative fistula mitigation strategies, HPB-trained surgeons were more likely to use stents, biologic sealants, and autologous tissue patches (p = 0.007, <0.001 and 0.001, respectively). Surgical oncology trainees reported greater autologous patch use (p = 0.003). HPB fellowship-trained surgeons were less likely to routinely use intraperitoneal drainage (p = 0.036) but more likely to utilize early (POD ≤ 3) drain amylase values to guide removal (p < 0.001). Finally, HPB fellowship-trained surgeons were more likely to use the Fistula Risk Score in their practice (29 vs. 21%, p = 0.008). CONCLUSION: Fellowship training correlated with significant differences in surgeon experience, operative approach, and use of available fistula mitigation strategies for PD.

A hypofractionated radiation regimen avoids the lymphopenia associated with neoadjuvant chemoradiation therapy of borderline resectable and locally advanced pancreatic adenocarcinoma.

BACKGROUND: Preclinical studies have shown synergy between radiation therapy and immunotherapy. However, in almost all preclinical models, radiation is delivered in single doses or short courses of high doses (hypofractionated radiation). By contrast in most clinical settings, radiation is delivered as standard small daily fractions of 1.8-2 Gy to achieve total doses of 50-54 Gy (fractionated radiation). We do not yet know the optimal dose and scheduling of radiation for combination with chemotherapy and immunotherapy. METHODS: To address this, we analyzed the effect of neoadjuvant standard fractionated and hypofractionated chemoradiation on immune cells in patients with locally advanced and borderline resectable pancreatic adenocarcinoma. RESULTS: We found that standard fractionated chemoradiation resulted in a significant and extended loss of lymphocytes that was not explained by a lack of homeostatic cytokines or response to cytokines. By contrast, treatment with hypofractionated radiation therapy avoided the loss of lymphocytes associated with conventional fractionation. CONCLUSION: Hypofractionated neoadjuvant chemoradiation is associated with reduced systemic loss of T cells. TRIAL REGISTRATION: ClinicalTrials.gov NCT01342224, April 21, 2011; NCT01903083, July 2, 2013.

Preoperative computed tomography scan to predict pancreatic fistula after distal pancreatectomy using gland and tumor characteristics.

BACKGROUND: Preoperative risk stratification for postoperative pancreatic fistula in patients undergoing distal pancreatectomy is needed. METHODS: Risk factors for postoperative pancreatic fistula in 220 consecutive patients undergoing distal pancreatectomy at 2 major institutions were recorded retrospectively. Gland density was measured on noncontrast computed tomography scans (n = 101), and histologic scoring of fat infiltration and fibrosis was performed by a pathologist (n = 120). RESULTS: Forty-two patients (21%) developed a clinically significant pancreatic fistula within 90 days of surgery. Fat infiltration was significantly associated with gland density (P = .0013), but density did not predict pancreatic fistula (P = .5). Recursive partitioning resulted in a decision tree that predicted fistula in this cohort with a misclassification rate less than 15% using gland fibrosis (histology), density (HU), margin thickness (cm), and pathologic diagnosis. CONCLUSIONS: This multicenter study shows that no single perioperative factor reliably predicts postoperative pancreatic fistula after distal pancreatectomy. A decision tree was constructed for risk stratification.

Total pancreatectomy for pancreatic ductal adenocarcinoma: review of the National Cancer Data Base.

BACKGROUND: Total pancreatectomy is infrequently performed for pancreatic cancer. Perceived operative mortality and questionable survival benefit deter many surgeons. Clinical outcomes, described in single-center series, remain largely unknown. METHODS: The National Cancer Database was queried for cases of pancreatic ductal adenocarcinoma undergoing total pancreatectomy (1998-2011). Univariate survival analyses were performed for 21 variables: demographic (8), tumor characteristics (5), surgery outcomes (6), and adjuvant therapy (2). The Log-rank test of differences in Kaplan-Meier survival curves was used for categorical variables. Variables with p < 0.05 were included in a multivariate analysis. Cox proportional hazards regression was used to analyze continuous variables and multivariate models. RESULTS: 2582 patients with staging and survival data made up the study population. 30-day mortality was 5.5%. Median overall survival was 15 months, with 1, 3, and 5-year survival rates of 60%, 22%, and 13%, respectively. Age, facility type, tumor size and grade, lymph node positivity, margin positivity, and adjuvant therapy significantly impacted survival in multivariate analysis. CONCLUSION: Although total pancreatectomy is a reasonable option for selected patients with pancreatic ductal adenocarcinoma, survival of the entire group is limited. Operative mortality is improved from prior reports. Greater survival benefits were seen in younger patients with smaller, node negative tumors resected with negative margins in academic research centers.