Introduction: A tool to predict the risk of postoperative pancreatic fistula (POPF) after distal pancreatectomy (DP) is currently lacking. It could be used for the selection of preventive strategies, for benchmarking across centers and stratifying patients by baseline risk in clinical studies and is ideally calculated prior to surgery. We aimed to develop and externally validate the first fistula risk score for POPF after DP. Method: Predictive variables for POPF were found using data of patients undergoing DP in two Italian centers (2014-2016) utilizing multivariable logistic regression. A prediction model was designed based on these variables. These data were pooled with the data of two US and three Dutch centers (2007-2016). Discrimination and calibration were assessed in an internal-external validation procedure. Results: Overall, 1336 patients after DP were included, of whom 291 (22%) developed POPF grade B/C. A preoperative risk score was developed, including two variables: pancreatic neck thickness (OR:1.14 [95% CI:1.11-1.17] per mm increase) and pancreatic duct diameter (OR:1.46; [95%CI: 1.32-1.65] per mm increase). The model performed well in the design cohort (AUC:0.80 (95% CI:0.76-0.84)) and after internal-external validation (AUC:0.73 (95% CI:0.70-0.76)). Three risk groups for POPF grade B/C were identified: low-risk (<10%, 238 pts [18%]), intermediate-risk (10-25% 684 pts [51%]),), and high-risk (>25% 414 pts [31%]) Conclusions: The Distal Fistula Risk Score (D-FRS) is the first externally validated risk score to successfully predict the risk of POPF after DP. It can be easily calculated prior to surgery using www.pancreascalculator.com. The three distinct risk groups may facilitate personalized treatment.

Distal pancreatectomy with celiac axis resection (DP-CAR) is a surgical procedure with high morbidity and mortality performed in patients with locally advanced pancreatic cancer. Preoperative embolization of hepatic artery (PHAE) has been postulated as a technical option to increase resection rate.comparison of morbidity and mortality at 90 days, operative time, hospital stay and survival between patients that performed DP-CAR with and without PHAE.Observational retrospective multicentre study. Inclusion criteria: patient operated in Spanish centers with DP-CAR for pancreatic cancer from April 2004 until 23 June 2018. Preoperative (PHAE, neodjuvant treatment), intraoperative (operative time and blood loss) and postoperative data (morbidity, hospital stay, R0 and survival) were studied. Complications were measured with Clavien classification at 90 days. Specific pancreatic complications were measured using ISGPS classifications. Data were analyzed using R version 3.1.3 (http://www.r-project.org). Level of significance was set at 0.05.41 patients were studied. 26 patients were not embolized (NO-PHAE group) and 15 patients received PHAE. Preoperative BMI and percentage of neoadjuvant chemotherapy were the only preoperative variables different between both groups. The operative time in the PHAE group was shorter (343 min) than in the non-PHAE group (411 min) (p < 0.06). Major morbidity (Clavien > IIIa) and mortality at 90 days were higher in the PHAE group than in the non-PHAE group (60% vs 23% and 26.6% vs 11.6% respectively) (p < 0.004). No statistical difference in overall survival was observed between both groups (p = 0.14).In our study PHAE is not related with less postoperative morbidity. Even more, major morbidity (Clavien III-IV) and mortality was higher in PHAE group.

BACKGROUND:Minimally invasive pancreatoduodenectomy (MIPD) is increasingly being performed because of perceived patient benefits. Whether conversion of MIPD to open pancreatoduodenectomy worsens outcome, and which risk factors are associated with conversion, is unclear.METHODS:This was a post hoc analysis of a European multicentre retrospective cohort study of patients undergoing MIPD (2012-2017) in ten medium-volume (10-19 MIPDs annually) and four high-volume (at least 20 MIPDs annually) centres. Propensity score matching (1 : 1) was used to compare outcomes of converted and non-converted MIPD procedures. Multivariable logistic regression analysis was performed to identify risk factors for conversion, with results presented as odds ratios (ORs) with 95 per cent confidence intervals (c.i).RESULTS:Overall, 65 of 709 MIPDs were converted (9.2 per cent) and the overall 30-day mortality rate was 3.8 per cent. Risk factors for conversion were tumour size larger than 40 mm (OR 2.7, 95 per cent c.i.1.0 to 6.8; P = 0.041), pancreatobiliary tumours (OR 2.2, 1.0 to 4.8; P = 0.039), age at least 75 years (OR 2.0, 1.0 to 4.1; P = 0.043), and laparoscopic pancreatoduodenectomy (OR 5.2, 2.5 to 10.7; P < 0.001). Medium-volume centres had a higher risk of conversion than high-volume centres (15.2 versus 4.1 per cent, P < 0.001; OR 4.1, 2.3 to 7.4, P < 0.001). After propensity score matching (56 converted MIPDs and 56 completed MIPDs) including risk factors, rates of complications with a Clavien-Dindo grade of III or higher (32 versus 34 per cent; P = 0.841) and 30-day mortality (12 versus 6 per cent; P = 0.274) did not differ between converted and non-converted MIPDs.CONCLUSION:Risk factors for conversion during MIPD include age, large tumour size, tumour location, laparoscopic approach, and surgery in medium-volume centres. Although conversion during MIPD itself was not associated with worse outcomes, the outcome in these patients was poor in general which should be taken into account during patient selection for MIPD.

Introduction: Minimally-invasive pancreatoduodenectomy (MIPD) has been suggested as an alternative to open pancreatoduodenectomy (OPD). However, large international multicenter studies comparing oncological outcome after MIPD and OPD for pancreatic cancer are lacking. Methods: A multicenter international propensity-score matched retrospective cohort study including all consecutive patients undergoing MIPD or OPD (January 2010 to July 2019). Patients who had grossly positive resection margins (R2) or metastatic disease were excluded. Patients after MIPD were matched on a 1:1 ratio to OPD controls using propensity scores based on age, sex, BMI, ASA, abdominopelvic surgical history, and tumor size. Propensity score matching is applied to reduce the effect of confounding by indication. Primary outcome was overall survival, secondary outcomes included margin negative resection rate (R0), total number of extracted lymph nodes and chemotherapy rate. Outcomes were addressed by Kaplan-Meier analyses with between group comparisons (log-rank tests) and a multivariable Cox survival model to adjust for remaining confounders unaccounted for by matching (T stage, N stage, tumor grade). Results: A total of 695 MIPD from 20 centers in 8 countries and 900 OPD from 5 centers in 5 countries were collected. Statistical analyses on survival and oncological outcomes will be finished by the end of February 2020. Conclusion: This is the first large international multicenter study comparing outcomes after MIPD and OPD for pancreatic cancer. Results will be available upon acceptance for IHPBA 2020.

Background: FOLFIRINOX chemotherapy is used increasingly as upfront treatment for patients with borderline resectable pancreatic cancer (BRPC) or locally advanced pancreatic cancer (LAPC). However, multicenter studies on survival and complications after pancreatectomy following FOLFIRINOX chemotherapy are scarce. We aimed to assess overall survival, surgical morbidity and mortality of pancreatic cancer resections after upfront FOLFIRINOX chemotherapy. Methods: Multicenter, international retrospective cohort study among centers of the European-African Hepato-Pancreato-Biliary-Association (E-AHPBA). Included were patients undergoing resection for pathology-proven or highly-suspected pancreatic cancer (2012-2016) after at least 2 cycles of FOLFIRINOX chemotherapy. The primary endpoint was overall survival. Secondary endpoints included response rates, surgical radicality (i.e. R0-margin defined as >1mm), major morbidity and 90-day mortality. Results: We included 495 patients from 29 centers in 22 countries. The majority of these patients had cancer in the pancreatic head (n=329, 66.3%). Preoperatively, patients received a median 6 (IQR 5- 8) cycles of FOLFIRINOX and 104 (21.0%) patients also received radiotherapy. At restaging, complete radiological remission was observed in 25 (5.0%) patients, a partial response in 278 (56.2%), stable disease in 174 (35.1%) and progression in 5 (1.0%) patients. Post-operative major morbidity (i.e. Clavien-DIndo grade ≥3) occurred in 89 (18.0%) patients, with mortality reported in 9 (1.8%) patients. An R0-resection was achieved in 290 (58.6%) patients, and adjuvant chemotherapy was administered to 331 (66.9%) patients. Median overall survival from diagnosis was 35.1 (95% CI 31.1-386) months, including 29.2 (95% CI 22.0-38.3) in patients with LAPC. Conclusion: This pan-European study found a 35 months survival and low 1.8% mortality after pancreatectomy following FOLFIRINOX chemotherapy in patients with BRPC and LAPC. Future randomized studies will have to confirm the survival benefit of this approach compared to treatment with chemotherapy-alone.

Background: A recent RCT on MIDP reported significant reduction in time to functional recovery and a nonsignificant 13% absolute risk reduction for major morbidity compared to ODP. This study aimed to predict the population impact of minimally invasive distal pancreatectomy (MIDP) on major morbidity compared to open distal pancreatectomy (ODP). Methods: Nationwide observational cohort study evaluating the association between surgical approach and composite major morbidity (CMM; death or severe complications) after elective distal pancreatectomy for tumors and cysts using ACS-NSQIP® (2014-2016) and external validation using Dutch Pancreatic Cancer Group (DPCG) data (2005-2016). Multivariable logistic regression assessed the impact of MIDP on CMM at varying implementation rates between (0%-100%), including conversion (0%-25%), using marginal effects. Results: Of 2921 ACS-NSQIP® patients, 1359 (47%) underwent ODP, and 1562 (53%) underwent MIDP with 18% conversion. MIDP was associated with reduced 30-day major morbidity (OR 0.50, 95CI 0.42-0.60, p<0.001), confirmed upon subgroup analysis and external validation (n=637; OR 0.47, 95CI 0.29-0.77, P = 0.003). Conversion mitigated this association in ACS-NSQIP® (OR 2.74, 95CI 2.05-3.65, P<0.001). The predicted absolute population risk reduction in ACS-NSQIP® was 12% (95CI 7.8-15) at 18% conversion and 15% (95CI 11-19%) at 0% conversion (Fig. 1). Conclusion: MIDP was associated with a 12% absolute reduction in major morbidity compared to open in two nationwide datasets, confirming findings of a recent RCT. Lower conversion rates may further reduce major morbidity, especially at higher MIPD implementation rates. When combined with recent evidence, this study suggests MIDP could become the new gold standard for pancreatic tail resection.

Biliary tract carcinoma (BTC) comprises gallbladder and intra-/extrahepatic cholangiocarcinoma (GBC, ICC, EHC), which are currently classified by anatomical origin. Better understanding of the mutational profile of BTCs might refine classification and improve treatment. We performed a systematic review of studies reporting on mutational profiling of BTC. We included articles reporting on whole-exome/whole-genome-sequencing (WES/WGS) and targeted sequencing (TS) of BTC, published between 2000-2017. Pooled mutation proportions were calculated, stratified by anatomical region and sequencing technique. A total of 25 studies with 1806 patients were included. Overall, TP53 was the most commonly mutated gene in BTC. GBC was associated with mutations in PFKFB3, PLXN2 and PGAP1. Mutations in IDH1, IDH2 and FGFR fusions almost exclusively occurred in ICC patients. Mutations in APC, GNAS and TGFBR2 occurred exclusively in EHC patients. In conclusion, subtypes of BTCs exhibit minor differences in mutational profile, which is likely influenced by the cell of origin.

Introduction: Biliary tract carcinomas (BTCs) are heterogeneous tumors in which analysis of genetic alterations using next-generation-sequencing (NGS) plays an increasing role in diagnosis and therapy. However, a systematic overview of BTC mutation profiles is lacking. We performed a systematic review including a weighted overview of BTC mutation profiles, to enable the construction of BTC-specific NGS panels. Methods: A systematic search was performed for articles on whole-exome/whole-genome sequencing (WES/WGS) and targeted sequencing (TS) of BTCs published between 2000 and June 2017. Supplementary data were requested when not available. For each study, the frequency of mutated and non-mutated genes was recorded. The weighted mutation-proportion was calculated per gene. These were stratified by tumor localization and NGS technique. Results: Out of 1332 studies, 23 studies were included in the final analysis. The most frequently mutated genes for intrahepatic cholangiocarcinoma were KRAS, IDH1, ARID1A, TP53, MCL1 and TP53, ARID1A, MCL1, PBRM1, MLL3, respectively. For extrahepatic cholangiocarcinoma these were MLL3, TGFBR2, TP53, KRAS, CDKN2A and ALB, BCL3, CDKN2A, TGFBR2, TP53, respectively. For gallbladder carcinoma these were TP53, C5DM1, EYS, ARID2, CSDM3 and TP53, CSDM1, CDKN2A, MCL1, EYS, respectively. Conclusion: The gene mutations in BTC varied between different localizations in the biliary tree. These differing mutational profiles may be the result of differences in the prevalence of molecular subtype for each location. NGS results further differed depending on the used NGS technique (WES/WGS or TS), underscoring the need for custom-made panels for TS based on WGS/WES results for future studies and the development of diagnostic tests.

Background: We assessed short-term outcomes after minimally invasive (laparoscopic, robot-assisted, hybrid) pancreatoduodenectomy (MIPD) versus open pancreatoduodenectomy (OPD) among European high-volume (>10 MIPDs and >20 PDs overall per year) centers. Methods: Multicenter propensity-score-matched (1:1) retrospective study on MIPD vs OPD for (pre-)malignant tumors or cysts in 14 European MIPD centers (2012-2017) and OPD data from Dutch and German pancreatic surgery registries (2014-2017). Propensity scores were based on age, sex, BMI, ASA, comorbidities, ECOG, tumor location, suspected cancer, organ involvement, and venous resection. Primary outcome: 30-day major morbidity (Clavien-Dindo 3a-5). Secondary outcomes: grade-B/C pancreatic fistula (POPF), R1-resection (< 1 mm) margin, hospital stay and 30-day mortality. Results: Of 4220 included patients, all 730 MIPD (413 laparoscopic, 184 robot-assisted, 130 hybrid) were matched to 730 OPD. Major morbidity (28% vs 29%, P>0.729), mortality (4.0% vs 2.9%, P=0.314), R1-resections (20 vs 23%, P=0.269), and hospital stay (mean 17 vs 17 days, P>0.99) were similar, POPF occurred more frequently after MIPD (23% vs 14%, P< 0.001). The association between MIPD and major morbidity (OR 0.95, P=0.686), mortality (OR 1.57, P=0.121), and POPF (OR 1.84, P< 0.001) remained similar after excluding, respectively, lower volume (10-20 MIPD/year) centers, first 10 MIPD cases, hybrid cases, hybrid and laparoscopic cases. Conclusions: In this largest propensity-score-matched study on MIPD vs OPD to date, we found no significant differences in major morbidity, mortality, R1-resection rate, and hospital stay, but more POPF after MIPD. MIPD can safely be performed in high-volume centers, but the increased rate of POPF should be addressed.

Objective: To evaluate short-term surgical outcomes for minimally invasive (MIS) vs open pancreaticoduodenectomy (PD) for which national studies are lacking. Methods: ACS-NSQIP targeted pancreas database queried for patients who underwent PD in 2014. Approach categorized as planned open vs intention-to-treat MIS. One-to-one propensity score matching used to account for selection bias in pre-operative and intra-operative characteristics. Results: 2400 patients were identified, of whom 195 (8.1%) underwent MIS PD. Patients who underwent MIS were more likely to be: white (85.6% vs 80.1%; p=0.047) and male (56.4% vs 51.2%; p=0.16), and less frequently preoperatively biliary stented (45.1% vs 51.9%; p=0.071). Gland texture, duct size, diagnosis (benign neoplasm, malignant, other), and 30-day mortality were not significantly different. Of note, no information on surgeon experience or hospital volume available. MIS more likely to result in pancreatic fistulae (24.6% vs 17.5%; p=0.013), longer operative time (median 435 vs 361 mins; p<0.0001), but shorter LOS (median 7 vs 8 days; p=0.034). After matching, cohort consisted of 390; pancreatic fistulae remained significantly higher for MIS (24.6% vs 13.3%; p=0.005). ISGPF grade not available; however, SSI, and mortality equivalent in matched cohort (Figure). Operative time remained longer (median 435 vs 357 mins; p<0.001). Trend observed for increased delayed gastric emptying in MIS (20.0% vs 12.8%; p=0.056). Conclusion: Matched comparison of MIS-PD using national data demonstrated prolonged operative time, higher rates of pancreatic fistulae and no difference in LOS. These results taken in context of more favorable single- and multi-institutional data may suggest deployment of MIS pancreaticoduodenectomy should be rigorously selective, based on criteria that might include surgeon, center, and patient metrics.

Introduction: Pancreatic cancer encasing the celiac axis is considered unresectable according to most guidelines. Nevertheless, in cases without involvement of the aorta, superior mesenteric- and gastro-duodenal arteries, distal pancreatectomy with celiac axis resection (DP-CAR) could lead to a radical (R0) resection. preoperative hepatic artery embolization (PHAE) may enhance the development of the pancreatic-arcade and reduce the risk of ischemic complications. We aimed to aggregate the evidence on DP-CAR for pancreatic cancer involving the celiac axis, including the impact of PHAE and adjuvant therapy. Methods: A systematic review was performed according to the PRISMA guidelines. PubMed, Embase, and Cochrane Library were searched to identify studies on DP-CAR published before May 27th 2015. Primary endpoint was overall survival; secondary endpoints included morbidity and radical resection rate. Results: A total of 20 retrospective studies, including 263 patients, were included. The methodological quality of the studies ranged from poor to moderate. The radical resection rate was 77%, major morbidity 27%, ischemic morbidity 10% and 90-day mortality 2.7%. Overall, 47% (74/157) of patients underwent PHAE and 42% (93/223) patients underwent portal vein (PV) resection. Ischemic morbidity rate was 13% (15/116) in series where >50% of patients underwent PHAE. Median overall survival was 15.4 months (range 9–48), whereas survival was 17.9 months (range 10–26) in series where >50% of patients received adjuvant therapy. Conclusions: In selected patients, DP-CAR for pancreatic cancer with celiac axis involvement seems safe with acceptable mortality and survival. The benefit of PHAE is unclear and should be determined by prospective studies.TableStudyDP-CARs (N)Age, median (Y)PHAE (%)PV Resection (%)R0 (%)Major morbidity (%)Ischemic morbidity (%)Length of stay, median (D)90d Mortality (%)Baumgartner et al. 201211610649190918Hirano et al. 200723638370914822NRNRMiura et al. 201450641006492541239NROkada et al. 20143766>502257NR5NRNRShimura et al. 20121469NR36NR3521NR0Takahashi et al. 2011166502556560386Wang et al. 2015156102010070267…...........................Overall2636347 (74/157)42 (93/223)77 (174/227)27 (26/96)10 (26/256)322.7 (3/113)[Outcomes after DP-CAR] Open table in a new tab [Outcomes after DP-CAR] [Survival vs. (neo-) adjuvant treatment]