Young Investigator (pre-recorded)
Other / Misc 
YI01 Young Investigator (pre-recorded) 
Selection of Presentations from Abstract Submissions

OYIA-01 Hepatocellular Adenoma in Men. A Nationwide Assessment
Alicia Furumaya, Netherlands

B. van Rosmalen1, A. Furumaya1, A.J. Klompenhouwer2, M. Doukas3, R. de Man2, J. Ijzermans4, C. Dejong5,6, T.M. van Gulik1, J. Verheij7, Dutch Benign Liver Tumour Group
1Surgery, Amsterdam UMC, University of Amsterdam, Netherlands, 2Gastroenterology, Erasmus University Medical Center, Netherlands, 3Pathology, Erasmus University Medical Center, Netherlands, 4Surgery, Erasmus University Medical Center, Netherlands, 5Surgery, Maastricht University Medical Center, Netherlands, 6Surgery, Universitätsklinikum Aachen, Germany, 7Pathology, Amsterdam UMC, University of Amsterdam, Netherlands

Introduction: In contrast to women, men with hepatocellular adenoma (HCA) are extremely rare (incidence of 4 in ten million). Survival-rate and risk of malignant transformation into hepatocellular carcinoma (HCC) is currently unknown. We correlated histopathology and clinical data of all Dutch male HCA patients.
Methods: Samples of all Dutch male HCA patients between 2000 and 2019 were collected using PALGA (Dutch Pathology Registry). Histopathology was revised by two expert pathologist, supplemented by immunohistochemistry. Additional molecular analyses were performed to confirm malignancy (hTERT) and/or subtype of HCA.
Results: Sixty-six patients from 22 medical centres, with 67 tumours initially diagnosed as HCA were included. After expert-revision 26/67(38.2%) HCA, 17/67(25%) borderline HCA/HCC and 24/67(35.3%) HCC were diagnosed. Diagnosis was concordant with initial diagnosis in 30/67(44.8%) patients, and discordant in 37/67(55.2%) patients. Diagnosis changed from HCA to HCC in 8/67(11.9%) patients, and from HCC to HCA in 1/67(1.5%) patient. Clinical data matching 46 expert-revision patients was available. Of 22 HCA patients, four (18.1%) developed HCC, of which two died from HCC (9.0%). Of 11 borderline HCA/HCC, two died (18.1%). Three of the four HCA patients with clinical HCC development were beta-catenin mutated-HCA, one was inflammatory-HCA.
Conclusion: HCA in men is difficult to diagnose and additional immunohistochemical stainings and/or molecular analyses are important; 35.3% were probable or certain HCC's after revision. If HCA is confirmed after expert revision, there is still 18.1% risk of developing HCC and 9.0% mortality. It's advisable to perform diagnostics in expert centres, and to approach this tumour as potential HCC.
OYIA-02 A Patient-derived Organoid and Artificial Intelligence-based Workflow for Optimising Chemotherapeutics to Improve Outcomes in Pancreatic Cancer
Claire Alexandra Zhen Chew, Singapore

C.A.Z. Chew1, S.L. Chan2, K. Madhavan1,2, A.W.C. Kow1, S.G. Iyer1, K.Y. Ho3, E.K.-H. Chow4, C.E. Chee5, G.K. Bonney1
1Surgery, National University Hospital, Singapore, 2Surgery, National University of Singapore, Singapore, 3Gastroenterology, National University Hospital, Singapore, 4Cancer Science Institute, National University of Singapore, Singapore, 5Haematology-Oncology, National University Hospital, Singapore

Pancreatic ductal adenocarcinoma is one of the most aggressive solid tumors. Despite numerous advances in surgical techniques, survival has remained largely unchanged over the last decade. As such, pancreatic cancer has been described as a 'systemic disease' - and one which chemotherapy plays a pivotal role both pre- and post-surgery. Tools that can predict the most effective chemotherapeutic regimens for patients are urgently needed to improve survival in this dismal disease so as to optimise biological effects while minimising side effects and fitness for surgery.
We have established a biobank of patient-derived organoids from endoscopic ultrasound biopsies of pancreatic adenocarcinoma. Using an optimised technique these are generated within a clinically applicable timeframe for high-throughput drug screening that can inform therapeutic decisions. By applying artificial intelligence, we are able to optimize drug combinations to maximise clinical efficacy while minimising toxicity. The inclusion of preclinical drugs in our screening panel in addition to standard therapeutics allows for the derivation of novel drug combinations. This work forms the basis for an interventional trial at our centre. We also employ the use of highly sensitive proteomics to profile patients that demonstrate similar therapeutic responses to identify predictive biomarkers of chemosensitivity.
We present a platform for personalized medicine in pancreatic cancer that employs organoid technology and artificial intelligence to predict chemosensitivity and therapeutic response for individual patients.
OYIA-03 Ablative Radiation Provides Comparable 1-year Survival to Surgery after Neoadjuvant Chemotherapy for Locally Advanced Pancreatic Adenocarcinoma
Joshua Jolissaint, United States

J. Jolissaint1,2, M. Reyngold3, J. Bassmann1, K. Seier4, M. Gonen4, J. Drebin1, W. Jarnagin1, C. Crane3, A. Wei1
1Surgery, Memorial Sloan Kettering Cancer Center, United States, 2Surgery, Brigham and Women's Hospital, United States, 3Radiation Oncology, Memorial Sloan Kettering Cancer Center, United States, 4Epidemiology and Biostatistics, Memorial Sloan Kettering Cancer Center, United States

Introduction: For patients with locally advanced pancreatic ductal adenocarcinoma (LA PDAC), ablative radiation to biologically effective doses of ≥ 100 Gy can provide locoregional control. In this study, we compared survival after ablative radiation versus resection in patients with LA PDAC.
Methods: Patients with LA PDAC treated with neoadjuvant chemotherapy between 2014-18 at our institution were identified. Patients with resected AJCC 7th Ed. ypT3 or ypT4 tumors, or those requiring vascular resection, were compared to unresected patients treated with ablative radiation. Overall survival (OS) was calculated from completion of chemotherapy using Kaplan Meier methods and the log rank test was used to compare groups.
Results: 170 patients were included for analysis: 70 underwent resection (39.1%) and 109 (60.9%) received ablative radiation. Age, gender, ECOG status, CA 19-9, and nodal status were not different between groups. Patients treated with ablative radiation had a larger tumor size (3.2cm [0.0, 10.9] vs. 2.6cm [0, 7.1], p< 0.001), and were more likely to have arterial and venous sencasement (all p< 0.001) (Table 1). There was no significant difference in median OS for radiation (20.6 months, 95% C.I. 17.1-32.4) compared to surgery (30.5 months, 95% C.I. 19.4-40.7) (p=0.22), and 1-year survival was similar between groups (77% for radiation, 75% for surgery) (Figure 1).
Conclusion: In patients with LA PDAC, ablative radiation is associated with a 1-year survival comparable to that achieved with resection. Ablative radiation may be a promising alternative to surgical resection in select subsets of PDAC and warrants further prospective investigation.
VariableAblative RadiationResectionp-value
Age, median (range) (years)67.6 (42.4-90.6)68.4 (45.3-85.1)0.71
Female gender, N (%)55 (50.5)35 (50.0)>0.95
ECOG performance status 2+, N (%)16 (14.7)4 (5.7)0.09
CA 19-9 at diagnosis, median (range) (u/mL)218 (0.0-9635.0)117 (0.0-80195.0)0.18
Nodal involvement, N (%)40 (36.7)22 (31.4)0.52
Celiac artery or hepatic artery encasement, N (%)48 (44)5 (7.2)<0.001
Superior mesenteric artery encasement, N (%)43 (39.4)6 (8.7)<0.001
Superior mesenteric vein or portal vein encasement, N (%)35 (32.1)5 (7.2)<0.001
Follow up time from chemotherapy completion, median (range) (months)15.9 (3.7-34.3)26.9 (2.3-48.1)N/A
[Table 1. Patient and tumor characteristics among patients undergoing ablative radiation or resection for locally advanced pancreatic adenocarcinoma]

[Figure 1. Comparison of overall survival between ablative radiation (RT) and surgery]
OYIA-04 Risk Factors for Postoperative Bile Leak and Outcomes in Patients Undergoing Minimally Invasive Hepatectomy: A Contemporary ACS NSQIP Analysis
Charles Vining, United States

C. Vining1, K. Kuchta2, Y. Berger1, P. Paterakos2, D. Schuitevoerder1, K. Roggin1, M. Talamonti2, M. Hogg2
1Surgery, University of Chicago, United States, 2Surgery, NorthShore University HealthSystem, United States

Introduction: Despite advances in surgical technique, bile leak remains a common complication associated with morbidity following hepatectomy. We sought to identify risk factors of biliary leak and outcomes in patients undergoing minimally-invasive hepatectomy.
Methods: An ASC-NSQIP multi-institutional retrospective cohort study from 2014-2017. Distribution of bile leak by surgical approach was identified. Univariate and multivariate factors associated with bile leak and outcomes were evaluated using chi-square and logistic regression.
Results: Of 13,955 patients who underwent hepatectomy, 10,190(73%) were open, 3438(24.6%) were laparoscopic and 323(2.3%) were robotic. A laparoscopic approach was associated with decreased bile leak for all hepatectomies (5.4% vs 8.4%; p=0.028) compared to robotic. Multivariate operative risk factors for bile leak in minimally-invasive hepatectomies included laparoscopic (vs robotic) approach (OR=0.51 [95% CI 0.33-5.06]; p=0.003), conversion (OR=2.28 [95% CI 1.62-3.10]; p< 0.001), biliary reconstruction (OR=4.11 [95% CI 2.14-7.90]; p< 0.001) and transfusion (OR=1.88 [95% CI 1.28-2.75]; p=0.001). Bile leak was associated with increased reoperation (0.8% vs 8.9%; p< 0.001), 30-day readmission (5.4% vs 27.2%; p< 0.001), 30-day mortality (0.6% vs 1.9%; p=0.042), increased median length of stay (3 days vs 6 days; p< 0.001), and any (15.2% vs 65.3%; p< 0.001), surgical (9.7% vs 59.6%; p< 0.001) and medical (6.9% vs 28.6%; p< 0.001) complications for patients who underwent minimally-invasive hepatectomy (Table 1).
Conclusion: Risk factors for developing bile leak in patients undergoing minimally-invasive hepatectomy were identified. Bile leaks were associated with multiple additional complications, and the robotic approach was a greater risk for bile leak than laparoscopic approach.
[Table 1. Multivariable Analysis of Factors Associated with Bile Leak]
OYIA-05 Pharmacological Activation of Nrf2 Enhances Metabolic Remodelling and Functional Liver Regeneration Following Major Hepatectomy
Benjamin Kai Yan Chan, United Kingdom

B.K.Y. Chan1,2, S.S. Forootan2, M. Elmasry2,3, T.M. Bunday2, H.Z. Malik1, C.E. Goldring2, I.M. Copple2, S.W. Fenwick1
1Department of Hepatobiliary Surgery, Liverpool University Hospitals NHS Foundation Trust, United Kingdom, 2MRC Centre for Drug Safety Science, University of Liverpool, United Kingdom, 3Department of Surgery, Oxford University Hospital, United Kingdom

Introduction: Post-hepatectomy liver failure (PHLF; 10% incidence) remains a risk following major liver resections. Two-stage surgeries attempt to address the inadequate future liver remnant but with inherent risks. Pharmacological enhancement of liver regeneration could improve outcomes and extend resectability. The transcription factor Nrf2 regulates a battery of metabolic and cell defence genes. We hypothesized that Nrf2 activation would enhance liver regeneration following hepatectomy.
Methods: The effects of an Nrf2 activator were evaluated in a mouse major hepatectomy model. Liver volume and function (indocyanine green clearance) were assessed using magnetic resonance imaging and multispectral optoacoustic tomography, respectively. Hepatocyte size was assessed through phalloidin immunofluorescence tissue staining. RNASeq and RT-qPCR were used to determine gene expression changes in mouse liver tissue and primary human hepatocytes.
Results: Nrf2 activation significantly enhanced post-hepatectomy regeneration of liver volume and reduced the loss of hepatic function in mice. This was associated with a significant increase in hepatocyte hypertrophy during early regeneration. RNASeq revealed that Nrf2 activation caused an upregulation of redox, cofactor and glutathione metabolism genes, congruent with increases in NADP/H content and glutathione levels, as well as the downregulation of lipid synthesis genes (Figure 1). Notably, such genes were similarly modulated in primary human hepatocytes following in vitro treatment with the Nrf2 activator.
Conclusions: Nrf2 activation can improve liver regeneration by enhancing metabolic remodelling and hepatocyte hypertrophy, which translates into an improvement in hepatic function. Our results suggest Nrf2 as a novel target in the treatment of liver pathology, especially the avoidance of PHLF.

[Figure 1]
OYIA-06 Natural History and Optimal Treatment Strategy of Intraductal Papillary Mucinous Neoplasm of Pancreas
Youngmin Han, Korea, Republic of

Y. Han, J.-Y. Jang, W. Kwon, H. Kim, Y.H. Byun, J.S. Kang, Y.J. Choi, M.Y. Oh
Seoul National University Hospital, Korea, Republic of

Background: With increasing detection of intraductal papillary mucinous neoplasms(IPMN), a tailored approach is needed. We explored the natural history of IPMN and suggest optimal treatment based on malignancy risk using nomogram and Markov decision model.
Methods: Patients with IPMN, who underwent surveillance or surgery, were included. Change in worrisome features/high-risk stigmata and malignancy conversion rate was calculated through radiologic and pathologic reviews. Life expectancy and quality-adjusted life year (QALY) were compared using a nomogram predicting malignancy.
Results: Overall, 2,006 patients with histologically confirmed or radiologically typical IPMN were enrolled. Of these, 1,773(88.4%) had branch duct(BD), 81(4.0%) had main duct and 152(7.6%) had mixed type at initial diagnosis. The cumulative risk of developing worrisome feature/ high-risk stigmata was 19.0% at 5-year and 35.0% at 10-year- follow-up. The progression of malignancy rate at 10-year follow-up was 79.9% for main and mixed type IPMN and 5.9% for BD-IPMN. Nomogram based malignancy risk prediction is well correlated with natural history based on pathologic biopsy and shows good stratification of the survival. Decision model recommends surgery to maximize overall survival and quality-adjusted life year for patients under 75-years old, especially those with over 35% malignancy risk
Conclusion: Compared to the high risk of malignancy (79.9%) in main and mixed type IPMN, that (5.9%) of BD-IPMN are very indolent. The nomogram-based decision model suggests surgery rather than surveillance for patients with high malignancy rate. Optimal treatment strategy between surgery and surveillance should consider patient's health status, malignancy risk, and centre's experience.
[Life Expectancy and Quality Adjusted Life Year according to treatment group in IPMN]
OYIA-08 Neoadjuvant FOLFIRINOX versus Neoadjuvant Chemoradiotherapy and Adjuvant Chemotherapy for Resectable and Borderline Resectable Pancreatic Cancer: Update on the Nationwide Multicenter PREOPANC-2 Randomized Trial
Quisette P. Janssen, Netherlands

Q.P. Janssen1, M.G. Besselink2, M.Y.V. Homs3, G. van Tienhoven4, J.W. Wilmink5, B. Groot Koerkamp1, Dutch Pancreatic Cancer Group (DPCG)
1Surgery, Erasmus University Medical Center, Netherlands, 2Surgery, Amsterdam UMC, University of Amsterdam, Netherlands, 3Medical Oncology, Erasmus University Medical Center, Netherlands, 4Radiotherapy, Amsterdam UMC, University of Amsterdam, Netherlands, 5Medical Oncology, Amsterdam UMC, University of Amsterdam, Netherlands

Background: Two randomized studies have found superior R0 resection rate with neoadjuvant chemotherapy compared to upfront surgery for (borderline) resectable pancreatic cancer. The PREOPANC-2 trial compares two neoadjuvant regimens in patients with resectable and borderline resectable pancreatic cancer ((B)RPC).
Material and methods: This multicenter randomized controlled trial includes patients with pathologically proven pancreatic cancer, WHO performance score < 2, and both ≤90 degrees arterial and ≤270 degrees venous abutment on imaging. Patients are randomized to receive 8 cycles of neoadjuvant FOLFIRINOX chemotherapy without adjuvant treatment (arm A), or 3 cycles of neoadjuvant gemcitabine combined with hypofractionated radiotherapy (36 Gy in 15 fractions), followed by surgery and adjuvant 4 cycles of gemcitabine, similar to the intervention arm of the PREOPANC-1 trial. The primary endpoint is OS by intention-to-treat. Secondary endpoints include progression-free survival, quality of life, and (R0) resection rate. Blood samples for biomarkers are collected at five time points. Sample size is 368 patients, with 2.5 years accrual and 1.5 years follow-up after last inclusion.
Results: All 15 centers performing pancreatic surgery in the Netherlands are open and 199 (54%) patients have been randomized. Last three monthly accrual rates were 15, 16, and 18 patients. More than 50% of all eligible patients in the Netherlands have been enrolled in the last months. Biomarkers have been collected for 90% of sampling moments.
Conclusion: Rapid accrual of patients in an RCT comparing two neoadjuvant regimens for (B)RPC is feasible. Results of the PREOPANC-2 trial are anticipated in the summer of 2022.
[Study overview PREOPANC-2 trial]
OYIA-09 The Immune Micro-environment of Treatment Naïve and Neoadjuvant Treated Pancreatic Ductal Adenocarcinoma Tissues from the PREOPANC-1 Randomized Controlled Trial
Diba Latifi, Netherlands

D. Latifi1, F. Grevers2, Y. Li2, W. de Koning2, M. Suker1, B. Groot Koerkamp1, C.H.J. van Eijck1, D.A.M. Mustafa2, DPCG PREOPANC-1 Study Group
1Surgery, Erasmus University Medical Center, Netherlands, 2Pathology, Erasmus University Medical Center, Netherlands

Objectives: To investigate the immune micro-environment in tumors of patients with pancreatic ductal adenocarcinoma (PDAC) after immediate surgery or neoadjuvant therapy and its relation to survival outcome.
Methods: We analysed tumor samples from 50 patients without and 53 patients with neoadjuvant chemo-radiotherapy. Immune profiling of FFPE tumor samples was performed using the NanoString PanCancer assay and the expression of 770 immune related genes was measured. We used the two sample T-test for comparison of gene expression levels between the two groups. The multivariate Cox proportional hazards model was performed to identify independent prognostic predictors. The Kaplan-Meier method with a log-rank test was applied to compare the survival distributions between groups.
Results: Total tumor infiltrating lymphocytes (TILs) was equally distributed between treatment groups. B cell infiltration was decreased significantly (p< 0,001) in the neoadjuvant group, whereas T cell infiltration did not vary between groups. Furthermore, CXCL12 expression, a chemokine involved in tumor proliferation and progression, was significantly (p< 0,0001) higher in neoadjuvant treated tumors. However, its receptor CXCR4 was down regulated. In addition, high expression of the ALCAM gene was strongly associated with poor prognosis in both treatment arms. Results from immune cell type ratios as prognostic factor will follow.
Conclusion: The tumor immune micro-environment after neoadjuvant treatment differs from treatment naïve tumors, indicating that neoadjuvant treatment affects the distribution of immune cells in PDAC tumors.
OYIA-10 Response to and Survival after First-Line FOLFIRINOX or Gemcitabine/Nab-Paclitaxel for Localized Pancreatic Ductal Adenocarcinoma
Giampaolo Perri, Italy

G. Perri1, L. Prakash2, G.R. Varadhachary3, E.J. Koay4, M. Kim2, N. Ikoma2, C. Tzeng2, J.E. Lee2, M.H. Katz2
1General and Pancreatic Surgery, Verona University, Italy, 2Surgical Oncology, MD Anderson Cancer Center, United States, 3Medical Oncology, MD Anderson Cancer Center, United States, 4Radiation Oncology, MD Anderson Cancer Center, United States

Introduction: We sought to evaluate radiographic and serologic measures of response to first-line chemotherapy with FOLFIRINOX or gemcitbine/nab-paclitaxel (GA) for localized pancreatic cancer (PDAC) and their association with survival
Methods: All patients diagnosed with localized PDAC at MD Anderson between 2010 and 2017 who received at least 3 cycles of first-line chemotherapy with FOLFIRINOX or GA were included. 495 patients were treated with FOLFIRINOX (n= 295; 60%) or GA (n= 200; 40%) as first-line chemotherapy. Main outcomes/measures reviewed were: resection rate, radiographic (RECIST 1.1/change in tumor volume/anatomic stage) and serologic (CA 19-9) metrics of response, overall survival.
Results: Patients treated with FOLFIRINOX were younger, had better ECOG performance status but more invasive tumors (all P< 0.01). Overall survival was similar (FOLFIRINOX vs GA: 48 months vs NR for resected, P= 0.8; 18 vs 17 months for non-resected, P= 0.2). 30% of patients (n= 101) with elevated pretreatment CA 19-9 had a normalization after chemotherapy, 11% (n= 56) of patients had RECIST partial response, 5% of BR and LAPC patients were anatomically downstaged. Decrease in tumor volume after therapy was seen in 67% of the patients (n= 331), with median decrease of 20% of the pretreatment volume. There were no differences between FOLFIRINOX and GA in terms of radiographic and serologic response to therapy. RECIST progressive disease, decrease in tumor volume and posttreatment CA 19-9 level were independent predictors of survival.
Conclusion: Radiographic and serologic metrics of response to first-line chemotherapy are important prognostic factors and similar between FOLFIRINOX and GA.
OYIA-11 Transatlantic Registries of Pancreatic Surgery in the USA, Germany, the Netherlands, and Sweden (GAPASURG): Comparing Variables, Patients, Treatment Strategies, and Outcomes
Tara Mackay, Netherlands

T. Mackay1, E.M. Gleeson2, U.F. Wellner3, O.R. Busch1, T. Keck3, H.C. van Santvoort4, B. Tingstedt5, H.A. Pitt6, M.G. Besselink1, Global Audits on Pancreatic Surgery Group (GAPASURG)
1Amsterdam UMC, AMC, Netherlands, 2Hahnemann University Hospital and Drexel University College of Medicine, United States, 3DGAV StuDoQ|Pancreas and Clinic of Surgery, UKSH Campus, Germany, 4Regional Academic Cancer Center Utrecht, University Medical Center Utrecht and St. Antonius Hospital Nieuwegein, Netherlands, 5Clinical Sciences Lund, Lund University, Skåne University Hospital, Sweden, 6Temple University Health System, Lewis Katz School of Medicine at Temple University, United States

Introduction: Pancreatic surgery registries facilitate both quality improvement and clinical research. We aimed to compare existing registries for variables collected, patient, tumor, and treatment characteristics, and outcomes.
Methods: Fifty-one core parameters for pancreatoduodenectomies (2014-2017) in registries from the USA, Germany, the Netherlands, and Sweden were compared with relative and absolute largest differences (RLD; ALD) between extremes (smallest vs. largest).
Results: Overall, 22983 pancreatoduodenectomies were included (15224, 3558, 2795, and 1406 in the USA, Germany, the Netherlands, and Sweden, respectively). Design of the registries varied as 18/51 (35.3%) core parameters were not available. Preoperative chemotherapy in patients with pancreatic ductal adenocarcinoma was administered most often in the USA (27.6%, 4.9%, 7.0%, and 3.4%, RLD 8.1, ALD 24.2%, p< 0.001). Minimally invasive procedures were performed most often in the Netherlands (7.8%, 4.5%, 13.5%, and unknown, RLD 3.0, ALD 9.0%, p< 0.001. Median length of stay was 8.0, 16.0, 12.0, and 11.0 days (RLD 2.0, ALD 8.0, p< 0.001). Reoperation was performed most frequently in Germany and Sweden (5.7%, 17.1%, 8.7%, and 11.2%, RLD 3.0, ALD 11.4%, p< 0.001). In-hospital mortality was 1.3%, 4.7%, 3.6%, and 2.7% (RLD 3.6, ALD 3.4%, p< 0.001).
Conclusion: Considerable differences exist in design, patients, treatment strategies, and outcomes in registries of pancreatic surgery among four Western countries. The absolute largest nationwide differences of 24.3% for preoperative chemotherapy, 9.0% for minimally invasive surgery, 11.4% for reoperation rate, and 3.4% for in-hospital mortality require further study and improvement. This analysis provides core parameters for future registries on pancreatic surgery.
OYIA-12 Adjuvant Therapy in Distal Cholangiocarcinoma Following Pancreaticoduodenectomy: A National Cancer Database Analysis
Sivesh Kamarajah, United Kingdom

S. Kamarajah1, T. Frankel2, F. Bednar2, C. Cho2, H. Nathan2
1Newcastle University Trust Hospitals, United Kingdom, 2Department of Surgery, University of Michigan, United States

Background: There is conflicting evidence for the benefit of adjuvant chemotherapy following resection of distal cholangiocarcinoma (dCCA), especially for node-negative (N0) resections. We aimed to evaluate the association of adjuvant chemotherapy with survival after surgical resection of dCCA.
Methods: Using National Cancer Database (NCDB) data from 2004 to 2015, we identified patients with resection of non-metastatic dCCA. Patients with neoadjuvant radiotherapy and chemotherapy were excluded. Propensity score matching on 16 hospital-, patient- and pathological-level variables was used to account for treatment selection bias. A multivariable Cox proportional hazards model was then used to analyze the association of chemotherapy with survival.
Results: Of 3,881 (38%) adjuvant chemotherapy and 4,342 (62%) non-adjuvant chemotherapy patients, 3,026 adjuvant chemotherapy and 3,026 non-adjuvant chemotherapy patients remained in the cohort after matching. Clinicopathologic and demographic variables were well balanced after matching. After matching, adjuvant chemotherapy was associated with higher survival (median 28.6 vs 18.7 months, p< 0.001). After multivariable adjustment, adjuvant chemotherapy remained associated with a survival benefit (HR 0.64, 95% CI 0.60 - 0.69, p< 0.001). Stratified and multivariable interaction analyses showed that this benefit was restricted to patients with node positive disease (HR: 0.84, CI95%: 0.74-0.96, p=0.013) and R1 resection (HR: 0.79, CI95%: 0.69-0.90, p< 0.001).
Conclusion: In this large retrospective cohort study, adjuvant chemotherapy after dCCA resection was associated with a survival benefit in patients with node-positive and margin positive disease. Adjuvant chemotherapy should be considered routinely after node-positive and margin positive resection of dCCA.
OYIA-13 Preoperative Prediction Score of Hepatocellular Carcinoma Recurrence in Living Donor Liver Transplantation: Validation of SNAPP Score Developed at Asan Medical Center
Seok-Hwan Kim, Korea, Republic of

S.-H. Kim1, D.-B. Moon2, G.-C. Park2, S.-G. Lee2, S. Hwang2, C.-S. Ahn2, K.-H. Kim2, T.-Y. Ha2, G.-W. Song2
1Surgery, Chungnam National University Hospital, Korea, Republic of, 2Division of Liver Transplantation and Hepatobiliary Surgery, Department of Surgery, Asan Medical Center/University of Ulsan College of Medicine, Korea, Republic of

The application of the previously proposed scoring systems is not readily available due to the lack of simplicity for predicting hepatocellular carcinoma (HCC) recurrence. We aimed to develop and validate the new score system, which can predict HCC recurrence after living donor liver transplantation (LDLT) by using morphologic and biologic data. Predictors for HCC recurrence after LDLT were developed (n = 627) and validated (n = 806) in 1433 patients between 2007 and 2016 at Asan Medical center (AMC) to create the SNAPP score [tumor Size and Number, alpha-fetoprotein (AFP), vitamin K absence-II (PIVKA-II), positron emission tomography (PET)]. On multivariable Cox proportional hazards regression, the SNAPP factors were independently associated with HCC recurrence. The SNAPP score was highly predictive of HCC recurrence (C statistic, 0.802), and 5-year post-LT recurrence rates were significantly different between low, intermediate, and high SNAPP score groups. The performance of the SNAPP score (C-index [95%CI], 0.903 [0.820-0.987]) on predicting tumor recurrence after LDLT was better than that of the NYCA, the RETREAT, and the MoRAL score. The SNAPP score provides excellent prognostication after LDLT for HCC patients. Hence, we can help voluntary patients' decisions about whether to undergo LDLT or not.