Citations

4091 Background: Optimal management of ECCA patients with elevated CA19-9 remains undefined. We hypothesized CA 19-9 elevation as a marker of aggressive biology in ECCA and that inclusion of CA 19-9 in the staging may improve OS discrimination. Methods: Patients with ECCA with CA 19-9 levels reported to the National Cancer Database (years 2004-2015) were included. The patients were classified based on their CA19-9 levels and a new staging system was proposed. Based on the current knowledge, we considered 37 U/ml as our cut off. Kaplan Meier method was used to compare OS between the groups. The net reclassification improvement (NRI) model was used to assess the predictive improvement in the proposed survival model. Results: A total of 2100 patients met the inclusion criteria: 601 (32%) and 1436 (68%) had normal and elevated CA19-9 levels, respectively. Rates of chemo (p=0.16) and radiation therapy (p=0.07) were similar between groups, but patients with elevated CA19-9 were less likely to undergo resection. Resected patients with CA19-9 elevation had higher 30-day mortality (p=0.02) and lower median OS (p<0.01). Patients with elevated CA 19-9 levels had decreased stage-specific survival in all stages (p<0.01). On adjusted analysis, CA19-9 elevation independently predicted poor OS (HR: 1.67 [1.42-1.97]) with impact resembling nodal metastasis, positive resection margin, lymphovascular invasion, and non-receipt of surgery or chemotherapy (p<0.01). CA 19-9 included proposed staging system (Table) had a better OS discrimination over AJCC 7th edition. The new staging system had a concordance of 60% as opposed to 58% for the AJCC staging, leading to an improvement of 2% (p<0.01). NRI of 46% (95% CI: 39-57) indicates that the new staging system is substantially effective at re-classifying events at 12 months as compared to AJCC staging. Conclusions: Elevated CA19-9 was found to be an independent risk factor for mortality in ECCA and its inclusion in the newly proposed staging system markedly improved OS discrimination.[Table: see text]

3101 Background: PDO is a promising translational tool that recapitulates the biology and drug response of donor cancer patient. However, an unmet need is to have PDO drug-screening data available for treatment decision making in clinic. We conducted a pilot study to determine whether PDO testing results will be available at critical treatment decision points in metastatic GI cancer patients. Methods: Metastatic GI cancer patients undergoing core-needle biopsy were eligible. Tumor cells isolated from ≤4 fresh biopsy tissues were grown in a Matrigel-based culture. PDO response to anti-cancer drugs were evaluated; and when available, correlated with donors’ clinical response to the same agent(s). PDO response was defined as IC50 < 0.1 × published Cmax of the drug clinically; stable as IC50 between 0.1 to 10 × Cmax. Radiographic response was per RECIST criteria. Results: We enrolled 27 refractory metastatic GI cancer patients (9 colorectal [CRC], 9 pancreas, and 9 biliary tract). Median lines of therapy were 4, 2, and 2; the success rate of organoid establishment was 89%, 44%, and 55%, respectively. The median time from biopsy to availability of drug-testing data was 64 days (range: 24 to 93 days). The median time from biopsy to next CT re-staging in donors was 64 days. The established PDOs shared histological and genomic features with donor clinical tissue. PDO and clinical responses to the same agent(s) were correlated in 2 CRC donors including (1) BRAFV600E-mutated PDO responded to vemurafenib + panitumumab, as did the donor who had partial response (PDO drug-testing data were available 55 days post-biopsy, 23 days prior to restaging scan); (2) KRAS/FGF-dual amplified PDO had stable disease status to regorafenib, as did the biopsied lesion from the donor (73 days post-biopsy, 5 days post-scan). Conclusions: We showed the feasibility of completing PDO drug sensitivity testing in metastatic GI cancer patients within a short time that could impact clinical decision making, particularly in CRC. PDO drug response showed correlation with clinical response. With further refinement, PDO can be a powerful tool for personalizing cancer therapy in metastatic GI cancer patients.

PURPOSE Recent advances in molecular diagnostic technologies have allowed for the evaluation of solid tumor malignancies via noninvasive blood sampling, including circulating tumor DNA (ctDNA) profiling. We sought to characterize the ctDNA genomic alteration landscape in patients with biliary tract cancers (BTCs). PATIENTS AND METHODS From January 2015 to February 2018, 124 patients with BTC at the Mayo Clinic Comprehensive Cancer Center underwent ctDNA testing using a clinically available assay. The majority of samples (n = 122) were tested using the 73-gene panel that includes somatic genomic targets, including complete or critical exon coverage in 30 and 40 genes, respectively, and in some, amplifications, fusions, and indels. RESULTS A total of 138 samples were included, with approximately 70% of patients having intrahepatic BTC. All patients had locally advanced or metastatic BTC. Samples with one or more alterations, when variants of unknown significance were excluded, numbered 105 (76%). Each sample contained, on average, three alterations with a median allelic fraction of 0.52%. The overall landscape of alterations is summarized in Figures 1 and 2. After excluding variants of unknown significance, therapeutically relevant alterations were observed in 76 patients (55%), including BRAF mutations, ERBB2 amplifications, FGFR2 fusions, FGFR2 mutations, and IDH1 mutations seen in 21% of patients. A different spectrum of alterations was observed in patients with early-onset BTC (younger than age 50 years) compared with older patients (older than age 50 years). CONCLUSION Data on ctDNA in BTC is currently limited. Our study, the largest cohort reported to date to our knowledge, demonstrates the feasibility of ctDNA testing in this disease. We provide a foundation upon which the field can continue to grow.

672 Background: Rectal squamous cell carcinoma (RSCC) is a rare form of gastrointestinal malignancy. Using the NCDB, we determined the prognostic factors and survival outcomes of RSCC in the United States. Methods: We identified histologically confirmed cases of RSCC from the National Cancer Data Base (2004-2015). Univariate and multivariable methods were used to assess factors associated with survival. Kaplan-Meier method and log-rank test were used to perform overall survival (OS) analysis. Results: Of the 5,527 cases included in our analysis, 67% were female. Median age at diagnosis was 61 years and did not differ by sex. The proportion of patients with stage 1, 2, 3 and 4 diseases were 22%, 16%, 20%, and 11%, respectively (30% unknown stage). Among the ones who received surgical resection of primary tumor, 41%, 30%, 20% and 8% are of stages 1, 2, 3 and 4 respectively. The rate of R0 resection was 54%, 63%, 55% and 38% in stage 1, 2, 3 and 4, respectively. The R0 resection rate was much higher in patients who received neoadjuvant chemo or radiation therapy or both (87%, 78%, 74%, and 57% in stages I, 2, 3 and 4, respectively) as compared to that of their counterparts. On stage wise sub-group OS analysis, stage 1-3 patients had OS benefit from surgery (as compared to no-surgery) (145 vs 90 months, p<0.001) as opposed to 4 disease (16 vs 11 months, p=0.06). Adjuvant radiation therapy improved the median OS in stage 1-3 patients (as compared to no-adjuvant radiation) in patients with positive surgical margins (not reached vs 40 months, p<0.0001). Patients with stage 4 disease treated with radiation therapy had a better median OS than those without (16 vs 6 months, p<0.0001). On therapy wise sub-group analysis, the patients who received surgery only had a median OS of 145 months; surgery + chemoradiation (adjuvant and neoadjuvant) = not reached; whereas patients with chemoradiation only had median OS of 80 months. Conclusions: This is the largest registry-based study on RSCC to date. RSCC had a diverse OS varied significantly according to stage of the disease at presentation and therapy received. Surgical resection of primary tumor was associated with improved OS as compared to that of patients who received chemoradiation.

442 Background: Neutrophil to lymphocyte ratio (NLR) has been used as an inflammation-based prognostic marker for various malignancies. The aim of our study was to determine whether NLR can independently predict the overall survival in patients with metastatic gallbladder cancer (GBC). Methods: We identified patients diagnosed with GBC who were treated at Mayo Clinic between the years 2000 and 2016. Patients who had nonmetastatic GBC were excluded along with the patients who did not have data for neutrophils and lymphocytes. Optimal cutoff point for NLR was identified by plotting martingale residuals against NLR and patients were divided into two groups, ≥ 5 or < 5. Demographic, follow-up data and outcomes were collected by retrospective review of electronic medical records. Fisher’s exact test was used to compare categorical variables, while The Mann- Whitney U test was used to compare continuous variables. Kaplan-Meier curves were plotted for NLR ≥ 5 and NLR < 5 and overall survival (OS) between the two groups were compared using log rank test. Multivariate survival analysis was performed using Cox-proportional hazard regression. Results: A total of 231 patients met our inclusion criteria, of which, 138 (60%) had NLR < 5 and 93 (40%) had NLR ≥ 5. Patients with NLR ≥ 5 were more likely to be older and have poor performance score, lower albumin level, higher alkaline phosphatase level, and higher platelet count. There were no significant differences noted in gender, race and administration of chemotherapy between the two groups. In univariate analysis, patients with NLR ≥ 5 at presentation had a significantly worse OS compared to those with NLR < 5 (Median survival: 3.6 vs. 8.8 months, p < 0.001). In multivariate analysis, adjusting for age, ECOG status, albumin, ALP, AST, ALT, bilirubin, platelet count and administration of chemotherapy, NLR of ≥ 5 was associated with a worse OS compared to NLR < 5 (HR: 1.70, 95% CI:1.20 – 2.39. p < 0.05). Conclusions: Our study demonstrates that NLR ≥ 5 is an independent predictor of poor prognosis in patients with metastatic gallbladder cancer.

433 Background: Current staging systems for gallbladder cancer (GBC) are inadequate, as they are based on surgical pathology, and therefore are not relevant for unresectable patients and patients undergoing neoadjuvant chemotherapy. Methods: Patients with a confirmed diagnosis of GBC who were seen at Mayo Clinic between the years 2000 and 2016 were included in this analysis. Data on demographic and tumor characteristics and outcomes were collected by retrospective review of electronic medical records. A model predictive of overall survival was developed using Cox proportional hazard regression analysis. Harrel’s C-statistic was calculated to evaluate the predictive accuracy of the model and compared with the TNM staging system. Results: A total of 523 patients were included in the final analysis, with a median age of diagnosis of 68 years. The median duration of follow up of the entire cohort was 12 months. In multivariate analysis, factors predictive of poorer overall survival were: ages 65-74 years (HR : 1.80, 95% CI: 1.33–2.43) and ages 75+ years (HR: 2.93, 95% CI: 2.12–4.06) compared to age <55 years; tumor size ≥ 5 cm by imaging (HR: 1.24, 95% CI: 1.01–1.55); nodal involvement by imaging (HR:1.61, 95% CI: 1.23–2.10); involvement of distant organs by imaging (HR: 2.85, 95% CI: 2.16–3.75); ECOG performance score of 2 or higher (HR: 1.78, 95% CI: 1.36–2.32) compared to ECOG 0-1; albumin level <3.5 g/dL (HR: 1.40, 95% CI: 1.08–1.81); and alkaline Phosphatase level ≥ 200 IU/L (HR: 1.49, 95% CI: 1.21–1.84). Using these seven predictive factors of survival we created a four-tier staging system. The median survivals of Stages I, II, III and IV created in our novel system were 64, 34, 20 and 7 months with corresponding hazard ratios of 1, 1.5, 2.5 and 8.5 respectively. The C-statistic for this novel staging system was 0.68 compared to C-statistic of 0.69 for the TNM staging system, indicating similar performance in predicting survival. Conclusions: We have created a novel clinically-based staging system for patients with GBC based on nonoperative information at the time of diagnosis which performs on par with the current surgical pathology based TNM staging system.

426 Background: Small bowel adenocarcinoma is a rare malignancy affecting approximately 2,000 patients per year. There is a paucity of evidence prognosticating patients with small bowel adenocarcinoma. We aimed to evaluate multiple factors in patients with resected small bowel adenocarcinoma to determine any association with survival outcomes. Methods: Ninety three patients who underwent resection for stage I-III small bowel adenocarcinoma were retrospectively identified utilizing the pathology database at a single tertiary referral institution. All patients had complete follow up data and were included in the survival analysis. JMP software was used for statistical analysis. Overall survival was performed utilizing Kaplan-Meier method, and log-rank tests were used for statistical comparisons. Cox proportional hazards were performed to control for age, gender, location of tumor, tumor size, tumor stage, and adjuvant therapy. Sensitivity analysis was performed to establish best cutoff points for continuous variables. All tests were two sided and a P value of < 0.05 was considered significant. Results: The median age at diagnosis was 65 years (range 32-90). 61% were male. Median tumor size was 4.5 cm. There were 20, 36, and 37 patients with stage I, stage II, and stage III disease, respectively. Median overall survival (OS) was 151 months, 104 months, and 44 months for stages I, II, and III disease. In a multivariate analysis, independent predictor factors included presurgical lymphocyte to monocyte ratio (LMR) > 4.0, with a Hazard Ratio (HR) 0.13 (95% CI 0.007-0.69, p = 0.01), presurgical neutrophil to lymphocyte ratio (NLR) < 8.0, HR 0.39 (95% CI 0.17-0.96, p = 0.04), and tumor size < 7.5 cm, HR 0.22 (95% CI 0.07-0.85, p = 0.03). Stage, age, T stage, and N stage influenced overall survival in univariate analysis, but were not statistically significant on multivariate analysis. Conclusions: LMR and NLR independently predict survival in patients with resected small bowel adenocarcinoma.

399 Background: PSCC is a rare form of exocrine pancreatic malignancy with a dismal prognosis. Using the NCDB, we determined the prognostic factors and survival outcomes of PSCC in the United States. Methods: We performed a retrospective analysis of patients with histologically confirmed PSCC from 2004-2015 using NCDB. Kaplan-Meier method and log-rank test were used to perform overall survival (OS) analysis. Hazard Ratios were calculated using the Cox-proportional hazard method. Results: Of the 654 cases included in our analysis, 46% were female. Median age at diagnosis was 70 years and did not differ by sex (p = 0.19). The proportion of patients with stage I, II, III and IV diseases were 5%, 18%, 12%, and 54%, respectively (10%, unknown stage). Among these, 23% (35 of 150) of stage I and II disease, 10% (8 of 78) of stage III, and 2% (7 of 353) received surgical resection of the primary tumor. The rate of R0 resection was 74% in stage I and II; 38% in stage III; 29% in stage IV disease. Median OS for the entire cohort was 4 months and was significantly higher in patients who received surgical resection of the primary tumor (17 vs. 4 months, p < 0.001). On stage wise sub-group analysis, stage I-II patients had OS benefit from surgery (21 vs. 5 months, p < 0.001) as opposed to stage III (7 vs. 6 months, p = 0.31) and IV disease (5 vs. 3 months, p = 0.17). Adjuvant chemotherapy had no role in prolonging survival in stage I-II disease (20 vs 24 months, p = 0.6). Stage IV patients treated with chemotherapy had a better median OS than those without (5 vs. 2 months, p < 0.0001). On Cox multivariable analysis, stage IV disease (HR: 1.92 CI: 1.46-2.52, p < 0.001) and advanced patient age (HR: 1.02; CI:1.01-1.03, p < 0.001 were associated with poor OS, whereas OS was not dependent on the sex, race, grade, insurance status, surgery, and chemotherapy. Conclusions: This is the largest registry-based study on PSCC to date. PSCC had a diverse OS varied significantly according to increasing age and stage of the disease at presentation. Surgical resection of primary tumor was associated with improved OS in stages I-II, whereas chemotherapy improved OS in stage IV disease. The results of our study may aid the prognostication of patients and in treatment decision making.

393 Background: Clinicopathological features and the outcomes of patients with fibrolamellar hepatocellular carcinoma (FLHCC) are not clearly defined. Methods: Data were collected by retrospective chart review on 42 patients with FLHCC treated between 1990 and 2017 at the Mayo Clinic. Clinicopathological characteristics, response to treatment, recurrence pattern and survival were analyzed. Results: Of 42 patients (17 males and 25 females; median age at diagnosis 22 years, range 15 to 39), 10 patients (23%) had stage I disease and 32 patients (77%) had stage II to IVB disease. All 10 patients with stage I disease and 21 of 32 patients with stage II-IVB disease underwent resection at presentation. In stage I patient group, 6 patients experienced recurrence with a median time to recurrence of 30.5 months, resulting in a 5 year overall survival (OS) of 86%. Patients with stage II to IVB disease who underwent resection (n=21) at presentation had a median OS of 32.5 months and 5 year OS of 44%. In the upfront surgery group, 71% of the patients experienced recurrent disease. The median OS of patients with unresectable disease (n=11) was 10 months. Systemic therapy was given to 17 patients which included sorafenib, FOLFOX (5-fluorouracil/leucovorin and oxaliplatin), gemcitabine plus oxaliplatin, single agent adriamycin or gemcitabine, capecitabine plus interferon alfa, gemcitabine plus cisplatin, cisplatin plus adriamycin and nivolumab . Sorafenib was given to 9 patients and 4 patients achieved stable disease (SD) with duration ranging from 5 months to 5 years. One programmed cell death receptor positive-1 patient had a durable complete response after 2 months of therapy with nivolumab. Conclusions: In FLHCC, surgical resection was associated with prolonged OS. However, recurrences were common after the surgery. Limited benefit was derived from the systemic treatment. In rare cases, therapy with a checkpoint inhibitor may provide a viable treatment option.

291 Background: Cholangiocarcinoma (CCA) has limited treatment options. Genomic analyses have led to development of targeted therapies now in clinical trials, and may enable discovery of new treatment options. However, biopsy often yields limited tissue, thus hampering tissue-based profiling opportunities. Comparative data regarding circulating tumor DNA (ctDNA) analysis and tissue based profiling in CCA are limited. Methods: We performed ctDNA NGS analysis along with tissue based profiling in pts with advanced CCA (January 2015- February 2018). ctDNA analysis was performed using Guardant 360 (Guardant Health, CA) which detects single nucleotide variants, amplifications, fusions, and specific insertion/deletion mutations in up to 73 different genes and the majority of tissue based profiling using Foundation One. The mutant allele fraction (MAF) for detected alterations was calculated relative to wild type in ctDNA. Therapeutic relevance was defined as alterations within OncoKB levels 1-3B and R1. The study was conducted in accordance with Mayo Clinic IRB requirements. Results: Among 124 pts and 139 total samples, ctDNA NGS revealed at least one genomic alteration (excluding variants of uncertain significance and synonymous mutations) in 89% of pts. Median number of alterations per pt was 3 [range, 1-15], with a median MAF of 0.42% (range, 0.1% - 94.2%). The total number of unique alterations was 321. The most commonly altered genes: TP53 (31%), KRAS (11%), FGFR2 (7%), APC and PIK3CA (each 5%) and ARID1A (3%). Amplifications were noted in 14 genes: BRAF, CCND1, CCND2, CCNE1, CDK4, CDK6, EGFR, ERBB2, FGFR1, FGFR2, MET, MYC, PDGFRA, and PIK3CA. Tissue-based profiling was available in 57 (46%) pts, with a median of 63 days between liquid and tissue biopsy. IDH1, FGFR2, TP53 and KRAS were most common gene mutations found in pts who had both liquid and tissue biopsy done. (Comparative results to be shown). Conclusions: ctDNA plasma profiling of pts with advanced CCA is a feasible alternative method to gather comprehensive genomic data. Further larger cohort studies comparing landscape of alterations seen on ctDNA versus tissue-based assays are needed.