Randomized phase II study of platinum and etoposide (EP) versus temozolomide and capecitabine (CAPTEM) in patients (pts) with advanced G3 non-small cell gastroenteropancreatic neuroendocrine neoplasms (GEPNENs): ECOG-ACRIN EA2142

4020

Background: High grade (G3) GEPNENs are a rare and heterogeneous disease entity for which there is little prospective treatment data. EP chemotherapy is the treatment standard but this may not be appropriate for all G3 GEPNEN pts. CAPTEM has demonstrated activity in G3 GEPNENs and may be a promising alternative. EA2142 aimed to determine if CAPTEM was superior to EP in pts with G3 GEPNENs. Methods: This was a multicenter, randomized (1:1) phase II trial for pts with a locally advanced and unresectable or metastatic well differentiated G3 neuroendocrine tumor (NET) or a poorly differentiated, non-small cell G3 neuroendocrine carcinoma (NEC) of suspected gastrointestinal origin and an ECOG PS of 0-2. Pathology must have demonstrated a Ki-67 of 20-100% or at least 10 mitoses/10 high powered field. Pts were randomized to receive capecitabine 750 mg/m2 orally every 12 hours on days 1-14 and temozolomide 200 mg/m2 orally once daily on days 10-14 of a 28-day treatment cycle (Arm A) or etoposide 100 mg/m2 daily on days 1-3 with either cisplatin 25 mg/m2 daily on days 1-3 or carboplatin AUC 5 on day 1 of a 21-day treatment cycle (Arm B). Restaging scans were performed every 8 weeks and toxicity monitored per CTCAEv4. Final statistical plan was to accrue 80 pts to detect a 67% improvement in progression free survival (PFS) (primary endpoint) with CAPTEM as compared to EP, 80% power and one-sided significance level of 0.10. A planned interim analysis for efficacy and futility was conducted. Results: A total of 67 pts were enrolled (Arm A, n=32; Arm B, n=35). Male 58%, African American 4%, Asian 3%. Mean age 61. Among 63 eligible pts, primary tumor site pancreatic 56%, non-pancreatic 43%. Poorly differentiated 57%, well differentiated 33%, unknown 10%. Mean Ki-67 48% (Arm A), 60% (Arm B). The study was closed prior to full accrual due to futility at 57.7% information time. In the interim analysis, among 62 eligible pts, PFS, overall survival and response rate with CAPTEM were 2.43 months (mos) (95% CI 2.04, 7.72), 12.6 mos, 9% respectively vs 5.36 mos (95% CI 2.14, 7.23), 13.6 mos and 10% with EP. Toxicity was evaluable in 57 pts with Grade (G) 3/4 events occurring in 29% of pts on Arm A, 66% of pts on Arm B. G3/4 events occurring in more than 5% of pts on Arm A—febrile neutropenia (n=2); abdominal pain (n=2); diarrhea (n=2); nausea (n=2); neutropenia (n=2); dehydration (n=2) and on Arm B—anemia (n=8); febrile neutropenia (n=2); fatigue (n=2); lymphopenia (n=2); neutropenia (n=12); thrombocytopenia (n=4); leukopenia (n=6). There was one G5 event on Arm A due to sepsis. Conclusions: CAPTEM does not appear to be superior to EP chemotherapy as front-line treatment for pts with G3 NENs but does demonstrate a more favorable toxicity profile. Studies assessing G3 NET independently of G3 NEC are needed. Clinical trial information: NCT02595424.

Telotristat ethyl with PRRT in the treatment of well differentiated neuroendocrine tumors

e16205

Background: Feasibility of telotristat ethyl, a tryptophan hydroxylase inhibitor in serotonin synthesis, in combination with PRRT/lutetium Lu 177 DOTATATE is unclear in well differentiated neuroendocrine tumors (WDNET). We explore real world treatment patterns in patients receiving telotristat and PRRT. Methods: A retrospective study was conducted on 67 patients with a histopathologic diagnosis of WDNET, treated with telotristat and/or PRRT at the Winship Cancer Institute in Emory University from 2018 to 2022. OS and PFS were assessed by Kaplan Meier curves. Descriptive analysis was performed with associations characterized by multi- (MVA) and univariate (UVA) Cox proportional hazards model, t-test, and ANOVA. Results: Of the 67 patients, 44 received PRRT alone and 23 received PRRT with telotristat for carcinoid symptoms. Median age was 69 (range 18-80) with male 57% and white race 65%. 16 (26%), 41 (67%), 4 (7%) had grades 1 (Ki67 < 3%), 2 (3-20%), and 3 (> 20%) disease respectively. The plurality (n = 31, 46%) had primary small bowel (SB) NETs and 40 (65%) underwent primary NET resection. 2 had gastrinomas, 1 VIPoma, and 1 insulinoma. All had liver and 23 (40%) had bone metastases. Most patients received PRRT as 3rd line therapy. 59 (88%) in the entire cohort and 17 (74%) in the telotristat group received 4 doses of PRRT. 15 (17%) in the entire cohort had delayed/dose reduced PRRT. Telotristat was associated with improved diarrhea and decreased serotonin (all p < 0.003). 16 (70%) on telotristat reported improvement in diarrhea and flushing. 5 (17%) had grade 3/4 adverse events while receiving telotristat. Of those, 1 developed hypotension, 1 hypokalemia, 1 myelodysplastic syndrome, 1 febrile neutropenia, and 1 anemia and pursued hospice care after a cerebrovascular event. Median(m) OS, PFS, and follow up were 216, 33, 67 months (mo) for the entire cohort respectively. 6 (26%) died in the telotristat/PRRT group vs 17 (39%) PRRT only. 10 (43%) in the telotristat group progressed vs 18 (41%) in PRRT only. mPFS was 33 mo in telotristat/PRRT and not reached (NR) in PRRT only (p = 0.684). mOS was NR in the telotristat/PRRT group vs 216 mo in PRRT only (p = 0.717). In MVA, OS was associated with primary NET resection (HR 0.232), age > 60 (HR 7.88), SB primary NET (HR 0.097), and PRRT completion (HR 0.062, all p < 0.05). In MVA, PFS was associated with diarrhea (HR 3.7) and small bowel primary NET (HR 0.264, all p < 0.05). In UVA, lower OS and PFS were associated with elevated chromogranin A (CGA) (upper limit 225, HR 3.34, p < 0.02). Initial neutrophil-eosinophil ratio (NER) was associated with OS (HR 21.2) and PFS (HR 10.8) in MVA with optimal cutoff of 35.28 ( both p < 0.02). Conclusions: Patients with primary SB WDNETs and carcinoid symptoms benefited from receiving telotristat/PRRT in our study. CGA, serotonin, and NER are potential biomarkers to assess disease patterns. Multi-institutional studies and further follow up are needed to confirm these findings.

Impact of metformin on clinical outcomes in advanced hepatocellular carcinoma treated with immune checkpoint inhibitors

4118

Background: Non-alcoholic steatohepatitis (NASH) is an emerging etiology for hepatocellular carcinoma (HCC) and contributes to the increasing incidence of HCC worldwide. Patients with NASH often have risk factors of metabolic syndrome including hypertension, obesity, and type 2 diabetes (T2DM). NASH induced HCC has been shown to be associated with less response to immune check point inhibitors (ICIs) in HCC. Anti-diabetic agent metformin has been shown to be associated with improved outcomes in patients treated with ICIs in melanoma and non-small cell lung cancer. However, the impact of metformin on the efficacy of ICIs is not well defined in HCC. The main purpose of this study was to examine the effect of metformin on clinical outcomes in patients with advanced HCC treated with ICIs. Methods: We performed a retrospective analysis of patients with advanced HCC treated with ICIs in first and later-line settings between 2015 and 2021. The primary endpoints were overall survival (OS), progression free survival (PFS), and objective response rate (ORR) as assessed according to Response Evaluation Criteria in Solid Tumors (RECIST) version 1.1. Patients were stratified based on their usage of metformin. OS and PFS were analyzed using Cox proportional hazard models and Kaplan-Meier analysis with log-rank test. Results: A total of 111 patients met inclusion criteria, 18 patients in the metformin group and 93 patients in the non-metformin group. Most common cause of HCC was viral hepatitis (52%), followed by NASH (29%), alcohol (8%) and other (11%). Baseline characteristics between the two groups were similar except all patients in the metformin group had a diagnosis of T2DM. ORR was 5.6%. (1 partial response) in the metformin group vs 22.6% (5 complete responses, 16 partial responses) in the non-metformin group. Median OS was 45.9 months in the non-metformin group vs 10.8 months in metformin group (HR 1.99, 95% CI 0.95-4.21, p = 0.064). Median PFS of 6.6 months vs 2.5 months (HR 1.75, 95% CI 0.93-3.29, p = 0.077). Moreover, metformin usage was associated with shorter median OS of 10.8 months (HR 1.96, 95% CI 0.75-5.09, p = 0.16) vs 20.9 months among patients with T2DM. OS was significantly worse in patients with poor ECOG performance status 2-3, MELD score 10-23, higher grade tumor histology, AFP > = 400, and use of IO in later lines of therapy. Conclusions: In this retrospective study metformin use was associated with worse clinical outcomes in advanced HCC patients treated with ICIs.

Characteristics and outcomes of patients with multiple synchronous colon cancer primaries

194

Background: Patients (pts) with multiple synchronous colon cancer primaries (MCPs) constitute a unique subset of pts with colon cancer. However, there are limited published studies about these pts. The objective of this study is to compare the characteristics and outcomes of pts with MCPs to those with single colon cancer primaries (SCPs) using the largest study population to date. Methods: Data was obtained from the Surveillance, Epidemiology, and End Results (SEER) database between 2004 and 2015. Pts with synchronous MCPs were included and were matched 1:3 with pts with SCPs based on the Coarsened Exact Matching method for age, gender, and race. Only patients with multiple synchronous primaries were included (time since index = 0 months). We excluded pts with a lag time since diagnosis of index primary of 1 month or more. Univariate (UNA) and multivariable (MVA) analyses were performed to identify factors associated with patient outcomes. Kaplan-Meier analyses and Cox proportional hazards models were used to assess the association between tumor/patient characteristics and overall survival (OS). Results: A total of 3322 pts with MCPs and 9966 pts with SCPs were identified. Median age was 71 years. Majority were male (51.5%) and White (80.1%). 73.4% and 69.6% of pts had 12 or more lymph nodes examined for the MCPs and SCPs cohorts, respectively. The SCPs cohort included more T4 stage and more well- and moderately-differentiated histology. OS was significantly shorter in MCPs compared to SCPs (HR 1.29; 1.22-1.36; p < 0.001), with a 5- and 10-year OS rate of 47.8% and 28.2% for the MCPs and 56.4% and 41.6% for the SCPs, respectively, for all stages combined. In the MCPs cohort, the use of adjuvant chemotherapy was associated with an improved survival in AJCC stages II, III, and IV but not stage I. Conclusions: This is the largest study evaluating the impact of MCPs on outcomes. Across stages II to IV, pts with MCPs have a shorter survival than those with SCPs. Pts with stage II MCPs who receive adjuvant chemotherapy derive a survival benefit. Current guidelines do not list multiple synchronous primaries as a high-risk feature for stage II.

Impact of local therapy on survival among patients with metastatic anal squamous cell carcinoma

4

Background: About 10-20% of patients with anal squamous cell carcinoma (SCCa) present with metastatic disease, and are usually treated with systemic chemotherapy. The role of local therapy to control the primary tumor is controversial in this setting. We evaluated survival impact of local therapy in metastatic anal SCCa. Methods: Data were obtained from all US hospitals that contributed to the National Cancer Database (NCDB) between 2004 and 2015. We excluded patients who did not receive palliative systemic chemotherapy. Univariate (UVA) and multivariable analyses (MVA) were performed to identify factors associated with patient outcome. Kaplan-Meier analysis and Cox proportional hazards models were used to assess the association between tumor/patient characteristics and overall survival (OS). Results: 1,160 patients were identified over 12 years. Median age was 57 years. Majority were female (64.9%), non-Hispanic Whites (79.1%) and had Charlson-Deyo Score of 0 (83.6%). Most common metastatic sites were liver (25.9%), lung (11.6%) and bone (8.5%). More than 79% of the patients received radiation to the primary site, and 10.4% underwent surgical resection for local control. Use of local therapy correlated closely with a significant improvement in OS on MVA (HR 0.66; 0.55-0.79; p < 0.001), with a 12-month and 5-year OS rates of 72.8% and 25.7% respectively, compared with 61.1% and 14.6% for patients treated with chemotherapy only. Poor prognostic factors included male gender (HR 1.44; 1.24-1.67; p < 0.001), age > 70 years (HR 1.28; 1.02-1.62; p = 0.034), lack of health insurance (HR 1.32; 1.02-1.71; p = 0.034), and cloacogenic zone location (HR 4.02; 1.43-11.30; p = 0.008). There was no benefit from abdominoperineal resection (mOS = 19.7mos; HR 1.05; 0.48-2.29; p = 0.909), but both local resection of the primary (mOS = 24.8mos, HR 0.48; 0.29-0.80; p = 0.005) and palliative radiation (mOS = 22.6 mos; HR 0.66; 0.55-0.79; p < 0.001) were associated with improved OS. Conclusions: This is the largest reported study on management of de novo stage IV SCCa. The data suggest that local control of the primary tumor through resection or radiation improved OS in patients with anal SCCa. Patients unlikely to benefit from local therapy include age over 70 years, male, lack of health insurance and cloacogenic carcinoma.

Phase Ib/II trial of siltuximab and spartalizumab in patients in metastatic pancreatic cancer

TPS626

Background: Interleukin-6 (IL-6) is associated with carcinogenesis, immune suppression, and poor prognosis in pancreatic adenocarcinoma (PDAC). Preclinical data demonstrated dual inhibition of IL-6 and (programmed death ligand-1) PD-L1 facilitates CD8+ T cell migration into pancreatic tumors and was effective in controlling tumor growth in syngeneic and genetically engineered PDAC mouse models. Siltuximab is a chimeric monoclonal antibody which targets the IL-6 molecule specifically and spartalizumab is a high-affinity ligand-blocking humanized IgG4 antibody against the PD-1 receptor. Based on this preclinical rationale, we developed a phase Ib/II trial to determine the recommended phase II dose (RP2D), evaluate the safety, toxicity profile, preliminary antitumor activity, and immunogenicity of the siltuximab and spartalizumab in patients with previously treated metastatic PDAC. Methods: The phase Ib trial design is standard 3+3. Primary endpoint is to determine RP2D. Siltuximab is administered intravenously (IV) in three dose levels of 6 mg/kg (DL1), 11 mg/kg (DL2), 9 mg/kg (only if 2 DLTs observed on DL2) every 3 weeks with spartalizumab at 300 mg IV every 3 weeks. Eligible patients must have stage IV PDAC who have failed at least one prior therapy age ≥18 years, ECOG PS 0-1, no prior anti PD-1 or anti-PD-L1 agent. After RP2D is established, an expansion phase will enroll 24 patients with PDAC. Pre and on-treatment biopsy will be performed in 24 patients in the expansion cohort for correlative analysis. Pre-treatment and on-treatment peripheral blood samples will be collected from all patients. In the expansion phase patients will receive initial cycle (every 3 weeks) treatment with either spartalizumab or spartalizumab plus siltuximab and then starting cycle 2 all patients receive the combination following the on-treatment research biopsy. This design will enable us to evaluate the immunological effects of spartalizumab alone versus the combination in the tumor microenvironment and peripheral blood. This study was activated in January 2020 and to date 12 patients were enrolled in dose escalation phase. The dose expansion phase has recently started accrual. Clinical trial information: NCT04191421.

A phase I study of pharmacokinetic (PK)-driven sequential dosing of rucaparib (RUB) with irinotecan liposome (nal-IRI) and fluorouracil (5FU) in metastatic gastrointestinal (mGI) and pancreas (PANC) cancers

563

Background: RUB is an oral PARP1,2,3 inhibitor that demonstrated efficacy in patients (pts) with ovarian and prostate cancers harboring deleterious BRCA mutations. RUB exerts synergistic anti-tumor effect with IRI preclinically though the combination has overlapping toxicities. We previously published on the population PK of nal-IRI (Adiwijaya, Ma et al, Clin Pharm Ther 2017). We conducted a phase I study to evaluate a novel sequential dosing of RUB with nal-IRI/5FU in mGI cancer pts. Methods: Eligible pts had incurable mGI cancer previously received > 1 line of therapy (rx), ECOG PS 0-1, had RECIST measurable disease, adequate organ reserves and not received IRI for metastatic disease. Previous PARPi rx was excluded. The endpoints included dose limiting toxicity (DLT), maximum tolerated dose (MTD) and toxicity profile. The dose escalation utilized the 3+3 design. RUB was given oral bid on Day 4 to 13 and 18 to 27 with nal-IRI i.v. and 5FU i.v. 2400 mg/m2 over 46 hr on Day 1 and 15, every 28 day. Planned dose levels were RUB 400 mg/nal-IRI 50 mg/m2 (DL1), 400 mg/70 mg/m2 (DL2) and 600 mg/70 mg/m2 (DL3). Adverse events (AEs) were scored per CTCAE v4.03. Molecular profile was evaluated by CLIA-certified NGS testing. Results: Eighteen pts including 11 colorectal (CRC), 6 PANC, 1 gastroesophageal (GE) were enrolled and 12 were evaluable for DLTs. DL2 was not tolerable (DLT: G3 diarrhea, nausea and vomiting) and DL2A was added (RUB 600 mg/nal-IRI 50 mg/m2). DL2A enrolled 6 pts with no DLT and was determined as the MTD. Of DLT-evaluable pts, G3 and worse treatment-related AEs from all cycles were diarrhea (33%), fatigue (25%), leukopenia (25%), neutropenia (25%), anemia (8%) and nausea (8%). Four of 12 response evaluable pts had partial response: 2 CRC (1 had ATM mut), 1 PANC ( ATM mut), 1 GE ( BRCA2 mut) whilst 3 responders previously had platinum (PLA). Five pts had stable disease beyond 16 weeks (range 18.9 to 100.7 weeks), and all had prior PLA. Conclusions: The study successfully determined the MTD of RUB in combination with nal-IRI and 5FU. Encouraging efficacy was observed in PLA-treated mGI cancers including responses in those harboring ATM and BRCA alterations. The study is proceeding to evaluate the efficacy of the combination in metastatic pancreas cancer pts with and without BRCA1/2 or PALB2 alterations. Clinical trial information: NCT03337087.

Clinical features and outcomes of colloid carcinoma of pancreas compared to pancreatic ductal adenocarcinoma

e16259

Background: Colloid carcinoma (CC) of the pancreas is a rare histopathological subtype of ductal adenocarcinoma (PDAC), with poorly defined prognostic factors and therapeutic outcomes. The aim of this study is to characterize the clinicopathological features and evaluate the overall survival (OS) and prognostic factors of patients with pancreatic CC using National Cancer Database (NCDB). Methods: Patients diagnosed with CC of the pancreas and PDAC between 2004 and 2016 were identified from the NCDB using ICD-O-3 morphology (8480/3 for CC and 8140/3 for PDAC) and topography codes (C25). Univariate and multivariable analyses were conducted and Kaplan-Meier analysis and Cox proportional hazards models were used to perform OS analysis. Results: A total of 56,846 patients met the inclusion criteria for the final analysis. Of the total population included, 2,430 patients (4.3%) had CC and 54,416 patients (95.7%) had PDAC. For both, CC and PDAC, there was a male preponderance (52.0%, 52.5%), Caucasians (85.1%, 84%), occurrence above the age of 70 (39.2%, 38.2%), and the most common primary site was the head of the pancreas (50.5%, 53%). For CC, the percentage of pathologic stage III colloid pancreas cancer appeared the lowest (3.5%, 85 patients), compared to stage I (16.7%), stage II (37.8%), and stage IV (42.1%). While in PDAC, the percentage of pathologic stage I (5.94%) and stage III (4.44%) patients was lower than stage II (37.21%) and IV (52.41%). CC and PDAC more frequently presented with < 5cm tumor, at academic or research cancer centers, and diagnosed between 2009 and 2013 compared to 2004–2008 ( p< 0.001). For both CC and PDAC, the majority underwent surgical resection (58%, 53%), systemic chemotherapy (57.8%, 63%) and did not receive radiotherapy (78.8%, 77.6%). A positive surgical margin on pathologic evaluation was associated with worse outcomes for CC and PDAC in both univariate and multivariate analysis (HR 1.61; 1.56–1.66; p< 0.001 and HR 1.43; 1.38–1.48, p< 0.001). CC had a better 1-year overall survival (OS) in all stages compared to PDAC (p < 0.001). In multivariate analysis, mucinous carcinoma histology, female sex, diagnosis between 2004 and 2009, well/moderately differentiated histology, chemotherapy, age at diagnosis less than 60, radiation therapy after surgery, and local surgical procedure of primary site and pancreatectomy (p < 0.001) were associated with better OS compared to PDAC. Colloid histology was associated with better 1-year overall survival (OS) in all stages compared to PDAC (p < 0.001). Conclusions: Colloid carcinoma of pancreas is associated with a better overall survival as compared to pancreatic ductal adenocarcinoma. This is the largest study to address the clinical features and outcomes of colloid carcinoma of pancreas.

Treatment outcomes for stage T1b-2 esophagogastric adenocarcinomas

e16085

Background: Treatment of localized esophageal, gastroesophageal junction (GEJ), and stomach cancer is neoadjuvant therapy with either chemoradiation or chemotherapy followed by surgery. Treatment for T1b-2 stage disease is not well evaluated and this stage is underrepresented in prospective studies. The aim of this study is to evaluate survival outcomes among the three treatment modalities (neoadjuvant chemotherapy (NACT), neoadjuvant chemoradiation (NACRT), and upfront surgery (US)) in this population using the National Cancer Database (NCDB). Methods: Patients (pts) with clinical stage T1b-2N0 and any pathological stage (excluding metastatic) adenocarcinoma of the esophagus, GEJ, and stomach treated with neoadjuvant therapy or upfront surgery, with or without adjuvant chemotherapy (AC), were identified between 2004 and 2015 in the NCDB. Univariate and multivariable analyses were conducted, and Kaplan-Meier analysis and Cox proportional hazard models were used to identify the association between the three treatment modalities and overall survival (OS). Results: A total of 2260 pts were analyzed. The median follow-up was 66.6 months. The median age was 67 years. Most pts were White (86%) and male (77%). 1018 (45%) had moderately-differentiated grade, while 946 (42%) had poorly-differentiated/undifferentiated grade. The most common site of disease was the lower third of esophagus (34.1%). 161 pts (7%) received NACT, of whom 45 pts received AC; 537 pts (24%) received NACRT, of whom 40 pts received AC. 1562 pts (69%) underwent US, of whom 146 pts received AC. US with AC was associated with the best survival, followed by NACT with AC; median OS was 90.1 and 86.8 months for surgery with AC and NACT with AC, respectively. NACRT was associated with the worst survival (39.5 and 40.2 months with and without AC, respectively). The 5-year OS rates were 59.8%, 58.5%, 52.1%, 44.9%, 37.3%, and 37.8%, for US, NACT, and NACRT, with and without AC, respectively. The rate of tumor upstaging was highest in the NACT group, followed by the NACRT group, and lowest in the US group. Postsurgically, 62 (39%) and 48 (30%) pts in the NACT group and 198 (37%) and 161 (30%) pts in the NACRT group had upstaging in their T and N stages, respectively, compared to 214 (13%) and 326 (21%) pts in the US group. For the 1107 pts who also had pathological T1b-2N0 stage disease following US, no difference in survival was observed with or without AC. Conclusions: Upfront surgery with adjuvant chemotherapy and perioperative chemotherapy are associated with the best survival compared to preoperative radiotherapy. This is the largest study to address the best approach for the treatment of T1b-2 stage disease.

Phase II randomized, double-blind study of mFOLFIRINOX plus ramucirumab versus mFOLFIRINOX plus placebo in advanced pancreatic cancer patients (HCRN GI14-198)

413

Background: Vascular endothelial growth factor (VEGF) and VEGF receptor (VEGFR)-mediated signaling and angiogenesis contribute to the pathogenesis and progression of pancreatic adenocarcinoma (PCA).VEGF is expressed in all PCA tumors. VEGF-A/VEGFR-2 signaling plays an important role in inducing invasion and migration of PCA cells. The pVEGFR-2 is significantly associated with invasion of the anterior capsule of pancreas and arteries. In preclinical studies, the anti-tumor activity of fluoropyrimidines, but not that of gemcitabine, caused the release of bone marrow derived circulating endothelial progenitor cells (CEPs) and Tie-2 expressing monocytes (TEMs) as well as the induction of pro-angiogenic growth factors. Methods: This phase II randomized, multi-center, and double-blinded trial was designed to compare the efficacy and safety of mFOLFIRINOX/ramucirumab (Arm A) versus mFOLFIRINOX/ placebo (Arm B) as front-line therapy in recurrent or metastatic PCA patients. The primary endpoint was progression free survival (PFS) at 9 months, and the secondary endpoints included overall survival (OS) and response rate. Results: A total of 86 subjects were enrolled, 82 were eligible (42 in Arm A v. 40 in Arm B). The mean age of the subjects in the two arms were comparable (61.7 v. 63.0, respectively); 43 male, 69 Caucasian. On the univariate analysis, there was no difference in distribution between the 2 arms for age, gender, race and ethnicity. The median PFS was 5.6 in Arm A compared to 6.7 months in Arm B (one-sided log-rank, p = 0.322). At 9 months, the progression free rates were 25.1% v. 35% for Arms A and B, respectively. The mFOLFIRINOX/ramucirumab combination was well tolerated. Patients in Arm A reported a slightly higher number of adverse events (AEs) encounters, most commonly diarrhea (29 vs 28), fatigue (25 v. 25), vomiting (24 v. 14), weight loss (23 v. 17), and abdominal pain (20 v. 15). Arm A had more SAEs than Arm B (43 v. 25), with sepsis most commonly reported in both arms (3 in each), vomiting (3 v. 2), diarrhea (3 v.1) and duodenal obstruction (3 v. 0). Arm B had a slightly higher response rate (22.58%) compared to Arm A (17.65%) that was not statistically significant. The median OS in Arm A was 10.3 compared to 9.7 months for Arm B (one-sided log-rank, p = 0.094). Conclusions: In this randomized phase 2 study, the addition of ramucirumab to mFOLFIRINOX did not improve PFS, response rate, or OS as initial therapy for metastatic pancreatic cancer. FOLFIRINOX/Ramucirumab combination was well tolerated in the treatment of PCA. Clinical trial information: NCT02581215.

Survival analysis of colorectal cancer patients treated with first-line modified FOLFOX6 with or without bolus fluorouracil

35

Background: The combination of oxaliplatin, bolus 5-FU (b5FU), infusional 5-FU (5-FUCI), and leucovorin (LV) is the preferred first line treatment option for mCRC. This study evaluates the impact of b5FU on survival in first line therapy for mCRC patients treated with mFOLFOX6. Methods: This was a retrospective chart review of patients ≥ 18 years old with mCRC receiving palliative first line mFOLFOX6 chemotherapy with or without b5FU/ LV from January 1, 2010 through June 1, 2019 at Winship Cancer Institute, Emory University. Data collection included the following: demographics (age, race, gender), disease characteristics (tumor sidedness), microsatellite status, KRAS status, BRAF status, addition of monoclonal antibodies (bevacizumab, panitumumab), ECOG PS, grade 3/4 neutropenic events, addition of growth factors, and treatment delays. The primary endpoint was PFS. The multivariable Cox proportional hazards model for PFS and OS was performed with selected covariates of interest. Results: A total of 252 patients with mCRC met the inclusion criteria. Median follow-up time was 2.4 years. 161 patients (64%) received mFOLFOX6 with b5FU/LV and 91 patients (36%) received mFOLFOX6 with no b5FU/LV. More cycles were delivered in the b5FU group as compared to the non-b5FU group (mean, 4.8 v. 3.8 cycles, respectively; p < 0.001). There were no differences in grade 3 and 4 neutropenic events between groups. Growth factor usage was numerically higher in the bolus group though not significantly different (p = 0.06). No difference was observed in treatment delays between groups (p = 0.83). There was no statistical difference in PFS between treatment groups (1.1 years in the b5FU/LV group v. 0.8 years (95% CI, 0.6-1.0) in the no 5-FU/LV bolus group; p = 0.076). The median OS was 2.5 years in the b5FU/LV group compared to 1.8 years in the no b5FU/LV group (p = 0.012). On univariate analysis, tumor sidedness and performance status were significantly different between groups. On multivariate analysis, none of the variables were significantly different between groups. Conclusions: The omission of the b5FU/LV from mFOLFOX6 does not significantly impact PFS, toxicity or treatment delays. However, OS is significantly shorter when the b5FU/LV is omitted suggesting the clinical importance of maintaining bolus administration with 5-FUCI in the first line palliative treatment of mCRC.

Survival outcomes of adjuvant chemotherapy in elderly patients with stage III colon cancer

89

Background: The survival impact of multi-agent (MAC) as compared to single-agent (SAC) adjuvant chemotherapy (AC) in elderly patients with stage III colon cancer (CC) remains controversial. The aim of this study is to evaluate the survival outcome comparing MAC to SAC in this population utilizing the National Cancer DataBase (NCDB). Methods: Patients 70 years and older with pathological stage III CC were identified between 2004 and 2015 from the NCDB using ICD-O-3 morphology and topography codes: 8140-47, 8210-11, 8220-21, 8260-63, 8480-81, 8490, and C18.0-18.8 (without C18.1). Univariate and multivariable analyses were conducted and Kaplan-Meier analysis and Cox proportional hazard models were used to identify the association between MAC vs. SAC and overall survival (OS). Results: A total of 41,707 elderly patients (≥70 years old) with stage III CC were identified. Around half of the patients (n = 20,257; 48.5%) received AC; the majority of whom (n = 12,923, 63.8%) received MAC. The median age was 79 (range 70-90). Of the patients who received AC, the majority were female (n = 11,201, 55.3 %), Caucasians (87.4%) and had a moderately differentiated tumor grade (n = 12,619, 62.3%). Tumor size more than 4 cm was identified in 11,785 (58.2%) patients and 18,496 (91.3%) had negative surgical margins. Low-risk stage III CC constituted 50.6% (n = 10,264) of the study population. High-risk stage III CC was associated with worse OS compared to low-risk disease (p < 0.001). MAC was associated with better 5-year OS compared to SAC (p < 0.001). High-risk stage III patients who received MAC had an OS of 4.2 v. 3.4 years in SAC (p < 0.001). In low risk stage III, patients who received MAC had median OS of 8.5 v. 7 years in SAC (p < 0.001). In Univariate, male sex, positive surgical margin, insurance and facility types, age, year of diagnosistumor size, and Charlson-Deyo Score of > 2 were associated with worse OS (p < 0.05). Conclusions: Multi-agent AC is associated with better survival in stage III CC patients 70 years and older compared to SAC. Enhanced benefit of MAC was shown for both low risk and high risk stage III CC.

Phase II trial of nivolumab and metformin in patients with treatment refractory microsatellite stable metastatic colorectal cancer

95

Background: Preclinical data suggests metformin can improve immune exhaustion of tumor infiltrating lymphocytes and potentiate the effects of PD-1 blockade. By normalizing the hypoxic TME, metformin was shown to improve cytotoxic T cell function and efficacy of anti-PD-1 antibody in highly aggressive B16 melanoma and MC38 colon adenocarcinoma tumor models. Based on this preclinical rationale we conducted a phase II study with nivolumab and metformin combination in treatment refractory MSS metastatic colorectal cancer (mCRC). Methods: Nivolumab 480 mg IV every 4 weeks and Metformin 1000 mg po twice daily was administered in 28-day cycles following a 14-day metformin only lead-in phase.Eligible patients included stage IV metastatic treatment refractory MSS mCRC (patients must have received oxaliplatin, irinotecan, and fluoropyrimidine), age ≥18 years, ECOG PS 0-1, adequate organ function, no prior anti PD-1 agent. The primary endpoint was overall response rate (ORR). Secondary endpoints included overall survival (OS) and progression free survival (PFS). Simon’s two-stage Minimax design was employed (H0: ORR =4%; H1: ORR=15%; alpha = 0.1; power =80%). If ≥1 objective response was observed in the first evaluable 18 patients, 10 additional patients would be included in the cohort. ≥3 objective responders in 28 patients would be required to be considered positive study. Pre-treatment and on-treatment research biopsies and correlative peripheral blood specimens were collected. Results: A total of 24 patients were enrolled, 6 patients were replaced per protocol, and 18 patients had evaluable disease. Of the 18 evaluable patients 11/18 (61%) were female, median age 58 [IQR 50-67]. 2 patients had prolonged stable disease (4 and 10 cycles). No patients had objective response based on RECIST 1.1. Median OS and PFS was 5.1 months [95% CI (2-11.7)] and 2.3 months [95% CI (1.7-2.4)], respectively. Most common grade 3 and 4 toxicities were anemia (n=2) and diarrhea (n=2). Conclusions: In treatment refractory MSS mCRCnivolumab and metformin combination was well tolerated. Two patients achieved stable disease, but no objective response was seen; therefore, the study did not proceed with the second stage of enrollment. Immunologic correlative analysis of this study is ongoing. Clinical trial information: NCT03800602.

Impact of primary tumor size/horizontal extent on survival in colorectal cancer

125

Background: Pathologic staging in colorectal cancer (CRC) is crucial in patient management. Data regarding the impact of size/horizontal tumor extent is limited, contradictory and currently excluded from the American Joint Committee on Cancer (AJCC) staging model. However, a previously published SEER analysis showed that AJCC stages I and IIIA have similar 2- and 5- year survival rates, and worse rates for stage II. Using the largest cohort to date, we report the impact of primary tumor size on CRC survival. Methods: Data were obtained from all US hospitals that contributed to the National Cancer Database (NCDB) between 2010 and 2015. Univariate and multivariate analyses were performed to identify factors associated with patient outcome. Kaplan-Meier analysis and Cox proportional hazards models were used to assess the association between tumor/patient characteristics and overall survival (OS). Results: A total of 61,145 patients were identified with a similar gender distribution (M/F:50.9%/49.1%). The mean age was 62.7years (SD+/-14.1) and 82% were non-Hispanic Whites. Majority had colon primary (82.7%) and 82.4% had microsatellite stable (MSS) disease. Distribution across stages I-IV was 20.1%, 32.1%, 34.7% and 13.2% respectively. Among the total study population, AJCC stage correlated closely with OS on multivariate analysis (HR 1.49, 2.29, 8.38 for stages II to IV compared to stage I), while the distinguishing power for tumor size was relatively mild (HR 1.19 and 1.33 for 5-10 cm and >5cm compared to <5cm). Among patients with stage II disease, tumors >10cm were associated with worse survival compared to those <5cm (HR 1.2; 1.03-1.39; p=0.22). Stage III disease also had differential survival rates; patients with tumors 5-10cm (HR 1.21; 1.14-1.28; p<0.001) and >10cm (HR 1.57; 1.37-1.80; p<0.001) had worse survival than those <5cm. Patients with stage II who did not receive adjuvant chemotherapy (CTX) had worse survival outcomes (HR 1.29; 1.08-1.55; p=0.005) compared to stage III disease who did. Accounting for tumor size, there was no statistically significant survival differences between stage I patients and stages II and III patients who received adjuvant chemotherapy. Conclusions: Tumors larger than 10cm have inferior outcomes among patients in the same AJCC stages. Stage II patients without adjuvant CTX did worse than stage III with CTX. Further studies are needed to clarify the role of tumor size in staging models. [Table: see text]

Adjuvant concurrent chemoradiotherapy in extrahepatic cholangiocarcinoma

4583

Background: Resected cholangiocarcinomas are rare and have high relapse rates. Adjuvant chemotherapy is the standard of care (BiLCAP Trial). Adjuvant radiation therapy benefit is not well defined. This study aims to evaluate survival outcomes of the effect of adjuvant chemoradiotherapy compared to chemotherapy in extrahepatic cholangiocarcinoma (EHC) using the National Cancer Database (NCDB). Methods: Patients with resected EHC between 2004 and 2013 were identified from the NCDB using ICD-O-3 histology and topography codes: 8140, 8160, 8161, 8162 and C24.0. Patients with neoadjuvant therapy were excluded from this analysis. Univariate and multivariable analyses were conducted, and Kaplan-Meier Curves were used to compare overall survival (OS) based on treatment received. Results: A total of 236 EHC patients were identified. Males comprised 60.6% and 88.1% were Caucasian. Median age was 64 (range, 31-84) years. The majority were distal (72.0%, N = 157) followed by perihilar (20.6%, N = 45), hilar (6.4%, N = 14) and cystic (0.9%, N = 2). Distribution across stages I-III was 28.8% (N = 68), 56.8% (N = 134), and 14.4% (N = 34), consecutively. Adjuvant chemotherapy was given in 37.7% (N = 89) and adjuvant chemoradiotherapy in 62.3% (N = 147). The median dose of radiation was 50.4 Gy. Adjuvant chemoradiotherapy was mostly given in regional node positive disease (p = 0.016) and negative surgical margin (p = 0.002) compared to regional node negative disease and positive surgical margin, respectively. The use of adjuvant chemoradiotherapy was associated with improved OS compared to chemotherapy alone in univariate (HR 0.64; 95% CI 0.44-0.93; p = 0.019) and multivariable analysis (HR 0.65; 95% CI 0.44-0.96; p = 0.030). Median survival and 1 year-OS for patients that received chemoradiotherapy was 33.8 months (95% CI 28, NA) and 87.7% (80.9%, 92.1%) compared to chemotherapy alone which was 23.8 months (95% CI 18.9, 35.4) and 75.5% (64.9%, 83.3%). Conclusions: Adjuvant chemoradiotherapy was associated with improved survival in patients with resected EHC compared to chemotherapy alone. This conclusion warrants further prospective studies to confirm these results.

A multicenter randomized phase II study of nivolumab in combination with gemcitabine/cisplatin or ipilimumab as first-line therapy for patients with advanced unresectable biliary tract cancer (BilT-01)

4582

Background: Patients (pts) with advanced biliary tract cancers (BTC) have poor prognosis with a median overall survival (OS) less than 12 months (mos). This randomized phase 2, multi-institutional, study was designed to investigate the role of combinational immunotherapy, using nivolumab (nivo) with gemcitabine (gem)/cisplatin (cis), or nivo with ipilimumab (ipi) in pts with untreated advanced BTC. Methods: Key eligibility criteria include histologically confirmed unresectable or metastatic BTC without prior systemic therapy, measurable disease per RECISTv1.1, ECOG PS 0-1, and absence of autoimmune disease or chronic steroid use. Arm A included gem 1000 mg/m2 and cis 25 mg/m2 d1, 8 Q3w + nivo 360 mg d1 Q3w for 6 mos followed by nivo 240 mg Q2w monotherapy for a total duration of 2 yrs; Arm B included nivo 240 mg Q2w and ipi 1 mg/kg Q6w for 2 yrs, in absence of disease progression. Primary endpoint is progression-free survival (PFS) rate at 6 mos with an alternative hypothesis of 80% (null hypothesis of 59%, one-sided alpha 0.05, power 80%) for each non-comparative arm. Secondary endpoints include overall response rate (ORR) per immune related (ir)RECIST, median PFS and OS and safety. Exploratory objectives include biomarker analysis using include sequential whole exome/transcriptome and immune cell subsets in tissue and blood. Results: 71 eligible pts (49% male, 83% Caucasian) with 35 in Arm and 36 in Arm B with a median age of 62 (range 20-80) yrs, and majority with metastatic disease (90%) were enrolled across 6 US sites. PFS rate at 6 mos was 70% in Arm A and 18.6% in Arm B. The median PFS was 8.8 mos (95% CI, 6.1 to 11.3) in Arm A and 4.1 mos (95% CI, 2.4-5.2) in Arm B. Ten patients on Arm A and 2 on Arm B remain on active treatment; additional 7 are in follow-up for OS. ORR, safety data and median OS evaluation are underway and will be presented at the meeting. Exploratory analyses are pending. Conclusions: The observed PFS rates at 6 mos in either arm are insufficient to reject the null hypothesis of 59% PFS at 6 months. While Arm B is inferior, Arm A appears to be as effective as standard of care although OS estimates are pending maturity. Clinical trial information: NCT03101566 .

Impact of tumor side on clinical outcomes in stage II and III colon cancer with known microsatellite instability status

4068

Background: Microsatellite instability high (MSI-H) status indicates better prognosis in early stage colon cancer (CC) compared to microsatellite stable (MSS). However, the impact of tumor side, left side (L) versus right side (R), is not described on clinical outcomes based on MSI status. Methods: Patients with pathological stage II and III primary adenocarcinoma of the colon between 2010 and 2015 were identified in the National Cancer Database (NCDB) using ICD-O-3 morphology and topography codes: 8140-47, 8210-11, 8220-21, 8260-63, 8480-81, 8490 and C18.0, 18.2,18.3, 18.5,18.6, 18.7. Univariate (UVA) and multivariable (MVA) survival analyses were conducted, and Kaplan-Meier Curves were used to compare overall survival (OS) based on tumor location and treatment received with Log-rank test. Results: A total of 35,071 patients with stage II (n = 17,629) and III (n = 17,442) CC were identified. 51.3% female; 81.5% Caucasian; median age 66 (range, 18-90). Majority of stage II and III tumors were R, 61.2% (n = 10,794) and 56.0% (n = 9,763). MSI-H was more common in stage II compared to III, 23.3% (n = 4,115) vs 18.2% (n = 3,171) (p < 0.0001). Survival was better in stage II MSI-H compared to MSS, 5 year-OS 75.1% vs 71.8% (p = 0.0057). However, stage III CC survival was better in MSS compared to MSI-H, 5-year OS 60.5% vs 58.0% (p < 0.001). In stage II MSI-H CC R was more common than left, 78.3 % (n = 3223) vs 21.7% (n = 892). There was no significant difference in survival between stage II MSI-H L vs R (5-year OS 76.2% vs 74.7%, p = 0.1578). Stage II MSS CC R was more common than L, 56.0% (n = 7571) vs 44.0% (n = 5943), and survival was better in L vs R (5-year OS 73.2% vs 70.8%, p = 0.0029). Stage III MSI-H CC was more common in R than L, 75.6% (n = 2397) vs 24.4% (n = 774) and survival was better in L (5-year OS 62.5% vs 56.5%, p = 0.0026). Stage III MSS CC was more common in R than L, 51.6% (n = 7366) vs 48.4% (n = 6905), and survival was better in L vs R (5-year OS 67.0% vs 54.4%, p < 0.001). Conclusions: Survival was better in left sided tumors compared to right in stage II MSS, stage III MSS and stage III MSI-H CC.

The impact of body mass index (BMI) on the safety and outcomes of small molecule inhibitors (SMI) in gastrointestinal cancer

144

Background: A number of studies have demonstrated that overweight and early obese states are associated with improved survival in renal and melanoma cancer patients. The purpose of this study was to investigate the impact of BMI on outcomes of GI cancers treated with SMIs. Methods: A retrospective chart review was conducted to compare outcomes between patients with BMIs ≥ 25 and < 25 who received treatment at Winship between 1/2010-8/2019. The primary objective was to determine the PFS, OS, and ADR rates of patients with BMI ≥ 25 treated with SMIs compared to patients with BMIs < 25. PFS and OS rates were estimated with the Kaplan-Meier method and compared between the groups using the log-rank test. The incidence of adverse events was estimated as frequency and percentage and logistic regression was used to estimate the impact of BMI on adverse effects. Results: 269 patients were included in analysis for PFS, 61 for OS and 281 for ADR rate analysis. Diagnoses included HCC (n = 123, 44%) and CRC (n = 158, 56%). There was no significant difference seen in PFS and OS between the BMI < 25 and BMI ≥ 25 groups (HR 1.17 (0.87-1.56), p = 0.291, HR 1.10 (0.66-1.84), p = 0.713, respectively). A significant difference was demonstrated in the rates of adverse reactions between the two groups (OR 0.31 (0.19-0.52), p<.001) with BMI < 25 having a lower rate of ADRs. After adjusting for covariates, HCC vs CRC diagnosis (p = 0.002) and not requiring dose reduction or delay due to toxicity (p<0.001) were significantly associated with better PFS. Conclusions: Based on the multiple regression analysis, patients with a BMI ≥ 25 do not have improved outcomes over those with a BMI < 25 when taking SMIs for their GI malignancies. Patients with BMI < 25 experienced less drug toxicity that required dose reduction, discontinuation, or treatment delays.

[Table: see text]

Impact of high-risk features for stage II adenocarcinoma of the appendix

795

Background: Clinico-pathological high risk features are frequently utilized in adjuvant chemotherapy (AC) decisions in stage II colorectal cancer and their utility in stage II appendiceal adenocarcinoma (AA) is less established. The aim of this study is to determine the impact of high risk features on clinical outcomes and whether high risk features are predictive of AC benefit in stage II AA. Methods: Patients with pathological stage II AA between 2010 and 2015 were identified from the National Cancer Database (NCDB) using ICD-O-3 morphology and topography codes: 8140, 8480 and C18.1. High risk stage II AA was defined as having at least one of the following clinicopathological features: T4 tumor, < 12 lymph nodes examined, poorly differentiated histology, positive margins, or lymphovascular invasion. Patients with none of these features were defined as low-risk. Results: A total of 1,040 patients were identified. 51.0% males, 84.5% Caucasian; median age 61 (range, 19-90). 46.4% were determined to have high-risk stage II AA. High-risk status was associated with worse OS compared to low-risk in univariate (HR 1.55; 95% CI 1.18-2.02; p = 0.001) and multivariable analyses (HR 1.36; 95% CI 1.03-1.79; p = 0.028). High-risk stage II AA patients had significantly worse 5-year OS compared to low-risk patients (67.1% vs. 74.5%, p = 0.0013). AC was administered in 34.4% (n = 166) of high-risk patients and in 36.5% (n = 203) of low-risk patients. Among high-risk patients, AC was not associated with better OS in univariate (HR 0.86; 95% CI 0.59-1.26; p = 0.722) and multivariable analyses (HR 1.35; 95% CI 0.90-2.04; p = 0.324) compared to no AC. Similarly, among low-risk patients, AC was not associated with better OS in univariate (HR 0.92; 95% CI 0.60-1.39; p = 0.813) and multivariable analyses (HR 1.27; 95% CI 0.81-2.02; p = 0.334) compared to no AC. For high-risk patients, 5-year OS was 68.3% in patients that received AC vs. 66.5% in patients that did not (p = 0.722). For low-risk patients, 5-year OS was 74.0% in patients that received AC vs. 76.3% in patients that did not (p = 0.813). Conclusions: High-risk stage II AA patients had significantly worse 5-year OS compared to low-risk patients. AC did not improve survival regardless of high risk features in stage II AA.

FIDES-01, a phase II study of derazantinib in patients with unresectable intrahepatic cholangiocarcinoma (iCCA) and FGFR2 fusions and mutations or amplifications (M/A)

TPS597

Background: Deregulation of the FGFR signaling pathway is implicated in various cancers. In iCCA, FGFR genetic aberrations include FGFR2 fusions and, less commonly, FGFR2 M/A. iCCA prognosis is poor, and chemotherapeutic and targeted treatment options are limited. While FGFR2 fusions are acknowledged oncogenic drivers, the oncogenic potential of FGFR2 M/A is less well defined. Derazantinib (DZB) is an investigational, oral small-molecule kinase inhibitor with activity against FGFR1, 2 and 3, which demonstrated antitumor activity in patients with unresectable iCCA with FGFR2 fusions. Based on preliminary efficacy data demonstrating durable responses of > 6 months and a clinically meaningful progression-free survival in a subset of iCCA patients harboring FGFR2 M/A (NCT01752920), the multicenter, multicohort open-label phase 2 study FIDES-01 is evaluating the effect of DZB in separate cohorts of iCCA patients with FGFR2 fusions or FGFR2 M/A. Methods: The new cohort evaluates 300 mg once daily dosing of DZB in patients with unresectable iCCA with FGFR2 M/A per liquid or tissue biopsy-based next generation sequencing and at least one previous systemic therapy. Treatment will continue until progressive disease, intolerance, withdrawal of informed consent, or death. Using a Simon’s two-stage design, the primary endpoint to assess the antitumor activity of DZB is the proportion of patients with PFS at 3 months (PFS3; per RECIST 1.1 central review). Secondary objectives are evaluation of median PFS, objective response rate, duration of response, safety profile, quality of life (incl., QLQ-C30, QLQ-BIL21, EQ-5D), and symptom response from baseline. Current status: The study was initiated in July 2019 with planned enrollment of 43 patients with confirmed FGFR2 M/A. Clinical trial information: NCT03230318.

Is adjuvant chemotherapy beneficial for stage II-III goblet cell tumors of the appendix?

796

Background: Goblet cell tumors (GCT) of the appendix are very rare tumors constituting 2.5%-5% of all primary appendiceal neoplasms. Role of adjuvant chemotherapy (AC) is not established for GCT. This study aims to evaluate the impact of AC in stage II-III appendiceal GCT. Methods: Patients with pathological stage II and III GCT who underwent surgical resection between 2006 and 2015 were identified from the National Cancer Database (NCDB) using ICD-O-3 morphology and topography codes: 8243/3, 8245/3 and C18.1. Patients treated with neoadjuvant systemic and/or radiation therapy and adjuvant radiation were excluded. Univariate and multivariable analyses were conducted, and Kaplan-Meier Curves were used to compare overall survival (OS) based on treatment received with Log-rank test. Results: A total of 1,046 patients were identified. 53.7% males and 89.0% Caucasian; median age 56 (range, 20-90) years. Distribution across pathological stages II-III was 83.6% (N = 874) and 16.4% (N = 172) consecutively. 8.3% (N = 73) of stage II and 50.6% (N = 87) of stage III patients received AC. In the total cohort, AC was not associated with better OS compared to no AC in univariate analysis (HR 1.84; 95% CI 1.26-2.67; p = 0.001) or multivariable analysis (HR 0.94; 95% CI 0.57-1.52; p = 0.790). For stage II patients, AC was not associated with better OS in univariate (HR 1.24; 95% CI 0.60-2.57; p = 0.562) or multivariable analyses (HR 1.67; 95% CI 0.76-3.64; p = 0.199). Similarly, in stage III patients, AC was not associated with better OS in univariate (HR 0.78; 95% CI 0.48-1.29; p = 0.340) or multivariable analyses (HR 0.55; 95% CI 0.28-1.04; p = 0.067). In the entire cohort 5-year OS for patients that received AC was 83.9% (80.3%, 86.9%) versus 70.7% (60.9%, 78.5%) (p = 0.001) with no AC. For stage II patients, 5-year OS was 77.3% with AC vs. 87.7% with no AC (p = 0.562). For stage III patients, 5-year OS was 64.8% with AC vs. 54.4% with no AC (p = 0.340). Conclusions: AC was not associated with improved 5-year OS in patients with pathological stage II and III GCT compared to no AC.

Clinical and molecular markers of immune checkpoint inhibitor (ICI) response in dMMR colorectal cancer (CRC) patients (pts)

225

Background: ICIs induce durable responses in dMMR CRC patients. However, clinical and molecular biomarkers of response to ICIs have not been well-established. In this study, we investigated impact of specific MMR gene loss, BRAF V600E mutation and clinical characteristics of pts on clinical outcomes of ICIs. Methods: Pts were eligible if they had confirmed dMMR CRC by IHC or MSI-H by PCR and received ICIs between 01/01/2012 and 05/01/2019 at Winship Cancer Institute of Emory University, Mayo Clinic, Vanderbilt or Stanford University. Due to the functional dependency, the groups were categorized as protein loss of MLH1+PMS2 vs MSH2+MSH6. Log-rank test, Cox hazard model and Fisher’s exact test were used for survival outcomes, the best response and the distribution of variables among the subgroups. Results: A total of 66 pts with dMMR CRC were identified and BRAF status was available for 41 pts. ORRs in MLH1+PMS2 and MSH2+MSH6 groups were 72.9% and 56.5% respectively (P = 0.189). At 2 years, PFS rates were 55.6% and 78.2% for MLH1+PMS2 and MSH2+MSH6 groups respectively (P < 0.001). Pts with BRAF V600E mutations had significantly worse outcomes as compared to pts with wild-type BRAF (2-year PFS rate of 35.0% and 73.3% respectively; P < 0.001). Notably pts < 65 had better 2-year disease control rates when compared to > 65 (71.1% and 41.5% respectively; P < 0.001). We also observed worse 2-year PFS rates in pts with liver metastases (P = 0.014). CRC side and tumor volume did not impact 2-year PFS rates in our cohort. Conclusions: Our data suggest that pts with loss of function in MSH2+MSH6 may have better 2 year-PFS rates compared pts with MLH1+PMS2 even though ORR favored MLH1+PMS2 group suggesting that ORR may not reflect the durability of ICI response in dMMR CRC patients. Consistently, pts with BRAF V600E mutation which is associated with MLH1 promoter methylation had significantly worse 2-year PFS rates. Overall, our findings suggest that BRAFV600E mutation, the affected MMR proteins, pt age, and site of metastasis may impact durability of ICI response in dMMR CRC patients.

Radiation as a single modality treatment in localized pancreatic cancer

703

Background: Locally advanced pancreatic cancer (LAPC) is managed with multimodality therapy. A subset of patients with LAPC are not good candidates for aggressive treatment. The aim here is to evaluate the outcomes of single modality radiation therapy for LAPC using the National Cancer Database (NCDB). Methods: Data was obtained between years 2004 and 2013. Pancreatic ductal adenocarcinoma (PDAC) patients with unresectable local disease were identified excluding patients who received chemotherapy or surgery. Univariate and multivariable analyses identified factors associated with patient outcome. Kaplan-Meier analysis and Cox proportional hazards models were used for patient characteristics and overall survival (OS). Results: A total of 6,590 patients were included; 480 (6.9%) received radiation therapy only and 6470 (93.1%) received no treatment. Mean age was 73.5 (range, 28‐90) years, with the majority being White (N = 5685; 83.2%) and female (N = 3779; 54.4%). Poorly differentiated histology and tumors ≥ 4 cm ( > T3 stage) accounted for 47.8% and 52.7%, respectively. The median dose of radiation was 39.6 Gy. Stereotactic body radiation (SBRT) was given in 64 patients and external-beam/Intensity modulated radiotherapy (IMRT) in 416 patients. Charlson-Deyo score of +1 was seen in 34.4% of patients who received no treatment, 32.8% of patients who received SBRT and in 29.8% of patients who received external-beam IMRT. Radiation therapy was associated with improved OS compared to no treatment in univariate and multivariable analyses controlling for sex, Charlson-Deyo score, age, tumor size, amongst other covariates. Median OS for patients who received SBRT, external-beam/IMRT or no radiation was 8.6, 6.7 and 3.4 months; respectively (P < 0.001). There is a significant difference in 12-month OS for the SBRT cohort (31.9%; 95% CI 20.9%-43.5%) compared to patients who received no radiation (15.1%; 95% CI 14.2%-16.0%), similarly seen on multivariable analysis (HR 0.50; 95% CI 0.38-0.65; P < 0.001). Conclusions: The current study is the first to evaluate the efficacy of radiation as single modality therapy in LAPC. The results suggest a potential benefit for radiation therapy alone, in comparison to no treatment.

A phase II study of niraparib in combination with EGFR inhibitor panitumumab in patients with advanced colorectal cancer

TPS269

Background: Panitumumab (Pmab) is a recombinant monoclonal antibody that binds specifically to the human epidermal growth factor receptor (EGFR), and is indicated for metastatic colorectal carcinoma (mCRC). EGFR inhibition induces synthetic lethality with poly ADP ribose polymerase inhibitors (PARPi) by attenuating DNA repair pathways. This susceptibility to PARPi-induced cell death by EGFR inhibition is associated with deficient Non-homologous end joining (NHEJ), and Homologous recombination (HR) mediated DNA repair and persistence of DNA damage. Furthermore, efficacy of PARPi (such as niraparib) is highly correlated with platinum sensitivity. Cancer cell sensitivity and resistance to both PARPi and platinum have been associated with loss and restoration of HR DNA repair, indicating similar mechanisms of anticancer activity and resistance. Platinum sensitivity in CRC could therefore predict for anticancer properties of PARPi when utilized in the setting of synthetic lethality. Combining PARP and EGFR inhibition has the potential to confer synergistic benefit, while ameliorating resistance mechanism to PARPi. This study aims to evaluate the activity of the combination of niraparib and Pmab in RAS wildtype (WT) mCRC. Methods: Eligible patients for the trial include advanced, RAS WT mCRC who have been intolerant of, progressed on, or failed at least one line of systemic chemotherapy. Those currently on first line oxaliplatin-containing regimen are allowed on the trial if they have remained stable or better (PR or CR) for at least 4 months on that line of treatment, and are being considered for maintenance therapy as standard of care. Patients must also be 18 years old, ECOG PS 0-1 and measurable disease per RECIST 1.1. A safety run-in cohort of 6 eligible patients, and additional 20 patients with the same inclusion criteria will be enrolled. Pmab dose - 6 mg/kg IV on days 1 & 15 of each 28-day cycle; Niraparib - 200mg or 300mg (based on body weight and platelet count) orally continuously. Primary endpoint: clinical benefit rate (CR +PR + SD). Biomarker analysis includes skin biopsies evaluated for p-Caspace-3, PARP, p-MAPK, Ki-67, and p27. The study was activated in Sept. 2019. Clinical trial information: NCT03983993.

Phase Ib trial of pembrolizumab and XL888 in patients with advanced gastrointestinal malignancies: Results of the dose-escalation phase

830

Background: XL888 is a selective inhibitor of heat shock protein 90 (HSP90). It modulates several oncogenic signaling pathways, and the tumor microenvironment. In preclinical models, XL888 potentiates efficacy of PD-1 inhibition. We report the results of the dose escalation (DE) portion of a phase Ib trial of combined XL888 and pembrolizumab (P) in advanced gastrointestinal adenocarcinomas. Methods: XL888 was administered orally (PO) in three dose levels of 45 (DL1), 90 (DL2), 60 (only if DLT on DL2) mg twice weekly with P 200 mg IV on day 1, in 21-day cycles. Eligible patients included stage IV or locally advanced unresectable gastrointestinal adenocarcinomas with at least one prior therapy (patients with colorectal (CRC) adenocarcinoma must have received oxaliplatin, irinotecan, and fluoropyrimidine), age ≥18 years, ECOG PS 0-1, adequate organ function, no prior anti-PD-1 or anti-PD-L1 agent. The primary endpoint was recommended phase II dose (RP2D), while secondary endpoints included safety and tolerability. Pre-treatment and on-treatment correlative peripheral blood specimens were collected. Results: A total of 14 patients were enrolled in the DE phase. 9 male, median age 66.5. Diagnoses included CRC (6), pancreatic adenocarcinoma (5), biliary tract cancer (1), ampullary (1), and duodenal (1). Two patients were ineligible for assessing the primary endpoint (DL2) due to biliary stent obstruction and sepsis. One DLT (grade 3 autoimmune hepatitis) was observed on DL2. We enrolled three patients on DL3. Five additional patients were subsequently enrolled on DL2 with no additional DLT. Three patients (1 duodenal, 2 CRC) had prolonged stable disease (6, 9 and 15 cycles). The most common treatment-related toxicities included autoimmune hepatitis (G3; n = 1), retinopathy (G2; n = 2), nausea (G2; n = 1), constipation (G2; n = 1), and diarrhea (G2; n = 3). Conclusions: The XL888 and pembrolizumab combination had an acceptable safety profile and the RP2D of XL888 was 90 mg twice weekly combined with P 200 mg, every 3 weeks. The dose expansion portion and a robust series of immunologic correlative laboratory studies for this study is ongoing. Clinical trial information: NCT03095781.

The impact of inflammatory biomarkers, BMI, and sarcopenia on survival in advanced hepatocellular carcinoma treated with immunotherapy

553

Background: Sarcopenia and inflammation are independently associated with worse survival in cancer patients. This study aims to determine the impact of inflammatory biomarkers, BMI and sarcopenia on survival in advanced hepatocellular carcinoma (HCC) patients treated with immunotherapy. Methods: We performed a retrospective review of advanced HCC patients treated with immunotherapy-based therapies at Winship Cancer Institute between 2015 and 2019. Baseline computed tomography and magnetic resonance imaging scans were collected at mid-L3 level, assessed for skeletal muscle density using SliceOmatic (TomoVision, version 5.0) and converted to skeletal muscle index (SMI) by dividing it by height (m)2. Gender-specific sarcopenia was defined by median value of SMI. The optimal cut for continuous inflammation biomarker was determined by bias-adjusted log-rank test. Overall Survival (OS) was set as primary outcome and Cox proportional hazard model was performed. Results: 57 patients were included; 77.2% male, 52.6% Caucasian, 58.5% ECOG PS 0-1, 80.7% Child Pugh A. Treatment was second line and beyond in 71.9%. The median follow-up time was 6 months. Sarcopenia cut-off for males and females was SMI of 43 and 39, respectively. 49.1% of patients had sarcopenia. Median OS was 5 vs. 14.3 months in sarcopenic vs. non-sarcopenic patients (p=0.054). Median OS was 5 and 17.5 months in patients with BMI <25 and BMI ≥25 respectively (p=0.034). Median OS was 3.6 and 14.3 months for patients with neutrophil to lymphocyte ratio (NLR) ≥ 5.15 vs. NLR < 5.15 (p<0.001). In multivariable Cox regression model, higher baseline NLR was associated with worse OS (HR: 4.17, 1.52-11.39, p=0.005). Gender specific sarcopenia showed a trend of worse OS (HR: 1.71, 0.73-4.00, p=0.215) but was not statistically significant. BMI<25 was associated with worse OS (HR: 2.73, 1.15-6.53, p=0.023). In the association with PFS, neither baseline BMI nor gender specific sarcopenia showed statistical significance. Conclusions: Baseline BMI and NLR may predict OS after immunotherapy treatment. After controlling for baseline Child Pugh Score and NLR, gender specific sarcopenia was not associated with OS significantly.

Evaluation of Ki67 and other predictors of survival in metastatic neuroendocrine tumor (NET) to the liver treated with Y90 radioembolization

e15687

Background: Ki67 index is a predictor of survival in patients with metastatic neuroendocrine tumor (NET). The purpose of this study is to evaluate Ki67 index and other potential predictors of overall survival (OS) in patients with NET metastases to the liver treated with Y90 radioembolization. Methods: In an institutional review board-approved retrospective study, consecutive patients with NET metastases to the liver who were treated with Y90 radioembolization from 2013-2018 at a single institution were evaluated. Patients with documented Ki67 index were stratified according to 2017 World Health Organization (WHO) grading based on Ki67 index (G1: < 3%, G2: 3-20%, G3: > 20%). Age, gender, and objective tumor response on post Y90 imaging were also evaluated as potential predictors of survival after Y90. Objective tumor response was evaluated at 1 and/or 3 months post Y90 with multiphase MRI utilizing Response Evaluation Criteria for Solid Tumors (RECIST). Overall survival (OS) from time of Y90 was analyzed using Kaplan-Meier estimation. Predictors of survival were evaluated using log-rank test with p < 0.05 as the statistically significant level. SPSS software v. 25 (IBM Corporation, Armonk, NY) was used for all statistical analysis. Results: A total of 77 patients were identified; 36 (47%) had a documented Ki67 index from either their primary tumor, liver metastasis, or both. Primary tumor site included pancreatic (n = 10), small bowel (n = 7), pulmonary (n = 5), gastric (n = 3), large bowel (n = 3), and renal (n = 1). A primary site was not identified in several patients (n = 7).G1 tumors comprised 31% (n = 11) of patients, while G2 and G3 tumors made up 50% (n = 18) and 19% (n = 7) of the cohort, respectively. Median overall survival (OS) of the entire cohort was 51.1 months. Median OS in patients was 63.0 months in G1 tumors, 51.1 months in G2 tumors, and 3.1 months in G3 tumors (p < 0.001). Objective response on initial MRI follow-up after Y90 radioembolization also predicted prolonged OS (51.2 months versus 17.9 months, p < 0.001). Age at time of diagnosis and gender were not predictors of survival after Y90 radioembolization. Conclusions: WHO grading based on KI67 index and objective tumor response appear to be predictors of prolonged survival in patients with metastatic NET to the liver treated with Y90 radioembolization.

Predictors of survival in chemorefractory colorectal liver metastases treated with Y90 radioembolization

e15044

Background: The aim of the study was to evaluate molecular signatures and timing of yttrium-90 (Y90) as prognostic factors of survival following Y90 radioembolization (RE) in patients with chemorefractory colorectal liver metastases (CRLM). Methods: In a retrospective single-institution study, patients with unresectable, chemorefractory CRLM who had undergone genomic analysis prior to Y90 RE from 2013 to 2018 were identified. Timing of Y90 (following failure of first line vs following multiple lines of chemotherapy) and oncogenic mutation status (MAPK and PI3K pathways) were evaluated as predictors of survival length following Y-90. Survival was measured from first Y90 using Kaplan-Meier estimation and predictors of outcome were identified using log-rank analysis. Results: Overall, 58 patients with chemorefractory CRLM who underwent genomic analysis prior to Y-90 were identified. Of these, 26 (44.8%) had MAPK mutations, and 2 (3.5%) had PI3K mutations. Median survival after first Y-90 RE was 12.9 months. There was significantly prolonged overall survival (OS) in patients receiving Y90 immediately following failure of the first line of chemotherapy (FOLFOX ± bevacizumab) versus patients who underwent multiple lines of chemotherapy prior to Y90 (median OS of 16.9 vs 11.4 months for others; p = 0.023). When stratifying the cohort based on the timing of Y90, patients with WT-MAPK (n = 8) had prolonged OS if they underwent Y-90 following failure of the first line of chemotherapy (median OS of 22.3 vs 12.9 mo; p = 0.019). Patients with WT-PI3K (n = 19) had prolonged OS if they underwent Y90 following failure of the first line of chemotherapy (median OS of 16.9 vs 12 mo; p = 0.037). When analyzing the entire cohort non-stratified by timing of Y90, genetic mutation was not found to be a predictor of prolonged OS. Conclusions: Y90 RE after failure of first line of chemotherapy in CRLM appears to confer prolonged survival specifically for WT-MAPK and WT-PI3K oncogenes. Conversely, the prognosis of more favorable genetic status appears to be negated if Y90 RE is performed after failure of second or third line chemotherapies.

Analysis of age, tumor-sidedness, and mismatch repair (MMR) genes with response to immune checkpoint inhibitors (ICIs) in MMR-deficient (dMMR) colorectal cancer (CRC) patients (pts): A multi-institutional study

e15029

Background: ICIs induce durable responses in dMMR CRC pts with overall response rates (ORR) of 30-50%. Even though the loss of expression of any MMR gene predicts ICIs response, it is unknown if ORRs are similar across all MMR genes (MLH1, PMS2, MSH2, and MSH6). In this study, we analyzed the impact of each specific MMR gene loss and clinical characteristics of pts with best response to ICIs. Methods: Pts were eligible if they had confirmed dMMR CRC by IHC or microsatellite instability-high (MSI-H) by PCR, and received ICIs between 01/01/2012 and 10/01/2018 at Winship Cancer Institute of Emory University, Mayo Clinic or Vanderbilt University Medical Center. Due to the pattern of frequent concurrent loss and functional dependency, the groups were categorized as MLH1 ±PMS2 vs. MSH2 ±MSH6. Cox proportional hazard model and Fisher’s exact test were used for the best response and the distribution of variable among the subgroups. Results: A total of 45 pts with dMMR CRC were identified. ORRs in MLH1 ±PMS2 and MSH2 ±MSH6 groups were 68% and 57.1% respectively without statistical difference (Table). Pts with age < 50 and 50-65 years old had better ORRs compared to pts with age >65 (58.3%, 85.7% and 42.1% respectively, P=0.036). Left-sided tumors had a trend toward higher ORRs compared to right-sided tumors (83.3% vs 51.5% P=0.086). Gender and BRAF status were not predictors of response. BRAF mutations were more common in right-sided tumors (29.6% vs 11.1% respectively) and in older patients. Conclusions: Our data suggest that MSI-H CRC pts aged 50-65 treated with ICIs, have improved ORR compared to pts > 65; pts with left-sided tumors have a trend toward improved ORR compared to those with right sided tumors. [Table: see text]

Derazantinib (DZB) provides antitumor efficacy regardless of line of therapy in patients (pts) with FGFR2-fusion positive advanced intrahepatic cholangiocarcinoma (iCCA)

e15607

Background: FGFR2 fusions are prevalent in 13-22% of iCCA and known oncogenic drivers. DZB is a kinase inhibitor with potent pan-FGFR activity. In a non-comparative Phase 2a study, DZB was administered to 29 pts with FGFR2-fusion positive advanced, inoperable iCCA, either as first-line (1L) (n = 2), 2L (n = 13), 3L (n = 10), 4L (n = 2) or 5L therapy (n = 2). The objective response rate (ORR) with DZB was 21%, disease-control rate (DCR) 83% and median PFS 5.7 months (Mazzaferro et al. 2018 BJC). Data from biliary tract cancer studies suggest decreasing treatment effects of chemotherapy with increasing lines of treatment. Here, we present a post-hoc analysis of outcomes of pts treated with DZB in 1L/2L (n = 15) compared to pts treated post-2L (n = 14). Methods: Pts received 300 mg DZB QD PO. Eligibility criteria included locally confirmed, positive testing of FGFR2 fusion expression (FISH or NGS), ECOG PS 0-1. Objective responses were determined using RECIST 1.1. Disease control rate was defined as CR, PR or SD. Results: The mean age of pts treated in 1L/2L was 66y and 55y in post-2L; 73% were females in 1L/2L and 50% in post-2L treatment; other demographic variables were balanced between groups (87% vs 86% of liver target lesions, median baseline lesion size of 97.5 mm vs 109.5 mm, ECOG PS0 was 60% vs 71%). Of 15 1L/2L group pts, 12 (80%) had prior platinum-based chemotherapy as compared to all 14 pts in the post-2L group. In the 1L/2L and post-2L groups, ORR was 20% and 21%, DCR was 80% and 86%, and a reduction in sum of the largest diameter of target lesions was observed in 60% and 64% of pts, respectively. Median PFS was 5.5 mo (95% CI, 1.9-11.9) and 6.2 mo (3.6-9.2) for the 1L/2L and post-2L groups, respectively. Types of drug-related adverse events were similar in 1L/2L and post-2L. Conclusions: Anti-tumor efficacy of DZB in iCCA patients measured either by ORR, DCR, tumor shrinkage or PFS was numerically similar irrespective of treatment line. These data suggest that DZB is an effective treatment option that can be applied early in the treatment continuum of iCCA patients or at later stages to offer anti-tumor efficacy and disease control. Clinical trial information: 01752920.

Role of resection of the primary in metastatic well/intermediate-differentiated neuroendocrine tumor (NET)

e15693

Background: Resection of the primary tumor in metastatic neuroendocrine tumor (NET) is controversial. The aim of this study is to evaluate survival outcomes and identify prognostic variables of surgical resection of the primary tumor in metastatic NET patients. Methods: Data were obtained from all US hospitals that contributed to the National Cancer Database (NCDB) between 2004 and 2013. Chi-square and ANOVA tests were done to identify factors associated with surgical modality. Univariate and multivariate cox proportional hazards models were used for association between patient characteristics and survival. Kaplan-Meier curves were generated and log-rank tests conducted to compare the survival difference of patient characteristics. Results: A total of 2,361 patients between 18 and 90 years of age with stage IV well/intermediate-differentiated NET were identified. The mean age was 62.1 years (SD±13), with an equal male to female ratio (50.0%). Majority of NET primaries were in the small intestine (33.0%), pancreas (26.3%), and lung (24.4%). The majority were well differentiated tumors (69.6%) and 42.5% of patients underwent surgery at the primary site. On multivariate analysis total surgical resection of the primary (HR 0.44; 0.22-0.90; p < 0.001), female sex, year of diagnosis 2010-2014, negative surgical margin, Charlson-Deyo score < 2, and age < 51 years at diagnosis were associated with better overall survival (OS). Conclusions: Resection of the primary in stage IV well/intermediate-differentiated NET was associated with improved 5-year OS compared to patients with no surgery in small intestine (60.1% vs 44.2%), lung (70.0% vs 20.2%), and pancreas tumors (59.3% vs 30.6%).

Blood-based next-generation sequencing analysis of neuroendocrine tumors

4110

Background: Neuroendocrine tumors (NET) comprise around 2% of all malignant tumors of the gastrointestinal system. The genomic landscape of NET has not been well studied. The aim of this study was to confirm the feasibility of next generation sequencing (NGS) using ctDNA in NET and characterize common alterations in the genomic profile. Methods: Molecular alterations in 114 plasma samples from 114 patients with NET using clinical-grade NGS of ctDNA (Guardant360Ò) across multiple institutions were evaluated. The test detects single nucleotide variants in 54-73 genes, copy number amplifications, fusions, and indels in selected genes. Results: A total of 114 NET patients were evaluated, of which 64 (56.1%) were female. Mean age was 59.7 years with a range between 23-89 years. ctDNA NGS testing was performed on 114 plasma samples; 1 patient had testing performed twice. Genomic alterations were defined in 94 (n = 94/114, 82.5%) samples with a total of 289 alterations identified after excluding variants of uncertain significance (VUSs) and synonymous mutations. Alterations were identified in at least one sample from 83 patients; TP53 associated genes were most commonly altered (n = 83/289, 28.7%), followed by KRAS (n = 22, 7.6%), PI3CA (n = 15, 5.2%), CCNE1 (n = 15, 5.2%), BRAF (n = 13, 4.5%), MYC (n = 12, 4.1%), ERBB2 (n = 11, 3.8%), APC (n = 10, 3.5%), EGFR (n = 10, 3.5%), MET (n = 10, 3.5%), PTEN (n = 9, 3.1%), RB1 (n = 9, 3.1%), CDK6 (n = 7, 2.4%), AR (n = 5, 1.7%), ARID1A (n = 5, 1.7%), FGFR1 (n = 5, 1.7%), and PDGFRA (n = 5, 1.7%). Other genomic alterations of low frequency, but clinical relevance included: CDK4 (n = 4, 1.3%), NF1 (n = 4, 1.3%), RAF1 (n = 4, 1.3%), GNAS (n = 3, 1.0%), KIT (n = 3, 1.3%), BRCA2 (n = 2, 0.7%), CCND2 (n = 2, 0.7%), CTNNB1 (n = 2, 0.7%), JAK2 (n = 2, 0.7%), NRAS (n = 2, 0.7%), SMAD4 (n = 2, 0.7%), and TERT (n = 2, 0.7%). Alterations in AKT1, ALK, ATM, BRCA1, CCND1, CDKN2A, FGFR2, MTOR, RHOA, SMO and STK11 were all reported once (n = 1, 0.3%). Conclusions: Evaluation of ctDNA is feasible among individuals with NET. Liquid biopsies are not invasive and can provide personalized options for targeted therapies in NET patients.

Perioperative therapy in patients with metastatic colorectal cancer

e18231

Background: Knowledge about perioperative systemic therapy in metastatic colorectal cancer (mCRC) is limited. We aim to describe the nationwide pattern of use and survival outcomes of patients with mCRC treated with surgical resection. Methods: Data were obtained from all US hospitals that contributed to the National Cancer Database (NCDB) between 2004 and 2013. Univariate and multivariate analyses was done to identify factors associated with patient outcome. Results: A total of 61,940 patients with stage IV CRC older than 18 years were identified. Mean age was 63.4 years (SD±14), with a male preponderance (54.8%). About 80% were Caucasian and 69.9% had colon cancer. Compared to medical treatment only, resection of both primary and metastatic sites (13.5%; HR 0.40; 0.37-0.44; p < 0.001), or primary site resection alone (49.2%; HR 0.52; 0.48-0.56; p < 0.001) were associated with improved overall survival (OS). Other co-variates associated with improved survival included younger age group, year of diagnosis (2009-2013), colon tumor location, and < 3 metastatic sites (Table). Five-year OS for resection of primary and metastatic site (28.2%) was higher than for primary site resection alone (14.9%) or no surgical treatment (4.7%). Conclusions: Resection of metastatic sites or primary tumor was associated with improved survival in patients with stage IV CRC.[Table: see text]

Genomic alterations in appendiceal carcinoma using circulating DNA

658

Background: Appendiceal cancers (AC) comprise around 0.5% of all gastrointestinal neoplasia. The genomic landscape of AC has not been well studied. The yield of circulating tumor DNA (ctDNA) from the plasma of patients with AC has not been reported. The aim of this study is to confirm the feasibility of NGS using ctDNA and characterize common alternations in the genomic profile of AC. Methods: The molecular alterations in 372 plasma samples from 303 patients with AC using clinical-grade NGS of ctDNA (Guardant 360) across multiple institutions, was evaluated. The test detects single nucleotide variants in 54 -73 genes, copy number amplifications, fusions, and indels in selected genes. Results: A total of 303 AC patients were evaluated; 169 female (56%). Median age was 56.8 (range: 25-83). ctDNA NGS testing was done on 372 plasma samples; 48 patients had testing performed twice, 9 three times, and 1 was tested four times. Genomic alterations were defined in 207 (55.6%) samples with a total of 288 alterations identified after excluding variants of uncertain significance (VUSs) and synonymous mutations. TP53 associated genes were most commonly altered (n = 96, 33.3%), followed by KRAS (n = 41, 14.2%), APC (n = 19, 6.6%), EGFR (n = 15, 5.2%), BRAF (n = 13, 4.5%), NF1 (n = 13, 4.5%), MYC (n = 9, 3.1%), GNAS (n = 8, 2.7%), PI3CA (n = 7, 2.4%), MET (n = 6, 2.08%), ATM in 6 (1.6%). Other genomic alterations of low frequency, but clinically relevant: AR (n = 4, 1.39%), TERT (n = 4, 1.39%), ERBB2 (n = 4, 1.39%), SMAD4 (n = 3, 1.04%), CDK4 (n = 2, 0.69%), NRAS (n = 2, 0.69%), FGFR1 (n = 2, 0.69%), FGFR2 (n = 2, 0.69%), PTEN (n = 2, 0.69%), RB1 (n = 2, 0.69%), and CDK6, CDKN2A, BRCA1, BRCA2, JAK2, IDH2, MAPK, NTRK1, CDH1, ARID1A, and PDGFRA were all reported once. Conclusions: Evaluation of ctDNA was feasible among individuals with AC. The frequency of genomic alterations in ctDNA testing is similar to those previously reported in tissue NGS. Liquid biopsies are non-invasive methods that can provide personalized options for targeted therapies in patients with AC.

Impact of genomic alterations (GAs) on outcomes and their distribution by age groups in metastatic colorectal cancer (mCRC) patients (pts)

560

Background: Although clinical outcomes has substantially improved over the last decade, long term survival in mCRC remains rare. Molecular profiling (MP) of CRC is routinely conducted to identify potential therapeutic targets. The aim of this project is to evaluate the impact of uncommon GAs on outcomes and characterize their distribution by age. Methods: Pts were eligible if they had mCRC (synchronous or metachronous) and underwent MP between 01/2013 and 05/2018. GAs were obtained from Foundation Medicine reports. Clinical data were collected by trained personnel by detailed chart review. Multivariable survival analyses (MA) with Cox model were conducted for survival outcomes and Fisher’s exact test was used to assess the differences among age groups (< 45, 45-60, > 60). The study was reviewed by Institutional IRB ( IRB00097021 ). Results: 161 patients with mCRC had MP, and 159 of those patients had survival data. The most commonly detected GAs were APC (133/161, 82.6%) TP53 (128/161, 79.5%). In univariate analyses mutations in BRCA 1/2, RB1, SOX9, CDK8, FLT3, and IRS2 amplification were associated with worse survival outcomes. In MA, including initial stage of disease, GAs in BRCA1/2, RB1, FLT3, SOX9, and IRS2 remained statistically significant (Table). When we performed MA by age groups, mutations SOX9 in age group < 45 and BRCA1/2 in age group 45-60 were significant predictors of worse outcomes. We also compared the frequency of mutations among age groups and FAM123B was significantly more common in age group 45-60 (P = 0.038). Conclusions: Our data suggests that GAs in BRCA1/2, RB1, FLT3, SOX9, and IRS2 may predict worse outcomes in mCRC. Therapeutic approaches targeting these pathways should be investigated. Differences in distribution and prognostic significance of mutations were observed based on age. [Table: see text]

Phase II randomized, double-blind, study of mFOLFIRINOX plus ramucirumab versus mFOLFIRINOX plus placebo in advanced pancreatic cancer patients hcrn GI14-198

TPS475

Background: The prognosis of pancreas adenocarcinoma (PCA) remains poor. A chemotherapy backbone is the current standard of care in PCA. The choice of a chemotherapeutic backbone may impact the efficacy of anti-angiogenic therapy in PCA. Ramucirumab has increased activity with fluoropyrimidines (5FU) because 5FU increases VEGF expression. Methods: This is a double-blind, placebo controlled Phase II study. Subjects will receive either Ramucirumab or a placebo followed by mFOLFIRINOX every two weeks of a 28 day cycle until progression or discontinuation for other reasons. The primary endpoint of this clinical trial is nine month PFS defined as the time from enrollment to the time of progression or death. Among the key inclusion criteria, subjects must have recurrent or metastatic pancreas adenocarcinoma (PCA) with no prior first-line systemic treatment, ECOG PS 0-1, adequate organ function, no DVT, PE or other thromboembolism within three months of randomization. Total number of patient enrolled as of September 19, 2018 is 48 of 85 at eight sites; 27 male (56%), 42 Caucasians (87.5%), three African American (6.2%), one Asian (2%). Median age is 63 (40 - 76). Majority of patients (41) had de novo metastatic disease and six with recurrent disease after surgery. Regimen has been tolerated well with no unanticipated events. Clinical trial information: NCT02581215.

Clinical outcomes of hepatocellular carcinoma variants compared to hepatocellular carcinoma

435

Background: There is no consensus regarding treatment for HCC variants. Clinical outcomes of HCC variants differ from pure HCC. The aim of this study is to compare clinicopathological characteristics, treatment, and outcomes of HCC variants with pure HCC. Methods: Patients with HCC and variants with 8170/3-8175/3 and 8180/3 ICD-O-3 codes were identified from National Cancer Database between 2004 and 2013. Univariate and multivariate survival analyses were conducted to analyze the association between histology and overall survival (OS). Results: 80,280 patients were identified; pure HCC 78,461 (97.7%), fibrolamellar (FLHCC) 310 (0.4%), scirrhous 161 (0.2%), spindle cell 72 (0.1%), clear cell 487 (0.6%), pleomorphic 23 (0.0%), and combined HCC and cholangiocarcinoma (mixed HCC) 766 (1.0%). 76.7% were male and 72% Caucasian. The mean age was similar in all except FLHCC (37.9 vs. 60.9-64.1 years, p < 0.001). Liver transplant was performed in 10.1% of pure HCC, 14.5% of mixed HCC, 16.2% of scirrhous, 6.9% of spindle cell, 8.8% of clear cell, 8.7% of pleomorphic, and 3.2% of FLHCC (p < 0.001). Pure HCC (10.57%) underwent surgical resection less often than variants; FLHCC (54.8%), clear cell (34.5%), mixed HCC (29.8%), spindle cell (33.3%), pleomorphic (34.8%), and scirrhous (9.9%) (p < 0.001). Ablation was performed in 9.8% of pure HCC, and in up to 8.7% of HCC variants. More than a third of all patients received chemotherapy; pure HCC (42.3%), mixed HCC (38.5%), scirrhous (31.1%), spindle cell (36.1%), clear cell (35.5%), pleomorphic (34.8%), and FLHCC (41.3%). FLHCC had the best 5-year OS (38.7%), spindle cell and pleomorphic had the worst (9.6% and 13.0%). In univariate and multivariate analyses, fibrolamellar histology, female sex, diagnosis between 2009 and 2013, treatment at academic center, well/moderately differentiated histology, early stage, and chemotherapy was associated with better OS compared to pure HCC, male sex, diagnosis between 2004 and 2008, treatment at community cancer program, poorly differentiated, late stage, and no chemotherapy (p < 0.001). Conclusions: HCC variants underwent surgical resection more often than HCC. FLHCC had the best 5-year OS. Liver transplant is commonly performed in HCC variants.

Altering the cytokine profile in the pancreatic cancer microenvironment with heat shock protein-90 inhibitors to enhance immunotherapy

One feature of pancreatic ductal adenocarcinoma (PDAC) is the fibrotic stroma, which is a result of interactions between tumor cells, immune suppressive cells and pancreatic stellate cells (PSC). Stroma-derived cytokines including IL-6 represent a significant barrier restraining immunity against PDAC. Targeting pathways in the tumor microenvironment (TME) that regulate cytokine secretion can be an innovative means to improve access of effector immune cells to PDAC. We hypothesize that targeting Hsp90 can modulate the TME, through its ability to target inflammatory signaling and cytokine production by PSC and enhance the efficacy of immunotherapy. The Hsp90 inhibitor XL888 led to a concentration-dependent inhibition of IL-6 and MCP-1 production by PSC in vitro. Concurrently, PSC displayed growth inhibition as determined by MTT assay, and reduced alpha-SMA expression, a marker of cell activation. Inhibition of Jak/STAT and MAPK signaling was confirmed via immunoblot. These data compliment our published results indicating that Hsp90 inhibitors modulate survival pathways and PDAC growth in vitro and in vivo. This has led to an investigator-initiated Phase Ib/II clinical trial of XL888 (Hsp90i) and pembrolizumab (anti-PD-1) now accruing at our institution. Extensive laboratory correlative studies are ongoing to examine the impact of Hsp90 inhibition on anti-PD-1 mediated T cell proliferation, cytokine production, and impact on PSC-derived cytokine signatures. This work is being performed in paired biopsy samples (pre- and on-treatment), and peripheral blood from expansion cohorts of patients with metastatic pancreatic cancer (n=16) or colorectal cancer (n=16).

Resection of pancreatic cancer following induction chemotherapy

406

Background: Survival of resectable pancreas cancer (RPC) treated with resection and adjuvant therapy is 22-28 months (mo). Locally advanced unresectable pancreatic cancer (LAPC) treated with combination chemotherapy have a median survival of 24 mo. The objective of this project is to evaluate the effect of neoadjuvant treatment on survival outcome of localized PC. Methods: Charts of localized PC patients treated at Emory University from 2009 to 2016 were reviewed. Information on demographics, stage and treatment was collected. Survival rates were estimated by Kaplan-Meier method and compared with log-rank test. A Cox proportional hazard model was fitted to estimate the adjusted effect of treatment on overall survival(OS). Results: A total of 415 patients were included; 144 RPC, 158 borderline resectable (BRPC) and 108 LAPC. Stage was determined at the multidisciplinary conference. The median age was 67.7 years (30-92); 49% male, and 63% Caucasians. The median OS for RPC, BRPC, and LAPC was 16.9, 14.6 and 10.9 mo, respectively. Stage, type of chemotherapy and age were significant predictors of OS after adjusting for gender, race, age, surgery, stage, chemotherapy, margins and radiation. Of the 144 RPC, 137 underwent surgery and 3 received neoadjuvant treatment; 73 RPC were followed in outside facility with missing follow up data. Of the 71 RPC treated at Emory; 91% received adjuvant gemcitabine. Of the 158 BRPC, 84 underwent surgery; 44 received FOLFIRINOX neoadjuvant therapy, 23 received gemcitabine/nab-paclitaxel, and 16 received gemcitabine single agent. BRPC patients who underwent resection had a median OS of 18.5 mo (95%CI: 14.2, 26.4), significantly longer than RPC (P = 0.044). Combination chemotherapy was significantly associated with improved OS at 36 mo (38.9%) when compared to single agent gemcitabine (6.3% at 36 mo) (p = 0.009). BRPC patients who received FOLFORINOX and surgery had a median OS of 31.5 mo. Conclusions: Overall survival of BRPC patients who undergo resection after FOLFIRINOX is significantly improved (more than doubled) compared to upfront resection for RPC. Preoperative therapy provides the best approach for systemic disease early in the course of treatment.

Phase Ib trial of pembrolizumab and XL888 in patients with advanced gastrointestinal malignancies

TPS526

Background: Heat shock protein 90 (HSP90) has a central role in modulating tumor microenvironment, inflammatory signaling pathways (NF-κB, HIF-1α and Jak-STAT), tumor antigen presentation and expression, PD-L1 expression and macrophage migration inhibitory factor (MIF) as well as cytokine production. Inhibitors of HSP90 have been shown in preclinical studies and in patient samples to exert anti-tumor effects and modulate signaling pathways. XL888 is a selective inhibitor of HSP90. Based on this preclinical rationale, we have developed a phase Ib/II trial to determine the recommended phase II dose, evaluate the safety, toxicity profile, preliminary antitumor activity, and immunogenicity of the XL888 and Pembrolizumab combination in previously treated patients with advanced gastrointestinal tumors. Methods: The phase Ib trial design is standard 3+3. XL888 is administered orally (PO) in three dose levels of 45 (DL1), 90 (DL2), 60 (only if DLT on DL2) mg twice weekly with pembrolizumab at 200 mg IV on day 1, in 21-day cycles. Eligible patients must have stage IV or locally advanced unresectable gastrointestinal adenocarcinomas who have failed at least one prior therapy (patients with colorectal adenocarcinoma must have previously received oxaliplatin, irinotecan, and fluoropyrimidine), age ≥18 years, ECOG PS 0-1, no prior anti PD-1 or anti-PD-L1 agent. After recommended phase II dose is established, an expansion phase will enroll 16 patients with pancreatic adenocarcinoma (Arm A) and 16 patients with colorectal adenocarcinoma (Arm B). Primary endpoint response rate. In the expansion phase patients will receive initial cycle (3 weeks) treatment with either pembrolizumab or pembrolizumab plus XL888 and then starting cycle 2 all patients receive the combination. Blood will be collected pre-treatment, post 1st and 2nd cycle. Eight patients in each arm will undergo pre and post treatment tumor biopsies. This design will enable us to evaluate the effects of pembrolizumab alone versus the combination. This study was activated in June 2017 and to date 4 patients were enrolled in dose escalation phase. The dose expansion phase is expected to start accrual in December 2017. Clinical trial information: NCT03095781.

Epidemiology and treatment of high-grade gastrointestinal neuroendocrine tumors (HG-GI-NETs)

421

Background: High grade neuroendocrine tumors of the gastrointestinal tract are rare tumors. Management strategies are modeled after small cell lung cancer (SCLC). Treatment patterns and outcomes have not been studied. Methods: Data were obtained from all US hospitals that contributed to the National Cancer Database (NCDB) between 2004 and 2013. Univariate and multivariate testing was done to identify factors associated with patient outcome. Kaplan-Meier analysis and Cox proportional hazards models were used to assess the association between patient characteristics and survival. Results: A total of 1,861 patients were identified for the 10 years of the study. The mean age was 63 years (SD±13), with a male preponderance (53.3%). The vast majority of patients (78.1%) were non-Hispanic Whites. The most common primary sites were pancreas (PNET = 19.4%), large intestine (18.1%), esophagus (17.8%) and rectum (15.5%). About 27.9% of the cases were resectable at the time of diagnosis, and distribution across stages 1-IV was 6.6%, 10.5%, 18% and 64.6% consecutively. Liver was a common site of metastases (50.4%), followed by bone (11.3%) and lungs (10.8%). Only 3.5% of the patients had brain metastases. On univariable analysis, age < 65years (HR 0.72; 0.66-0.8; p < 0.001) and treatment at an academic center (HR 0.88; 0.79-0.99; p < 0.034) were associated with improved survival. Multivariable analysis confirmed prognostic advantage of treatment at an academic center. Patients treated with chemotherapy had a median overall survival (OS) of 11.2 months, compared with 1.7 months for those who did not. The median OS for high grade PNET was 6 months, compared to 9.9 months for other HG-GI-NETs. One year and 5-year survival rates were 27.5% vs. 41% and 4.5% vs. 12.3% respectively. Conclusions: This is the largest series of HG-GI-NET. Almost two-thirds of the cases present with metastatic disease. Pattern of metastasis differs from SCLC. Survival is short. Treatment at high volume academic center, younger age and use of chemotherapy are associated with improved survival.

A Phase 1 study of ARQ 087, an oral pan-FGFR inhibitor in patients with advanced solid tumours

Background:

ARQ 087 is an orally administered pan-FGFR inhibitor with multi-kinase activity. This Phase 1 study evaluated safety, pharmacokinetics, and pharmacodynamics of ARQ 087 and defined the recommended Phase 2 dose (RP2D).

Methods:

Patients with advanced solid tumours received ARQ 087 administered initially at 25 mg every other day and dose-escalated from 25 to 425 mg daily (QD) continuous dosing. FGF19, 21, 23, and serum phosphate were assessed as potential biomarkers of target engagement.

Results:

80 patients were enrolled, 61 in dose-escalation/food-effect cohorts and 19 with pre-defined tumour types in the expansion cohort. The most common ARQ 087-related adverse events were fatigue (49%), nausea (46%), aspartate aminotransferase (AST) increase (30%), and diarrhoea (23%). Four patients (5%) experienced grade 1 treatment-related hyperphosphataemia. Dose-limiting toxicity was reversible grade 3 AST increase. The RP2D was 300 mg QD. Pharmacokinetics were linear and dose-proportional from 25 to 325 mg QD, and were unaffected by food. Statistically significant changes (P-value<0.05) suggest phosphate and FGF19 levels as markers of target engagement. In 18 evaluable patients with FGFR genetic alterations, 3 confirmed partial responses (two intrahepatic cholangiocarcinomas (iCCA) with FGFR2 fusions and one urothelial cancer with FGFR2 and FGF19 amplification) and two durable stable disease at ⩾16 weeks with tumour reduction (FGFR2 fusion-positive iCCA and adrenocortical carcinoma with FGFR1 amplification) were observed.

Conclusions:

ARQ 087 had manageable toxicity at the RP2D of 300 mg QD, showed pharmacodynamics effects, and achieved objective responses, notably in patients with FGFR2 genetic alterations.

A phase Ib/II study of cancer stemness inhibitor napabucasin (BBI-608) in combination with gemcitabine (gem) and nab-paclitaxel (nabPTX) in metastatic pancreatic adenocarcinoma (mPDAC) patients (pts)

4106

Background: Cancer stem cells are fundamentally important for resistance to therapy, recurrence and metastasis. Napabucasin is a first-in-class cancer stemness inhibitor in development identified by its ability to inhibit STAT3-driven gene transcription and spherogenesis of cancer stem cells (Li et al, PNAS 112(6):1839, 2015). Preclinical studies suggest that napabucasin sensitizes heterogeneous cancer cells to chemotherapy and targeted agents. Methods: A phase Ib/II multi-center study in mPDAC pts was performed to confirm the RP2D, PK profile and evidence of anticancer activity of napabucasin in combination with nabPTX and Gem. Pts received napabucasin 240 mg BID with weekly nabPTX 125 mg/m2 and gem 1000 mg/m2for 3 out of every 4 weeks until disease progression (PD) or other discontinuation criterion. Results: Of 71 intent to treat (ITT) pts enrolled, 49 (69%) were treatment-naïve and 22 (31%) received neoadjuvant treatment. There were no significant PK interactions, dose-limiting or unexpected toxicities. Most common adverse events (AEs) included grade 1 diarrhea/cramping, nausea and fatigue with grade 3 AEs noted in 12 pts: fatigue (8), electrolyte imbalance (2), diarrhea (1), dehydration (1), nausea (1) and weight loss (1). Among pts who received RECIST evaluation (60), disease control (DCR; CR+PR+SD) was observed in 55 (92%), with 1 CR (2%) and 26 PR (43%) (31 - 78% regression). Of 11 pts with non-evaluable disease, treatment stopped due to compliance (4), consent withdrawal (3), clinical PD (1), toxicity (1), insurance (1) and death (1). Among 71 ITT pts, DCR was observed in 55 (77%), with 1 CR (1.4%) and 26 PR (37%). Maturing median progression free survival and overall survival (OS) in ITT pts is >7.1 and >10.4 m, respectively. Conclusions: This study showed that napabucasin can be combined with nabPTX and gem, with encouraging signs of efficacy in mPDAC now being confirmed in a phase 3 study. Clinical trial information: NCT02231723. [Table: see text]