Search

Background Venous thromboembolism (VTE) can be the first manifestation of cancer; however, the current incidence of malignancy in unselected patients with first unprovoked VTE needs to be confirmed. Material and methods Between March 1st, 2013 and February 28th, 2015 we included and followed-up all patients living in the Brest district, France, who were seen in hospitals or the community for a first symptomatic unprovoked VTE event. The primary study outcome was the one-year incidence of cancer. Results 526 patients, mean age 66.6 ± 18.1 years, 246 (46.8%) men, were included in the study. In the year following VTE, 26 patients were diagnosed with cancer, corresponding to a one-year cumulative incidence of cancer of 5.06% (95% CI 3.47–7.35). Age ≥60, smoking and pulmonary embolism were significantly associated with cancer diagnosis in multivariate analysis. Fifty percent of cancers were patent at the time of VTE diagnosis, mostly detected on CTPA (Computed Tomographic Pulmonary Angiography) performed for pulmonary embolism assessment. After excluding patients with patent cancer at VTE diagnosis, the one-year incidence of cancer was 2.65% (95% CI: 1.55–4.52); in multivariate analysis, only current smoking was independently associated with a significant 5.4-fold increased risk for cancer diagnosis (HR 5.40; 95% CI 1.31–22.27). No cancer was diagnosed in patients aged 50 years or younger. Conclusion The one-year incidence of cancer after a first unprovoked VTE was 5.06%. Half of the cancers were diagnosed during the diagnosis procedure for pulmonary embolism using CTPA.

Background In 2008, a study of the characteristics of hospitalised patients led to the development of a prognostic tool that distinguished three populations with significantly different 2-month survival rates. The goal of our study aimed at validating prospectively this prognostic tool in outpatients treated for cancer in terminal stage, based on four factors: performance status (ECOG) (PS), number of metastatic sites, serum albumin and lactate dehydrogenase. Patients and methods PRONOPALL is a multicentre study of current care. About 302 adult patients who met one or more of the following criteria: life expectancy under 6 months, performance status ≥ 2 and disease progression during the previous chemotherapy regimen were included across 16 institutions between October 2009 and October 2010. Afterwards, in order to validate the prognostic tool, the score was ciphered and correlated to patient survival. Results Totally 262 patients (87%) were evaluable (27 patients excluded and 13 unknown score). Median age was 66 years [37–88], and women accounted for 59%. ECOG PS 0–1 (46%), PS 2 (37%) and PS 3–4 (17%). The primary tumours were: breast (29%), colorectal (28%), lung (13%), pancreas (12%), ovary (11%) and other (8%). About 32% of patients presented one metastatic site, 35% had two and 31% had more than two. The median lactate dehydrogenase level was 398 IU/l [118–4314]; median serum albumin was 35 g/l [13–54]. According to the PRONOPALL prognostic tool, the 2-month survival rate was 92% and the median survival rate was 301 days [209–348] for the 130 patients in population C, 66% and 79 days [71–114] for the 111 patients in population B, and 24% and 35 days for [14–56] the 21 patients in population A. These three populations survival were statistically different (P Conclusion PRONOPALL study confirms the three prognostic profiles defined by the combination of four factors. This PRONOPALL score is a useful decision-making tool in daily practice.

5-Fluorouracil (5-FU)–based treatments can lead to early-onset severe (4%–5%) even fatal (0.3%) toxicities in patients with dihydropyrimidine dehydrogenase (DPD) deficiency. This multicenter prospective cohort study aimed to assess the clinical benefit of pretherapeutic screening for DPD deficiency using a multiparametric approach. Two parallel cohorts of patients treated with 5-FU–based chemotherapy for colorectal carcinoma were compared in a prospective nonrandomized study. In arm A, patients had DPD deficiency screening before treatment, whereas in arm B no pretherapy screening was performed. Dosing was based on 5-FU administration guidelines of each institution. DPD deficiency screening was performed using a combined multiparametric approach (5-FUODPM Tox). The frequency of early grade 4–5 toxic events potentially induced by 5-FU was compared in the two groups. At total of 1,142 patients (n = 1,116 evaluable) were enrolled. In arm A, out of 718 evaluable patients, nine grade 4 early toxicities potentially related to 5-FU were reported in nine patients (1.2%) with no toxic death despite one complete DPD deficiency and 24 partial deficiencies. The 24 patients with partial deficiency had safe pharmacokinetics (PK)-monitored 5-FU. In arm B, among 398 evaluable patients, 17 grade 4–5 toxic early events potentially related to 5-FU were reported in 12 patients (4.2%). The incidence of early severe toxicity was significantly higher in arm B (P = .0019), confirming the positive impact of pretherapeutic DPD assessment. The percent of patients with a toxicity grade 3 or higher observed in arm A was 10.8% (n = 78) compared to 17.55% (n = 69) in arm B (P = .0497). The percentage of death was reduced from 2.5/1,000 in arm B to 0 in arm A. The time to occurrence of all grade ≥3 toxicities was determined in both arms and the difference between the two arms was significant (P = .047). Overall, one patient with complete DPD deficiency confirmed retrospectively died within 13 days from grade 5 multivisceral toxicity. Enrollment was prematurely closed after external experts’ decision. In conclusion, multiparametric pretherapeutic DPD deficiency screening significantly lowered the risk of early severe toxicity and avoided an early toxic death. This approach should be used for safe administration of 5-FU–based treatments.

Triplet chemotherapy (FOLFOXIRI) + bevacizumab regimen is indicated as first-line treatment of BRAF-mutated metastatic colorectal cancer (mCRC). Nevertheless, its proven therapeutic efficacy in clinical trials was solely based on partial morphologic responses assessed by CT. To date, only 1 case of complete response assessed by FDG PET/CT was reported in literature in BRAF-mutated mCRC, but treated with doublet chemotherapy (FOLFIRI) + cetuximab regimen. We report a complete metabolic response assessed by FDG PET/CT, maintained over time (13 months) in a 60-year-old woman with BRAF-mutated mCRC treated by FOLFOXIRI-bevacizumab. This also confirms that FDG PET/CT is emerging as a useful approach for therapeutic assessment of targeted drugs in mCRC.

Background Regorafenib prolonged overall survival (OS) versus placebo in patients with treatment-refractory metastatic colorectal cancer (mCRC) in phase III trials. We conducted an observational study of regorafenib for patients with mCRC in real-world clinical practice. Methods The international, prospective, CORRELATE study recruited patients with mCRC previously treated with approved therapies, for whom the decision to treat with regorafenib was made by the treating physician according to the local health authority approved label. The primary objective was safety, assessed by treatment-emergent adverse events (TEAEs; National Cancer Institute Common Terminology Criteria for Adverse Events, version 4.03). Results A total of 1037 patients were treated. The median age was 65 years (range: 24–93); 87% of patients had Eastern Cooperative Oncology Group performance status 0–1, 56% of patients had KRAS, 7% had NRAS and 4% had BRAF mutations. The initial regorafenib dose was 160 mg/day in 57% of patients. The most common grade III or IV drug-related TEAEs were fatigue (9%), hand–foot skin reaction (7%) and hypertension (6%). Drug-related grade V (fatal) TEAEs occurred in 1% of patients. Dose reductions for drug-related TEAEs occurred in 24% of patients. Median OS was 7.7 months (95% confidence interval [CI]: 7.2–8.3), and median progression-free survival (PFS) was 2.9 months (95% CI: 2.8–3.0). Conclusions In this real-world, observational study of patients with mCRC, the regorafenib toxicity profile was similar to that reported in phase III trials. The starting dose for almost half of patients was less than the approved 160-mg dose, and the median OS and PFS were in the range observed in phase III trials. Trial registration: NCT02042144.

e16063 Background: One accepted treatment for patients (pts) with advanced HER2-negative G/GEJ is mFOLFOX6 (5-FU, folinic acid, oxaliplatin). Despite treatment options, 5-year survival is poor, and limited biomarkers exist to inform treatment selection. Claudin 18.2 (CLDN18.2), a tight junction protein normally confined to gastric mucosa of healthy tissue, is often retained in G/GEJ. Zolbetuximab, a chimeric IgG1 monoclonal antibody, binds to CLDN18.2 and mediates cancer cell death via antibody-dependent cellular cytotoxicity (ADCC) and complement-dependent cytotoxicity (CDC). Nonclinical results showed that cytotoxic chemotherapy increased CLDN18.2 expression, improving ADCC/CDC activity of zolbetuximab. Phase 2 results (NCT01630083, FAST) showed prolonged survival with zolbetuximab+EOX (epirubicin, oxaliplatin, capecitabine) vs EOX in G/GEJ. This study assessed antitumor activity and safety/tolerability of first-line zolbetuximab+mFOLFOX6 (Cohort 2 in NCT03505320) in G/GEJ with high CLDN18.2 expression. Methods: Cohort 2 of this multicohort study enrolled adult pts with metastatic or locally advanced unresectable G/GEJ. Patients had measurable disease (RECIST v1.1), HER2-negative disease, and high CLDN18.2 expression (≥75% of tumor cells demonstrating moderate-to-strong membranous staining by central IHC testing). Patients received zolbetuximab 800 mg/m2 IV on Cycle 1 Day 3 then 600 mg/m2 Q3W. Zolbetuximab and mFOLFOX (Q2W from Cycle 1 Day 1) were administered in 42-day cycles. Key endpoints were safety/tolerability and objective response rate by independent central review (ORRICR RECIST v1.1). Results: As of Jan 26, 2021, Cohort 2 had enrolled 21 pts; median age was 63 years (range, 36-74), 57% were male, 43% were Asian, and 38% were white. Of 19 evaluable pts, 12 had confirmed partial responses; ORRICR was 63.2% (95% CI: 38.4-83.7) (Table). Median PFS was 13.7 months (95% CI: 7.4-not estimable); 12-month PFS rate was 58%. Common adverse events (AEs) were nausea (90.5%; grade 3, 4.8%) and vomiting (61.9%; grade 3, 9.5%). Common grade 3/4 AEs were decreased neutrophil count (33.3%) and neutropenia (28.6%). There were no fatal AEs. Conclusions: Results suggest promising antitumor activity with zolbetuximab+mFOLFOX6 in metastatic or locally advanced G/GEJ. The safety profile was manageable and no new safety signals were identified. Clinical trial information: NCT03505320. [Table: see text]

3109 Background: A subset of gastric cancers exhibits platinum sensitivity and genomic instability that is characteristic of homologous recombination deficiency (HRD). Cells with HRD are sensitive to poly (ADP-ribose) polymerase (PARP) inhibition. PARP inhibitor maintenance therapy following platinum-based chemotherapy has been a successful treatment strategy in pts with ovarian cancer. Pamiparib is an orally administered selective PARP protein 1 and 2 (PARP1/2) inhibitor that has shown potent DNA-PARP trapping activity and crosses the blood brain barrier in preclinical studies. In early phase clinical studies (NCT02361723; NCT03333915), pamiparib showed an acceptable safety profile and promising antitumor activity. PARALLEL 303 compared the efficacy and safety of pamiparib vs placebo as maintenance therapy in pts with inoperable locally advanced or metastatic gastric cancer that responded to platinum-based 1L chemotherapy. Methods: The primary endpoint of this double-blind, randomized, global phase 2 study (NCT03427814) was progression-free survival (PFS) as determined by the investigator per RECIST Version 1.1. Key secondary endpoints included time to subsequent treatment, objective response rate, duration of response, time to response, overall survival (OS) and safety. At the time of this analysis, OS data were immature due to the short duration of study. Data presented here will focus on PFS and safety. Results: 136 pts were randomized 1:1 to receive pamiparib 60 mg orally (PO) twice daily (BID) (n=71) or placebo PO BID (n=65) in 28-day cycles. The median PFS was longer with pamiparib vs placebo, but did not reach statistical significance (3.7 months; 95% CI, 1.94–5.26 vs 2.1 months; 95% CI, 1.87–3.75 months); hazard ratio 0.799 (95% CI, 0.534–1.193; P=0.1428). Treatment-emergent adverse events (TEAEs) of ≥ Grade 3 were experienced by 29 pts (40.8%) in the pamiparib arm, and 20 pts (30.8%) in the placebo arm. The most common TEAEs of ≥ Grade 3 were blood and lymphatic system disorders in the pamiparib arm, and gastrointestinal disorders in the placebo arm. TEAEs leading to treatment discontinuation were: 8 pts (11.3%) in the pamiparib arm and 2 pts (3.1%) in the placebo arm. TEAEs leading to death were: 2 pts (2.8%; 1 pneumonia, 1 unexplained) in the pamiparib arm, and 2 pts (3.1%; 1 hepatic rupture, 1 sepsis) in the placebo arm. Conclusions: Although pamiparib did not meet statistical significance for superiority vs placebo for its primary endpoint, it was generally well tolerated with few treatment discontinuations due to TEAEs. No new safety signals were identified with pamiparib, and its safety profile was consistent with that of other PARP inhibitors. Clinical trial information: NCT03427814.

4096 Background: Regorafenib (R) has shown promising efficacy in patients (pts) with BTC refractory to standard chemotherapy. Anti-PD1/PD-L1 antibodies have only limited clinical activity. Synergy between R and anti–PD-1/PD-L1 antibodies has been shown in pre-clinical solid tumor models. Methods: This is a single-arm open-label multicentric phase II trial (Bayesian adaptive design) assessing the efficacy and safety of R (160 mg QD 3weeks/4) + avelumab (A) (10 mg/kg every 2 weeks) combination in BTC pts. The primary endpoint was the objective response rate under treatment, based on central review according to RECIST 1.1. Secondary endpoints included: 1-year progression free survival (PFS), 1-year overall survival (OS), and Safety using NCI-CTCAE v5.0. Correlative studies were planned from pts tumor samples obtained at baseline. Results: Between Nov. 2018 and Nov. 2019, 34 BTC pts were enrolled in 4 centers. Median age was 63 (range 36 – 80). Median follow-up was 9.8 months. Median number of previous treatment lines for metastatic or locally advanced disease was: 2 (range 1 – 4). Twenty-nine (85.3%) pts experienced at least 1 dose modification or treatment interruption of R or A due to an adverse event (AE) related to the treatment. The most common grade 3/4 AEs were : Hypertension (17.6%), Fatigue (14.7%), and maculo-papular rash (11.8%). No death was related to the treatment. Among the 29 pts with at least one imaging tumor assessment, 4 (13.8%) achieved a partial response, and 11 (37.9%) demonstrated stable disease including 10 (34.5%) pts with tumor shrinkage. Fourteen pts (48.3%) had progressive disease. The median PFS and OS were 2.5 months (95%CI 1.9 – 5.5) and 11.9 months (95%CI 6.2 – NA) respectively. Baseline tumor samples were available for 27 pts. High IDO and PD-L1 expression at baseline was associated with better outcome. Conclusions: The R+A combination is associated with significant anti-tumor activity with promising survival rates in this heavily pre-treated population. Full Biomarkers analyses will be presented at the meeting. Clinical trial information: NCT03475953.

168 Background: In the randomized, international, double-blind, placebo-controlled KEYNOTE-590 (NCT03189719) study, pembrolizumab (pembro) + chemotherapy (chemo) provided statistically significant and clinically meaningful improvement in OS, PFS, and ORR vs placebo + chemo as first-line therapy for patients (pts) with locally advanced/unresectable or metastatic adenocarcinoma or esophageal squamous cell carcinoma (ESCC) or Siewert type 1 esophagogastric junction adenocarcinoma (EGJ). Here we report HRQoL outcomes in KEYNOTE-590. Methods: 749 pts were randomized 1:1 to pembro 200 mg or placebo Q3W for up to 2y + chemo (cisplatin 80 mg/m2 Q3W [d1; 6 doses] + 5-FU 800 mg/m2 on d1-5 Q3W). EORTC QLQ-C30, EORTC QLQ-OES18, and EQ-5D-5L questionnaires were administered at baseline, every 3 weeks (Q3W) up to week 24, and then Q9W up to 1 year or end of treatment, and at the 30-d safety follow-up visit. HRQoL was assessed in all treated patients who completed ≥1 HRQoL assessment (N = 711: 356 for pembro + chemo; 355 for chemo). Change from baseline to week 18 in EORTC-QLQ-C30 global health status (GHS)/QoL and physical functioning, and in QLQ-OES18 scores were prespecified secondary endpoints. Change from baseline to week 18 in EQ-5D scores was an exploratory endpoint. Time to deterioration (TTD) was evaluated for all endpoints. Least square mean (LSM) change from baseline (95% CI) was compared using a constrained longitudinal data analysis model. TTD was compared using a stratified log rank test and Cox proportional hazards model. P-values are nominal and two-sided. Results: QLQ-C30, QLQ-OES18 and EQ-5D-5L compliance was ≥90% in both arms at baseline and at week 18. There was no significant difference in least squares mean (LSM) change from baseline to week 18 in GHS/QoL status between arms (LSM difference [95% CI] -0.10 [-3.40-3.20]; P = 0.9530). Median TTD in GHS/QoL was similar between arms (HR, 0.86 [95% CI, 0.66-1.13]; P = 0.2864). Outcomes were similar in ESCC PD-L1 CPS ≥10, ESCC, and PD-L1 CPS ≥10 patient populations. LSM change from baseline to week 18 for QLQ-OES18 pain subscale was better for pembro + chemo (-4.78) vs chemo (-1.85 ) (-2.94, -5.86 to -0.02; P = 0.0487). There was no significant difference in LSM change from baseline to week 18 between arms for reflux (-1.19; -4.49-2.10; P = 0.4781) or dysphagia (-2.35; -7.78-3.07; P = 0.3945). VAS LSM change from baseline to week 18 was similar between arms (1.20, -1.61-4.01; P = 0.4016). Conclusions: HRQoL was stable and similar over 18 weeks in the pembro + chemo and chemo arms. Together with superior OS, PFS, and ORR and a manageable safety profile with pembro + chemo, these results support the clinically meaningful benefit of pembro + chemo in patients with advanced esophageal cancer including EGJ adenocarcinoma. Clinical trial information: NCT03189719.

To determine the recommended dose and antitumor activity of single-agent elisidepsin as a 24-h intravenous (i.v.) infusion fortnightly [biweekly, d1 and 15 every 4 weeks (q4wk); Arm A, dose-intensity strategy] or as a 3-h i.v. infusion weekly (d1, 8, 15 and 22 q4wk; Arm B, dose-density strategy) in adult patients with unresectable, locally advanced or metastatic pretreated esophageal, gastroesophageal junction and gastric cancer. Patients were randomized to one of two elisidepsin dosing schedules. Phase Ib starting doses were 8.0 mg flat dose (FD) in Arm A and 3.0 mg FD in Arm B. Phase II subsequently explored antitumor activity of both dosing schedules at the respective recommended doses. Forty-four patients received elisidepsin: 12 in stage Ib and 32 in stage II. The recommended doses were defined as 10 mg FD (Arm A) and 3.75 mg FD (Arm B). Both schedules were well tolerated. Most adverse events were mild or moderate, reversible and predictable with no meaningful differences between schedules. The pharmacokinetic profiles of both schedules were similar to those reported previously in patients with solid tumors treated with a comparable dose. An interim analysis found tumor control in one patient receiving elisidepsin fortnightly, and in none given elisidepsin weekly; patient accrual was therefore discontinued due to lack of efficacy. Both schedules at the recommended doses presented an acceptable safety profile, but lack of response means that we do not recommend further evaluation of single-agent elisidepsin as chemotherapy for unresectable, locally advanced or metastatic gastroesophageal cancer.

Purpose: Although controversial, assessment of epidermal growth factor receptor (EGFR) expression is required for the approved indications of Cetuximab in metastatic colorectal cancer (mCRC). With the objective of improving patient selection, “ERBITUX-OUEST” study aimed at analyzing EGFR status in a large cohort of mCRC patients who received cetuximab without preliminary EGFR screening, and assessing the correlation between EGFR status and response to treatment retrospectively. Patients and methods: 332 patients treated with Irinotecan Cetuximab based regimen after progression on irinotecan or oxaliplatin therapy were included. EGFR status was assessed using three available immunohistochemistry (IHC) tests and in situ hybridization in case of negativity. Clinical outcomes of EGFR-positive and EGFR-non-detected (or considered as negative with at least one test) patients were compared. Results: Of the 332 samples centrally screened, 194 were classified as full-positive (i.e., EGFR-positive for all three tests), 86 as full-negative, and 52 as discordant. One third of the 131 negative samples with FDA approved test should be reclassified as positive with at least one of the two others tests. Regarding results from FDA approved test only, neither objective response rate (ORR), progression-free survival (PFS) nor overall survival (OS) differed significantly between EGFR-negative and EGFR-positive patients (P = 0.788, 0.326 and 0.888, respectively). Similarly, comparison of full-negative to other groups did not show any significant difference in terms of ORR (P = 0.507), PFS (P = 0.222) or OS (P = 0.686). Conclusion: These data strongly argue against mCRC patients selection for Cetuximab treatment based on EGFR expression as measured by currently available IHC technics.

Supplemental Digital Content is available in the text, To evaluate the diagnostic performance of net influx rate (Ki) values from a whole-body dynamic (WBD) 68Ga-DOTATOC-PET/CT acquisition to differentiate pancreatic neuroendocrine tumors (pNETs) from physiological uptake of pancreatic uncinate process (UP). Patients who were benefited from a WBD acquisition for the assessment of a known well-differentiated neuroendocrine tumor (NET)/suspicion of disease in the prospective GAPET-NET cohort were screened. Only patients with a confirmed pNET/UP as our gold standard were included. The positron emission tomography (PET) procedure consisted in a single-bed dynamic acquisition centered on the heart, followed by a whole-body dynamic acquisition and then a static acquisition. Dynamic (Ki calculated according to Patlak method), static (SUVmax, SUVmean, SUVpeak) parameters, and tumor-to-liver and tumor-to-spleen ratio (TLRKi and TSRKi (according to hepatic/splenic Ki)), tumor SUVmax to liver SUVmax (TM/LM), tumor SUVmax to liver SUVmean (TM/Lm), tumor SUVmax to spleen SUVmax (TM/SM), and tumor SUVmax to spleen SUVmean (TM/Sm) (according to hepatic/splenic SUVmax and SUVmean respectively) were calculated. A Receiver Operating Characteristic (ROC) analysis was performed to evaluate their diagnostic performance to distinguish UP from pNET. One hundred five patients benefited from a WBD between July 2018 and July 2019. Eighteen (17.1%) had an UP and 26 (24.8%) a pNET. For parameters alone, the Ki and SUVpeak had the best sensitivity (88.5%) while the Ki, SUVmax, and SUVmean had the best specificity (94.4%). The best diagnostic accuracy was obtained with Ki (90.9%). For ratios, the TLRKi and the TSRKi had the best sensitivity (95.7%) while the TM/SM and TM/Sm the best specificity (100%). TLRKi had the best diagnostic accuracy (95.1%) and the best area under the curve (AUC) (0.990). Our study is the first one to evaluate the interest of a WBD acquisition to differentiate UP from pNETs and shows excellent diagnostic performances of the Ki approach.

Background: Fluoropyrimidines (FPs) carry around 20% risk of G3-5 toxicity and 0.2-1% risk of death, due to dihydropyrimidine dehydrogenase (DPD) deficiency. Several screening approaches exist for predicting toxicity, however there is ongoing debate over which method is best. This study compares 4 screening approaches. Method: 472 patients treated for colorectal, head-and-neck, breast, or pancreatic cancers, who had not been tested pre-treatment for FP toxicity risk, were screened using: DPYD genotyping (G); phenotyping via plasma Uracil (U); phenotyping via plasma-dihydrouracil/uracil ratio (UH2/U); and a Multi-Parametric Method (MPM) using genotype, phenotype, and epigenetic data. Performance was compared, particularly the inability to detect at-risk patients (false negatives). Results: False negative rates for detecting G5 toxicity risk were 51.2%, 19.5%, 9.8% and 2.4%, for G, U, UH2/U and MPM, respectively. False negative rates for detecting G4-5 toxicity risk were 59.8%, 36.1%, 21.3% and 4.7%, respectively. MPM demonstrated significantly (p < 0.001) better prediction performance. Conclusion: MPM is the most effective method for limiting G4-5 toxicity. Its systematic implementation is cost-effective and significantly improves the risk-benefit ratio of FP-treatment. The use of MPM, rather than G or U testing, would avoid nearly 8,000 FP-related deaths per year globally (500 in France), and spare hundreds of thousands from G4 toxicity.

4617 Background: Eryaspase is composed of L-asparaginase encapsulated in erythrocytes. It has demonstrated significant efficacy in combination with chemotherapy in a randomized phase 2 trial in second-line in patients with advanced pancreatic adenocarcinoma. We assessed, in this study, the prognostic and predictive value of circulating tumor DNA (ctDNA) in plasma samples of patients included in the eryaspase phase 2 trial. Methods: Samples prospectively collected pre-treatment at each 28-day cycle were centrally analyzed by next-generation sequencing (BPER method). Prognostic values of baseline ctDNA and ctDNA early changes between day 0 and 28 were assessed in both arms combined on objective response rate (ORR), progression free survival (PFS) and overall survival (OS). We conducted interaction test between ctDNA positivity and treatment arm, and the predictive value of ctDNA for eryaspase efficacy was investigated. Results: Patients with at least one available plasma sample have been included (n = 122/141). The presence of ctDNA at baseline was identified in 68% (77/113) of patients and was an independent negative prognostic factor for OS (4.6 vs 8.8 months; p = 0.0025) and PFS (1.6 vs 3.3 months; p = 0.00043). Early change in ctDNA levels was assessed by separating patients into three categories (one without detectable ctDNA, and two according to radio median value between day 0 and day 28) that were significantly correlated with ORR, PFS and OS. A significant interaction was observed between the presence of ctDNA and eryaspase efficacy. In patients with ctDNA detectable at baseline, eryaspase was associated with better PFS (HR = 0.53; 95% CI: 0.3-0.94) and OS (HR = 0.52; 95% CI: 0.29-0.91). Conclusions: We confirm from a prospective randomized trial that 1/ the presence of ctDNA at baseline is a major prognostic factor, 2/ the early change of ctDNA correlates with treatment outcome and 3/ the ctDNA could be a predictive biomarker of eryaspase efficacy. Clinical trial information: NCT02195180 .

4019 Background: Regorafenib (R) has been shown to modulate anti-tumor immunity by different mechanisms including reduction of tumor-associated macrophages (TAMs). Synergy between R and anti–PD-1/PD-L1 antibodies has been shown in pre-clinical models. Methods: This is a single-arm open-label multicentric phase II trial assessing the efficacy and safety of R (160 mg QD 3weeks/4) + Avelumab (A) (10 mg/kg every 2 weeks) combination in non MSI-H mCRC patients (pts). The primary endpoint was the confirmed objective response rate, based on central review according to RECIST 1.1. Secondary endpoints included: 1-year progression free survival (PFS), 1-year overall survival (OS), and Safety using NCI-CTCAE v5.0. Correlative studies were planned from pts tumor samples obtained at baseline and C2D1. Results: Between Nov. 2018 and Oct. 2019, 48 pts were enrolled in 4 centers. Median age was 61.8 (range: 26.3-78.7). Median follow-up was: 7.2 months. Median number of previous treatment lines was: 3 (range: 1-7). 41 (87.2%) pts experienced at least 1 dose modification or treatment interruption. The most common grade 3/4 adverse events were palmar-plantar erythro-dysesthesia syndrome (29.8%), hypertension (23.4%) and diarrhea (12.8%). No death was related to the treatment. Among 40 pts who had at least one imaging tumor assessment, 12 (30%) had reduction in tumor burden. Best response was stable disease for 23 pts (57.5%) and progressive disease for 17 pts (42.5%). The median PFS and OS were 3.6 months (CI95%: [1.8 – 5.4]) and 10.8 months (CI95%: [5.9 – NA]) respectively. Baseline tumor samples and paired biopsies were available for 24 and 15 pts respectively. High infiltration by TAMs at baseline was significantly associated with adverse outcome (PFS: 1.9 vs 3.7 months, p=0.045; OS: 4.8 months vs NR, p=0.027). Increased tumor infiltration by CD8+ at C2D1 compared to baseline was significantly associated with better PFS (p=0.011). Combining low TAMs infiltration and low tumor cells to CD8+ T cells distance enabled the identification of a subgroup of pts (n= 6/24, 25%) more likely to benefit from R+A combination: median PFS: 5.3 vs 1.9 months (p=0.037); median OS: NR vs 5.3 months (p=0.02). Conclusions: The R+A combination achieved PFS and OS that compared favourably with historical data of R alone in this clinical setting. High-resolution analysis of tumor samples identified a composite score based on TAMs infiltration and tumor cell to CD8+ T cells distance which could be used as a biomarker in further studies investigating this approach in mCRC pts. Clinical trial information: NCT03475953 .

439 Background: The benefit of checkpoint blockade in advanced gastric cancer is limited and biomarkers related to response are needed. Novel gene expression analysis software was used to identify Hallmarks of Cancer associated with clinical benefit following nivolumab treatment in >2nd line advanced gastroesophageal adenocarcinoma (GEA). Methods: RNA-sequencing data from baseline GEA patient diagnostic tumor samples (103 from NCT02862535; 5 from NCT02862535) were analyzed using the claraT platform (V2.0.0, Almac Diagnostic Services). 62 gene signatures were quantified representing 6 key Hallmarks of Cancer (Avoiding Immune Destruction, Activating Invasion and Metastases, Sustaining Proliferative Signaling, Inducing Angiogenesis, Resisting Cell Death and Genome Instability and Mutation). Clinical benefit (CB) was defined as tumor response or overall survival (OS) > 1 year. HER2 status was from medical records. Survival analyses used cox proportional hazards models. Results: Gene expression signatures (GES) identified 5 molecular subgroups (C1-C5). Rate of CB in each molecular subtype are outlined in Table. C3 and C4 had significantly improved OS compared to C2, (HR = 0.45; p= 0.02 and HR = 0.42; p= 0.02). Greater proportions of HER2+ subjects were present in C4 and C3 vs. C2, with C3 statistically significant (60% vs. 14%; p= 0.012). Gene expression characterized by chromosomal instability (CIN) and homologous recombination repair deficiency (HRD) were associated with HER2(+) (wilcox p= < 0.05). Patients selected by only using CIN & HRD had significant improvement in OS (HR = 0.63; p= 0.03). Conclusions: Interferon-based GES did not predict benefit from immune checkpoint blockade. GES representing HRD and activation of HER2, EGFR and MAPK (each enriched in CIN) were associated with improved survival upon checkpoint blockade in advanced GEA patients. Clinical trial information: NCT02862535 . [Table: see text]

TPS4666 Background: Second-line treatment options for advanced pancreatic adenocarcinoma are currently limited. Eryaspase, asparaginase (ASNase) encapsulated in red blood cells (RBCs) is an investigational product under development. Following infusion, asparagine and glutamine are actively transported into RBCs where they are hydrolyzed by the encapsulated ASNase. We have recently reported the outcome of a randomized Phase 2b study inpatients with advanced pancreatic cancer whose disease progressed following first-line treatment(NCT02195180). Eryaspase in combination with gemcitabine monotherapy or FOLFOX combination therapy improved overall survival (OS) and progression free survival (PFS). The safety profile of eryaspase was acceptable. The results of this Phase 2b study provided a rationale for initiating this confirmatory Phase 3 pivotal trial (TRYbeCA-1). Methods: TRYbeCA-1 is a randomized, open-label Phase 3 trial (N = ~500) of eryaspase combined with chemotherapy in patients with adenocarcinoma of the pancreas who have failed only one prior line of systemic anti-cancer therapy for advanced pancreatic cancer and have measurable disease. Patients are randomized in a 1:1 ratio to receive gemcitabine/Abraxane or irinotecan-based therapy (FOLFIRI [FOLinic acid-Fluorouracil-IRInotecan regimen] or irinotecan liposome injection/5-fluorouracil/leucovorin) with or without eryaspase, administered as IV infusion on Day 1 and Day 15 of each 4-week cycle. Key eligibility criteria include performance status 0 or 1; stage III-IV disease; documented evidence of disease progression; available tumor tissue; and adequate organ function. The primary endpoint is OS. Key secondary endpoints include PFS and objective response rate, safety, quality of life, pharmacokinetics and pharmacodynamics, and biomarker research. A hazard ratio in OS of 0.725 is being targeted which represents a conservative estimate based on the Phase 2b data and is viewed as being highly clinically relevant. An IDMC is established to review safety at regular intervals andto review efficacy data at the planned interim and final analyses. IDMC last reviewed the trial in October 2019 and suggested the trial continue as planned. Clinical trial information: NCT03665441 .

Background: Second-line treatment options for advanced pancreatic adenocarcinoma are currently limited. Eryaspase, asparaginase (ASNase) encapsulated in red blood cells (RBCs), is an investigational product under development. Following infusion, asparagine and glutamine are actively transported into RBCs where they are hydrolyzed by the encapsulated ASNase. We have recently reported the outcome of a randomized phase 2b study in patients with advanced pancreatic cancer whose disease progressed following first-line treatment (NCT02195180). Eryaspase in combination with gemcitabine monotherapy or FOLFOX combination therapy improved overall survival (OS) and progression-free survival (PFS). The safety profile of eryaspase was acceptable. The results of this phase 2b study provided a rationale for initiating this confirmatory phase 3 pivotal trial (TRYbeCA-1). Methods: TRYbeCA-1 is an international, randomized, open-label phase 3 trial (N= ~500) of eryaspase combined with chemotherapy in patients with adenocarcinoma of the pancreas who have failed only one prior line of systemic anticancer therapy for advanced pancreatic cancer and have measurable disease. Patients are randomized in a 1:1 ratio to receive gemcitabine/abraxane or irinotecan-based therapy (FOLFIRI [FOLinic acid-Fluorouracil-IRInotecan regimen] or irinotecan liposome injection/5-fluorouracil/leucovorin) with or without eryaspase, administered as IV infusion on Day 1 and Day 15 of each 4-week cycle. Key eligibility criteria include performance status 0 or 1, stage IV disease, documented evidence of disease progression, available tumor tissue, and adequate organ function. The primary endpoint is OS. Key secondary endpoints include PFS and objective response rate, safety, quality of life, pharmacokinetics and pharmacodynamics, and biomarker research. A hazard ratio in OS of 0.725 is being targeted, which represents a conservative estimate based on the phase 2b data and is viewed as being highly clinically relevant. An IDMC will be established to review safety at regular intervals and to review efficacy data at the planned interim and final analyses. Citation Format: Pascal Hammel, Rosanna Berardi, Eric Van Cutsem, Jaime Feliu, Richard Greil, Harpreet Wasan, Jean-Philippe Metges, Peter Nygren, Pia Österlund, Marcus Noel, Thomas Seufferlein, Geert-Jan Creemers, Anu Gupta, Sophie Salesse, Nigel Biswas-Baldwin, Hedy Dion, Hagop Youssoufian, Iman El-Hariry, Manuel Hidalgo. Trybeca-1: A randomized, phase 3 study of eryaspase in combination with chemotherapy versus chemotherapy alone as second-line treatment in patients with pancreatic adenocarcinoma (NCT03665441) [abstract]. In: Proceedings of the AACR Special Conference on Pancreatic Cancer: Advances in Science and Clinical Care; 2019 Sept 6-9; Boston, MA. Philadelphia (PA): AACR; Cancer Res 2019;79(24 Suppl):Abstract nr B22.

FOLFIRI-aflibercept chemotherapy is indicated in second-line treatment of metastatic colorectal cancer (mCRC) after progression under FOLFOX (±bevacizumab). Nevertheless, its proven therapeutic efficacy in clinical trials was based on partial responses. And to date, only 2 cases of complete response assessed by CT were reported in literature. We report a complete metabolic response assessed by FDG PET/CT in a 50-year-old woman with mCRC treated by FOLFIRI-aflibercept. This also confirms that FDG PET/CT is emerging as a useful approach for therapeutic assessment of targeted drugs in mCRC.

International audience; We conducted a multicenter proof of concept phase II trial in patients with advanced colorectal cancer receiving FOLFIRI-cetuximab regimens to explore individual drug tailoring using pharmacogenetics and pharmacokinetics (PK) monitoring. Patients were stratified by their pharmacogenetic/phenotypic status: the irinotecan dose was adjusted according to the number of TA tandem repeats in the UGT1A1 promoter, while the 5-fluorouracil (5-FU) dose was initially adjusted according to dihydropyrimidine dehydrogenase (DPD) activity at initial screening (5-FU[ODPM Tox]) followed by PK-guided dose optimization (5-FU[ODPM Protocol]). An advanced cetuximab PK/pharmacodynamics (PD) study was performed but dosage remained unchanged. Eighty-five patients receiving second-line chemotherapy were enrolled. Mean irinotecan doses at 3 months were 247 ± 50, 210 ± 53 and 140 ± 21 mg/m[2] for those with 6/6 (33), 6/7 (26), and 7/7 (7) TATA repeats in the UGT1A1 promoter region, respectively. The 5-FU dose was initially reduced in four patients with DPD deficiency, but mean 5-FU dose at 3 months was 2,412 ± 364 mg/m2 (1,615-3,170 mg/m[2]). Grade 4 toxicities were not encountered and grade 4 neutropenia occurred in 6.8%, 5.9%, and 0% of patients with 6/6, 6/7, and 7/7 UGT1A1 genotypes. The objective response rate was 25.8% among the 85 patients, 57.3% in patients with tumors wild type (WT) for KRAS, and 25% in those whose tumor harbored a mutant-KRAS. Secondary resection of hepatic metastases was performed in 31.7% of patients. Median progression-free survival (PFS) for all 85 patients was 181 days and 200, 132, and 121 days for patients with 6/6, 6/7, and 7/7 UGT1A1 genotypes, respectively; these differences were not statistically different. In parallel, a strong relationship was shown between cetuximab AUC and regimen efficacy. We conclude that personalized drug tailoring when administering in FOLFIRI + cetuximab allows for safe and efficient individual dose intensification.

En 2006, bevacizumab-FOLFIRI represente la therapie ciblee administrable des la premiere ligne chez les patients porteurs d’un cancer colorectal metastatique non operable. Une serie homogene de 111 patients colliges en region Bretagne et Pays de la Loire ayant recu du bevacizumab- FOLFIRI en premiere ligne en 2006 revele les resultats suivants: 51 reponses, 29 stabilites, 21 progressions et 10 toxicites avant evaluation. La mediane de survie globale (OS) est de 25,1 mois et la mediane de survie sans progression (PFS) de 10,2 mois. Dans le cas d’une chirurgie secondaire, l’OS median triple de 18,8 mois chez les patients non reseques versus 59,2 mois ceux reseques. En comparant les sujets âges de plus et de moins de 70 ans, aucune difference n’a ete mise en evidence en termes de benefice ou de risque. Bevacizumab-FOLFIRI pourrait etre administre en pratique courante chez les personnes âgees sous couvert d’une evaluation geriatrique et d’une approche multidisciplinaire.

e15690 Background: NeuroEndocrine Tumors (NETs) remain a poorly known entity, especially in older patients. The therapeutic management is multidisciplinary combining surgery, interventional radiology, chemotherapy, cancer supportive care and also geriatrics oncology. The aim of this study performed by the OncoGeriatry Coordination Unit (UCOG) of Brittany, in collaboration with the National Reference Network for the management of NET (RENATEN), was to analyze the current management of NETs in older patients (more than 75 years old) and to develop recommandations to optimize the management of elderly patients with NENs. RENATEN’s guidelines mean systematic discussion of the cases in Regional Multidiscplinary Tumor Board (MTB) dedicated to NETs (RENATEN MTB). Methods: This study is a retrospective observational study including patients aged 75 years and over, with a NET that has been diagnosed in an pathological laboratory or whose file has been discussed in a MTB (usual MTB or RENATEN MTB) between 2014 and 2017 in the Brittany area. Results: 51 patients with a median age of 79 years [75-92] have been included up until now, among whom 77% had digestive NET, 12% urological, 8% dermatological (Merkel tumor), 4% pulmonary, 2% gynecological and 2% NETs of unknown origin. Tumor grade was 1, 2, 3 and unkown in 20%, 14%, 37% and 29% respectively. Overall, 33% of the tumors were well differentiated and 63% were metastatic (synchronous metastases :78%). Only 20% of patients benefited from a geriatric oncology consultation to help the clinician in the therapeutic decision. File of 53% of patients has been discussed one time in a specialized RENATEN MTB and 57% in another MTB. No proof of discussion in MTB was found in 12%. The treatment proposed in the RENATEN MTB was effective in 85% of patients (2 patients died immediately after RENATEN MTB, 2 patients did not wish to be treated). Only 14% of patients received initial surgery, 10% radiotherapy and 41% chemotherapy. The overall survival of this cohort is 11.3 months [0.3 -89] (for grade 3, 7.9 months [0.8-21.7]). Conclusions: The rate of multidisciplinary meeting presentation has to be improved, as well as the geriatric oncology management. A cohort of more than 100 patients will be presented at the meeting and treatment response and survival data based on age class, tumor grade, histological subtypes and therapeutic strategy will be shown.

4010 Background: The phase 3 KEYNOTE-181 study compared pembrolizumab (pembro) vs chemo as second-line therapy for patients (pts) with advanced/metastatic squamous cell carcinoma (SCC) and adenocarcinoma (ACC) of the esophagus (NCT02564263). Methods: Eligible pts were randomized 1:1 to pembro 200 mg Q3W for up to 2 years or choice of paclitaxel, docetaxel, or irinotecan. Randomization was stratified by histology (SCC vs adenocarcinoma) and region (Asia vs rest of world). Primary end points were OS in the SCC, PD-L1 combined positive score (CPS) ≥10, and the ITT. Secondary endpoints included PFS, ORR, safety; exploratory endpoints included health-related quality of life (HRQoL) in CPS ≥10. Results: 628 pts were randomized (401 with SCC; 222 with CPS ≥10). As of Oct. 15, 2018, median follow-up was 7.1 mo (pembro) vs 6.9 mo (chemo). In CPS ≥10, OS was superior with pembro vs chemo (median 9.3 vs 6.7 mo; HR 0.69; 95% CI 0.52-0.93; P= 0.0074). In CPS ≥10 SCC, median OS was 10.3 mo vs 6.7 mo and in CPS ≥10 ACC, median OS was 6.3 mo vs 6.9 mo; 12-mo OS rates were higher with pembro vs chemo (Table). In SCC, median OS was 8.2 mo vs 7.1 mo; HR 0.78; 95% CI 0.63-0.96; P= 0.0095. In the ITT, median OS was 7.1 mo vs 7.1 mo; HR 0.89; 95% CI 0.75-1.05; P= 0.0560. Updated OS will be presented. Grade 3-5 drug-related AEs (≥10% incidence in either arm) included decreased white blood cells (0% vs 10%), decreased neutrophils (0.3% vs 10%). In CPS ≥10, HRQoL improved with pembro vs chemo only for EQ-5D VAS (difference in LS mean change from baseline 5.57; 95% CI 0.58-10.56). Conclusions: Pembro significantly improved OS vs chemo as second-line therapy for advanced esophageal cancer with PD-L1 CPS ≥10, with a more favorable safety profile and stable and similar QOL. These data support pembro as a new second-line standard of care for esophageal cancer with PD-L1 CPS ≥10. Clinical trial information: NCT02564263. [Table: see text]

3529 Background: Severe, sometimes fatal, toxicity can occur during the 1st or 2nd course of chemotherapy using fluoropyrimidines (FPs), and poses a serious public health problem. FPs carry a 3-5% risk of grade ≥ 3 early toxicities and 0.2% risk of death linked to Dihydropyrimidine Dehydrogenase (DPD) deficiency. Methods: Of 29,000 patients screened since July 2000, 472 were referred to us due to severe toxicity during 1st round 5-FU, or because pre-screening was done too late. Toxicity evaluation was performed according to the NCI scale of adverse reactions to cancer drugs (0=none, 5=death). Patients were previously 5-FU naïve, had different cancers, and received various protocols, eg. 5FU or Capecitabine; bolus±continuous or per os. The reliability of the following 4 pre-treatment screening tests to predict grade≥ 4 toxicity was assessed: 1) DPYD genotype mutation (*2A,*2B,*7, 13, HapB3) 2) Plasma uracil (U) level, 3) Plasma dihydrouracil/uracil ratio (UH2/U) 4) a multiparametric approach with genotyping, UH2/U ratio and key patient factors (age, sex, etc.). McNemar’s test with Bonferroni correction was used for statistical analysis. Results: Of the 472 referred patients, 169 had grade 4 or 5 toxicity, of which 41 died from toxicity. 98 had one or plus DPYD mutation: 42(42.9%)*2A; 43(43.9%)*2B; 3(3%)*7; 4(4%)*13; 8(8.16%) HapB3; 1 was homozygous *2A. Data below compare the 4 screening methods for predicting grade 4-5 toxicity. Conclusions: The multiparametric approach is statistically (p

4048 Background: KEYNOTE-181 (NCT02564263) is an open-label, randomized, phase 3 trial in ACC and SCC of the esophagus that evaluated IV pembro 200 mg Q3W for up to 2 years vs investigator choice of single-agent paclitaxel/docetaxel/irinotecan (control). Pembro was superior to control for OS in patients with PD-L1 CPS ≥10 (N = 222; median 9.3 vs 6.7 months; P= 0.0074). Here we present results of prespecified HRQoL analyses in this population. Methods: The EORTC QLQ-C30 and EORTC QLQ-OES18 were administered at baseline; weeks 2, 3, 4, 6, 9, 12, 18; every 9 weeks up to 1 year/end of treatment; and 30-day safety follow-up visit. Data from patients receiving ≥1 dose of study treatment and completing ≥1 HRQoL assessment were analyzed. Least squares mean (LSM) score change from baseline to week 9, 95% CI, and nominal P values were calculated. Time to deterioration (TTD) (≥10-point decline from baseline) was assessed by Kaplan-Meier method and Cox regression model. HRs, 95% CIs, and nominal P values are provided. Results: The HRQoL population included 218 PD-L1 CPS ≥10 patients (107 pembro, 111 control). QLQ-C30 compliance at week 9 was 88.9% for pembro and 83.9% for control. There was no significant difference in LSM between arms (3.68; 95% CI –2.28, 9.64; P= 0.2248) in global health status (GHS)/QoL score. Week 9 QLQ-OES18 compliance was 88.4% for pembro and 83.3% for control. QLQ-OES18 scores were not significantly different between arms. TTD for pain (HR 1.02; 95% CI 0.58, 1.81; P= 0.5282), reflux (HR 1.69; 95% CI 0.83, 3.47; P= 0.9254), and dysphagia (HR 1.81; 95% CI 0.97, 3.37; P= 0.9693) subscales were not significantly different between arms. Conclusions: Over 9 weeks, patients treated with pembro had stable GHS/QoL scores similar to those of patients treated with single-agent docetaxel/paclitaxel/irinotecan. Combined with the superior OS and lower rate of treatment-related AEs seen with pembro, these data support clinically meaningful benefit of pembro in esophageal cancer patients with PD-L1 CPS ≥10. Clinical trial information: NCT02564263.

118 Background: Preclinical studies suggest that MMP9 inhibition relieves immune suppression and promotes T cell infiltration to potentiate checkpoint blockade. To test this hypothesis, patient tumor samples obtained in a phase 2, open-label, randomized study (NCT02864381) of previously treated advanced gastric cancer comparing the MMP9-specific inhibitor andecaliximab (ADX) plus nivolumab (N) vs N alone were evaluated for T cell biomarkers. Methods: CD8 and PD-L1 (28-8 DAKO) were assessed by immunohistochemistry. IFNg, Teffector (Teff), and activated CD8+ T cell (ActT) gene signatures (GS) were assessed by RNASeq in archival baseline (BL) and biopsies obtained between weeks 5 and 9 (on-treatment; OT). Results: For both N and ADX/N, intratumoral CD8+ cells were significantly increased in OT biopsies relative to BL. The CD8+ OT increase was significantly greater for the ADX/N treatment relative to N in the PD-L1+ subgroup. In the ADX/N group only, IFNg, Teff, and ActT GS were significantly higher in OT biopsies relative to BL. OT change from BL of ≥ 300% vs

626 Background: First line of RAS wild-type (WT) unresectable metastatic colorectal cancer (mCRC) can be doublet chemotherapy with an anti-VEGF (Vascular Endothelial Growth Factor), or an anti-EGFR (Epidermal Growth Factor Receptor). Waiting for RAS status, many oncologists initiate chemotherapy and add the anti-EGFR secondly. The objective was to compare the delayed introduction of the anti-EGFR to the immediate introduction of the anti-VEGF in first-line treatment of RAS WT mCRC. Methods: This was a retrospective cohort analysis from 2013 to 2016, multicentric with 28 health care centers. We included patients with RAS WT unresectable mCRC treated between 2013 and 2016 by a doublet chemotherapy with the anti-VEGF introduced immediately or with the anti-EGFR introduced at C2 or C3. Progression free survival (PFS), overall survival (OS) and response rate (RR) for the two cohorts were compared. Hazard ratios (HR) with 95% confidence interval (95%CI) were estimated with cox regression models weighted on propensity score to deal with potential confounders. Results: A total of 262 patients were included, 129 in the immediate anti-VEGF group and 133 in the delayed anti-EGFR group. Median follow-up was 37.9 months. Ninety-two patients had the anti-EGFR introduced at C2, 40 at C3. The median delay of RAS analysis was 19 days (q1-q3: 13-26). Patients treated with anti-VEGFs were more likely men (68% versus 56%), with more metastatic sites ( > 2 sites: 15% versus 9%). A propensity score including the number of metastatic sites and a possible previous treatment was built. Delayed anti-EGFRs were associated with longer PFS compared to immediate anti-VEGFs: 13.8 versus 11.0 months, p = 0.0244. After weighting, delayed anti-EGFRs were still associated with better PFS: HR 0.74, 95%CI [0.61 – 0.90], p = 0.0024. OS was not different between the two arms (30.5 for anti-VEGF versus 29.9 months, p = 0.3934), even after weighting (HR 0.86, 95%CI [0.69 – 1.08], p = 0.2024). There was a better RR with delayed anti-EGFRs: 66.7% versus 45.6%, p = 0.0007. Conclusions: Our findings suggest that, while waiting for RAS status, the delayed introduction of the anti-EGFR is a valid option, compared to the immediate introduction of the anti-VEGF.

TPS471 Background: Second-line treatment options for advanced pancreatic adenocarcinoma are currently limited. Eryaspase, asparaginase (ASNase) encapsulated in red blood cells (RBCs) is an investigational product under development. Following infusion, asparagine and glutamine are actively transported into RBCs where they are hydrolyzed by the encapsulated ASNase. We have recently reported the outcome of a randomized Phase 2b study inpatients with advanced pancreatic cancer whose disease progressed following first-line treatment(NCT02195180). Eryaspase in combination with gemcitabine monotherapy or FOLFOX combination therapy improved overall survival (OS) and progression free survival (PFS). The safety profile of eryaspase was acceptable. The results of this Phase 2b study provided a rationale for initiating this confirmatory Phase 3 pivotal trial (TRYbeCA-1). Methods: TRYbeCA-1 is an international, randomized, open-label Phase 3 trial (N = ~500) of eryaspase combined with chemotherapy in patients with adenocarcinoma of the pancreas who have failed only one prior line of systemic anti-cancer therapy for advanced pancreatic cancer and have measurable disease. Patients are randomized in a 1:1 ratio to receive gemcitabine/abraxane or irinotecan-based therapy (FOLFIRI [FOLinic acid-Fluorouracil-IRInotecan regimen] or irinotecan liposome injection/5-fluorouracil/leucovorin) with or without eryaspase, administered as IV infusion on Day 1 and Day 15 of each 4-week cycle. Key eligibility criteria include performance status 0 or 1; stage IV disease; documented evidence of disease progression; available tumor tissue; and adequate organ function. The primary endpoint is OS. Key secondary endpoints include PFS and objective response rate, safety, quality of life, pharmacokinetics and pharmacodynamics, and biomarker research. An HR in OS of 0.725 is being targeted representing a conservative estimate based on the P2b data and is viewed as being highly clinically relevant. An IDMC will be established to review safety at regular intervals and to review efficacy data at the planned interim and final analyses. Clinical trial information: NCT03665441.

539 Background: The anatomical location of the primary tumor has been associated with different prognosis and outcomes in mCRC. REG significantly improved overall survival (OS) versus placebo in patients with mCRC who progressed on standard therapies in phase 3 trials. We report the final results of outcomes in patients with mCRC with left-sided (L) or right-sided (R) primary tumors who were treated with REG in the real-world CORRELATE study. Methods: CORRELATE (NCT02042144) was a prospective, observational study designed to characterize the safety and effectiveness of REG in patients previously treated with approved therapies and for whom the decision to treat with REG was made by the treating physician according to the local health authority label. Patients with tumor location unknown or tumors in both regions were excluded (n = 62). OS was analyzed using the Kaplan–Meier method. Results: Overall, 975 patients were included in the analysis: L = 768 (79%) and R = 207 (21%). Most patients were ECOG PS 0–1 (L: 86%; R: 87%). Median time from initial diagnosis and from diagnosis of metastatic disease to REG treatment was slightly longer in L versus R tumors (34 vs 30 months and 27 vs 23 months, respectively). A higher proportion of patients with L versus R tumors, respectively, had prior radiotherapy (34% vs 15%) and a lower proportion had a partial colectomy (44% vs 75%). Both groups received a median of three prior regimens. REG treatment duration was similar in the two groups (median duration: L 2.6 months vs R 2.3 months), as was median progression-free survival (95% CI): 2.8 months (2.6, 2.9) for L tumors versus 2.7 months (2.5, 3.1) for R tumors. Median OS (95% CI) was 7.4 months (6.7, 8.0) for L tumors versus 8.2 months (6.6, 9.3) for R tumors; one-year OS rates were 32% versus 34%, respectively. Conclusions: Results from this observational study suggest that OS is similar in patients with L and R tumors treated with REG. Clinical trial information: NCT02042144.

75 Background: Andecaliximab (ADX) is a monoclonal antibody that inhibits matrix metalloproteinase 9 (MMP9). Preclinical studies suggest that MMP9 inhibition relieves immune suppression and promotes T-cell infiltration to potentiate checkpoint blockade. Methods: Phase 2, open-label, randomized study of the efficacy and safety of ADX + nivolumab (NIVO) vs. NIVO alone in patients with pre-treated metastatic gastric or gastroesophageal junction (GEJ) adenocarcinoma. Patients were randomized to either ADX 800 mg IV + NIVO 3 mg/kg IV, or NIVO 3 mg/kg IV alone, and stratified by tumor PD-L1 status. Treatment was administered every 2 weeks. Re-staging CT scans were performed every 8 weeks to evaluate response. Primary endpoint: objective response rate (ORR). Secondary endpoints: progression-free survival (PFS), overall survival (OS), and adverse events (AEs). Results: Of the 144 patients randomized, 141 were treated, 109 (76%) completed tumor assessment. 81% of patients were white, with 69% male and a mean (SD) age of 59 (12) years. ORR (95% CI) was 11.1% (4.9–20.7%) in patients receiving ADX + NIVO, and 6.9% (2.3–15.5%) in those receiving NIVO alone, p = 0.6. Kaplan-Meier estimated median (95% CI) PFS was 1.8 (1.8–2.0) months in patients receiving ADX + NIVO, and 1.9 (1.7–1.9) months in those receiving NIVO alone, p = 0.2. Kaplan-Meier estimated median (95% CI) OS was 7.2 (5.2–9.1) months in patients receiving ADX + NIVO, and 5.9 (3.5–8.6) months in those receiving NIVO alone, p = 0.4. AEs leading to treatment discontinuation occurred in 1 patient in the ADX + NIVO group, and in 1 patient in NIVO-only group. PD-L1 and mismatch repair deficient subgroup analyses will be presented. Exploratory biomarker analyses will be submitted separately. Conclusions: Addition of ADX to NIVO did not improve ORR, PFS, or OS compared with NIVO alone in patients with pre-treated metastatic gastric or GEJ adenocarcinoma. Combination of ADX with NIVO had a favorable safety and tolerability profile. Clinical trial information: NCT02864381.

2 Background: Patients with advanced esophageal cancer after first-line chemotherapy (chemo) have a poor prognosis and limited treatment options. We present results of the phase 3 KEYNOTE-181 study of pembrolizumab vs investigator’s choice chemo as second-line therapy for patients (pts) with advanced/metastatic squamous cell carcinoma (SCC) and adenocarcinoma of the esophagus or Siewert type I adenocarcinoma of the EGJ (NCT02564263). Methods: Eligible pts were randomized 1:1 to pembrolizumab 200 mg Q3W for up to 2 years or investigator’s choice chemo of paclitaxel, docetaxel, or irinotecan. Randomization was stratified by histology (SCC vs adenocarcinoma) and region (Asia vs rest of world). Primary end points were OS in the SCC, PD-L1 combined positive score (CPS) ≥10, and ITT populations. Results: 628 pts were randomized including 401 with SCC, and 222 with CPS ≥10. As of October 15, 2018, the median follow-up was 7.1 mo (pembrolizumab) vs 6.9 mo (chemo). Pembrolizumab was superior to chemo for OS in CPS ≥10 (N=222; median 9.3 vs 6.7 mo; HR 0.69; 95% CI 0.52-0.93; P=0.0074). The 12-mo OS rate in pts with CPS ≥10 was 43% vs 20%. There was clinically meaningful improvement in OS with pembrolizumab vs chemo in pts with SCC, but this was not statistically significant per prespecified boundaries (N=401; 8.2 mo vs 7.1 mo; HR 0.78; 95% CI 0.63-0.96; P=0.0095). In the ITT population, while directionally favorable, the difference in OS was not statistically significant (N=628; 7.1 mo vs 7.1 mo; HR 0.89; 95% CI 0.75-1.05; P=0.0560). Fewer pts had any-grade (64% vs 86%) or grade 3-5 (18% vs 41%) drug-related AEs with pembrolizumab vs chemo. Conclusion: Pembrolizumab significantly improved OS compared with chemo as second-line therapy for advanced esophageal cancer with PD-L1 CPS ≥10, with a more favorable safety profile. These data support pembrolizumab as a new second-line standard of care for esophageal cancer with PD-L1 CPS ≥10. The phase 3 KEYNOTE-590 study of pembrolizumab plus chemo as first-line therapy for advanced esophageal cancer is ongoing (NCT03189719). Clinical trial information: NCT02564263.

Optimizing the care management of patients is a major issue for our society. In Brittany-Pays-de-la-Loire (almost 10% of French population), an observatory of cancer has been created in 2003. Its main objective was the follow-up of expensive drugs. The knowledge of the use of these drugs in clinical practice has led to development of a thesaurus of good use. Thus, regular exchanges between clinicians have almost totally reduced not medically justified prescriptions by the thesaurus after and before administration to patient. The thesaurus has given away to national guidelines from 2007. For example, in these two regions, optimization of the use of gemcitabine and bevacizumab has allowed to save respectively 2.5 millions euros between 2005 and 2008 and 3 millions euros between 2009 and 2010 (breast cancer only). Optimizing the use of anticancer drugs has allowed a real health economy without any bad impact on patient management. Respecting medical ethics, the main objective remains to optimize health care. This highly participation of clinicians currently allows to reflect together on the relevance of the last chemotherapy.

International audience

Trastuzumab in combination with capecitabine or 5-fluorouracil and cisplatin has been approved by the European Medicines Agency (EMEA) for the treatment of patients with human epidermal growth factor receptor 2 (HER2)-positive (immunohistochemistry [IHC] 3+ or IHC 2+/ fluorescence in situ hybridization [FISH]-positive or IHC 2+/ silver in situ hybridization [SISH]-positive) metastatic adenocarcinoma of the stomach or gastro-esophageal (GE) junction. HER2 testing in gastric cancer (GC) differs from testing in breast cancer (BC) due to major differences in the tumor biology; as the disease is progressing rapidely, we recommend to test every GC at diagnosis and to offer a rapid testing (less than five days) in the metastatic setting. IHC should be the initial testing methodology and FISH or SISH should be used to retest IHC 2+ samples. As GC more frequently shows incomplete membrane staining and focal staining for HER2, HER2 testing guidelines have been adapted from BC protocols. The scoring system is slightly different in respect to the characteristics of GC. For in situ hybridization, SISH should be used in order to identify heterogeneous staining with a higher accuracy than FISH. Enrollment in training and quality assurance programs is highly recommended. In case of negativity on biopsy, it is recommended to retest for HER2, when possible, on surgical specimens and/or metastasis. This will ensure accurate and consistent HER2 testing results, which will allow the appropriate selection of patients eligible for treatment with trastuzumab.

4010Background: NAPA is an oral investigational agent, hypothesized to inhibit cancer stemness pathways, including STAT3 pathway implicated in cancer stem-cell viability. Synergistic antitumor acti...

e15157Background: Immunotherapy is a new paradigm with EMA approval in melanoma and lung cancer. However, recent study has underlined potential drug interactions between immunotherapy and antibioti...

723 Background: The US FDA approved pembro for treating pts with 1) recurrent locally advanced or metastatic G/GEJ adenocarcinoma, whose disease has progressed on or after ≥2 prior therapies and whose tumors express PD-L1 (combined positive score [CPS] ≥1), and 2) unresectable or metastatic, microsatellite instability-high (MSI-H) solid tumors that have progressed after prior therapy and who have no fitting options. We report Asian subgroup analyses from cohort 1 of KEYNOTE-059 (NCT02335411), a global, phase 2 study in advanced G/GEJ cancer. Methods: Eligible pts had measurable recurrent or metastatic G/GEJ adenocarcinoma whose disease has progressed on ≥2 prior chemotherapy regimens. Pts received pembro 200 mg Q3W up to 2 y. PD-L1+ tumors had a CPS ≥1. Primary end points were ORR (RECIST 1.1, by central review) and safety. Results: Cohort 1 enrolled 259 pts; 57% had PD-L1+ tumors. MSI status was evaluable in 174 tumor samples; of these, 7 were MSI-H. At data cutoff (4/21/2017), median (range) follow-up was 6 mo (1-25). Overall ORR was 12% (95% CI, 8-17) and median (range) DOR was 14 mo (2-19+). PFS6-mo rate was 15% and OS6-mo rate was 46%. In pts with PD-L1+ tumors, ORR was 16% (95% CI, 11-23) and median (range) DOR was 14 mo (3+-19+). In pts with PD-L1+ tumors, PFS6-mo rate was 20% and OS6-mo rate was 50%. In pts with MSI-H tumors, ORR was 57% (95% CI, 18-90) and median (range) DOR was not reached (5-14+ mo). In cohort 1, 41 pts were Asian and 218 pts were non-Asian. PD-L1+ tumors occurred in 42% of Asian pts and 60% of non-Asian pts. ORR was 12% (95% CI, 2-36) in Asian pts with PD-L1+ tumors and 17% (95% CI, 11-24) in non-Asian pts with PD-L1+ tumors. One of 7 pts with MSI-H tumors was Asian; this pt had CR. Grade 3-5 treatment-related AEs occurred in 17% and 18% of Asian and non-Asian pts, similar to the overall cohort. Conclusions: Pembro showed durable clinical benefit in previously treated pts with advanced G/GEJ cancer, especially those with PD-L1+ or MSI-H tumors. Safety and efficacy were similar in Asian and non-Asian pts. These findings highlight pembro as a standard treatment option in Asian and non-Asian pts with advanced G/GEJ cancer. Clinical trial information: NCT02335411.

De la biologie devenue strategique en clinique aux nouveaux outils d’evaluation diagnostique, des nouvelles therapeutiques ciblees couplees a une chimiotherapie aux effets secondaires controles a la radiotherapie toujours plus precise associee a une chirurgie de plus en plus optimisee, la recherche clinique en cancerologie digestive est devenue clairement multidisciplinaire. Son challenge va etre de developper ces avancees ensemble et avec un souci d’evaluation constant, d’essayer d’etendre les progres connus dans le cancer colorectal metastatique aux autres localisations digestives au pronostic beaucoup plus defavorable. Dans ce cadre, la prevention et l’optimisation des traitements par l’aide de facteurs predictifs, le developpement de l’oncogeriatrie et la dimension medicoeconomique seront parmi les grands challenges a relever ensemble.

This dose escalation study was designed to determine the maximum tolerated dose (MTD) and recommended doses (RDs) of 5-fluorouracil (5FU), folinic acid and oxaliplatin (FOLFOX) with concomitant radiotherapy in inoperable/metastatic oesophageal squamous cell carcinoma or adenocarcinoma. Patients received three courses of LV5FU2 regimen (folinic acid 200 mg m(-2), bolus 5FU 300-400 mg/m(2), continuous infusion 5FU 400-600 mg m(-2) on days 1 and 2) and escalating doses of oxaliplatin 50 to 100 mg m(-2) on day 1 (FOLFOX). This regimen was repeated every 2 weeks, concomitant to a 50-gray radiotherapy per 5 weeks. Three more cycles were delivered after completion of radiation therapy. Three to six patients were allocated to each of the five dose levels until MTD was reached. Thirty-three patients were enroled and 21 had metastatic disease. Maximum tolerated dose was oxaliplatin 100 mg m(-2), and continuous infusion 5FU was 600 mg m(-2) day(-) (level 5). The most common toxicities were neutropenia, dysphagia and oesophagitis. The RDs were those of FOLFOX-4 regimen (oxaliplatin 85 mg m(-2) and full doses of LV5FU2). The overall response was 48.5%, including 12% complete response. Response rate on primary tumour was 62.9%. This FOLFOX-4 regimen was reasonably well tolerated and effective in inoperable/metastatic oesophageal carcinoma and warrants additional investigation.

Resume Le pronostic du cancer de l’œsophage depend essentiellement du stade initial de la maladie. Nos objectifs ont ete d’evaluer la valeur pronostique de la TEP realisee dans le bilan initial de ce cancer et notamment d’etudier le lien entre la survie globale et l’intensite de fixation du FDG au niveau de la tumeur primitive d’une part, et la presence d’adenopathies et de metastases visibles en TEP d’autre part. Cinquante-deux patients porteurs d’un cancer de l’œsophage ayant beneficie d’une TEP au FDG dans leur bilan d’extension initial ont ete inclus. Le SUV maximum (SUV max ) a ete mesure pour chaque lesion primitive et nous avons note la presence eventuelle d’adenopathies et de metastases. La survie globale a ete etudiee en utilisant la methode de Kaplan-Meier. La valeur pronostique des donnees de la TEP a ete determinee par une analyse de regression de Cox. Une chirurgie a visee curative a ete proposee a la moitie des patients (52 %). La chirurgie, un SUV max superieur a 9, la presence d’au moins deux anomalies en TEP et la presence de ganglions fixant le FDG se sont averes etre des facteurs pronostiques de la survie globale en analyse univariee. En analyse multivariee, seul un SUV max superieur a 9 et la presence de ganglions fixant le FDG etaient des facteurs pronostiques pejoratifs independants. Cette etude prospective montre que la TEP au FDG apporte des informations pronostiques, ce qui peut apporter une nouvelle raison d’inclure cet examen dans le bilan d’extension initial des cancers de l’œsophage.

Angiopoietin-2 (Ang-2) and vascular endothelial growth factor (VEGF) contribute to gastric cancer aggressiveness by up-regulating the expression of proteases. We evaluated the expression and the prognostic significance of angiogenic factors and proteases in 148 patients with R0-resected gastric cancer. Expression of VEGF, Ang-2, cyclooxygenase-2 (COX-2), urokinase-type plasminogen activator (uPA) and its inhibitor PAI-1, matrix metalloproteinases (MMP)-1 and -9 were assayed by immunohistochemistry. After a mean of 63 +/- 4 months, 81 out of 148 patients had died due to disease. The probability of being free of recurrence was 62, 48, and 42% at 2, 5, and 10 years, respectively. Single bivariate analysis identified VEGF, Ang-2, COX-2, PAI-1, and MMP-9 expression, along with several clinicopathological parameters (grade of curability, lymph node ratio, pTNM, pT, pN), as variables associated with both decreased disease-specific survival and recurrence. On multivariate analysis, after adjusting for significant clinical covariables, positive VEGF immunostaining was the primary prognostic factor, and no other tumor marker variable could add any significant improvement for the prediction, for both disease-specific survival (p = 0.001; HR, 3.27; 95% CI, 1.76 to 6.10) and tumor recurrence (p = 0.002; HR, 2.81; 95% CI, 1.48 to 5.35). Our study suggests that VEGF alone may be clinically useful for establishing therapeutic decisions in gastric cancer patients.

Resume Le carcinome epidermoide primitif du pancreas est une tumeur rare. Nous rapportons I’observation d’un carcinome epidermoide pancreatique metastatique chez un homme de 49 ans. Le diagnostic histologique a ete etabli a partir de I’analyse d’une biopsie hepatique, ce qui ne nous permet pas d’eliminer formellement une tumeur adenosquameuse dont seul le contingent epidermoide aurait metastase au foie. Deux lignes de chimiotherapie ont pu etre realisees, la premiere ayant ete associee a de la radiotherapie a visee antalgique. Successivement ont ete administres du 5 fluoro uracile (5FU) et cisplatine (CDDP) puis de la gemcitabine. La gemcitabine a permis une reponse objective avec amelioration de I’etat general, ce qui n’avait jamais ete rapporte auparavant. Cependant le pronostic de ce type de tumeur reste mauvais, la survie de notre malade n’a ete que de 8 mois.

This study was undertaken to determine the value of tumour microvessel density (MVD) and the expression of p53 and vascular endothelial growth factor (VEGF) as prognostic markers in patients with gastric cancer operated on for cure. In all, 156 patients with curatively resected gastric cancer constituted the basis of this blinded retrospective evaluation. Patients were treated with either surgery alone (n=53) or surgery plus adjuvant chemotherapy (n=103). Tumour MVD, p53 expression, and VEGF expression were assayed using immunohistochemical techniques. After a mean follow-up of 43 months, 64 (41%) patients had died and 55 (35%) patients developed tumour recurrence. Positive correlations between MVD and both p53 (P=0.005) and VEGF (P=0.005) expression were observed. Both MVD ⩾100 (P=0.05) and positive VEGF expression (P

Introduction Severe, even fatal toxicities which occur during the first course of fluoropyrimidines chemotherapy treatments pose a serious public health problem. This family of chemotherapy molecules is used in over 60% of protocols, in adjuvant and metastatic settings. In adjuvant therapies, using fluoropyrimidines engenders a toxic (even fatal) risk to a patient who is possibly in remission, if not cancer-free. Even protocols using orally-administered molecules (capecitabine, UFT, others) cause the same toxicities. The results of more than 11,000 patients, assessing risk in the prevalence of severe toxicities by the search for mutations in the DPD and phenotyping are included in this study. Methods From April 2001 to December 2011, 11,351 patients were screened for the risk of severe toxic reaction to fluoropyrimidines. Chemotherapy protocols were 5-Fluorouracil based (e.g., FEC, LV5FU2, FOLFOX, FOLFIRI), as well as orally- delivered fluoropyrimidines (capecitabine and UFT.) 11,104 patients were screened before treatment, as has been the regular protocol in our institution. 247 blood samples were received from other institutions from patients who, after having received their first treatment, showed signs of severe (grade III-IV) toxicity, or had died. The evaluation of risk associates genotyping for mutations on the DPYD gene (1) and phenotyping (dihydrouracil and plasmatic uracil) (2). The calculations were made via the ODPM Tox (ODPM, Angers, France), and every patient whose screening before treatment showed risk received an individually-adapted dose via ODPM Protocol (ODPM, Angers, France). Results The frequencies of mutations observed in the total population are presented in Table I. The frequencies of mutations involved in serious toxicities with fluoropyrimidine are shown in Figure 1. In the patient population screened after toxicity occurred (247 patients): 27 patients (11%) died from toxicity. 16 patients (59%) had one or more genetic mutations and 11 presented with no mutation, but a phenotype characteristic of DPD deficit (ODPM Tox). 220 patients (89%) showed grade III – IV toxicity after first treatment (hematological, diarrhea, etc). Among these patients, 66 (30%) presented with genetic mutation and 148 (67%) showed altered phenotype (ODPM Tox). In population screened before treatment (11, 104): no patients died and none presented with grade III-IV toxicity. Among this population, 264 (2.4%) patients carried one or more risk-related mutations and 956 showed a phenotype characteristic of DPD deficiency. All of these patients received chemotherapy treatments using the ODPM Tox dose-adjustment and all subsequent rounds were dose-adjusted using ODPM Protocol. Conclusion The dual approach to DPD deficiency screening using genotype and phenotype (ODPM Tox) before treatment avoided all severe toxicities. This system has been used in our institution for the past 10 years on a routine basis. Only 8 patients (.07%), of which 2 presented with cardiac toxicity, neither carried a known mutation nor showed a phenotype for risk of toxicity. Thus, the technique has over 99.9% reliability. All of the patients who carried a known risk for toxicity were either treated by another protocol without fluoropyrimidines, or the 5-Fluorouracile treatments were dose-adjusted (ODPM Protocol). Download : Download full-size image Figure 1 . Frequencies of mutations involved in serious toxicities with fluoropyrimidne. Table I . Mutation frequency

Loss of heterozygosity (LOH) on chromosome 9 and p16 (MTS1/CDKN2) gene mutations have been reported in various human cancers. The present study aimed to determine the prevalence of LOH in 100 oesophageal squamous cell carcinomas (OSCCs) by typing microsatellite loci and mutations of the p16 gene. The methods used included denaturing gradient gel electrophoresis (DGGE) and DNA sequencing of exon 2. LOH was found in 14.7% of the OSCC cases. Six gene alterations were identified in exon 2. They consisted of three deletions and the same polymorphism in three samples. The relatively low rate of p16 mutation compared with the frequency of LOH suggests the possible involvement of another tumour suppressor gene located on chromosome 9 in oesophageal carcinogenesis.

4003 Background: Pembro has shown promising antitumor activity and manageable safety in a phase 1 study of pts with previously treated advanced gastric cancer. We conducted a global, multicohort, phase 2 study of pembro in pts with advanced gastric or gastroesophageal junction (G/GEJ) cancer (KEYNOTE-059;NCT02335411). Methods: Cohort 1 enrolled 259 pts, aged ≥18 y with measurable recurrent or metastatic G/GEJ adenocarcinoma who had progressed on ≥2 prior chemotherapy regimens and had ECOG PS 0-1. Pts received pembro 200 mg Q3W up to 2 y or up to disease progression, investigator/pt decision to withdrawal, or unacceptable toxicity. PD-L1+pts had expression in ≥1% tumor or stromal cells using IHC (22C3 antibody). Primary end points: ORR (RECIST 1.1, by central review), safety, and tolerability. Results: Of 259 pts in cohort 1, 76.4% were men; median age was 62.0 y. 51.7% and 48.3% received pembro as 3rd-line (3L) and 4L+ therapy, respectively. 57.1% had PD-L1+ tumors. At data cutoff (Oct 19, 2016), median duration of follow-up was 5.4 mo (range, 0.5 to 18.7). Overall ORR (CR + PR) was 11.2% (95% CI, 7.6-15.7); 1.9% of pts (95% CI, 0.6-4.4) had CR, 9.3% had PR (95% CI, 6.0-13.5), 17% (95% CI, 12.6-22.1) had SD, and 55.6% (95% CI, 49.3-61.7) had PD. Median DOR was 8.1 mo (range, 1.4+ to 15.1+). ORR was 14.9% (95% CI, 9.4-22.1) in 3L pts and 7.2% (95% CI, 3.3-13.2) in 4L+. In PD-L1+ pts, ORR was 15.5% (95% CI, 10.1-22.4) with 2.0% (95% CI, 0.4-5.8) CR and 13.5% (95% CI, 8.5-20.1) PR; in PD-L1– pts, ORR was 5.5% (95% CI, 2.0-11.6), with 1.8% (95% CI, 0.2-6.5) CR and 3.7% (95% CI, 1.0-9.1) PR. In 3L pts with PD-L1+ tumors, ORR was 21.3% (95% CI, 12.7-32.3), with 4.0% (95% CI, 0.8-11.2) CR; in 3L pts with PD-L1– tumors, ORR was 6.9% (95% CI, 1.9-16.7), with 3.4% (95% CI, 0.4-11.9) CR. Grade 3-5 treatment-related AEs (TRAEs) occurred in 43 pts (16.6%). TRAEs led to discontinuation in 2 pts (abnormal LFT, bile duct stenosis) and were fatal in 2 pts (acute kidney injury, pleural effusion). Conclusions: Pembro showed encouraging efficacy and manageable safety after ≥2 prior lines of therapy in pts with advanced G/GEJ cancer in this large phase 2 trial. Survival and additional biomarker data, including MSI status, will be presented. Clinical trial information: NCT02335411.