Journal Logo


ASCO 2017

Gastrointestinal Cancer Updates

Varadhachary, Gauri R. MD

doi: 10.1097/01.COT.0000524337.34983.fc
  • Free
gastrointestinal cancer; ASCO
gastrointestinal cancer; ASCO:
gastrointestinal cancer; ASCO

This commentary highlights select early and mature trials updates in gastrointestinal malignancies. No surprise, immunotherapy and biomarker enriched studies led the way.

Gastric & Gastroesophageal Cancer (G/GEJ)

Resectable disease: The FLOT4-AIO, phase III trial of 716 patients showed perioperative chemotherapy with docetaxel, oxaliplatin, and fluorouracil/leucovorin (FLOT) significantly improved progression-free survival (PFS) and overall survival (OS) among patients with resectable G/GEJ cancers compared with epirubicin, cisplatin, and fluorouracil or capecitabine (ECF/ECX) (Abstract 4004). Median OS was 50 months with FLOT versus 35 months with ECX/ECF (HR 0.77, p = 0.012). Median PFS was 30 months with FLOT versus 18 months with ECX/ECF (HR 0.75; p = 0.004). Periop complications were similar (~50%). The authors concluded that perioperative FLOT improved outcomes in patients with resectable G/GEJ cancers and established a new standard of care.

Advanced disease: Researchers presented updated results of CheckMate-032 study of nivolumab (N) +/- ipilimumab (I) in patients with metastatic chemorefractory G/GEJ cancers (Abstract 4014). One hundred and sixty patients were randomized to receive N 3 mg/kg Q2W (N3), N 1 mg/kg + I 3 mg/kg Q3W (N1+I3), or N 3 mg/kg + I 1 mg/kg Q3W (N3+I1). The primary endpoint was ORR; 24 percent had PD-L1+ (≥ 1%) tumors. ORR was 12 percent in N3, 24 percent in N1+I3, and 8 percent in N3+I1. In patients with PD-L1 ≥ 1 percent, ORR was 19 percent (3/16) in N3, 40 percent (4/10) in N1+I3, and 23 percent (3/13) in N3+I1; in patients with PD-L1 < 1 percent, ORR was 12 percent (3/26), 22 percent (7/32), and 0 percent (0/30), respectively. Median OS was 6.2 months in N3, 6.9 months in N1+I3, and 4.8 months in N3+I1. The authors concluded that N ± I led to durable responses and long-term OS in Western patients with advanced G/GEJ cancer, which is consistent with the ONO-12 study in Asian patients.

Data was presented on cohort 1 of Keynote-059, which reported on the efficacy (ORR) and safety of pembrolizumab in patients with treated advanced gastric cancer (Abstract 4003). A total of 259 patients who had progressed on ≥2 prior regimens were enrolled. PD-L1+ patients had expression in ≥1 percent (composite positive score, CPS) tumor or stromal cells using IHC (22C3 antibody). Fifty-seven percent had PD-L1+ tumors. Overall RR was 11.2 percent (95% CI, 7.6-15.7), and in PD-L1+ patients ORR was 15.5 percent (95% CI, 10.1-22.4). Four percent of patients had MSI and RR was 57 percent in MSI-H patients. Eighteen gene T-cell inflamed gene expression profile score showed significant association with response to pembro (p=0.014).

Researchers reported on cohort 2 of Keynote-059, which evaluated efficacy of pembro plus 5-FU and cisplatin for first-line treatment of HER2 G/GEJ cancers (Abstract 4012). Of the 25 enrolled patients, 64 percent had PD-L1+ tumors. ORR was 60 percent (95% CI, 38.7-78.9) in all patients, 69 percent (95% CI, 41.3-89.0) in PD-L1+ patients, and 37.5 percent (95% CI, 8.5-75.5) in PD-L1 patients. Median PFS was 6.6 months (95% CI, 5.9-10.6); median OS was 13.8 months (95% CI, 7.3-not estimable).

The SWOG 1201 study showed no difference in OS using ERCC1 as a biomarker for platinum use for treatment selection for advanced metastatic E/G/GEJ cancers (Abstract 4009). The ABSOLUTE phase III study evaluated nanoparticle albumin-bound paclitaxel (nab-PTX) versus solvent-based paclitaxel (sb-PTX) in patients with pre-treated advanced gastric cancer and showed non-inferiority for Q1w nab-PTX in terms of OS and QOL (Abstract 4010). This study does not establish a new SOC given the non-inferiority question.

Hepatocellular Carcinoma (HCC)

METIV-HCC was a second-line, phase III study with tivantinib (ARQ 197, T, oral MET inhibitor) versus placebo in patients with MET-high HCC (MET ≥2+ in ≥50% of tumor cells) by centralized IHC (Abstract 4000). A total of 589 MET-high patients were randomly assigned 2:1 to oral tivantinib or placebo. Dose had to be decreased due to high neutropenia. Median OS was 8.4 months in tivantinib arm and 9.1 months in placebo, HR = 0.9, P = 0.81. Despite phase II efficacy signal, the phase III tivantinib study did not show improvement, as has been the case for other HCC studies with the exception of regorafenib, which showed improvement in OS in the phase III RESOURCE study.

Cheng reported on the data on the phase III trial of lenvatinib (LEN) versus sorafenib (SOR) in first-line treatment of patients with unresectable HCC. A total of 954 patients were enrolled and randomized 1:1 (LEN: 478; SOR: 476). Thirteen percent of LEN-treated and 9 percent of SOR-treated patients discontinued due to adverse events. Median OS was 13.6 months for LEN and 12.3 for SOR (HR, 0.92). The authors concluded that LEN is noninferior to SOR. Given that LEN was not less toxic or improved OS, it has not really added a novel option for HCC.

The SIRveNIB study evaluated the role of selective internal radiation therapy using SIR-Spheres yttrium-90 microspheres, Y90 (SIRT) versus sorafenib in locally advanced HCC in a phase III, multi-center, open-label, randomized controlled trial of 360 patients (Abstract 4002). OS in the Y90 and sorafenib arms was 8.54 and 10.58 months, respectively (HR 1.17, p =0.203). Tumor response rate was 16.5 percent and 1.7 percent (p < 0.001), respectively. This study was negative for OS and PFS and the role of Y90 remains undetermined. Another study reported on a smaller single institution study evaluating the safety and efficacy of sorafenib and Y-90 for advanced HCC (Abstract 4083). Thirty-eight patients were treated with sorafenib followed (after 4 weeks) with Y90 and the estimated mOS was 18.46 (95% C.I. 12.29 – NA); PFS was 12.29 months (5.72 - 18.79).

Data was reported on nivolumab in sor-naïve and treated patients with advanced HCC (CheckMate 040 study) (Abstract 4013). Overall, patients (n=262) had a median follow-up of 12.9 months, and 98 percent had Child-Pugh scores 5-6. In sor-naïve patients (n=80), the ORR was 23 percent, with 44 percent of responses (8/18) ongoing. The DCR was 63 percent; 40 percent of patients had stable disease ≥6 months. In sor-treated patients (n=182; 91% progressed on sor), the ORRs were 16-19 percent. Overall, responses occurred regardless of etiology or tumor PD-L1 expression.

Biliary Cancers (BTC)

Resected disease: The BILCAP randomized trial reported on the role of adjuvant capecitabine for BTC and 447 patients were randomized to capecitabine (n = 223) or observation (n = 224) (Abstract 4006). Primary sites were 84 (19%) intrahepatic, 128 (28%) hilar, 156 (35%) extrahepatic CCA, and 79 (18%) muscle-invasive gallbladder cancers. By ITT analysis (n = 447), median OS was 51 months for capecitabine and 36 months for observation, HR 0.80. This study confirmed the benefit of adjuvant capecitabine and is a reasonable standard of care.

Front-line advanced disease: Researchers presented encouraging results of the single-arm, phase II trial of gemcitabine (G), cisplatin (C), and nab-paclitaxel (N) in advanced biliary tract cancers (aBTC) (Abstract 4018). In the initial 51 patients treated, mPFS = 11.4 months (95% CI: 6.1, not reached) and mOS not reached (estimated > 20 months, 1-year survival rate 66.7%) and these were higher than historical control. Dose adjustments were needed due to grade 3 hematologic toxicity. These results merit evaluating GC +/-N in a randomized controlled study. The results of GN doublet were not significantly different from historical control—the median PFS and OS were 6.5 and 10.3 months, respectively. The triplet combination may have value if clinically meaningful OS and manageable toxicity can be achieved in a larger randomized study.

Subsequent lines: There is currently no SOC regimen and the SWOG S1310 randomized phase II study was stopped early with negative results reported for single agent MEK inhibitor trametinib versus 5-FU or capecitabine in refractory aBTC (Abstract 4016). OS for trametinib was inferior to 5FU (4.3 months vs. 8 months HR of 2.02, p=0.05). There were encouraging results in the biomarker-enriched trials given the rich NGS yield of BTC and likely driver mutations with IDH1 and FGFR alterations as potential therapeutic targets in 20-25 percent of intrahepatic cholangiocarcinomas (ICC). In addition, research was presented on the escalation and expansion cohorts of a phase I study of AG-120, an IDH1 mutant enzyme inhibitor (Abstract 4015). Among the 72 efficacy evaluable, mIDH1 ICC patients, 6 percent (n = 4) had a confirmed partial response and 56 percent (n = 40) experienced stable disease. A global, phase III, randomized, placebo-controlled study of AG-120 in mIDH1 CC has been initiated (ClarIDHy).

Data were reported on ARQ 087, an oral pan-fibroblast growth factor receptor (FGFR) inhibitor, in patients with advanced ICC with FGFR2 genetic aberrations (Abstract 4017). Thirty-five (119 total enrollment) ICC patients with FGFR2 genetic aberrations were treated with ARQ 087 daily. Thirty patients had at least one post-treatment radiographic assessment: six (20%) had PR (32-47% tumor reduction, all FGFR2 fusion positive), 17 SD (seven had tumor reduction between 10% and 25%), and seven had PD. Both studies (IDH1 and FGFR inhibitors) showed manageable toxicity and prolonged stable disease as likely best outcome for these agents.

Pancreatic Cancer (PDAC)

Resected disease : Results of the randomized phase II portion of NRG Oncology/RTOG 0848 evaluating the addition of erlotinib to adjuvant gemcitabine for patients with resected pancreatic head adenocarcinoma showed no benefit for erlotinib added to gemcitabine as adjuvant therapy for PDAC (Abstract 4007). Accrual to the trial is continuing to answer the phase III radiation question. Of note, the median OS for the gemcitabine arm was approximately 30 months (21.7-33.4), which has been a steady improvement over the past 2 decades and likely related to better staging, patient selection, and perioperative care.

Advanced disease: Researchers presented the results of the randomized, phase II study of PEGPH20 plus nab-paclitaxel/gemcitabine (PAG) versus AG in patients with untreated, metastatic PDAC (Abstract 4008). PEGPH20 degrades hyaruloran (HA), increasing the access and therapeutic index of anticancer agents. In stage II, randomization was 2:1 to PAG versis AG. Tumor HA was tested using a novel assay. Primary endpoints were PFS (evaluable patients) and TE event rate (stage II). Secondary endpoints were PFS by HA level and ORR. Of 246 patients with HA data, 84 (34%) were HA-High. The primary PFS endpoint was statistically significant for PAG versus AG (HR 0.73, mPFS 6.0 vs. 5.3). PFS in HA-High patients was also statistically significant in the PAG versus AG arm and was 9.2 months versus 5.2 months (HR 0.51; p = 0.048). OS in HA-High patients was 11.5 months (PAG) and 8.5 months (AG) (HR 0.96). The authors concluded that these data support HA as a potential predictive biomarker for patient selection of PEGPH20.

Colorectal Cancer

Resected disease: A shorter duration oxaliplatin adjuvant therapy would save patients significant neurotoxicity/time and reduce the health care costs. The rationale for the IDEA collaboration was to determine whether the standard 6 months of oxaliplatin-based chemotherapy (FOLFOX or CAPOX) is necessary (Abstract LBA1). A prospective, pre-planned pooled analysis of six concurrently conducted, randomized phase III trials (>12,000 patients) was performed to evaluate the non-inferiority (NI) of 3 months compared with 6 months of adjuvant FOLFOX/XELOX. The primary endpoint was DFS.

To demonstrate NI, the investigators reached consensus that a 12 percent relative risk increase in DFS in the 3-month arm compared with the 6-month arm was acceptable. This required a NI margin of 1.12 for the DFS hazard ratio (HR). Patient characteristics varied across the six studies; two studies enrolled a higher percentage of patients with T4 disease, and CAPOX use varied from 0 percent to 75 percent across the studies. The estimated 3-year DFS in the 3-month arm was lower than that in the 6-month arm by 0.9 percent (HR 1.07, 95% CI [1.00, 1.15]).

Neither TOSCA (Abstract 3501) nor the IDEA France study (Abstract 3500) were able to demonstrate that 3 months was as efficacious as 6 months, and it seemed that the benefit in IDEA was driven by the results from SCOT study (Abstract 3502) of 6,088 patients (4,107 patients on XELOX), which showed that 3 months adjuvant treatment is not inferior to 6 months treatment. The consensus recommended 3 months of adjuvant chemotherapy for patients with low-risk disease, defined as T1-3N1 tumors (~60% of stage III patients). For high-risk patients, defined as patients with T4 or N2 tumors, decisions on use of the shorter course should be individualized. It was noted that the IDEA collaboration highlights some of the drawbacks of pooled analyses of varied independent trials. The analysis of 3-month versus 6-month data did not meet its preplanned primary endpoint of noninferiority for DFS and comparison of low-risk versus high-risk disease was an unplanned post-hoc analysis. With that in mind, clinicians should make a decision for use of the shorter course based on an individual assessment of tolerability, risk, and choice of regimen.

Preliminary data was reported from two ongoing phase I studies of a novel carcinoembryonic antigen (CEA) CD3 T-cell bispecific antibody, alone (31 patients) and in combination with the PD-L1 inhibitor atezolizumab (36 patients), in patients with heavily pretreated, mostly microsatellite stable mCRC (Abstract 3002). The flexible bispecific antibody (CEA-TCB) simultaneously binds to CD3 on T lymphocytes and CEA on tumor cells, activating T cells in close physical proximity to the tumor cells. Bispecific antibody was used weekly and atezolizumab was administered IV Q3W. Encouraging results with stable disease and some PRs were noted. These results suggest promising activity of CD3 T cell bispecific antibody atezolizumab in MSS mCRC.

Interestingly, primary colon cancer tumor location has emerged as an independent prognostic factor when adjusted for age, gender, synchronous/metachronous, CMS, MSI, and BRAF status. Researchers reported on the sidedness (primary tumor location) data from CALGB/SWOG 80405 (Alliance), which had found no difference in OS when bevacizumab or cetuximab was added to first-line FOLFOX or FOLFIRI in All RAS wild-type mCRC patients. Left versus right mOS: 32.9 months versus 19.6 months (p < 0.0001).

Abstract 3504 reported on the somatic DNA mutations, MSI status, and mutational load (ML): same cohort of patients. Authors concluded that BRAF is a strong negative prognostic factor in mCRC, even when sidedness is taken into account, and in mutated patients longer OS is observed in bevacizumab arm versus cetuximab arm (p 0.041); in wild-type patients no arm difference is observed (p 0.291). OS did not differ between MSI-H and MSI-S patients (p 0.450). In MSI-H patients, longer OS is observed in bevacizumab arm versus cetuximab arm (p 0.002); in MSI-S patients no difference is observed (p 0.305). Excluding the hypermutated MSI-H in a subset of 205 patients, patients with ML>5 (N=93) showed a longer OS than patients with ML≤5 (N=112).

Researchers reported on the results of the randomized trial of irinotecan and cetuximab without (IC) or with vemurafenib (VIC) in BRAF-mutant mCRC (SWOG S1406) (Abstract 3505). Eligible patients had ECOG PS ≤1 and had received 1 or 2 prior regimens with no prior anti-EGFR agents. Randomization was stratified for prior irinotecan. The primary endpoint was PFS and 106 patients were enrolled, 39 percent with prior irinotecan therapy. PFS was improved with the addition of vemurafenib (HR 0.42) with median PFS of 4.4 months versus 2.0 months. Response rate was 16 percent versus 4 percent (P = 0.08), with disease control rate of 67 percent versus 22 percent. In patients with no prior irinotecan, median PFS was 5.7 months in the VIC arm versus 1.9 months in the IC arm. Authors concluded that results demonstrate the clinical benefits of the VIC triplet in patients with treatment-refractory BRAFV600 mutated mCRC, and support VIC as a potential new treatment option in this molecular subset.

The SUNSHINE study was a randomized, double-blind phase II trial of vitamin D supplementation in patients with previously untreated mCRC (Abstract 3506). One hundred thirty nine patients randomized to HiVitD experienced longer PFS compared to those randomized to LowVitD (median PFS, 12.4 vs. 10.7 months, respectively; log rank P=0.03). After multivariate adjustment for prognostic variables, HR was 0.66 (95% CI, 0.45-0.99, 2-sided P=0.04). A larger confirmatory phase III randomized trial appears warranted.

A prospective study of 992 patients from CALGB 89803 with stage III colorectal cancer (Abstract 10006) found that those who reported healthy (based on the American Cancer Society Nutrition and Physical Activity Guidelines) during and following adjuvant treatment had a 42 percent lower chance of death and trend toward improved DFS than those who had less healthy lifestyles.

Observational studies indicate that increasing nut intake is associated with lower risk of type 2 diabetes, metabolic syndrome, and insulin resistance. In a prospective, observational study of 826 patients with stage III, compared to patients who abstained from tree nuts, those who consumed ≥ two servings of nuts per week had an adjusted HR of 0.54 for DFS and 0.47 for OS (Abstract 3517). Association of total nut intake with improved outcomes was maintained across other known or suspected predictors of recurrence and mortality, including across common genomic alterations (microsatellite instability, KRAS mutation, BRAF mutation, and PIK3CA mutation).

The FOXFIRE, SIRFLOX, and FOXFIRE-Global (FF-SF-FFG) randomized studies evaluated the efficacy of combining first-line chemotherapy for metastatic colorectal cancer (mCRC) with selective internal radiotherapy (SIRT) using yttrium-90 resin microspheres in patients with liver metastases (Abstract 3507). Arm A was oxaliplatin-based chemotherapy ± investigator-chosen biologically-targeted agent. Arm B was the same systemic therapy (oxaliplatin dose modification) + single treatment SIRT with cycle 1/2 of chemotherapy. A total of 1,103 patients were randomized in 14 countries. There was no difference in OS or PFS between arms.

Fruquintinib, an oral kinase inhibitor selectively targeting VEGF receptors, showed efficacy in a phase III registration trial, FRESCO, in third-line mCRC Chinese patients (Abstract 3508). A total of 416 patients were randomized, 2:1 to fruquintinib versus placebo. Fruquintinib significantly improved OS comparing to placebo with an HR of 0.65. Median OS was 9.30 months [95% CI 8.18-10.45] in the fruquintinib group versus 6.57 months [95% CI 5.88-8.11] in the placebo group.

GAURI VARADHACHARY, MD, is Professor and Center Medical Director at the Gastrointestinal Center and Executive Medical Director of Ambulatory Operations at the University of Texas MD Anderson Cancer Center, Houston.

Gauri Varadhachary, MD
Gauri Varadhachary, MD:
Gauri Varadhachary, MD


  • FLOT is the new and improved modified-DCF regimen and can be expected to be the new standard of care in perioperative treatment of patients with G/GEJ cancers. Although, doublets (e.g., mFOLFOX) may suffice, with better tolerability and need to be tested in this setting. Until more data is available, a triplet regimen's routine use in the perioperative setting warrants continued caution.
  • There is considerable interest in anti-PD1/PDL1 therapy in untreated and treated cohorts of G/GEJ cancers. The role of PDL1 staining as a predictive biomarker of response remains to be seen (currently using CPS, IHC 22C3 antibody) and in the context of gene expression profiling, mutational load, and MSI status.
  • Tivantinib failed to improve overall survival (OS) compared with placebo as a second-line therapy for patients with MET-high inoperable HCC.
  • Given that lenvatinib was not shown to be less toxic or better than sorafenib in first-line treatment of patients with unresectable HCC, it has not really added a novel option for HCC.
  • The SIRveNIB study comparing Y90 (SIRT) versus sorafenib in locally advanced HCC was negative for OS and the role of Y90 remains undetermined. Future randomized phase III studies are warranted to assess sorafenib +/-Y90 in advanced HCC.
  • Ongoing studies with encouraging results with checkpoint blockade therapies in HCC suggest an emerging role for front and subsequent line anti-PD1 therapy for HCC over the next few years.
  • BILCAP study confirmed the benefit of adjuvant capecitabine in resected biliary cancer and should become standard of care.
  • Encouraging results from front line gemcitabine + cisplatin + nab-paclitaxel in advanced biliary cancers, merit evaluating gemcitabine/cisplatin +/- nab-paclitaxel in a randomized controlled study.
  • Encouraging early results were noted in the biomarker-enriched trials in biliary cancers with IDH1 and FGFR inhibitors with prolonged stable disease as likely best response.
  • Higher PFS and OS in the HA-high subgroup supports HA as a potential predictive biomarker for patient selection of PEGPH20 in metastatic pancreatic cancer. The ongoing global phase III HALO 301 mPC study has PFS and OS as co-primary endpoints.
  • Interestingly, primary tumor location has emerged as an independent prognostic factor when adjusted for age, gender, synchronous/metachronous, CMS, MSI, and BRAF status (left-sided colorectal cancer have better OS than right-sided colorectal cancer).
  • The IDEA consensus recommended 3 months of adjuvant chemotherapy for patients with low-risk disease, defined as T1-3N1 tumors, (~60% of stage III patients). For high-risk patients, defined as patients with T4 or N2 tumors, decisions on use of the shorter course should be individualized. Non-inferiority was not established for the entire IDEA cohort and this underscores the drawbacks of pooled analyses of varied independent trials.
  • There was clinical benefit noted of the VIC triplet (vemurafenib, cetuximab, and irinotecan) in patients with treatment-refractory BRAFV600 mutated mCRC, and support VIC as a potential new treatment option in this molecular subset.
  • Despite higher response rates and improved liver-specific PFS, the addition of SIRT to first-line oxaliplatin-fluorouracil chemotherapy for patients with liver-only and liver-dominant mCRC did not improve OS or PFS.
  • Healthy lifestyle and tree nut consumption and high-dose vitamin D seem intuitively a good practice and now prospective studies do show objective benefits of these interventions in treated and metastatic CRC, respectively.
Wolters Kluwer Health, Inc. All rights reserved.
    Home  Clinical Resource Center
    Current Issue       Search OT
    Archives Get OT Enews
    Blogs Email us!