GAURI VARADHACHARY, MD, is Professor in the Department of Gastrointestinal Medical Oncology at the University of Texas MD Anderson Cancer Center.
What have we learned about advances in GI cancers from the 2013 publications? The literature in the past year suggests a continued emphasis on evaluating the role of predictive markers and understanding cancer biology/heterogeneity.
A brief review cannot do justice to the vast number of important publications, so I highlight here five significant papers from 2013—I have chosen one (or more with a similar theme) for each gastrointestinal disease site that signifies the advances, drawbacks, and additional work planned ahead.
RAS Mutations and Management of Colorectal Cancer—Looking beyondKRASExon 2 Mutation; Updated PRIME Study(Douillard et al: Panitumumab-FOLFOX4 treatment and RAS mutations in colorectal cancer. NEJM 2013;369:1023-1034)
The original report from the PRIME study (JCO 2010;28:4697-4705) concluded that panitumumab-FOLFOX4 was well tolerated and significantly improved progression-free survival (PFS) in patients with wild-type (WT) KRAS tumors. Consistent with the results from other studies, patients with KRAS mutations in exon 2 (codons 12, 13) did not benefit from the addition of anti-epidermal growth factor receptor (EGFR) monoclonal antibody (mAb) therapy in the PRIME study.
Approximately 40 percent of tumor specimens from patients with metastatic colorectal cancer have KRAS mutations involving codons 12 and 13 (exon 2) or rarely codon 61 (exon 3).
In the updated analysis from the PRIME study, 639 patients who had metastatic colorectal cancer without KRAS mutations in exon 2 had tumor tested for at least one of the following: KRAS exon 3 or 4; NRAS exon 2, 3, or 4; or BRAF exon 15.
The overall rate of ascertainment of RAS status was 90 percent. A total of 108 patients (17%) with nonmutated KRAS exon 2 had other RAS mutations. These mutations were associated with inferior progression-free and overall survival (OS) with panitumumab-FOLFOX4 treatment, which was consistent with the findings in patients with KRAS mutations in exon 2. BRAF mutations were a negative prognostic factor.
Among 512 patients without (any) RAS mutations, progression-free survival was 10.1 months with panitumumab-FOLFOX4 versus 7.9 months with FOLFOX4 alone (hazard ratio for progression or death with combination therapy, 0.72; 95% confidence interval [CI], 0.58 to 0.90; P=0.004). In this group of patients, overall survival was 26.0 months in the panitumumab-FOLFOX4 group versus 20.2 months in the FOLFOX4-alone group (hazard ratio for death, 0.78; 95% CI, 0.62 to 0.99; P=0.04).
This study by Douillard et al evaluated the role of other activating RAS mutations as negative predictive biomarkers for anti-EGFR therapy, suggesting an urgent need for an expanded panel of RAS mutations (beyond KRAS codons 12, 13, 61) to be tested in patients embarking on cetuximab or panitumumab therapy. These tests will identify additional patients who will not benefit from anti-EGFR mAb therapy and guide clinicians to make the right decisions for their patients.
Gemcitabine and Nab-Paclitaxel for Metastatic Pancreatic Cancer—Finally, Some Options?(Von Hoff D et al: Increased survival in pancreatic cancer with nab-paclitaxel plus gemcitabine. NEJM 2013;369:1691-1703)
After two decades of disappointing Phase III trials and years of single-agent gemcitabine therapy, there are new front-line options for patients with metastatic pancreatic cancer.
Nab-paclitaxel was approved by the FDA for use in metastatic pancreatic cancer in September 2013 after the MPACT results demonstrated that the gemcitabine/nab-paclitaxel combination was superior to single-agent gemcitabine.
A total of 861 patients were randomly assigned to nab-paclitaxel plus gemcitabine (431 patients) or gemcitabine (430). The median overall survival was 8.5 months in the nab-paclitaxel–gemcitabine group compared with 6.7 months in the gemcitabine group (hazard ratio for death, 0.72; 95% CI, 0.62 to 0.83; P<0.001). The survival rate was 35 percent in the nab-paclitaxel–gemcitabine group versus 22 percent in the gemcitabine group at one year, and nine versus four percent at two years.
The most common adverse events of grade 3 or higher were neutropenia (38% in the nab-paclitaxel–gemcitabine group vs. 27% in the gemcitabine group), fatigue (17% vs. 7%), and neuropathy (17% vs. 1%). In the nab-paclitaxel–gemcitabine group, neuropathy of grade 3 or higher improved to grade 1 or lower in a median of 29 days. The authors concluded that the combination is superior to gemcitabine alone but causes more myelosuppression and peripheral neuropathy, which appear to be reversible.
Going forward, there are questions pertaining to selecting appropriate first-line therapy—choosing between FOLFIRINOX and gemcitabine/nab-paclitaxel. Cross-trial comparisons of progression-free and overall survival are hazardous, although there are some differences between FOLFIRINOX and gemcitabine/nab-paclitaxel that have practical implications.
The FOLFIRINOX study excluded patients older than 75 and included only those with an Eastern Cooperative Oncology Group (ECOG) performance status (PS) of 0 or 1. In MPACT, 10 percent of patients were over age 75 and about seven to eight percent of patients had a poorer PS, corresponding to an ECOG PS of 2 (the study allowed a Karnofsky PS of greater than 70).
The FOLFIRINOX study had more patients with carcinomatosis compared with the gemcitabine/nab-paclitaxel trial (19% vs. 4%). A randomized trial may (or may not) select a “better” front-line regimen, but without predictive biomarkers, we are far from the promise of personalized medicine and being able to match the right combination with the right patient.
The mechanism of action of nab-paclitaxel in this setting is the subject of ongoing research. Studies have shown that nab-paclitaxel improves intratumoral concentration of gemcitabine in murine models of pancreatic cancer, either through stromal depletion or by decreasing the primary gemcitabine-metabolizing enzyme, cytidine deaminase.
In the Phase I/II study of gemcitabine/nab-paclitaxel, secreted protein acidic and rich in cysteine (SPARC) was evaluated in 36 patients. A significant increase in OS was observed for patients in the high-SPARC group (n = 19) compared with those in the low-SPARC group (n = 17); median OS was 17.8 vs. 8.1 months, respectively (P = .0431).
The SPARC level remained a significant predictor of OS in a multivariate Cox regression model after adjusting for clinical covariates, including sex, race, age, treatment, and baseline CA19–9 level (P = .041). In addition, stromal SPARC was significantly correlated with OS (P = .013) but not SPARC in tumor cells (P = .15). At this time, we eagerly await tissue correlative data from the MPACT study.
Personalized Therapy for Advanced Hepatocellular Carcinoma (HCC)—Will cMET Inhibitors Get Us There?(Santoro A et al: Tivantinib for second-line treatment of advanced hepatocellular carcinoma: a randomised, placebo-controlled phase 2 study. Lancet Oncol 2013;14:55-63)
Currently, sorafenib is the only drug approved for advanced HCC with limited survival benefit. Preclinical animal and tissue-based studies and early Phase 1b studies suggest an emerging role for therapeutic inhibitors of the HGF-cMET axis in HCC. Binding of HGF to c-MET induces activation of downstream PI3K, Ras/Raf/MEK/ERK, and Cdc42/Rac1 signaling pathways, leading to cell proliferation, survival, and angiogenesis.
This presents the opportunity for the development of several -mumabs and -tinibs targeting the HGF-cMET pathway (including anti-HGF antibodies, anti c-MET mAb, selective oral small molecule inhibitors of c-MET, and non-selective multikinase inhibitors).
In the Lancet Oncology paper, Santoro and colleagues studied tivantinib (ARQ 197), a selective oral inhibitor of MET, in a multicenter, 2:1 randomized, placebo-controlled, double-blind, Phase II study—all patients had advanced HCC and Child-Pugh A cirrhosis, and had progressed on or were unable to tolerate first-line systemic therapy. The tivantinib dose was amended to 240 mg twice-daily because of a high incidence of treatment-emergent grade 3 or worse neutropenia.
The primary endpoint was time to progression, according to independent radiological review in the intention-to-treat population. Tumor samples were assessed for MET expression with immunohistochemistry (MET-high was defined as ≥2+ in ≥50% of tumor cells).
In this study, 71 patients were randomly assigned to receive tivantinib (38 at 360 mg twice-daily and 33 at 240 mg twice daily) and 36 to placebo. Although the time to disease progression was numerically better for patients treated with tivantinib (1.6 months) compared with placebo (1.4 months), these results are not clinically meaningful.
Interestingly, for patients with MET-high tumors, median time to progression (TTP) was 2.7 months with tivantinib (22 patients) vs. 1.4 months with placebo (15 patients). Median overall survival was 7.2 months (95% CI 3.9–14·6) for patients with MET-high tumors who received tivantinib versus 3.8 months (2.1–6.8) for MET-high patients who received placebo (HR 0.38, 95% CI 0.18–0.81; p=0.01).
There were no differences in efficacy between the tivantinib and placebo groups in the MET-low subgroup. Comparison of outcomes by MET status, independent of treatment, in the placebo group showed that patients with MET-high tumors had significantly shorter survival than those in the MET-low subgroup (3.8 vs. 9.0 months, respectively).
This trial is the first to personalize therapy for advanced HCC. It suggests that cMET is a prognostic and predictive biomarker and that patients with MET-high tumors who have failed or are intolerant to sorafenib might benefit from a targeted MET inhibitor. Based on these results, a large Phase III study is planned for patients with cMET-high HCC tumors.
Role of Anti-EGFR Therapy in Gastroesophageal Cancers—Do We Have a Final Verdict after REAL-3 and EXPAND?(Waddell T et al: Epirubicin, oxaliplatin, and capecitabine with or without panitumumab for patients with previously untreated advanced oesophagogastric cancer (REAL3): a randomised, open-label phase 3 trial. Lancet Oncol 2013;14:481-489); Lordick F et al: Cetuximab and cisplatin with or without cetuximab for patients with previously untreated advanced gastric cancer (EXPAND): a randomised, open-label phase 3 trial. Lancet Oncol 2013;14:490-499)
REAL-3 was a randomized, open-label Phase III trial that evaluated the addition of the anti-EGFR monoclonal antibody panitumumab to epirubicin, oxaliplatin, and capecitabine (EOC) in patients with unselected, untreated locally advanced, or metastatic gastroesophageal adenocarcinoma. A total of 553 patients were randomly allocated to receive up to eight 21-day cycles of open-label EOC alone or modified-dose EOC plus panitumumab (mEOC+P).
After a preplanned independent data-monitoring committee review, trial recruitment was halted and panitumumab was withdrawn. There was no difference in median overall survival (11.3 vs. 8.8 months) or PFS in patients treated with EOC vs. m-EOC+P, the latter showing a detrimental effect on OS and PFS.
Patients with rash did better than those with no rash in the mEOC+P arm (OS was 10.3 vs. 4.3 months, respectively). The authors concluded that the addition of panitumumab to EOC chemotherapy does not increase overall survival and cannot be recommended for use in an unselected population with advanced gastroesophageal adenocarcinoma.
REAL-3 is not alone. The EXPAND trial assessed the addition of cetuximab to a capecitabine-cisplatin doublet in 904 patients with previously untreated adenocarcinoma of the stomach and gastroesophageal junction, and did not meet its primary endpoint of improved PFS and showed no benefit in OS. Additionally, the COG trial assessed the anti-EGFR tyrosine-kinase inhibitor gefitinib compared with placebo in the second-line treatment of 450 patients with esophageal and gastroesophageal junction cancers. This trial also did not meet its primary endpoint, with no improvement in overall survival.
Should we be surprised by these results? No. Given the trend of targeted therapy results in unselected patient populations of GI cancers, the pretest probability of failure is rather high even with well-powered studies. The REAL-3 authors discuss the role of dose intensity (mEOC+P got less chemotherapy), and one can argue that is likely not a significant contributor. The possibility of a negative interaction between panitumumab and one or more of the EOC components has been raised.
Most importantly, it's back to the drawing board. At this time, we are far from the promise of personalized treatment using antiEGFR mAb therapy in gastroesophageal cancer (some may wonder if there is any role at all).
EGFR modulation is likely very different in gastroesophageal cancers than in colon cancer or pancreatic cancer. In addition, there may be significant differences between localized vs. metastatic disease and untreated vs. treated disease. A very small number of patients had KRAS mutations in the REAL-3 study, and no BRAF mutations were noted.
To plan smarter trials, we need to have a better understanding of the molecular taxonomy of gastroesophageal cancers, understand driver vs. passenger genes and signaling pathways, and evaluate the various cross-talking functional networks as a whole and determine how they change with stress (chemotherapy and immune system)—and unfortunately, toxic biologics in unselected patients will not get us there.
Intrahepatic Cholangiocarcinoma and Dysregulated Chromatin Remodeling—Some Light Shed on the Neglected One?(Jiao Yet et al: Exome sequencing identifies frequent inactivating mutations in BAP1, ARID1A, and PBRM1 in intrahepatic cholangiocarcinomas. Nat Genet 2013;45:1470-1473); Chan-On et al: Exome sequencing identifies distinct mutational patterns in liver fluke-related and non-infection-related bile duct cancers. Nat Genet 201;45:1474-8)
Most studies on molecular alterations in biliary tract cancers have focused on a small set of genes, often those known to be altered in pancreatic ductal adenocarcinomas. The studies discussed herein are of significant interest since it describes the results of exome sequencing in a series of clinically and pathologically well-characterized intrahepatic cholangiocarcinomas and gallbladder cancers with an impact on therapeutic targets.
In the U.S. study (Jiao et al), patients had no known risk factors for the development of bile duct cancers, and in the study from Asia and Europe (Chan-On et al), half the subjects had bile duct cancers caused by liver fluke.
In the discovery screen, Jiao et al sequenced the exomes of 32 intrahepatic cholangiocarcinomas and nine gallbladder cancers. The researchers discovered frequent inactivating mutations in multiple chromatin-remodeling genes, including BAP1 (25%), ARID1A (19%), and PBRM1 (17%). Mutation in one of these genes occurred in almost half of the carcinomas sequenced.
These findings highlight for the first time the key role of chromatin remodeling in this cancer. Patients with a mutation in any one of the three chromatin-remodeling genes (BAP1, ARID1A, or PBRM1) trended toward worse survival compared with subjects in whom all three genes were wild type (three-year survival rate of 47.1 percent for subjects with mutations compared with 93.3 percent for those without mutations), but these results were not statistically significant (P = 0.1672 by log-rank test). This group also identified somatic mutations at previously reported hotspots in the IDH1 and IDH2 genes encoding metabolic enzymes in intrahepatic cholangiocarcinomas (19% total).
Somatic mutations were identified in FGFR2 in intrahepatic cholangiocarcinoma (13%) and in several components of the PI3K pathway (22%) representing possible therapeutic targets for intrahepatic cholangiocarcinomas. In their prevalence screen, the group sequenced a subset of the potential driver genes in an independent set of 32 intrahepatic cholangiocarcinoma samples. As was observed in the discovery screen, multiple mutations were identified in chromatin-remodeling genes (34%) and in the IDH1 and IDH2 hotspots (22%).
When both the discovery and prevalence screens were combined for gallbladder cancers (17 samples total), TP53 was the most frequently altered gene in the series (44%) and the overall mutation prevalence in gallbladder carcinoma for BAP1, ARID1A, and PBRM1 was six percent, six percent, and 25 percent, respectively. No mutations were identified in IDH1 or IDH2 in all the gallbladder carcinomas analyzed.
In a separate paper, Chan-On and colleagues studied the impact of different carcinogenic exposures on the specific patterns of somatic mutations in bile duct cancers. The researchers profiled 209 cholangiocarcinomas from Asia and Europe, including 108 cases caused by infection with the liver fluke (Opisthorchis viverrini) and 101 cases caused by non-liver fluke related etiologies.
Whole-exome sequencing (n = 15) and prevalence screening (n = 194) identified recurrent somatic mutations in BAP1 and ARID1A. Comparisons between intrahepatic liver fluke-related and non-liver fluke-related cholangiocarcinomas demonstrated statistically significant differences in mutation patterns: BAP1, IDH1, and IDH2 were more frequently mutated and TP53 less frequently mutated in non-liver fluke cholangiocarcinomas. Functional studies demonstrated tumor suppressive functions for BAP1 and ARID1A, establishing the role of chromatin modulators in cholangiocarcinoma pathogenesis. These findings indicate that different causative etiologies may induce distinct somatic alterations, even within the same tumor type.
These two studies suggest that chromatin remodeling is an important area of investigation. Comparison of somatic mutation data for cholangiocarcinoma and gallbladder carcinoma suggests that, although both tumor types arise from biliary epithelium and that they are genetically distinct. We need larger studies to better understand the biological and clinical significance of these findings.