Targeted Therapy for Gastric Cancer
- First Online:
- Cite this article as:
- Smyth, E.C. & Cunningham, D. Curr. Treat. Options in Oncol. (2012) 13: 377. doi:10.1007/s11864-012-0192-6
- 1.1k Views
For patients with advanced gastric cancer, traditional double or triplet cytotoxic chemotherapy regimens result in a median survival of 9–11 months. As combination therapy is associated with increased survival, but also increased toxicity in a patient population whose performance status often compromised by their malignancy, development of more effective and less toxic treatment choices is mandated. Emerging data from gene expression profiling suggests that differences in pathological appearance and clinical behavior may be due the presence of unique molecular phenotypes. Characterization of the gastric cancer genomic landscape reveals the presence of multiple alterations in expression of receptor tyrosine kinases, which in conjunction with their ligands and downstream effector molecules represent potentially druggable pathways for future drug development. Treatment of HER2 positive gastric cancer with trastuzumab has led to significant gains in overall survival, and further manipulation of this pathway using the novel anti-HER2 directed agents pertuzumab and T-DM1 in addition to dual EGFR/HER2 blockade with lapatinib may yield positive results. In contrast, targeting of the EGFR pathway in combination with chemotherapy in unselected patients has not been fruitful to date, with no significant gains over standard chemotherapy yet demonstrated. Similarly, use of the anti-angiogenic monoclonal antibody bevacizumab was not successful in a large global randomized trial; however intriguing regional variations were seen with respect to efficacy of this drug, leading to calls for a second, regionally stratified study. Careful selection of patient subsets will become a key factor in future clinical trials, as novel targeted agents such as those targeting the MET/HGF and FGFR axes move forward into clinical development. It is hoped that treatment of patients in such molecularly defined groups is will lead to significant gains in survival compared to current treatment paradigms.
KeywordsGastric cancerTargeted therapyMolecular classificationHER2EGFRTrastuzumabBevacizumabFGFRMET
Selected trials of targeted agents in advanced gastric cancer
Anti-HER2 directed therapy
Bang et al. [12••]
Trastuzumab + cisplatin/5FU
Pishviain et al. 
Lapatinib + capecitabine
28% at 5 m
Iqbal et al. 
Anti-EGFR directed therapy
Pinto et al. 
Cetuximab + FOLFIRI
Moehler et al. 
Cetuximab + FUFIRI
Pinto et al. 
Cetuximab + docetaxel/cisplatin
Han et al. 
Cetuximab + FOLFOX
Kim et al. 
Cetuximab + XELOX
Woll et al. 
Cetuximab + oxaliplatin/irinotecan
Enzinger et al. 
Cetuximab + ECF
Cetuximab + cisplatin/irinotecan
Cetuximab + FOLFOX
Richards et al. 
Cetuximab + DOCOX
Rao et al. 
Matuzumab + ECX
ECX chemotherapy alone
Dragovich et al. 
Erlotinib (1st line)
Janmaaet et al. 
Gefitinib (2nd line)
Ferry et al. 
Gefitinib (1st/2nd line)
Anti-angiogenic monoclonal antibody therapy
Shah et al. 
Bevacizumab + cis/irinotecan
Shah et al. 
Bevacizumab + docetaxel/cis/5-FU
El-Reyes et al. 
Bevacizumab + doce/oxali
Enzinger et al. 
Bevacizumab + docetaxel/cis/irinotecan
Ohtsu et al. [41••]
Bevacizumab + cis/capecitabine
Placebo + cis/capecitabine
Antiangiogenic tyrosine kinase inhibitors
Sun et al. 
Sorafenib + doce/cisplatin (1st line)
Kim et al. 
Sorafenib + cisplatin/cape (1st line)
Bang et al. 
Sunitinib (2nd line)
Moehler et al. 
Sunitinib (2nd line)
Gastric cancer is a globally and biologically heterogeneous disease
Gastric cancer exhibits both geographical and molecular heterogeneity which may potentially impact on the development of targeted therapy for this disease. In Western countries, due to changes in environmental influences, a decrease in distal gastric cancer has been demonstrated, with a concurrent surge in numbers of patients diagnosed with proximal gastroesophageal junction type tumours, which remain relatively uncommon in Asia . Both distal and proximal intestinal type tumours are associated with dietary factors and chronic inflammation (H. pylori associated gastritis and reflux oesophagitis respectively), and with precursor lesions of metaplasia (Barrett’s oesophagus) and dysplasia preceding the development of overt malignancy . Diffuse gastric cancer, in contrast, is associated with similar incidence rates worldwide, no known environmental factors, and no pathological precursor lesion.
Developments in the molecular profiling of gastric cancer
These clinical and geographical differences are also apparent at the molecular level. Tan et al. have recently demonstrated two distinct subtypes of gastric cancer based on RNA expression profiling on 37 gastric cancer cell lines, which were then validated in 531 patients from four independent cohorts [8••]. These subtypes, G-INT and G-DIFF, were associated with highly distinctive gene expression profiles, and prognoses. Although the subtypes were significantly associated with the Lauren’s histological intestinal and diffuse subtypes, concordance was imperfect at 64%, reflecting the limitations of the current histological classification system. A similar approach by Shah et al., using a supervised (a priori determined) clustering analysis in a 36 patient dataset yielded three distinct subtypes consistent with diffuse, and proximal and distal intestinal type clinical and pathological classifications .
The genomic landscape of gastric cancer has also recently been described, which has identified common alterations and mutations which may serve as targets for molecularly guided therapy. In a study of 193 gastric cancer patient samples, 93 matched normal samples, and 40 gastric cell lines, 22 discrete areas commonly associated with high levels of copy number gain or loss were identified [10•]. Known amplifications such as EGFR, ERBB2/HER2 and CCND1 were confirmed, and novel amplifications of the transcription factors GATA6 and KLF5 were detected, as were deletions of PARK2, PDE4D, CSMD1 and GMDS. Analysis of inter-target relationships in this study demonstrated distinct patterns of co-amplification and mutual exclusivity within these amplifications and deletions, implying complex intracellular functional relationships with consequences for targeted therapy development. Potentially druggable receptor tyrosine kinase (RTK) alterations occurred in 37% of this patient cohort; the most frequently amplified RTK component was FGFR2 (9.3%), followed by KRAS (8.8%), EGFR (7.7%) and ERBB2 (7.2%). In Cox multivariate analysis and independent of chromosomal instability, RTK amplification status was shown to be an independent poor prognostic predictor for survival (HR 1.966, 95% CI 1.180 to 3.279, p = 0.01). Furthermore, although activating mutations of druggable pathways such as those of BRAF, KRAS and PIK3CA are rare in sporadic gastric cancer, it has recently reported that these mutations may be much more common in the microsatellite subset of gastric cancer. In a study of 63 microsatellite unstable gastric cancer specimens, Corso et al. demonstrated the presence of a mutation of either EGFR, KRAS, PIK3CA and MLK3 oncogenes in 56% of cases , implying that genotyping this subset of gastric cancer patients may prove fruitful when identifying patients who may benefit from targeted therapy.
HER2 directed therapy in gastric cancer
Lapatinib (Tykerb ©GlaxoSmithKline) is an oral available dual inhibitor of HER2 and EGFR and has demonstrated efficacy in the treatment of HER2 positive trastuzumab resistant breast cancer. In an unselected, previously untreated gastric cancer population, lapatanib as a single agent demonstrated minimal activity, with a response rate of 7% . In this study (SWOG S041), median overall survival was five months, less than that seen with conventional cytotoxic chemotherapy. In a molecularly selected population (EGFR or HER2 positive by IHC or FISH) undergoing second line treatment, lapatanib demonstrated an ORR of 0% and stable disease was seen in 8% of patients . More encouraging results have been seen with lapatinib in conjunction with capecitabine in the first line HER2 positive metastatic gastric cancer setting; in a preliminary report of 67 patients treated with this combination response and stable disease rates were 22% and 45% respectively . Lapatinib is also under evaluation for the first line treatment of gastric cancer in conjunction with ECX (NCT01123473) and CapeOx chemotherapy (LOGIC NCT00680901), whilst results from the second line TYTAN (NCT00486954) study (paclitaxel ± lapatinib) which has completed recruitment are awaited.
EGFR directed therapy for gastric cancer
The epidermal growth factor receptor (EGFR) is constitutively expressed at multiple sites, including skin, gut and renal tissue. Overexpression of this growth factor receptor is seen in 27–64% of gastric cancers, with increased rates in more proximal tumours [18, 19]. EGFR overexpression in gastric cancer is associated with older age, more aggressive histology and higher stage disease, and in addition is an unfavourable prognostic marker in multivariate analysis .
In contrast to NSCLC, were the use of the small molecule tyrosine kinase inhibitors erlotinib and gefitinib has met with substantial success, these agents have exhibited very limited activity in gastric cancer. Single agent erlotinib was associated with a response rate of 0% in the first line gastric cancer setting (9% for gastroesophageal adenocarcinomas) . This was reflected in overall survival of 4 and 7 months respectively for these groups. Gefitinib has demonstrated comparable response rates and overall survival in both treatment naive and refractory settings [21, 22]. Activity may be associated with increased EGFR expression; it is hoped that biomarker results from the phase III randomised COG trial (NCT01243398) will shed light on this issue.
Cetuximab (Erbitux©, Imclone Systems), the partially humanised murine anti-EGFR monoclonal antibody, has been the most extensively examined anti-EGFR therapy in gastric cancer. This agent has minimal activity as a single agent; a phase II study of 35 previously treated patients with metastatic gastric cancer yielded a response rate of 3%, a stable disease rate of 6% and a median overall survival of 3.1 months . Six non-randomised first line phase II studies investigated the addition of cetuximab to doublet chemotherapy [24–28]. Response rates in these studies ranged from 41–63%, and median overall survival from 9 m–16.6 m. A randomised study comparing the addition of cetuximab to three discrete chemotherapy backbones was presented at ASCO 2010. In the ECOG 1206 study anti-EGFR cetuximab (C) therapy was added to ECX, FOLFOX and irinotecan/cisplatin combination regimens. FOLFOX-C was associated with the lowest rates of grade ≥3 toxicity, and in common with ECF-C demonstrated a response rate of >50%. ECF-C and FOLFOX-C also demonstrated comparable rates of overall survival at 10 months each. In contrast, irinotecan/cisplatin plus cetuximab was associated with a modest response rate of 38% and an overall survival of 8.6 m and the highest rate of treatment modifications making this the least attractive option for future study . A further randomised phase II study comparing the addition of cetuximab to docetaxel and oxaliplatin chemotherapy was recently presented . This 150 patient study demonstrated an increase in response rate for the cetuximab containing arm (from 24% to 29%), but no increase in progression free or overall survival. Results of the large phase III randomised EXPAND study (NCT00678535) which examined the addition of cetuximab to cisplatin and capecitabine chemotherapy are expected to be presented in 2012.
The addition of the anti-EGFR monoclonal antibodies matuzumab and panitumumab to cytotoxic chemotherapy for gastric cancer has been associated with increased toxicity and the necessity for potentially deleterious dose reductions in the standard therapy arm. Matuzumab (EMD72000, Merck), was investigated in the phase II randomised MATRIX trial in combination with ECX chemotherapy . In this study, response rate was significantly lower in the experimental arm (31% vs. 58%) and both progression free and overall survival demonstrated a non-significant trend towards inferiority in the matuzumab-ECX arm (4.8 m vs. 7.1 m and 9.4 m vs. 12.2 m) respectively). It was postulated that the inferior results seen for matuzumab treated patients in this study was potentially because of elevation of the biologically required active dose of matuzumab beyond the maximally tolerated dose of this agent due to the addition of ECX chemotherapy. A third fully humanised anti-EGFR monoclonal antibody, panitumumab (Vectibix©, Amgen) has been evaluated in combination with EOC (epirubicin, oxaliplatin and capecitabine) chemotherapy in advanced gastric cancer in the large phase II randomised phase III REAL-3 study (NCT00824785). An initial phase II randomised safety run demonstrated a high incidence of grade 3 diarrhoea and neutropenia using standard dose EOC with panitumumab (37.5% each) . One toxic death was associated with EOC-P in this 16 patient cohort. This led to a reduction in the dose of EOC used in conjunction with panitumumab in the Phase III trial proper which has closed to recruitment and the results of which are awaited.
Biomarkers associated with response to cetuximab
Unlike metastatic colorectal cancer, were the presence of a KRAS mutation is common and associated with lack of response to anti-EGFR therapy, this mutation is both rare in gastric cancer, and does not appear to be associated with de novo resistance to this class of drugs [26, 28, 33]. Correlative biomarker analyses have been conducted for multiple studies of anti-EGFR therapy in gastric cancer. Luber et al. performed correlative analyses on 75% of the 52 patients treated with FOLFOX plus cetuximab in a non-randomised phase II study. Increased EGFR gene copy number (≥ 4.0), was significantly associated with better OS (HR 0.2, 95% CI: 0–0.8; P = 0.022) . Response rate and time to progression were significantly better in patients without evidence of EGFR phosphorylation; however this did not lead to a significant overall survival benefit. Moehler et al. examined EGFR, PTEN, HER2, VEGF-D, and VEGFR3 expression and correlated this with outcome in a study of 5-FU, irinotecan and cetuximab . Higher levels of EGFR expression were associated with increased response rates, but not with time to progression or overall survival. Expression of PTEN was significantly associated with both improvements in response rate and progression free and overall survival (OS mean: 23.8 versus 14.3 months, log-rank P = 0.0127) and PFS times (mean: 14.0 versus 6.8 months, log-rank P = 0.035). This contrasts with the biomarker analysis conducted on the Phase II DOCETUX study, in which negative/low EGFR expression and high ERK expression were associated with response to therapy . It is clear that small numbers and lack of a control arm in these studies impedes the ability to discern whether these are truly important either prognostic or predictive markers.
Targeting gastric cancer angiogenesis
Anti-angiogenic strategies have demonstrated efficacy in colorectal cancer, renal cell carcinoma, NSCLC, glioblastoma and ovarian epithelial cancers. As high tumour and serum VEGF levels are associated with poor prognosis in both resectable and advanced gastric cancer, it was hypothesised that the addition of bevacizumab, the anti-VEGF monoclonal antibody would enhance survival in this disease [35, 36]. Initial phase II single arm studies appeared promising; bevacizumab in combination with first line chemotherapy in advanced gastric cancer (largely US patients) was well tolerated, with median overall survival of almost 17 months achieved [37–40]. These impressive results were not sustained in the global, phase III randomised AVAGAST study, which randomised 774 patients to cisplatin/fluropyrimidine combination chemotherapy with or without bevacizumab [41••]. Although response rate and progression free survival were both improved in the bevacizumab arm (37% vs. 46% and 5.3 m vs. 6.7 m respectively), this was not reflected by an improvement in overall survival. Median survival was 10.1 m in the chemotherapy arm vs. 12.2 months in the bevacizumab containing arm (HR 0.87, p = 0.1002). Toxicity was acceptable, with no increase in rates of ≥ grade 3 haemorrhage and rates of gastrointestinal perforation in the bevacizumab group commensurate with those seen in colorectal cancer (2.3%). Regional variation in outcomes were seen, with notably different hazard ratios seen for response rate, progression free survival and overall survival in Asian, European and American (North and South) patient cohorts. At this time, it is unclear whether this relates to differences in treatment patterns or underlying disparities in disease biology and biomarker data are eagerly awaited.
Bevacizumab is currently under investigation in the MRC ST03 trial, which will randomise 900 patients to the combination of ECX (epirubicin, cisplatin and capecitabine) chemotherapy plus or minus bevacizumab in the peri-operative setting (NCT00450203). An initial safety report of the first 104 patients randomised to this study did not demonstrate any increase in the rates of gastrointestinal bleeding or wound healing complications for the experimental arm . Additionally, no excess of cardiac events or decreases in left ventricular ejection fraction were detected for the anthracycline containing ECX and bevacizumab combination. This study is likely to complete accrual in early 2013. Bevacizumab is also being studied in a randomised phase II trial in the perioperative setting for gastroesophageal junctional and oesophageal cancer in junction with FOLFOX chemotherapy (NCT01212822).
A second mechanism by which the VEGF axis is targeted is by the anti VEGFR-2 monoclonal antibody ramicurumab (IMC-1121B, ImClone) which is currently being evaluated in three randomised trials in advanced gastric cancer, both alone in comparison to best supportive care (NCT00917384), in conjunction with paclitaxel in the second line setting (NCT01170663) and with FOLFOX chemotherapy (NCT01246960) for previously untreated patients.
Alternate anti-angiogenic strategies using small molecule tyrosine kinase inhibitors have produced mixed results in gastric cancer. The multi-targeted TKI sorafenib (Nexavar©, Bayer) has activity against VEGFR −1 and −2. In a preliminary report of a second line study, one patient treated with single agent sorafenib in the second line setting exhibited a durable complete response of more than eleven months duration . This agent has been examined in combination with cisplatin and capecitabine in a phase I dose escalation study for patients with advanced gastric cancer. This trial demonstrated a response rate of 63%, a progression free survival of 10 months and overall survival of 15 months . In a phase II study of sorafenib in combination with cisplatin and docetaxel in the first line setting resulted in a RR of 41%, and overall survival of 13.6 months . Grade ≥ 3 neutropenia was seen in 64% of patients. In contrast, a second oral TKI with antiangiogenic properties, sunitinib (Sutent©, Pfizer) has been studied in a two second line gastric cancer patient cohorts, yielding disappointing response rates of 3-4%, progression free survival was of 1.3–2.6 m and overall survival was 5.8–6.8 months [46, 47]. Several sunitinib plus chemotherapy combination studies (NCT00555672—sunitinib plus cisplatin/5FU; NCT01020630—sunitinib plus FOLFIRI) are ongoing or completed and awaiting results.
Novel gastric cancer targets: PI3K/mTOR, HGF/MET and FGFR
Activation of the PI3K/mTOR pathway in gastric cancer has been demonstrated to be activated in gastric cancer in preclinical studies, and this pathway has been successfully exploited in the treatment of neuroendocrine tumours and renal cell carcinoma . Two single arm phase II studies of single agent everolimus (Afinitor©, Novartis) in the treatment of second line gastric cancer were associated with response rates of 0–4%, progression free survival of 1.7–2.7 months and overall survival of 8.3–10.1 months [49, 50]. In one of these studies, baseline expression of phosphorylated S6Ser240/4 was a significant predictor of progression free survival (HR 0.246, 95% CI 0.078–0.777) p = 0.017 ). However, the recently presented phase III randomised double blind GRANITE-1 study, which compared the use of everolimus 10 mg/d or placebo plus best supportive care in the second line treatment of advanced gastric cancer did not demonstrate any benefit in term of overall survival for the everolimus treatment arm (4.3 m placebo vs. 5.4 m everolimus HR 0.9, p = 0.1244) . Modest differences in six month progression free survival estimates were seen at 12% for everolimus and 4% for placebo, possibly indicating a subgroup treatment effect, and biomarker results are pending.
The MET receptor and its ligand hepatocyte growth factor/scatter factor (HGF/SF) are key mediators of the epithelial to mesenchymal transition (EMT) during embryogenesis, and in tumour invasion and metastasis. Overexpression of MET occurs in 50–60% of gastric cancers, and does not appear to be associated with prognosis, however, MET copy number gain ≥5x, which is less common (10%) has been independently associated in multivariate analysis with inferior outcomes in resected patients . Amplification of MET is similarly rare (2%), and is also a negative prognostic factor in advanced gastric cancer . When foretinib (GSK1363089, GlaxoSmithKline), a oral, small-molecule inhibitor of cMET and VEGFR2/KDR was examined the second line advanced gastric cancer setting in a non-molecularly selected patient cohort no responses were seen . However, in a randomised study of first line ECX chemotherapy with or without the anti- hepatocyte growth factor (HGF) antibody rilotumumab (AMG102, Amgen) a trend was demonstrated towards increased response rates for the combination arm for patients with high levels of expression of c-MET . Onartuzumab (MetMab©, Genentech), a humanised anti-METantibody has been tested in early phase studies with an anecdotal report of sustained response occurring in gastric cancer . The potential utility of MET as a target is further emphasised by reports of objective sustained clinical responses in two of four patients with MET copy number gain ≥5 treated with crizotinib, a tyrosine kinase inhibitor which targets both ALK and MET tyrosine kinases .
Abnormalities of fibroblast growth factor (FGF) signalling are frequently present in gastric cancer, in particular FGFR-2 amplification, which is detected in up to 9% of all gastric cancers with an increased frequency in diffuse type cancer [10•]. Numerous FGFR inhibitors are in clinical development, having demonstrated efficacy in cell lines and in human xenograft models in vivo [57, 58]. Brivanib, the selective dual inhibitor of FGF and VEGF was studied in a in 34 molecularly unselected chemotherapy refractory patients with advanced gastric cancer, and led to an 18% disease control rate at 12 weeks . The selective FGFR 1,2 and 3 inhibitor AZD4547 (Astra Zeneca) is currently being compared to paclitaxel in the second line setting for patients with FGFR polysomy or amplification (NCT01457846, SHINE), with multiple other studies planned.
With growing understanding of the molecular heterogeneity underlying gastric cancer, treatment paradigms are likely to move away from current “one-size fits all” cytotoxic chemotherapy regimens. The success of the TOGA trial, were survival in the most accurately molecularly targeted subgroup was almost double that previously seen with standard regimens, provides an excellent illustration of rational trial design. In order to best evaluate novel targeted agents, careful selection of possibly small patient subsets which may benefit from treatment using molecular characterisation is necessary. The addition of a molecular classification system to the current pathological one may aid in this aim. The failure of the AVAGAST study provides a useful counterpoint to this success – when designing global cancer trials, consideration must be given to regional differences in treatment patterns and disease biology. Failure to do so may lead to rejection of potentially efficacious drugs. It is also necessary to be vigilant with respect to toxicity resulting from the addition of targeted agents to standard cytotoxic chemotherapy, which may lead to dose reductions and inferior survival outcomes. However, with the emergence of selective inhibitors of novel targets which have demonstrated encouraging efficacy in early phase trials, future treatment options for gastric cancer appear bright.
D Cunningham: Has grants/grants pending from Amgen, Roche, and Merck Serono; EC Smyth: none.