You are here




Several clinical trials assessing the benefit of chemotherapy have been performed, including several that compared chemotherapy to best supportive care. Six trials assessed the potential benefit of early forms of chemotherapy for advanced gastric cancer compared to an untreated control group. 237-242 In the largest of these trials, 193 patients were randomly assigned to either 5-FU + MeCCNU or to no treatment. 238 The median survival time was 22 weeks for control subjects and 25 weeks for those who received at least one 6-week course of chemotherapy. Patients who died in the first 2 months were excluded from these figures. Median survival, estimated from survival curves, was in the 8- to 10-week range for all the patients, with no apparent difference between the two groups. It is of interest that a quality-of-life analysis was done in this trial and that it was published in 1978. Pain, well-being, and performance status were assessed at 8 and 16 weeks. Results of this quality-of-life analysis slightly favored the patients treated with chemotherapy, but few patient results were available for the 16-week time point. Subsequent trials with chemotherapy versus best supportive care have in general demonstrated a trend toward improved survival and quality of life with the use of chemotherapy. 237,239-241

The initial trials with chemotherapy reported response rates of 10% or higher and included agents such as 5-FU, mitomycin C, doxorubicin, epirubicin, cisplatin, BCNU (carmustine), methotrexate, etoposide, chlorambucil, and hydroxyurea. More recently, a variety of new chemotherapy drugs have become available. Some have shown promising response rates in advanced gastric cancer, as discussed in the following sections, including docetaxel, irinotecan, and oxaliplatin. Multiple phase II trials of combination chemotherapy have built upon the promising activity seen with a variety of single agents. Many of the combinations have shown promising activity based on initial results, only to be shown to be less active and more toxic in subsequent phase II and III trials. It has therefore been important to interpret the results of initial phase II trials with caution. With some of the more recent combinations showing promise, using drugs such as docetaxel or oxaliplatin, older regimens such as FAM, FAMTX, and the combination of etoposide, leucovorin, and 5-FU are rarely used.



Both intravenous (5-FU) and oral fluoropyrimidines (capecitabine, S-1) have been evaluated in trials for stomach or esophagogastric cancers. A number of trials with 5-FU have been performed over the last several decades culminating in a recent series of phase III trials (Table 75-12). 243-245 In a trial that mixed 254 patients with either adenocarcinoma, squamous cell carcinoma, or undifferentiated carcinoma of the esophagus or stomach, patients were randomly assigned to either protracted-infusion 5-FU or protracted-infusion 5-FU with mitomycin C. 244 The overall response rates were 16% and 19%, respectively, with median survivals of 6.3 and 5.3 months (P = 1.0), respectively. In a phase III trial performed in Japan, 280 patients with advanced stomach cancer were randomly assigned to receive one of three protocols: protracted-infusion 5-FU, continuous-infusion 5-FU combined with cisplatin, or the oral fluoropyrimidine UFT combined with mitomycin C. 243 This trial also showed no advantage of the two combinations over the benefits obtained with 5-FU alone. In a third phase III trial, the potentially promising combinations of etoposide + leucovorin + 5-FU, infusional 5-FU + cisplatin, and FAMTX were compared. 245 A total of 399 patients with advanced carcinoma of the stomach were randomly assigned. All three regimens showed modest activity, with no significant difference in the primary outcome measures.

The recent use of oral fluoropyrimidines in gastric cancer has focused primarily on capecitabine and S-1. As a single agent in previously untreated patients, capecitabine provided an overall response rate of 23% using a 3-week treatment schedule in a trial performed in Japan. 246 OS was 10 months. Various combinations of capecitabine and other chemotherapy drugs, such as cisplatin or docetaxel, have been evaluated. 247-250 The majority of recent trials, including several phase III trials 251,252 as described below, have evaluated capecitabine as a potential replacement for 5-FU.

The drug S-1 is an oral fluoropyrimidine in which the 5-FU prodrug tegafur is combined with two 5-FU-modulating substances, 5-chloro-2,4-dihydroxypyridine (gimeracil) and potassium oxonate (oteracil), at a molar ratio of 1 : 0.4 : 1. The dose-limiting side effects of S-1 are primarily nonhematologic and mainly diarrhea, nausea, and vomiting. There is also a difference in metabolism of the tegafur component of S-1 in Asians and whites related to polymorphic differences in the CYP2A6 gene. 253 This has led to different tolerable doses of S-1 in studies conducted in Asia compared with those conducted in Western populations. A variety of trials using S-1, alone or in combination with other chemotherapy drugs, have been performed in Asia, Europe, and North America. S-1 as a single agent has a reported response rate of 26% to 32%. 254 Based on the early evidence of activity, S-1 has been evaluated in combination with cisplatin in several phase II trials. In these trials, a response rate of 49% to 55% and OS of approximately 11 months were reported. 255-257

S-1 was directly compared with continuous-infusion 5-FU in a randomized phase III trial in Japan. 258 This trial was designed to evaluate the noninferiority of S-1 (n = 234) versus 5-FU (n = 234), using median OS as the primary end point. S-1 was noninferior to 5-FU, supporting its potential use as a substitute to 5-FU.



Taxanes have been evaluated in a variety of phase II trials. Two North American trials of single-agent paclitaxel have been reported. Paclitaxel (210 mg/m 2 over 3 hours every 3 weeks) as a single agent in previously untreated patients with metastatic gastric cancer provided a response rate of 11% and a median survival of 5.8 months. 259 In a separate trial, assessing two different infusion schedules (200 mg/m 2 over 3 hours vs. 24 hours every 3 weeks), an overall response rate of 17% was seen, with a higher rate occurring in those receiving a 24-hour infusion. The median survival in this trial was 8 months. 260 In several trials performed in Japan using paclitaxel (210 mg/m 2 over 3 hours every 3 weeks) as a single agent in the treatment of previously untreated patients with advanced gastric cancer, response rates of 23% and 28% were reported, with a median survival of 7.7 and 11.2 months. 261-263 Subsequent trials have assessed paclitaxel in combination with other chemotherapy agents. Phase II trials assessing paclitaxel in combination with cisplatin have given response rates of 33% to 46%, with median survivals of 8.9 to 13.8 months. 264,265 The use of low-dose paclitaxel in combination with cisplatin did not seem to alter outcome, with a response rate of 44% and OS of 12.1 months in a phase II trial performed in Korea. 266 A response rate of 33% and a median survival of 7.5 months have been reported in a small phase II trial with paclitaxel and carboplatin. 267 Paclitaxel in combination with 5-FU and leucovorin has been assessed in several phase II trials. In two trials using a 24-hour infusion of 5-FU, response rates of 46% to 48% and median survivals of 11 months were reported. 267,268

Docetaxel as a single agent has been evaluated in several phase II trials. In several single-institution trials of docetaxel 100 mg/m 2 every 3 weeks, a response rate of 17% and 24% was reported along with a median survival of 7.8 months in one of the trials. 269,270 Other trials assessed lower doses of docetaxel (60 or 75 mg/m 2 ) and have reported similar results, with response rates of 18% to 24% and a median survival of 11 months with the 75-mg/m 2 dose. 271-273

Subsequent trials have evaluated docetaxel in combination with several other chemotherapy agents. The combination of docetaxel, 5-FU, and leucovorin has been evaluated in several phase II trials from Europe and Korea. This combination led to a response rate of 26% to 28% and a median survival of 7.7 to 9.7 months. 274-276 In a randomized phase II trial, the combination of docetaxel and 5-FU was compared with paclitaxel and 5-FU. 277 Similar response (33% and 42%) and median survival (9.3 and 9.9 months) rates were reported for the two regimens. Several trials have assessed the combination of capecitabine and either weekly or every-3-week docetaxel and reported response rates of 39% to 60% and median survivals of 9.4 to 12 months. 249,278

The combination of docetaxel and cisplatin has been evaluated in several clinical trials as either the two-drug combination or in combination with other chemotherapy agents. With the two-drug combination in phase II trials, the reported response rate was 28% to 46% and median survival of 10.4 to 11.5 months. 279-282 Given the apparent promising activity of this combination, other trials explored the added benefit of 5-FU and reported response rates of 40% to 43% and median survivals of 9.0 to 9.7 months. 283,284 A multicenter randomized phase II trial, involving 158 patients, assessed docetaxel and cisplatin with or without 5-FU as potential support for a phase III trial. 285 In this trial, the combination of docetaxel and cisplatin had a response rate of 26% and a median survival of 10.5 months. The addition of 5-FU increased the response rate to 43%; however, the median survival was only 9.6 months.

In a subsequent phase III trial (see Table 75-12) the combination of docetaxel, cisplatin, and 5-FU was compared with cisplatin and 5-FU. 286 OS was significantly longer with docetaxel, cisplatin, and 5-FU (P = 0.02). The response rate was also higher with this combination (69% vs. 59%, P = 0.01). However, the combination of docetaxel, cisplatin, and 5-FU was also more toxic; treatment-related grade 3 or 4 adverse events developed in 69% of patients, compared with 59% of patients receiving cisplatin and 5-FU. Neutropenia and neutropenic fever were both significantly higher with the three-drug combination. A modified regimen of docetaxel, cisplatin, and 5-FU showed improved tolerability and led to a subsequent phase II trial in combination with bevacizumab. 287 A separate trial of a low-dose weekly regimen produced minimal toxicity, but only modest clinical benefit. 288


Platinum-containing regimens have been the focus of many different trials. Many of the trials conducted in the past focused on the use of cisplatin, whereas more recent trials have begun to evaluate the potential use of oxaliplatin. The combination of 5-FU and cisplatin has been evaluated in several trials and provided a response rate of 41% to 45% and a median survival of 7 to 11 months. 289-293 Given the apparent activity of this combination, a variety of phase III trials have compared 5-FU and cisplatin with other regimens (see Table 75-12). The response rates in these larger trials were 20% to 51%, and the median survival was 7.2 to 8.6 months. 245,286,294,295

Building on this two-drug combination, a variety of trials have assessed the combination of either 5-FU, cisplatin, and doxorubicin (FAP) or EAP. The FAP combination resulted in response rates of 31% to 50% and median survivals of 9 to 12 months in phase II trials. 296-298 A phase III study, however, could not show an advantage for FAP over 5-FU alone. 299 Less favorable outcomes were noted for the combination of EAP in phase II trials. 300,301 This combination also caused significant toxicity, with a toxic death rate averaging 12%. The combination of cisplatin, epirubicin, leucovorin, and 5-FU (PELF) was compared to FAMTX in a randomized phase III trial. Although both the overall and CR rates with PELF were significantly higher (overall, 39% vs. 22%; CR, 13% vs. 2%), only a nonsignificant improvement in OS was seen with PELF (see Table 75-12). 302

The ECF combination has been recognized as a potentially more promising cisplatin-containing regimen. Based on the results of phase II trials, this regimen showed activity at least comparable to other platinum regimens, but with less toxicity. 303-306 A subsequent phase III trial of patients with previously untreated locally advanced or metastatic gastroesophageal cancer compared ECF with the combination of FAMTX. 307 The overall response rate (45% vs. 21%, P = 0.0002) and median survival (8.9 vs. 5.7 months, P = 0.0009) were both significantly higher with ECF. The global quality-of-life scores were also better in patients receiving ECF.

Both carboplatin and oxaliplatin have been evaluated as potential alternatives to cisplatin. Only limited trial data are available for carboplatin. Carboplatin as a single agent has minimal activity. 308,309 However, the combination of carboplatin and paclitaxel has a reported response rate of 33% and median survival of 7.5 months with only moderate toxicity. 267

Oxaliplatin was initially developed for colorectal cancer but has been evaluated as a potential treatment option for gastric cancer when used in combination with 5-FU and other chemotherapy agents. Although the use of 5-FU and leucovorin alone result in response rates of 5% to 12% with short median survivals, 310,311 the addition of oxaliplatin seems to significantly improve response and OS. A variety of schedules using the combination of 5-FU, oxaliplatin, and leucovorin have been evaluated in phase II trials, including FLO, FUFOX, FOLFOX, and XELOX. These regimens have generally shown similar results with response rates of 38% to 65% and median OSs of 8.6 to 11.5 months. 312-317 In a phase III trial of 5-FU and leucovorin with either oxaliplatin or cisplatin, equivalent outcomes were seen. 318 With 220 patients randomly assigned, a nonsignificant trend in progression-free survival (PFS) and OS was noted.

Given the potentially promising activity of ECF, as described previously, a recent phase III trial for gastroesophageal cancer was conducted to assess the potential benefit of replacing cisplatin with oxaliplatin and infusional 5-FU with capecitabine using a 2 × 2 design (see Table 75-12). 319 This trial enrolled 1002 patients from 61 centers, and the results have been presented in a meeting abstract. Using a noninferiority statistical design for the trial, the combination of epirubicin, oxaliplatin, and capecitabine (EOX) was shown to be the most active of the four regimens evaluated. Compared with ECF, patients receiving EOX had both an improved 1-year survival (46.8% vs. 37.7%) and a statistically improved median survival (11.2 months vs. 9.9 months; HR: 0.80, 95% CI: 0.65, 0.97, P = 0.02). 251 In a smaller trial of 316 patients randomly assigned to either capecitabine and cisplatin or 5-FU and cisplatin, the combination of capecitabine and cisplatin showed significant noninferiority for PFS (5.6 vs. 5.0 months, HR: 0.81, 95% CI: 0.63, 1.04, P < 0.001). 252 A metaanalysis of these two trials concluded that the capecitabine combinations result in a significantly improved OS. 320

Two phase III trials have assessed the activity of S-1 in combination with cisplatin in patients with advanced stomach or esophagogastric adenocarcinoma. In the SPIRITS trial conducted in multiple centers in Japan, 150 patients were randomly assigned to receive S-1 alone and 148 to S-1 and cisplatin. 321 The addition of cisplatin resulted in a significant improvement in PFS (6.0 vs. 4.0 months, HR: 0.57, 95% CI: 0.44, 0.73, P < 0.0001) and OS (13.0 vs. 11.0 months, HR: 0.77, 95% CI: 0.61, 0.98; P = 0.04). The addition of cisplatin did result in a higher rate of grade 3 and 4 hematologic toxicity, nausea, and anorexia. In a separate trial (FLAGS) conducted in 24 countries and involving 1,053 patients (1% Asian), S-1 and cisplatin was compared with 5-FU and cisplatin. 322 Although the combination of S-1 and cisplatin had significantly less toxicity, no difference was seen in median OS.



The topoisomerase-I inhibitor irinotecan has activity in gastric cancer, though limited as a single agent. In a phase II trial, it provided a median survival of 6.4 months with a 14% response rate. 323 Irinotecan has been evaluated in combination with other potentially active chemotherapy drugs. In combination with docetaxel, a response rate of 46% and OS of 8.2 months has been reported. 324 Similar response rates and OS have been reported for irinotecan and cisplatin, irinotecan and capecitabine, irinotecan and mitomycin C, and irinotecan and oxaliplatin. 325-328 The addition of the antiangiogenic agent bevacizumab to the combination of irinotecan and cisplatin in a phase II trial did increase the response rate to 65% and the OS to 12.3 months. 329 The combination of irinotecan with leucovorin and infusional 5-FU (FOLFIRI) in a randomized phase II trial also showed promising results, with a response rate of 40% and OS of 11.3 months. 330 Phase III trials with these combinations have not been reported.

Continuous infusion 5-FU was compared to irinotecan and cisplatin in a randomized phase III trial in Japan. 258 This trial was designed to evaluate the superiority of irinotecan and cisplatin, using median OS as the primary end point. With 236 patients randomly assigned to irinotecan and cisplatin and 234 to 5-FU, 5-FU provided an equivalent outcome. The median survival for those receiving 5-FU was 10.8 months versus 12.3 months for those receiving irinotecan and cisplatin (HR: 0.85, 95% CI: 0.70, 1.04, P = 0.0552). A randomized trial compared the combination of irinotecan, 5-FU, and leucovorin with 5-FU and cisplatin. 331 Equivalent outcomes were seen as measured by OS and time to progression. However, the irinotecan-containing regimen was better tolerated.

Targeted Therapy


A variety of molecularly targeted therapies have been evaluated in patients with esophagogastric cancers. In general, the targeted therapies are intended to enhance and not replace chemotherapy. Multiple studies have indicated that markers of enhanced angiogenesis are correlated with worse outcomes. Many studies in particular have shown that high levels of vascular endothelial growth factor (VEGF) are predictive of poor survival. 332-334 Based on this observation, several clinical trials have evaluated the benefit of bevacizumab added to chemotherapy. Several phase II trials have suggested benefit to the use of bevacizumab. 335,336 However, a multinational randomized phase III trial (Avastin in Gastric Cancer [AVAGAST]) for patients with advanced stomach or GEJ cancers of capecitabine or 5-FU combined with cisplatin (n = 387) alone or in combination with bevacizumab (n = 387) failed to show any significant benefit in OS. 337 The median survival for patients receiving chemotherapy alone was 10.1 months compared to 12.1 months for those receiving bevacizumab (HR: 0.87, 95% CI: 0.73, 1.03, P = 0.1002).

Trials focusing on targeting Her-2 overexpression, however, have shown benefit in appropriately selected patients. Approximately 20% of patients with esophagogastric cancers show overexpression of Her- 2. 338 In the ToGA trial patients with adenocarcinoma of the stomach or GEJ, overexpressing Her-2, were randomly assigned to either chemotherapy alone, with capecitabine or 5-FU combined with cisplatin (n = 296), or chemotherapy with trastuzumab (n = 298). 339 Overexpression of Her-2 was based on preset criteria specific to stomach cancer. Median OS increased from 11.1 months to 13.8 months (HR: 0.74, 95% CI: 0.60, 0.91, P = 0.0046). A significant improvement in PFS and response rates were also seen.

A number of phase II trials have also explored the use of epidermal growth factor receptor (EGFR) inhibitors in esophagogastric cancers. The EGFR inhibitor cetuximab appears to have minimal activity as a single agent. 340 With no completed phase III trials of an EGFR inhibitor having been reported, the benefit of an inhibitor combined with chemotherapy remains uncertain. In one randomized phase II trial of matuzumab combined with chemotherapy (epirubicin, cisplatin, capecitabine) compared with chemotherapy alone, no apparent benefit was seen to the addition of matuzumab. 341

As with other targeted therapies, the ability to appropriately select patients for use of a specific agent will likely require molecular markers that correlate with response. Targeted therapies will undoubtedly have an increasing role in the treatment of esophagogastric cancers. Their role continues to be reliant on the development and completion of phase III trials specifically evaluating their use.



In summary, it is clear that chemotherapy may result in response rates of up to 50% or more in selected groups of patients with advanced esophagogastric cancer. There is now growing evidence from both recent phase II and III trials indicating that chemotherapy seems to prolong survival over what would be expected with best supportive care. In recent trials, median survival times have begun to clearly increase, with subsequent improvement in outcome measures including OS for treated patients compared to what had been observed in the prior several decades of clinical trials. However, it is not clear that there yet exists a standard treatment for advanced gastric cancer. Investigation of new agents and combinations must continue, and any “new standards” should be tested in controlled clinical trials looking at survival, quality of life, and cost analysis endpoints.

For patients with metastatic disease, the activity of the chemotherapy needs to be balanced with the toxicities of the regimen used. Although three-drug regimens have shown higher response rates and improved survival, their use is limited by their toxicity. Consideration should be given to the use of two-drug combinations in older patients or patients with a less robust performance status. In patients with tumors overexpressing Her-2, the addition of trastuzumab to a two-drug regimen may provide improved survival while limiting toxicity.