In 2003, an estimated 22,400 individuals in the United States will be diagnosed with gastric cancer, and 12,100 will die from the disease.[ 1] The worldwide incidence is approximately 550,000 cases per year, with 405,000 deaths.[2] Despite attempts at curative resection, survival for this disease remains poor (Table 1).[3] As a result, investigators have been evaluating adjuvant therapy to prevent recurrences and improve on overall mortality. Studies of the benefits of adjuvant therapy for gastric adenocarcinoma have yielded conflicting results. The more favorable results reported in Asian countries compared to Western countries may relate to ethnicity factors,[4] although differences in tumor location, prevalence of earlystage disease, and extent of preoperative staging evaluation may also account for some of these results.[5] Such discrepancies make it difficult to translate the results of studies in one of these populations to the other. Treatment Options Adjuvant and Neoadjuvant Radiotherapy
Gastric cancer has generally been considered a relatively radioresistant carcinoma, with limited benefits in symptom palliation for advanced disease.[ 6] However, the high incidence of locoregional recurrence after surgery for early disease has sparked interest in radiation as adjuvant or neoadjuvant therapy. Among patients who undergo a curative resection, 38% to 67% will develop a clinically evident locoregional recurrence, with nodal stage and number of positive lymph nodes predictive of risk.[7] Investigators have hypothesized that radiotherapy may sterilize the local bed and prevent the growth of presumed residual microscopic disease after surgery. Few trials have studied the impact of radiation alone as adjuvant therapy. The British Stomach Cancer Group randomized 436 patients who had undergone curative resection to no further therapy, adjuvant radiotherapy, or adjuvant chemotherapy. The 5-year overall survival for the 153 patients randomized to adjuvant radiotherapy was 12%, compared to 20% for those assigned to surgery alone (P > .2).[8] In China, researchers randomized 370 patients with locally advanced gastric cancer of the cardia to either immediate surgery or preoperative radiation (40 Gy) followed by surgery.[ 9] Patients who received neoadjuvant radiation had a significantly improved 5-year overall survival (30% vs 20%, P = .009) and a lower local recurrence rate (39% vs 52%, P < .025) compared to those who were treated with surgery alone. Although these results are intriguing, the authors did not provide information on preoperative staging, and the study was limited to patients with cardiac and gastroesophageal junction cancers. Recently, Skoropad et al reported on a trial of preoperative radiotherapy with 20 years of follow-up.[10] Of 152 patients who underwent exploratory laporatomy for staging, 102 were randomized to a short course of preoperative therapy (20 Gy over 5 days) followed by surgery, or surgery alone. The authors did not find any significant differences in survival between the two treatment groups (P = .56). Moreover, subset analyses by tumor location or stage of disease did not demonstrate any significant survival differences between treatment arms. Further studies with radiation alone are unlikely, given the renewed interest in the combination of chemotherapy and radiotherapy in the adjuvant setting. Intraoperative Radiotherapy
Intraoperative radiotherapy (IORT) may be an attractive option for achieving local control in gastric cancer due to its ability to deliver a single, large fraction of radiation with minimal damage to normal tissue.[9] Although most studies of IORT have been nonrandomized and conducted by single institutions, several experiences are notable. In Japan, a retrospective case-control comparison in 242 patients demonstrated that those with stage II-IV disease who received IORT achieved a 10% to 20% improvement in 5-year survival.[11] In contrast, Sindelar and colleagues at the National Cancer Institute (NCI) randomized 41 gastric cancer patients to either externalbeam radiation or IORT after surgery.[ 12] Although locoregional failure was less common among patients in the IORT arm (44% vs 92%, P < .001), overall survival did not differ between treatment arms. A randomized trial in Germany compared surgery alone to surgery with IORT and also found no statistical differences in overall survival or cancerrelated death.[13] Systemic Therapy
Approaches to adjuvant chemotherapy for gastric cancer have been based largely on the results of regimens evaluated in the setting of metastatic disease. Unfortunately, the use of chemotherapy in patients with advanced gastric cancer has produced disappointing results. Complete responses are rare, and partial responses to single-agent therapies have been limited, with results ranging from 0% to 30%.[14] Fluorouracil(Drug information on fluorouracil) (5-FU) remains the most widely used agent, with response rates to singleagent therapy in previously untreated patients ranging from 21% to 30%.[15,16] Other agents with activity as single-agent therapy include cisplatin(Drug information on cisplatin), doxorubicin(Drug information on doxorubicin), epirubicin(Drug information on epirubicin) (Ellence), docetaxel (Taxotere), paclitaxel(Drug information on paclitaxel), and irinotecan(Drug information on irinotecan) (CPT-11, Camptosar).[14] Combination chemotherapy regimens for metastatic gastric cancer have been tested extensively. Initial trials have often demonstrated striking results, with the majority of patients achieving a response. However, confirmatory trials in larger populations of patients have resulted in more sobering findings.[17] Many combination regimens have not been shown to improve overall survival compared to single-agent therapy when tested in prospective, randomized trials.[18-21] Recently, the combination of epirubicin, cisplatin, and 5-FU (ECF) produced improved response and overall survival rates with lower toxicity, compared to a prior standard of care, FAMTX (5-FU, doxorubicin [Adriamycin], methotrexate(Drug information on methotrexate)).[22] However, the median overall survival among patients treated with ECF was still only 8.7 months, with a 2-year survival rate of 14%. Other promising combination regimens include irinotecan or a taxane and a platinum agent,[14] but randomized, multicenter trials must be conducted to discern the true efficacy of these newer drugs in gastric adenocarcinomas. Adjuvant Chemotherapy
The administration of chemotherapy shortly after complete resection of gastric tumors in patients at high risk of recurrence has been assessed as a means of treating clinically undetectable micrometastases. A number of studies have investigated whether adjuvant chemotherapy can reduce the incidence of the disease recurrence and mortality following a curative resection. To date, four meta-analyses have been published that examined the impact of postoperative adjuvant chemotherapy on survival after surgical resection (Table 2).[23-26] Hermans et al analyzed 11 randomized trials, with a total of over 2,000 patients, that compared postoperative chemotherapy to surgery alone.[24] Most of these trials used older 5-FU-based regimens, including FAM (5-FU/doxorubicin/mitomycin [Mutamycin]), 5-FU/mitomycin, and 5-FU/semustine (methyl-CCNU). The investigators were not able to demonstrate a significant survival advantage for adjuvant chemotherapy beyond that achieved with surgery alone. However, the analysis was later criticized for a lack of sufficient statistical power and the studies chosen for inclusion. Earle and Maroun compiled a similar meta-analysis, limited to non- Asian studies and with stricter inclusion criteria.[23] The authors reported an odds ratio for death in the treated group of 0.80 (95% confidence interval [CI] = 0.66-0.97), corresponding to a relative risk of 0.94 (95% CI = 0.89-1.00). They calculated that in a group of patients similar to those included in the analyzed trials, 65% would have a recurrence and die of gastric cancer following treatment with surgery alone, compared to 61% following treatment with surgery and adjuvant chemotherapy. This absolute risk reduction of 4% indicates that 25 patients would need to receive adjuvant therapy to prevent one death (ie, the "number needed to treat" = 25). An Italian group led by Mari performed a meta-analysis of 20 trials published between 1982 and 1999.[25] Using time-to-event data with censoring, they found a favorable hazard ratio of 0.82 (95% CI = 0.75-0.89) for postoperative adjuvant chemotherapy over surgery alone. In a subgroup analysis, they did not detect any statistical differences between trials, with or without the use of anthracycline-based therapy. Most recently, Panzini and colleagues reported a meta-analysis similar to that of Mari et al, although trials that did not require complete resection of disease were excluded.[26] Of the 17 trials included, 14 overlapped with those in the analysis by Mari et al. The odds ratio for death for patients receiving adjuvant therapy in this analysis was 0.72 (95% CI = 0.62-0.94). Despite the modestly positive results of more recent meta-analyses, enthusiasm for adjuvant systemic chemotherapy remains limited. These analyses are tempered by the usual restraints, including publication bias, differences in patient populations and entry criteria of the trials, and a nonrandomized design. Indeed, no individual, prospective randomized trial has provided convincing evidence in favor of adjuvant chemotherapy. In addition, the authors of each of these four meta-analyses argue that larger, randomized trials are warranted and that adjuvant chemotherapy after curative resection for gastric cancer should be considered investigational. Neoadjuvant Chemotherapy
The use of neoadjuvant chemotherapy to downstage locally advanced disease and improve the resectability of the primary tumor in gastric cancer patients has attracted the attention of researchers. Phase I/II studies have demonstrated the feasibility of preoperative chemotherapy, with downstaging achieved in up to 50% of patients and complete pathologic responses in up to 10%.[27-30] In addition, operative morbidity and mortality were not increased with the use of preoperative therapy. However, prospective randomized trials have not demonstrated a benefit for neoadjuvant chemotherapy. The Dutch Gastric Cancer Group randomized 56 patients to preoperative FAMTX or immediate surgery.[ 31] In this small trial, 44% of patients did not complete the planned four cycles of FAMTX due to pro- gressive disease, and a nonsignificant improvement in resection rates was seen in the surgery-only arm. Moreover, patients randomized to preoperative FAMTX did not experience any survival benefit compared to those receiving surgery alone. Preliminary reports of neoadjuvant chemotherapy trials conducted in Asia have also failed to demonstrate improved survival rates for preoperative therapy.[ 32,33] Several large, randomized trials are currently being conducted in Europe to evaluate the role of neoadjuvant chemotherapy in the treatment of gastric cancer (Table 3).[14] Adjuvant Chemoradiotherapy
Gastric adenocarcinoma has a tendency to recur both locoregionally and distantly following curative resection (Table 4).[34-37] As a result, many investigators have explored the combination of chemotherapy and radiotherapy to prevent tumor recurrence. Several phase II trials have demonstrated favorable long-term survival with the use of chemoradiotherapy after curative resection, with 5-year survival rates ranging from 21% to 55%.[38] Previous randomized trials comparing surgery alone to chemoradiotherapy after surgery, however, were fraught with design flaws and did not convincingly support adjuvant treatment.[ 39,40] Dent et al randomized 142 patients with all stages of gastric cancer, including T4 and M1 disease, to surgery alone or postoperative therapy including 2,000 cGy of radiotherapy delivered in eight fractions and chemotherapy with either 5-FU or thiotepa(Drug information on thiotepa).[39] No differences in survival rates (P > .5) or quality of life emerged between the control and treatment arms. Of note, the radiation and chemotherapy doses in this trial would now be considered suboptimal. Subsequently, Moertel and colleagues randomized 62 patients with resectable gastric cancer to surgery alone or postoperative radiotherapy with 3,750 cGy plus 5-FU.[40] Due to a later-realized design flaw, informed consent was obtained after randomization and ~25% of patients in the postoperative therapy arm re- fused treatment. Intent-to-treat analysis demonstrated an improvement in 5-year survival in the adjuvant therapy arm (23% vs 4%, P < .05). However, when the treatment group was subdivided according to those who actually received treatment and those who refused it, 5-year survival rates between these subgroups and the control group were not statistically different. Of note, treatment did reduce locoregional recurrences (39% for treated patients compared to 54% for those who received no treatment). The authors concluded, "this study does not establish 5-FU plus radiation as effective surgical adjuvant therapy for gastric cancer but suggests this approach as a possible fruitful area for continued research."[40]

• INT 0116 -Ultimately, between 1991 and 1998, a large, US intergroup adjuvant chemotherapy and radiation therapy trial (INT 0116) was initiated by the Southwest Oncology Group (SWOG).[37] Investigators randomized 556 patients with resected stage IB-IV, M0 gastric and gastroesophageal adenocarcinoma to either no further therapy or postoperative chemoradiotherapy to investigate the potential benefit of adjuvant chemoradiotherapy. Patients were stratified by tumor stage and nodal status. Those randomized to chemoradiotherapy received one course of 5-FU at 425 mg/m²/d and leucovorin at 20 mg/m²/d on days 1 through 5, followed by 45 Gy of radiation (to the original tumor bed, regional lymph nodes, proximal and distal resection margins plus 2 cm) concurrently with bolus 5-FU at 400 mg/m² and leucovorin at 20 mg/m² on each of the first 4 and last 3 days of radiotherapy. Two additional cycles of 5-FU and leucovorin were administered 1 month after radiation therapy. A single radiation protocol coordinator reviewed all radiotherapy treatment plans. Of note, more than one-third of the initial plans were revised upon review. Patients treated with postoperative chemoradiation achieved a significant improvement in both disease-free and overall survival at 3 years (Table 5). These results translated into a 52% improvement in relapse-free survival and a 35% improvement in overall survival for patients in the intervention arm, after adjustment for extent of invasion through the stomach wall and nodal status.
• INT 0116 Limitations -Despite these encouraging findings, the study has generated several criticisms.[41] First, many patients underwent a suboptimal examination for lymph node metastases; more than one-half (54%) had a D0 dissection (fewer than seven lymph nodes identified in the surgical specimen), whereas only 10% had a D2 dissection (extended lymphadenectomy). However, there is no definitive evidence that a D2 dissection offers any survival advantage over a less radical resection.[42] Furthermore, rates of D0, D1, and D2 dissections in the intergroup trial are comparable to those in general practice, suggesting that the findings can be generalized to the population atlarge.[ 3] In addition, INT 0116 included patients with a wide variety of pathologic stages. This criticism questions the benefit of and need for adjuvant therapy in patients with the less-represented extreme stages of disease, particularly T1 and N0. Much of the benefit of chemoradiation in this trial appears to relate to improved local control rather than prevention of distant disease, and although patients treated with postoperative chemoradiotherapy experienced fewer locoregional recurrences, distant recurrences were equivalent between the two arms. Finally, postoperative adjuvant chemoradiotherapy after gastrectomy can be difficult to administer in full; in INT 0116, only 65% of patients in the treatment arm completed all prescribed therapy.
• Subsequent Research -Based on the results of INT 0116, curativeintent surgery followed by adjuvant chemoradiotherapy has become the standard of care for gastric cancer in North America. However, because the benefits associated with postoperative chemoradiotherapy were principally limited to improvements in locoregional control, investigators have examined the use of systemic regimens that are potentially more active than 5-FU/leucovorin. Fuchs and colleagues conducted a multicenter, pilot study of ECF plus continuous-infusion 5-FU administered during external-beam irradiation (Figure 1).[43] Among 21 patients receiving therapy, the toxicity profile was tolerable and comparable to that of INT 0116. At a median follow-up of 8.1 months (range: 1-27 months), only 14% of patients experienced a documented relapse. Because this pilot study demonstrated that postoperative ECF and continuous- infusion 5-FU with radiation was well tolerated, an NCI-sponsored trial was initiated that will randomize 825 patients to this regimen or the INT 0116 regimen after a curative resection of gastric or gastroesophageal adenocarcinoma (Figure 2). In a similar effort, the Radiation Therapy Oncology Group (RTOG) is conducting a randomized, phase II trial of two cisplatin- and paclitaxelcontaining regimens for adjuvant treatment of gastric or gastroesophageal cancer. Patients with stage IB- IIIB cancer will be randomly assigned to one of two groups. Patients in group 1 will receive a 5-day continuous infusion of 5-FU and a 1-hour infusion of cisplatin once a day for 5 days in weeks 1 and 5. They will also receive a 24-hour continuous infusion of paclitaxel during weeks 1 and 5. At least 3 weeks later, patients will undergo radiation therapy 5 days per week and will receive a 5-day continuous infusion of 5-FU and a 3-hour infusion of paclitaxel once per week for 5 weeks. Patients in group 2 will receive a 3-hour infusion of paclitaxel and a 1-hour infusion of cisplatin in weeks 1 and 5. Between 3 and 6 weeks later, they will undergo radiation therapy plus a 96-hour continuous infusion of paclitaxel and a 1-hour infusion of cisplatin once per week for 5 weeks. The planned accrual of this trial is 94 patients. The results of this trial will help investigators determine which chemoradiation regimen is sufficiently efficacious and tolerable to warrant a phase III comparison with the INT 0116 regimen.