Despite an overall rise in the incidence of gastrointestinal malignancies in the United States, there has been a significant decrease in the incidence of adenocarcinoma of the stomach over the past few decades. Nevertheless, gastric carcinoma remains the eighth leading cause of cancer death in the United States . In 1995, an estimated 22,800 new cases of gastric cancer will occur in the United States, 14,000 of which will occur in men, and approximately 14,700 patients will die of this disease. Unfortunately, only a small fraction of patients with gastric carcinoma present with localized disease . The 5-year survival rate of less than 20%  has not changed significantly during the past 30 to 40 years.
The incidence of gastric carcinoma varies widely throughout the world. Countries such as Japan and Chile have incidence rates as high as 78/10,000 and 70/100,000 population, respectively . In contrast, in the United States, the rate is only 10/100,000 persons . In addition, studies among migrants have shown that emigrants from high-incidence countries to low-incidence locations often experience a decreased risk of developing gastric carcinoma. This reduction in the risk was seen in subsequent generations and to a lesser degree in the first generation . Such findings strongly suggest that environmental factors play an important role in the etiology of gastric cancer and that exposure to risk factors occurs early in life.
Over the past 30 years, the incidence of gastric carcinoma in the United States has decreased by approximately 20%, whereas the mortality rate has decreased by 30% . Although the reason for the decline in incidence is not entirely known, it probably is related to changes in dietary habits and food preservation.
Predisposing Factors and Premalignant Disease
Chemical carcinogens have been thought to represent a major environmental etiologic factor in the pathogenesis of gastric cancer. Sugimura and Fujimura  reported that N-nitroso compounds formed by the interaction of dietary nitrite and amide compounds could induce gastric carcinogenesis in experimental animals. However, other studies have shown that increased consumption of processed, smoked, or salted meat and fish, which are high sources of N-nitroso compounds, is not consistently associated with an increased risk of gastric carcinoma . Diets low in vegetables, fruits, milk, and vitamin A and high in fried food, processed meat, and fish and alcohol(Drug information on alcohol) have been associated with an increased risk of gastric carcinoma in several cohort studies . Diets low in citrus fruit show the strongest association with gastric carcinoma. The protection afforded by vegetables and fruits is most likely related to their vitamin C content, which is thought to reduce the formation of carcinogenic N-nitroso compounds inside the stomach. Cooked vegetables, however, do not show the same protective effect as uncooked vegetables .
There is also evidence that fiber-rich foods  can decrease the formation of N-nitroso compounds. Calcium and vitamin A are postulated to protect the gastric mucosa against carcinogenesis by chemicals. Although no consistent association has been found between alcohol and tobacco consumption and an increased risk of gastric carcinoma , a number of prospective studies have linked cigarette smoking with an increased risk of this cancer .
Gastric resection also has been implicated as a predisposing factor for gastric carcinoma. Giarelli et al  reviewed autopsy results of 480 patients who had undergone gastric resections for benign disease and found that 31 (6.5%) of these patients had gastric-stump carcinomas. The mechanism is believed to be related to duodeno gastric bile reflux . Other epidemiologic studies of patients who underwent gastric operations for benign disease [15,16] suggest that the achlorhydria and atrophic gastritis that often occur after such procedures induce premalignant changes. However, the drawback to these studies is the absence of complete matching between the study and control groups for other environmental and life-style factors besides gastric surgery.
At least two premalignant conditions—intestinal metaplasia and pernicious anemia—may lead to gastric carcinoma. Intestinal metaplasia has a higher incidence in countries with a higher incidence of gastric carcinoma , and the former has been shown to precede gastric carcinoma . The estimated risk of developing gastric carcinoma in a patient with pernicious anemia is 20 times higher than the risk in age-matched controls .
The role of chronic Helicobacter pylori infection in gastric carcinogenesis remains controversial. Current evidence shows an association between serologic positivity for Helicobacter infection and the subsequent risk of developing gastric cancer . It is hypothesized that early life acquisition of H pylori increases the risk of developing both gastric cancer and gastric ulcer .
More than 95% of malignant gastric neoplasms are adenocarcinomas; the remaining 5% consist of lymphomas, leiomyosarcomas, and, infrequently, carcinoid tumors, carcinosarcomas, and squamous-cell carcinomas . Gastric cancer in humans is characterized by two histopathologic patterns that have demonstrated value in terms of epidemiologic parameters of demographic distribution and survival. The most common variant in populations at high risk for such cancer is the so-called intestinal type, in which malignant cells are united with each other to form glandular structures that somewhat resemble the glands of the gastrointestinal tract. The overriding etiologic factors in this type are of an environmental nature and are related to diet and infection. Diffuse carcinomas are relatively more common in populations at low risk of developing gastric cancer. Environmental factors appear to be of less etiologic significance than genetic influences are in diffuse carcinomas [22a].
Gastric carcinoma tends to invade through the gastric wall early and can involve adjacent structures, such as the transverse colon, pancreas, greater and lesser omentum, biliary tract, liver, and peritoneal ligaments. Even if gastric tumors do not involve adjacent structures, it is not uncommon for a T3 lesion to shed in the peritoneal cavity .
A common site of involvement in peritoneal seeding is Blumer's rectal shelf. In addition to regional spread, metastases can spread through the submucosal and subserosal lymphatic channels into the regional lymph nodes. The more commonly involved lymph nodes include those in the gastrohepatic ligament, celiac, and gastroduodenal region. The liver is the most common site of hematogenous metastasis , followed by the lungs and bones.
Tumor penetration, nodal metastases, location in the stomach, multicentricity, and distant metastases (TNM stage) have been the most important guides to prognosis in patients with gastric cancer . Certain pathologic features of gastric carcinoma, such as the gross appearance of the tumor have prognostic significance. More than one third of stomach carcinomas present as ulcers and show extensive submucosal infiltration that often involves the serosa . One fourth of tumors are scirrhous, with diffuse infiltration of the stomach wall leading to a marked fibrotic reaction. The 5-year survival rate for patients with scirrhous gastric carcinoma after gastric resection is only 2% , whereas the mucosal or polypoidal type is associated with a better prognosis.
The location of the tumor in the stomach also affects prognosis. A Gastrointestinal Tract Study Group (GITSG) study showed that lesions that occur in the cardia or esophagogastric junction have a poorer prognosis than do more distal lesions . Of note, cancer incidence data show a rise in the incidence of cardia carcinomas , which currently account for one-half of all gastric carcinomas. The reasons for this rise in incidence, however, is unknown. Speculation is that the decrease in distal gastric carcinomas might be linked to a decrease in the rate of H pylori gastritis.
The histologic grade of gastric cancer provides no additional prognostic information to the TNM stage. Recently, however, the Goseki histologic grading system was evaluated and showed good correlation with the pattern of tumor spread at necropsy . This grading system, which relies on tubular differentiation and mucous production, also identifies subgroups of patients who have a poorer prognosis than is predicted by TNM staging alone.
The majority of gastric cancers are thought to be caused by environmental factors that result in damage to the mucosa and that inhibits its ability to repair itself [29a]. This response is regulated, in part, by inhibitory and stimulatory factors that are products of proto-oncogenes and tumor-suppressor genes . The molecular basis of the progression of normal gastric epithelial cells through invasive cancer has been slowly emerging in recent years . The development of proper biomarkers associated with specific stages of multistep carcinogenesis as an intermediate study endpoint is strongly needed.
Up to now, the chromosomal changes found in gastrointestinal tumors remain poorly defined, and traditional cytogenetic techniques have been limited by the low number of mitotic cells obtained directly from the tumor and by the degree of karyotypic complexity observed in such mitoses.
Similar to colonic carcinogenesis, multiple genetic aberrations have been shown to involve both oncogenes and tumor-suppressor genes. The latter involve the loss of heterozygosities of several chromosomal loci and mutations in p53  and DCC genes . These mutations, the most common in human malignancies, appear at a variable frequency in patients with gastric cancer . Histologic type, stage, and study size and methodology account for the majority of these variations among the different studies. Some investigators have demonstrated a positive correlation between p53 mutations and the depth of tumor invasion, lymph-node metastases, and poor clinical outcome . APC gene mutations were identified in adenomatous precursors of gastric lesions, which suggests that such mutations play a role in early carcinogenesis .
Epidermal growth factor (EGF), its related peptide transforming growth factor-alpha, and their common receptor epidermal growth factor receptor (EGFR) have been implicated in the control of cell proliferation and differentiation in the gastrointestinal epithelium and may play an important role in gastric carcinogenesis . The coexpression of this receptor together with its ligand EGF in the same tumor correlates with a poor prognosis . Abnormalities of EGFR, EGF, and HER-2/neu also have been demonstrated in precursor lesions, which suggests that all have a role in early carcinogenesis . Overexpression of the HER-2/neu oncogene has been demonstrated in a significant proportion of gastric cancers with preferential amplification in well-differentiated cancers  and independent of EGFR expression . In general, HER-2/neu expression is associated with large tumors, lymphatic invasion, metastases, and shorter survival after curative resection [40,41]. In contrast to other gastrointestinal tumors, most studies have found that mutations in K-ras and c-myc are rare in gastric cancers .
Microsatellite instability and the development of the mutator phenotype are represented by alterations in DNA repeats and are detected in up to one third of gastric cancers, especially in poorly differentiated tumors. They were much more common in advanced-stage disease and were associated with chromosomal losses at 5q and 17p loci but not with tumor-suppressor gene mutations .
The diffuse growth patterns of gastric cancer have been associated with mutations in the E-cadherin gene, which encodes a cell-surface adhesion molecule . Preliminary evidence shows that gastric cancer cells express basic fibroblast growth factor and angiogenin, which function as angiogenic factors during the process of neovascularization .
Previously, the keystone of diagnosis of gastric carcinoma was an upper gastrointestinal barium study or x-ray. However, a study comparing radiographic findings with endoscopic biopsy results suggested that 9% to 40% of endoscopically positive lesions may be missed by barium studies . Endoscopy and biopsy of all lesions should be mandatory, even for lesions that appear to be benign on radiographic examination. The success rate of a single endoscopic biopsy in correctly identifying malignant gastric carcinoma has been reported to vary widely . Factors affecting the outcome include the tumor's gross appearance, size, and location, as well as the number of biopsies. The more biopsies performed, the higher the yield. In difficult situations, brush cytology may improve the diagnostic yield , and when combined with biopsy, it increases the sensitivity to 96.2% . In ulcerative, infiltrative, and submucosal lesions, fine-needle aspiration enables sampling of submucosal tissues. The biopsy may not be positive in patients with linitis plastica.
In addition to the aforementioned diagnostic tools, endoscopic sonography also may be used in the diagnostic workup and staging of patients. Lightdale et al  demonstrated that endoscopic sonography provided a more accurate assessment of the depth of tumor invasion and the spread to regional nodes than did chemotherapy. Serum tumor markers, such as carcinoembryonic antigen and CA19-9, have not proved useful for diagnosing gastric carcinoma because of their low specificity.
The most commonly used system for staging gastric cancer is the TNM system, which is based on postgastrectomy pathologic staging. A recent modification of this system defined a T4 lesion as any tumor invading adjacent structures and eliminated the N3 category. Table 1 lists the American Joint Committee on Cancer's TNM staging of gastric cancer.
|Primary tumor (T)|
|TX||Primary tumor cannot be assessed|
|T0||No evidence of primary tumor|
|Tis||Carcinoma in situ|
|T1||Tumor invades lamina propria or submucosa|
|T2||Tumor invades the muscularis or the subserosa|
|T3||Tumor invades muscularis propria|
|T4||Tumor invades adjacent structures|
|Regional lymph nodes (N)|
|NX||Regional lymph node(s) cannot be assessed|
|N0||No regional lymph-node metastasis|
|N1||Metastasis in perigastric lymph node(s) within 3 cm of the edge of the primary tumor.|
|N2||Metastasis in perigastric lymph node(s) more than 3 cm from the edge of the primary tumor or in lymph nodes along the left gastric, common hepatic, splenic, or celiac arteries|
|Distant metastasis (M)|
|MX||Presence of distant metastasis cannot be assessed|
|M0||No distant metastasis|
|Reprinted, with permission, from Soh LT, Ajani JA: Gastric carcinoma, in Pazdur R (ed): Medical Oncology: A Comprehensive Review, p 167. Huntington, NY, PRR Inc, 1993.|
The management of gastric carcinoma is determined primarily by the extent of the disease and ranges from surgical resection to surgery plus adjuvant radiotherapy and chemotherapy for patients with resectable disease to palliative therapy for patients with advanced carcinomas.
Resectable Gastric Carcinoma
Surgery: Surgical resection of gastric cancer is only feasible for tumors below stage T4. Curative resection, which involves removal of the primary tumor and regional lymph nodes with free margins, is useful only in patients with stage T1-2N0M0 tumors. Only 40% of patients who undergo exploratory laparotomy have a curative resection [50,51], and the majority of these patients eventually develop distant metastases. Even when curative resection is technically feasible, local and regional treatment failures are common. The median survival of patients who undergo curative resection for gastric cancer is 24 months, and the 5-year survival rate varies from 20% to 30% [52,53]. A study of recurrence patterns in patients with resected gastric carcinoma emphasized the high local-regional failure rate for this disease .
Planning the extent of surgical resection in gastric cancer remains an area of controversy because improved outcome has not been linked conclusively with more radical surgery . A new system designates gastric resections as D-0, D-1, or D-2, depending on the extent of nodal resection. D-0 refers to gastrectomy with incomplete resection of N1 nodes; D-1 and D-2 indicate complete resection of the regional lymph nodes in and outside the perigastric region, respectively. In addition, D-2 resection may also involve resections of other organs, which increases the operative risk.
In Western countries, D-1 resection is the most common operation performed in patients with gastric cancer. In Japan, a systematic approach has been developed to guide the extent of lymph-node dissection. Lymph nodes are classified as N1 through N4, depending on their relation to the primary tumor. N1 and N2 nodes represent regional disease, and N3 and N4 nodes are distant nodal metastases. An improvement in the survival rate (from 33% to 58%) was noted in a large series , as the extent of resection increased from D-1 to D-3. However, the significance of this study was limited by its retrospective nature. In contrast, a review  focusing on patients with gastric carcinoma treated between 1936 and 1963 found that survival rates decreased and operative mortality rates increased during the period when extended lymph-node dissection was being performed.
The prognostic relevance of systematic lymph-node dissection was evaluated in a prospective multicenter study of 2,394 patients in Germany . Radical dissection, defined as dissection of 26 or more lymph glands, was compared with standard dissection of fewer than 26 lymph nodes. Multivariate analysis identified radical dissection as an independent prognostic factor in the subgroups of patients with tumor stages II and IIIA as designated by the International Union Against Cancer (UICC). There was no survival advantage in patients with pN2 tumors. There was a significant difference in morbidity and mortality rates between radical and standard lymph-node dissection .
Recently, Bunt et al  investigated the effect of the extent of lymph-node resection on the pathologic TNM stage in 473 patients who underwent curative resection for gastric cancer. Tumor upstaging with D-2 resections occurred in 30% of patients. The extent of resection (D-2) together with the diligence of the surgeon in the number of nodes examined accounted, at least partially, for the superior stage-specific survival rates after D-2 resections compared with D-1 resections, without a real survival benefit in individual patients. Bonenkamp et al  reported a prospective trial comparing D-1 with D-2 dissections in 996 Dutch patients with gastric cancer. D-2 patients had a higher operative mortality rate than D-1 patients (10% vs 4%, P = .004) and experienced increased complications (43% vs 25%; P < .001). D-2 patients also had longer postoperative hospital stays (median, 25 vs 18 days). Two prospective randomized studies comparing D-1 with D-2 resections have been completed. Accrual and results currently are pending.
An important issue regarding resection of gastric carcinomas and potential for cure is that of resection-line involvement with the cancer, which also may determine whether additional postoperative therapy is indicated. Data from the British Stomach Cancer Group  indicate that of the operations considered potentially curable, 13% involved one or both resection lines, rendering the surgery palliative. Only 9% of patients with stages I to III disease who had resection-line involvement survived beyond 5 years compared with 27% of patients with clear lines.
Adjuvant Therapy: The high incidence of local and distant tumor recurrence after curative surgery for gastric cancer  stimulated interest in adjuvant (postoperative) therapy in the hope of improving the long-term outcome for these patients.
Adjuvant Radiotherapy: Radiation treatment alone has been shown to have curative potential in only a small percentage of patients who have residual disease following surgery or in patients with localized unresectable disease . However, available data suggest that radiotherapy might be effective in reducing local-regional recurrence rates and increasing the recurrence-free survival rates . The drawback of external-beam radiotherapy is the sensitivity of the gastric bed to radiation treatment, which limits the radiation dose to between 45 and 50 Gy. To achieve a greater effect from irradiation, fluorouracil(Drug information on fluorouracil) has been given concurrently as a radiosensitizer. Sixty-two patients with resectable gastric carcinoma but a poor prognosis were treated with adjuvant fluorouracil (given as three 15-mg/kg intravenous boluses) plus radiation (3,750 cGy over 24 fractions) initiated 3 to 6 weeks after surgery. The 5-year survival rate was superior in the treated group (23%) to that in patients treated with surgery alone (4%).
Another nonrandomized trial conducted by the Eastern Cooperative Oncology Group (ECOG) also showed improvement in survival with postoperative adjuvant chemotherapy and radiotherapy . The current Intergroup protocol addresses the issue of postoperative radiation treatment and chemotherapy in a randomized design. Patients undergoing curative resection of a stomach tumor are randomized to receive either four cycles of fluorouracil and leucovorin (folinic acid) or follow-up without any therapy. Radiotherapy is administered concurrently with the second cycle of chemotherapy. The projected accrual is 550 patients, of whom 350 patients have been enrolled so far.
Adjuvant Chemotherapy: Several controlled adjuvant chemotherapy studies have been conducted in patients with gastric carcinoma [27,66-73]. Of all randomized adjuvant chemotherapy trials, only two have shown a survival benefit for the treated patients. The findings of the two positive trials [27,73] have not been confirmed by another European or North American study. In the GITSG study , 142 patients were randomized to receive postoperative fluorouracil plus semustine (MeCCNU) or surgery alone. The chemotherapy group showed an improved 5-year survival rate (47% vs 33%). Grau et al  showed a survival advantage for patients treated with adjuvant mitomycin(Drug information on mitomycin) (Mutamycin), 20 mg/m², compared with patients treated with surgery alone.
The use of fluorouracil, doxorubicin(Drug information on doxorubicin) (Adriamycin, Rubex), and mitomycin, or FAM, was evaluated by the International Collaborative Cancer Group in 281 patients who were randomized to either adjuvant FAM or no treatment . Although the treated patients showed no improvement in survival, there was a statistically significant difference in the survival rate (41.4% vs 22.8%; P < .04) favoring treated patients with T3-4 disease. In another study of gastric carcinoma patients with stages IB, IC, II, and III disease conducted by the Southwest Oncology Group, adjuvant treatment with FAM did not improve survival over the rate achieved without adjuvant treatment . FAM2, which is FAM modified by increasing the drug doses and reducing the interval between treatment cycles, was studied by the European Organization for Research and Treatment of Cancer (EORTC) as an adjuvant therapy . Although the control group had a higher gastric cancer recurrence rate, there was no statistically significant difference in recurrence rates between the treated and untreated groups. The British Stomach Cancer Group trial of surgery alone vs either adjuvant chemotherapy (FAM) or radiotherapy in 436 patients showed no survival advantage at 5 years for patients receiving either form of adjuvant therapy compared with those undergoing surgery alone .
Since the late 1950s, adjuvant chemotherapy has been incorporated routinely into postoperative therapy for patients with gastric carcinoma in Japan. The drug most commonly used in the Japanese regimens is mitomycin. In a randomized study of 2,000 patients followed up for more than 10 years, only the group treated with a medium dose of mitomycin showed a significant survival advantage at 8 years compared with controls (73.6% vs 53.9%). When only patients with stage II disease were considered, the difference in survival rates between the treated and control groups was even greater (75% vs 42%). In contrast, four other randomized studies [77-80] using mitomycin combined with fluoropyrimidine or cytarabine(Drug information on cytarabine) showed no survival benefits except in subgroups with positive lymph nodes and serosal involvement.
A meta-analysis of randomized trials of adjuvant chemotherapy for gastric cancer confirms the finding that adjuvant chemotherapy regimens, although effective in phase II studies, do not significantly improve survival . At present, postoperative chemotherapy cannot be considered standard therapy in patients with curatively resected gastric cancer. New trials of adjuvant treatment of gastric cancer must include a nonchemotherapy control arm unless testing a potentially very active chemotherapy regimen.
Adjuvant Chemoimmunotherapy: Randomized trials of adjuvant chemoimmunotherapy in the management of gastric carcinoma have been reported. Several immunostimulators have been used in patients with gastric cancer. They include bacterial extracts from Schizophyllum commune,  Nocardia rubra, [83,84] Streptococcus species [85,86] fungal extracts such as Streptomyces olivoreticuli  and Coriolus versicolor [86,88], chemicals such as levamisole(Drug information on levamisole) [89,90], and protein-bound polysaccharide . Only five of these trials, however, were randomized, and all treated groups showed a survival benefit [84,86,91-93]. Immunotherapy not only improved survival but also decreased the infection rate associated with chemotherapy. Patients with advanced disease who had undergone curative and even noncurative resections benefited most from chemoimmunotherapy, whereas patients who underwent early curative resection or late palliative resection benefited least. Thus, the results obtained to date with chemoimmunotherapy appear more encouraging than those achieved with other adjuvant systemic therapies.
Newer Approaches in Adjuvant Therapy: A newer approach in adjuvant therapy includes the use of intraperitoneal treatment in an attempt to improve local relapse rates. Atiq et al  reported results of adjuvant intraperitoneal and systemic therapy after curative resection for T2N1-2M0 or T3-4 any N M0 in 35 patients. Adjuvant intraperitoneal cisplatin(Drug information on cisplatin) (Platinol), 25 mg/m², and fluorouracil, 750 mg, were administered daily for 4 days with simultaneous intravenous fluorouracil, 750 mg/m², as a continuous intravenous infusion for 24 hours. Five cycles were repeated at monthly intervals. Seven (25%) of 28 patients had disease recurrence; after a median follow-up of 24 months, 51% of patients remained alive and free of disease. Of 16 patients who had recurrence, 13 had an intra-abdominal component. The major two toxic effects were neutropenia and peritoneal fibrosis. The latter was treated with surgical lysis of adhesions.
A more recent study using postoperative intraperitoneal cisplatin showed a pattern of relapse and survival similar to that expected from a comparable population with gastric cancer . Sautner et al  reported a randomized study on patients with advanced gastric cancer (stages III and IV) who received intraperitoneal cisplatin postoperatively. No survival advantage was noted; median disease-free survival was 12.7 months compared with 9.7 months in patients treated with surgery alone. Intraperitoneal therapy did not influence the pattern of disease recurrence.
Localized Unresectable Gastric Carcinoma
Approximately two thirds of locally confined gastric tumors are considered to be locally advanced. Such tumors have a poor prognosis, particularly when they are bulky (> T3), located in the cardia, or involve local-regional lymph nodes. The aim of initial therapy is either to downstage the tumor to facilitate resection or to reduce the tumor size for palliative purposes.
External-Beam Radiotherapy: Moderate-dose (35 to 40 Gy) external-beam irradiation as a single modality has value in pain palliation but does not improve survival . However, when used concurrently with chemotherapy, it may prolong survival. Moertel et al  compared fluorouracil plus radiotherapy at 3,500 to 4,000 cGy with radiotherapy alone in the treatment of patients with locally unresectable gastric carcinoma. There was a 6-month survival advantage favoring patients who received both chemotherapy and radiotherapy. In another study, the GITSG randomized 90 patients with locally advanced gastric carcinoma either to combination chemotherapy, consisting of fluorouracil plus semustine, or to split-course radiotherapy, with an intravenous infusion of fluorouracil given during the first 3 days of two radiation courses of 25,000 cGy, separated by a 2-week break, followed by maintenance therapy with fluorouracil plus semustine. In the first 26 weeks, mortality was higher in the combined-modality group. At 3 years, the probability of survival plateaued in the combined-modality arm but continued to fall in the chemotherapy-alone arm .
Intraoperative Radiotherapy: Intraoperative irradiation may be used to deliver effective doses safely to unresectable lesions by moving the sensitive viscera from the radiation field. However, intraoperative irradiation has some disadvantages, which include uncertainty about the dose that can be given as a single fraction and the inability to sterilize bulky tumors with a single treatment. Abe et al  and Abe and Takahashi  reported a nonrandomized study in which survival was improved in 194 patients with locally advanced gastric carcinoma treated with single-fraction (2,800 to 4,000 cGy) intraoperative radiation in addition to gastric resection . A prospective, randomized, three-arm study compared surgical resection plus intraoperative radiotherapy with gastrectomy alone in patients with stage I or II disease. The third arm included patients with disease that extended beyond the gastric wall (stages III and IV) who received postoperative external-beam radiotherapy to the upper abdomen. In this small study of 100 patients, there was a significant reduction in the local failure rate in patients who received intraoperative radiotherapy . Further studies are needed to determine the role of intraoperative radiotherapy in the management of unresectable gastric cancer.
Preoperative Chemotherapy: Trials utilizing preoperative chemotherapy aim not only to increase resectability rates but also to improve survival rates and to reduce relapse rates in patients with gastric cancer that can be resected surgically. However, there have been few mature reports of neoadjuvant therapy in patients with gastric cancer, and most of the early reports included patients whose tumors were initially unresectable. Stephens  reported that 11 of 27 patients were disease free 1 to 5 years after preoperative FAM plus carmustine(Drug information on carmustine) (BiCNU) given intra-arterially. Wilke et al  reported on 34 patients with locally advanced, unresectable disease established by initial laparotomy who were treated with etoposide(Drug information on etoposide) (VePesid), doxorubicin, and cisplatin (EAP) . Twenty-three (70%) of 33 patients demonstrated a major response after EAP, with 21% having a clinically complete response. Nineteen of the 23 responders subsequently underwent gastric resection. Five clinically complete responses were pathologically confirmed; 10 patients with clinically partial responses were rendered free of disease after resection. After a median follow-up of 20 months, the relapse rate was 60% in patients who were pathologically free of disease. The median survival time for the entire group of patients was 18 months and for disease-free patients was 24 months. Another two-institution trial treated 48 patients with three cycles of EAP chemotherapy followed by surgery and two additional postoperative cycles . Of 48 patients, 6 achieved a clinically complete response preoperatively, and 9 had partial responses. No pathologically complete responses were noted at the time of resection. Seventy-seven percent of the group achieved a curative resection. The median survival of this group was 15.5 months.
Two recent studies combined preoperative chemotherapy and postoperative intraperitoneal adjuvant therapy. In one study, 38 patients with resectable gastric cancer received neoadjuvant fluorouracil by protracted continuous infusion with cisplatin and weekly leucovorin over 4 weeks . Thirty-five (92%) patients underwent laparotomy, 33 (87%) of whom had gastric resection, with 76% (29) having a total resection with disease-negative margins. Twenty-six (68%) of 38 patients received postoperative intraperitoneal therapy. Four (14%) of 29 patients have had recurrence at a median follow-up of 19 months. The median survival has not been reached at more than 17 months.
Investigators at Memorial Sloan-Kettering Cancer Center reported on 29 patients with high-risk gastric cancer (T3-4 any N M0) . The patients were treated with three preoperative cycles of fluorouracil, doxorubicin, and methotrexate(Drug information on methotrexate) (FAMTX) followed by surgical resection and postoperative intraperitoneal and intravenous fluorouracil-based therapy. Of 23 patients who completed therapy, tumors in 18 (78%) patients were operable and tumors in 16 (70%) patients were resectable. Thirteen of 23 patients had curative resections with disease-negative margins. At 6-month median follow-up, 9 (39%) patients remained disease free.
A phase II trial of preoperative chemotherapy recently was reported combining continuous intravenous infusion of fluorouracil (1,000 mg/m² for 5 days) and cisplatin (100 mg/m² on day 2) repeated every 4 weeks . Thirty patients with locally advanced gastric cancer were entered into this study. One patient achieved a complete response and 14 a partial response. D-0 resections were feasible in 60% of patients, mainly after objective response. No pathologically complete responses were seen in this study.
It should be noted that the rate of complete pathologic responses remains very low in the studies previously described. As such, the impact on survival is limited. No standard preoperative therapy exists for patients with locally advanced or unresectable gastric cancer, and the search for new therapies, possibly with newer agents, should continue. Preoperative chemotherapy also appears to be an attractive tool for clinical investigations in patients with earlier stages of gastric cancer.
Newer Approaches: Preoperative radiotherapy with local microwave hyperthermia has been evaluated in patients with locally advanced gastric cancer . A randomized trial of three groups has been reported of 293 patients who received surgery alone, surgery preceded by preoperative irradiation, or surgery followed by preoperative irradiation and hyperthermia treatment. Preoperative radiotherapy (20 Gy) did not improve the 3- or 5-year survival rate in gastric cancer patients compared with patients who received surgery alone. Local hyperthermia in combination with radiotherapy followed by surgery produced a statistically significant improvement in both the 3- and 5-year survival rates (22.1% and 21.3%, respectively).
Metastatic Gastric Cancer
Advanced gastric cancer encompasses patients with metastatic disease who are not curable with any of the current treatment modalities. Included in this category are patients with localized disease associated with extensive local-regional metastases.
Chemotherapy: As with other gastrointestinal tumors, gastric carcinoma responds to fluorouracil, but with an objective single-agent response of less than 20% . Other agents that achieve approximately a 20% response rate when used as single agents include the anthracyclines , mitomycin , etoposide , and cisplatin . However, most objective responses to these agents are partial and short-lived and have no survival benefit. In general, better responses are seen in patients who have a small tumor volume and a good performance status and in patients who have not been pretreated with other forms of chemotherapy.
Active drugs with nonoverlapping toxic effects can be combined to improve efficacy at tolerable toxicity levels. To date, however, few combination regimens have been shown to be effective against gastric carcinoma. The FAM regimen was reported first for inpatients with advanced gastric carcinoma . The response rate in that study was 42% (all partial responses), and the median duration of survival was 9 months. Subsequent studies using FAM have yielded varied results, with response rates ranging from 17% to 55%, but no complete responses were achieved . A phase III study by the North Central Cancer Therapy Group compared FAM with fluorouracil alone and with fluorouracil plus doxorubicin and found no significant survival difference among the three regimens . These findings suggest that combination chemotherapy may be superior to single-agent fluorouracil.
The FAMTX regimen was based on the in vitro evidence for synergy between fluorouracil and methotrexate. Use of this regimen in the treatment of 187 patients with metastatic gastric carcinoma was reported first by Klein et al . The overall response rate was 43%, with complete responses in 11% of patients. In a study reported by the EORTC , FAMTX achieved an objective response rate of 33% and a complete response rate of 13.4% in 71 patients with advanced gastric cancer; median survival in these patients was 6 months. However, toxicity in the EORTC study was significant, and there were four deaths, three of which were due to protocol violation. In an EORTC phase III trial comparing FAM with FAMTX , the FAMTX-treated group showed a superior response rate (41% vs 9%) and a superior median survival (40 vs 29 weeks). FAMTX, however, is a complex and expensive regimen with a 50% incidence of severe neutropenia, and therefore it has been abandoned.
The fluorouracil, doxorubicin, and cisplatin regimen has been evaluated by several investigators [119-121]. In these studies, response rates varied between 29% and 55%, but there were no complete responses. The median survival ranged from 4 to 12 months.
Interest in the EAP combination chemotherapy regimen was triggered by reports of possible synergism between cisplatin and etoposide [122,123] and by the reported synergism between cisplatin and doxorubicin . EAP originally was reported to yield an objective response rate of 64%, including a 21% clinically complete response rate when used in the treatment of 67 patients with locally advanced or metastatic gastric carcinoma . In the 12 patients who had advanced local-regional disease, 5 had pathologically complete responses and 7 had partial responses. In the patients with metastatic disease, the objective response rate was 56%, including a 15% complete response rate. The median survival of the entire group was 9 months, but among the complete responders, the median survival was 17 months.
In another series by Wilke et al , 33 patients with locally advanced gastric carcinoma treated with EAP had a response rate of 70%, with a 21% complete response rate (15% pathologically complete responses). In yet another series of patients treated with EAP , the response rate was only 33% (including an 8% clinically complete response rate) but with significant myelotoxicity and an unacceptably high treatment-related mortality rate of 11%. In a prospective, randomized comparison of FAMTX and EAP performed by investigators at Memorial Sloan-Kettering Cancer Center , response rates and median survival durations were similar in the two arms. However, toxicity was more severe in the EAP arm, with four treatment-related deaths. Use of EAP has been abandoned largely because of lower-than-expected response rates and unacceptably high levels of toxicity.
The combination of etoposide, leucovorin, and fluorouracil also has been tested in patients with advanced gastric cancer [128,129]. Reported objective response rates were 48% to 49%, including a 6% to 12% complete response rate. Median survival was 11 to 12.4 months and toxicity was acceptable. This regimen is easier to administer, and results of an EORTC-sponsored study comparing it with fluorouracil and FAMTX are awaited.
Modification of the schedule of fluorouracil administration has been used to improve the efficacy of chemotherapy. A continuous infusion of fluorouracil in combination with other agents has been used in a number of small trials. Response rates have varied from 8% to 66% [130-132]. Furthermore, the predominant toxic effect associated with these continuous infusion regimens was more severe mucositis but less myelosuppression and toxicity to other organs.
Despite all the foregoing attempts to devise an effective chemotherapeutic approach to gastric cancer, there is at present no standard chemotherapy regimen.
Palliative Procedures: Surgery has been used in the palliation of advanced unresectable gastric carcinoma. However, a surgical procedure would be justified only in selected cases of advanced disease. Fujimoto et al  treated 30 patients with advanced gastric carcinoma with debulking surgery followed by intraperitoneal hyperthermic perfusion. The catheter for infusion of the perfusate was inserted into the Douglas pouch and upper abdominal cavity during surgical treatment. The perfusate contained mitomycin, and special attention was paid to the effect on cardiorespiratory function. Temperatures at the inflow point and at the Douglas pouch were maintained at 45.0 to 46.3ºC and 43.5 to 45.1ºC, respectively. The temperature at the pulmonary artery was measured using a Swan-Ganz catheter and was kept below 41ºC. Following such treatment, Fujimoto et al reported 2-year survival rates of 45% in the group with intraperitoneal seeding and 56.5% in the group without peritoneal seeding.
The failure pattern after curative resection suggests that there is a high rate of micrometastases before surgery and, thus, a need for adjuvant therapy. However, at present there is no effective chemotherapeutic regimen with tolerable toxicities available for the treatment of metastatic disease. There is a need, in particular, to develop treatment strategies that consistently will result in high complete response rates to alter significantly the natural history of the disease. The type of adjuvant chemotherapy, its timing and scheduling, and route of administration should be considered systematically in future trials if an assumed therapeutic gain is to be demonstrated by adjuvant treatment of gastric cancer.
The plateau reached in treating patients with advanced gastric cancer underscores the need to develop new drugs in the treatment of this disease. Improvements in our understanding of the molecular events that initiate early gastric carcinogenesis and the cellular changes with disease progression will provide potential targets for the development of new treatment strategies. Examples of such approaches include gene therapy to introduce tumor-suppressor genes or to block the expression of oncogenes. Other strategies include targeting critical biochemical steps in tumor cells with antibody- or ligand-guided therapies. Circumventing cytotoxic-drug resistance in tumor cells may be one approach to improve the response to currently available cytotoxic drugs. For example, recent evidence suggests that overexpression of glutathione transferase-pi may significantly alter sensitivity to cisplatin and may therefore provide a rational target for biochemical modulation.
Another important issue in the treatment of gastric cancer relates to primary and secondary prevention of the disease. The role of H pylori infection is currently under intense investigation. Screening for early premalignant and malignant lesions has been undertaken successfully in Japan and has resulted in significant reduction in the incidence of the disease and marked downstaging of the tumor at presentation . The economic value of such an approach should be determined in Western societies that have seen a significant drop in the incidence of gastric cancer over the past several decades.