The incidence of gastric cancer has decreased in the United States and worldwide. There were 22,710 new cases of gastric carcinoma in United States in year 2003.1 Because a screening program is not cost effective, most patients with gastric cancer present at an advanced stage.

Surgical resection remains the main curative treatment for gastric cancer. Although the role of palliative gastrectomy in patients with advance gastric cancer is unclear, several studies have suggested that resection may provide some survival benefit.2–5 However, the extended survival after palliative gastrectomy in other studies was associated with significant postoperative morbidity, prolonged hospital stay, and poor quality of life.6–8 The aims of palliative resection are to enable oral food intake by improving obstructive symptoms, bleeding, or pain. The aim of the present study was to evaluate patient outcome after palliative gastric resection for metastatic gastric cancer.

PATIENTS AND METHODS

Between 1990 and 2000, 210 cases of histologically proven gastric adenocarcinoma were identified. Of those 210 patients, 104 (49.5%) underwent gastric resection. Stage IV gastric cancer was defined based on American Joint Commission on Cancer (AJCC, 6th edition), as M1 or T4N1–3M0.9

This study was approved by the Institutional Review Board of each participating institution individually. All patients were treated at the University of Kentucky, University of Tennessee, or Providence Hospital, Southfield, Michigan. Because of the small number of patients, no comparison was made of institutional difference in outcome or survival.

This is retrospective analysis of 105 patients with stage IV disease identified during this period; 81 (77.1%) received either no operation or a bypass procedures with or without adjuvant treatment, and 24 (22.9%) underwent palliative gastric resection (Table 1). The patients were divided into four groups based on treatment (Table 1). The indications for palliative gastrectomy were judged individually by attending surgeons based on patients’ general health, performance status, symptoms, extent of disease, and feasibility of resection.

Table 1 Treatment groups.
Full size table

Patient clinicopathological features, extent of disease, survival, operative mortality and morbidity, length of hospitalization, and readmissions were recorded. Hospital mortality was defined as death within 30 days or during the hospital stay.

For descriptive statistics and survival analyses the SPSS PC software was applied. The independent Students’ t-test was used to test the differences between groups, and the chi-squared test of association was used for categorical data. The Mann-Whitney U-test was used for length of hospitalization comparison. Cumulative survival rates were calculated according to the Kaplan-Meier analysis. Survival differences were compared using log-rank test; P values less than 0.05 were considered statistically significant.

RESULTS

This series consists of 105 patients with stage IV gastric adenocarcinoma. Of these 105 patients, 24 (22.9%) underwent resection and 81 (77.1%) did not. The characteristics of patients are shown in Tables 2 and 3. In the non-resection group, 34 (41.9%) received adjuvant therapy and 47 (58.1%) received no treatment.

Table 2 Patients with stage IV gastric carcinoma.
Full size table
Table 3 Clinicopathological characteristics of the patients with palliative gastrectomy.
Full size table

The mortality after palliative gastrectomy and laparatomy/bypass was 8.3% (2/24) and 8.5% (3/35) respectively. Because of the retrospective nature of this study, complete information on performance status was not available. In addition, because some of the patients did not received any treatment or resection, the clinicopathological features of some of the tumors were based on CT or endoscopy findings.

The mean survival in patients who had palliative gastrectomy (13.2 months) was longer than in patients who underwent no resection (5.5 months), 2-year survival for the two groups was 25% and 0, respectively. The 5-year survival in the resection group was 5%. There was a significant difference in survival between patients treated with and without gastrectomy (Fig. 1). The mean survival of the patients with stage IV gastric cancer based on treatment modality is shown in Table 4 and Figure 2.

Figure 1
figure 1

Survival of patients with stage IV gastric cancer with and without palliative resection.

Full size image
Table 4 Patients overall survival based on treatments modality.
Full size table
Figure 2
figure 2

Survival of patients with stage IV gastric cancer based on treatment modality.

Full size image

A statistically significant survival advantage was present comparing those patients who had resection plus adjuvant therapy with other groups (log-rank test, P = 0.014). However, the survival rates in groups 2, 3, and 4 were not statistically significant. Most of the chemotherapy regimens that patients received were 5-FU (5-fluorouricil) based. No specific protocol was followed for chemotherapy regimen, and the dose and duration of the regimen was based on the opinion of the medical oncologist in each case.

The mean survival among patients who underwent palliative gastrectomy for T4, N1–3, M0 disease (n = 5) and M1 disease with (n = 17) distant metastasis was 14.4 months (3.6–25.1, 95% confidence interval) and 11.2 months (3.5–18.9, 95% confidence interval), respectively. As shown in Figure 3, this was not statistically significant (log-rank test. P = 0.37). The patients who did not have distant metastasis were classified as stage IV, based on AJCC, 6th edition, as having T4, N1–3, M0 disease. T4 indicates tumor invading adjacent structures such as pancreas, colon, or spleen. Although these patients have M0 disease, their survival rate is comparable with that of patients with M1 disease, as shown in Figure 3.

Figure 3
figure 3

Survival in patients with stage IV gastric cancer who underwent palliative gastrectomy in the presence or absence of distant metastasis.

Full size image

Table 5 and Figure 4 show survival of patients who had distant metastasis (M1) base on treatment modality.

Table 5 Survival in patients with distant metastasis (M1, excluding T4N1– 3, M0) based on treatment modality.
Full size table
Figure 4
figure 4

Survival, based on treatment modality, of patients with stage IV gastric cancer who had distant metastasis.

Full size image

As shown in Figure 4, group 1 patients (resection plus adjuvant therapy) had a significant survival advantage over the other three groups (log-rank test, P = 0.02). The survival rates in groups 2, 3, and 4 were not different.

Mortality, morbidity, length of hospitalization, and readmissions in patients who underwent palliative gastrectomy for stage IV gastric cancer compared to patient outcomes in patients who underwent curative gastrectomy during the same period are shown in Table 6.

Table 6 Early outcome is patients who underwent palliative versus curative gastrectomy.
Full size table

DISCUSSION

In the countries where screening for gastric cancer is not routine, this disease is often far advanced at the time of diagnosis. Except for early gastric cancer, patient survival is poor. Based on surveillance, epidemiology, and end results (SEER) data, the average survival for patients with stage IV gastric cancer is 16 weeks, with only 25% living 34 weeks or longer.10 Alternatively, gastric cancer response to conventional chemotherapy is poor, especially in patients reported in Western studies.11,12 Glimelius et al., randomized 61 patients with advanced gastric cancer to receive either chemotherapy plus supportive care or supportive care only. Overall survival was longer in the chemotherapy group (median 8 months versus 5 months), although the difference was not statistically significant (P = 0.12).13 In our series, the mean survival in the patients who only received adjuvant therapy was 5.9 months compare to 5.2 months in those with no treatment.

Radical resection is the primary treatment for gastric cancer, but for patients with advanced or metastatic disease the benefit of resection is unclear. The role of palliative gastrectomy in patients with metastatic gastric cancer is matter of debate and controversy. Although some studies have shown that palliative gastrectomy is associated with survival benefit,2–5,14–18 other studies show that there is no role for palliative gastrectomy in patients with stage IV gastric cancer.6–8

The rationale for offering palliative gastrectomy to patients with stage IV gastric carcinoma is that primary tumor will result in gastric obstruction, perforation, bleeding, or debilitating ascites. Resection should therefore be associated with removal of gross disease with the goal to improve function and quality of life by removing a bulky symptomatic tumor. However, if a significant proportion of the tumor load is removed (cytoreduction surgery), perhaps the disease may be more responsive to adjuvant treatment.19 In theory, if the majority of the tumor can be debulked safely in combination with chemotherapy and radiation, the results may improve. It is clear that volume and tumor burden reduction diminishes the metabolic demands made on the patient by the tumor. In addition, because the tumor itself can produce immunosuppressive cytokine, reducing the tumor burden may also have an immunologic benefit.20

In our study, the overall survival of patients who were treated with multimodality approach (surgery plus adjuvant treatment) was far better than for those who were not (Table 3 and Figure 2), which supports the above argument. Alternatively, the survival advantage of palliative gastrectomy was also seen in patients with distant metastasis (Figure 4). Fujisaki et al.,21 have shown that gastric resection and degree of hepatic involvement are the only prognostic factors that predict survival in gastric cancer with concomitant liver metastases. Our earlier studies support this concept.16,17 However, Chow et al.,18 found that palliative gastrectomy did not improve survival in patients with gastric cancer and liver metastases.

Despite the survival advantage of palliative gastrectomy, some studies show high mortality and morbidity associated with the operation.6–8,14 In a Dutch Gastric Cancer Trial, 285 (26%) of 1078 randomized patients were classified as noncurative at laparotomy, and 156 patients had a palliative resection.14 The overall median survival of patients who underwent palliative resection was 8.1 months compared to 5.4 months in patients who had no resection (P < 0.001). However, the morbidity (38% versus 12%) and length of hospitalization (15 versus 10 day) was significantly higher in the resection group. Recently, progress in anesthesia and surgical techniques has significantly reduced the risk associated with radical surgical procedures. In our study the mortality and morbidity of palliative gastrectomy was 8.7% and 33.3%, respectively, which was not statistically different from mortality and morbidity of curative gastrectomy. However, length of hospitalization was significantly greater (22 versus 16 days). Owing to the nature of the study, no information on quality of life was available.

We believe that selection of appropriate patients with stage IV gastric cancer for resection is the key. These patients might be individuals with acceptable performance status, in whom resection of the tumor is technically feasible despite the fact that they might have metastatic disease. Removing the primary tumor might reduce tumor burden and leave the patients more responsive to adjuvant modalities.23–25 Recently, there have been multiple trails which show an improvement in outcome when treating patients with metatstaic gastric cancer with chemotherapy.26–29

Our findings showed that even in stage IV gastric cancer, palliative gastrectomy might be associated with survival advantage in selected group of patients, especially if combined with adjuvant therapy as part of multimodality treatment. As long as the primary tumor can be removed with reasonable risk of mortality and morbidity, every attempt should be made to resect the tumor.30 Leaving the tumor not only involves high risk of poor oral intake, pain, obstruction, and bleeding, it also greatly decreases the likelihood of a response to adjuvant treatment.