Prognostic Factors for Stage IV Gastric Cancer

Gastric cancer is the fourth leading cause of cancer in the world with the incidence of nearly 1 million cases (7.8% of the total). Regionally, half of the world cases occur in Eastern Asia. It is a second leading cause of cancer death worldwide (738,000 deaths, 9.7% of the total).¹ Although early diagnosis and standardized treatment strategies have contributed to the improvement of survival, stage IV patients still suffer from poor prognosis. Previous reports from Japan and Korea have shown the 5-year survival rate to be 15.3% and 18%, respectively.^2,3

Guidelines published by the Japanese Gastric Cancer Association (JGCA)^4,5 recommend selection from among 5 treatments for stage IV patients in clinical practice: surgery (extended or palliative), chemotherapy, radiation therapy or palliative care. Principles are still controversial, and clinical study is advocated. The nature of the disease does not allow us to conduct prospective studies; however, retrospective study plays an important role by means of clinical study.

This study is a retrospective study based on clinical practice and aims to clarify the favorable factors contributing to better prognosis of stage IV gastric cancer patients. Furthermore, we aim to assess the optimal treatment strategies.

Patients and Methods

Between April 2005 and March 2011, a total of 123 patients with pathologically confirmed stage IV gastric cancer at Saitama Medical Center were investigated according to the data collected from the chart. Clinicopathologic factors evaluated in this study were age, sex, performance status (PS), tumor location, gross tumor appearance, histologic type, carcinoembryonic antigen (CEA), carbohydrate antigen 19-9 (CA19-9), tumor depth (T), regional lymph node metastasis (N), liver metastasis (H), peritoneal dissemination (P), and peritoneal lavage cytology (CY). Distant metastasis besides H, P, and CY are relatively rare, except for lymph node metastasis beyond regional nodes [M (LYM)] We categorized these rare M factors as “other M” and included them as one of the variables in this study. Removal of the primary tumor or not and introducing systemic chemotherapy or not was also included as one of the variables for statistical analysis.

The staging process was based on the tumor-node-metastasis (TNM) staging system adopted by Union for Cancer Control (UICC; 7th edition, 2010).⁶ To avoid unnecessary terminological confusion, we followed the terminology defined in the English version of guidelines for diagnosis and treatment of carcinoma of the stomach,⁷ published by JGCA. PS was evaluated on admission in every patient using the method proposed by the Eastern Cooperative Oncology Group (ECOG).⁸ Diagnosis of stage IV was made by preoperative radiologic findings, such as computed tomography (CT), magnetic resonance imaging (MRI), and ultrasonography (US), otherwise at laparotomy or staging laparoscopy. Peritoneal lavage cytology (CY) was performed at the time of the laparotomy, and diagnostic information was given from the pathologic department during the operation. Pathologic reports of the specimens were investigated by the board-certified pathologists at our institution.

JGCS defines stage IV patients as having no absolute treatment strategy in the treatment guideline.² Therefore, we do not have a particular treatment algorithm against stage IV patients at this time. However, we weigh patients' clinical symptoms for decision making. Gastric outlet syndrome (GOS) is a most concerning and irritating symptom for patients, and if recognized, surgery tends to be our first choice. If not, chemotherapy can be an option. If poor patient status is remarkable, best supportive care is considered. Patients may start with chemotherapy and go on to have surgery with the emergence of GOS or bleeding. Whatever the treatment order may be, all patients in this retrospective study provided written informed consent before starting anything.

Whether to introduce chemotherapy first or not and the most effective surgical procedure were briefly discussed with patients and their family members. Decision making was based on the patients' wishes and objective judgment by the physician in charge. Attempt to carry out extended surgery often ended up with removal of the primary tumor after all, achieving only R1 surgery. Especially, when palliative surgery was anticipated, sufficient informed consent to the patient was made, and the operation was carried out on the premises. Chemotherapy regimen was based on tegafur, gimeracil, and oteracil potassium combined antitumor drug (S-1), an oral pyrimidine derivative, which is the major anti-gastric cancer agent today.^9–11 Combination with cisplatin (CDDP) is considered to be the first choice regimen today.¹² However, other agents were used in practice on a daily basis, such as irinotecan (CPT-11) and taxans (paclitaxel and docetaxel), and these patients were also included in the study.^13–15

Statistical Analysis

Statistical analyses were performed between the groups using the χ² or Mann-Whitney U tests where appropriate. Survival curves were analyzed using the Kaplan-Meier method, and differences were determined by the log-rank test. The multivariate Cox proportional hazards model was used to identify independent prognostic factors. For determining the independent prognostic factors, a factor with a P value < 0.1 by univariate analysis was introduced into multivariate analysis with a forward stepwise selection. Overall survival was defined as a period starting from the initial treatment (chemotherapy, surgery, or best supportive care) until death of any reason. Tests were considered significant at a P value < 0.05.

Results

Demographic and clinicopathologic characteristics are shown in Table 1. There were 91 men and 32 women, with a median age of 69 years (range, 26–88). PS 0 and 1 combined counted 102 patients (82.9%). The number of patients whose tumor location involved the proximal area of the stomach was 55 (44.7%). Sixty-eight patients (55.3%) had their tumor localized at the distal part of the stomach. Gross appearance of the tumor varied from Borrman type 0 to type 5, counting 1, 6, 18, 57, 33, and 18, respectively. All tumors were pathologically diagnosed as adenocarcinoma. Thirty-five (28.5%) of them were tubular type, and 88 (71.5%) were nontubular type, suggesting that adenocarcinoma that were potentially invasive in nature were dominant in stage IV. The number of T4 patients was 82 (66.7%) and N3 patients, 67 (54.5%). Distant metastasis defined by H, P, and CY counted 46 (37.4%), 62 (50.4%), and 39 (45.3%) patients, respectively. The number of CY1 patients being less than P1 seems odd, but this phenomenon can be explained by understanding the actual peritoneal lavage cytology (CY) carried out upon operation. According to the chart review, CY was routine for P0 patients (n = 61) and among them, 14 were CY1. The rest of the patients (n = 62) being P1, CY was not a routine procedure. The total number of patients with CY being 86, 25 patients were P1 and naturally enough CY1. Therefore, the sum of 14 patients (P0CY1) and 25 patients (P1CY1) was the number of total CY1 patients (n = 39). If we consider all P1 patients as CY1, the number of CY1 would sum up to 76, instead of 39, CY1 being more than P1. Other M counted 55 (44.7%) patients, and they were all a result of the extensive lymph node involvement, which can be classified as M (LYM). The median of CEA and CA19-9 was 6.05 ng/mL and 15.5 ng/mL, respectively.

Table 1  Clinicopathological features of stage IV gastric cancer patients

View Fullscreen

Figure 1 shows the flow diagram of the treatments undertaken in stage IV patients. A total of 78 patients underwent a surgical procedure. Among them, 37 patients underwent reduction surgery, and 41 patients had palliative surgery. The decision whether to resect the primary tumor was entrusted to the surgical team on laparotomy. Thirteen patients did not receive postoperative chemotherapy mostly because of the worsening of their performance status. Others wished to have best supportive care after all. Thirty-eight patients did not have surgery and received chemotherapy alone or went to have best supportive care (n = 7). Chemotherapy was carried out with S-1 based regimen (n = 99). Among them, patients with S-1 only, or in combination with CDDP, taxans, and CPT11 were 45, 45, 8, and 1, respectively (Table 2).

View Figure

• Download as Powerpoint
• Download Figure

Table 2  Treatment data of patients

View Fullscreen

Univariate analysis of prognostic factors for overall survival is listed in Table 3. Age (70 or more versus under 70; P = 0.03), performance status (PS2 or more versus PS0 or 1; P < 0.01), liver metastasis (H1 versus H0; P = 0.02), other M (positive versus negative; P < 0.01), CA19-9 (37 ng/mL or more versus less than 37 ng/mL; P < 0.01), reduction surgery (nonresection versus resection; P < 0.01), chemotherapy (No versus Yes; P < 0.01) were found to be significant prognostic factors for overall survival. As also listed in Table 3, PS, other M, reduction surgery, and chemotherapy were calculated as independent prognostic factors. Considering these results, the survival curve of the subgroup of stage IV patients (i.e., better PS patients without other M) who underwent reduction surgery plus systemic chemotherapy showed significantly better survival curve than those who lack any one of these 4 favorable variables (Fig. 2).

Table 3  Univariate and multivariate analysis

View Fullscreen

View Figure

• Download as Powerpoint
• Download Figure

Discussion

We have clearly demonstrated in this study that on the premises of conducting both reduction surgery and systemic chemotherapy, stage IV patients with good PS and M factors limited within H1, P1, and CY1 can prolong survival significantly. Conducting reduction surgery and systemic chemotherapy for poor prognostic patients seems too aggressive when comprehensive treatment strategy is still yet to be defined. Therefore, treatment strategy for stage IV gastric cancer patients is always a controversial issue in clinical practice owing to their poor prognosis. However, our results may be good news for these patients. Furthermore, it may be a cornerstone for optimal treatment strategy for future stage IV gastric cancer patients. To our knowledge, we are the first to conclude that certain subgroup of stage IV patients can benefit from conducting both reduction surgery and chemotherapy.

Treatment options for stage IV gastric cancer patients are limited since the spread of the disease is far beyond its origin. According to the JGCA data, cumulative 5-year survival of stage IV patients was 15.3%,² and Korean data also showed similar results with 18%.³ Moreover, among all stage IV patients, 4% to 9% of them are said to be too advanced with no chance of curative surgery and their prognosis is absolutely poor showing less than 10% at 3 years after diagnosis.¹⁶

JGCA suggests extended surgery or palliative surgery as clinical practice and reduction surgery as a clinical study for stage IV patients (T4 and N3 patients are classified as stage IV in this suggestion). Past literatures have proven that when stage IV is diagnosed by only one M factor, conducting resection may prolong survival.^17–22 Therefore, removing the primary tumor is an option even if responsive chemotherapy is available today.

S-1 has gained an ultimate position as a standard antitumor agent for gastric cancer treatment. Together with other agents such as CDDP, CPT-11, and taxans (paclitaxel and docetaxel), the response rate of chemotherapy has improved dramatically. They have provided better prognosis for stage IV patients.^9–11 S-1 with CDDP was proven to have a high response rate and is chosen as the first line chemotherapy.¹²

Adjuvant chemotherapy trials against advanced gastric cancer have shown promising results.^23,24 On the other hand, chemotherapy is said to have no influence on survival for the elderly.²⁵ Nevertheless, S-1 alone can be used safely for the elderly and poor PS patients, which most stage IV patients are.^26,27 Neoadjuvant chemotherapy settings have been discussed lately to give an adequate amount of antitumor agent and if possible, downstage the non-curative tumor to perform R0 operation. The European group has reported recently that R0 surgery increased but did not affect overall survival.²⁸ Kochi and his colleagues have reported the efficacy of neoadjuvant S-1+CDDP.²⁹ Before any conclusive results come out by the well-organized studies, we should be careful in adopting neoadjuvant chemotherapy in the clinical settings.

In this study, our patient treatment strategy was nothing but practice based, and the decision whether to operate first or begin chemotherapy first was not randomly selected. We always tried to provide surgical treatment not only for better oncologic outcome but also for the settlement of the uncomfortable gastrointestinal symptoms caused by the tumor often described as GOS. Moreover, ensuring the passage by surgery may contribute to achieve better compliance of S-1 treatment along with the increase of oral diet. If the patients had a strong will to start chemotherapy rather than surgery, we respected their decision and carried it out without any delay.

Nevertheless, 13 patients who had surgery as initial treatment failed to move on to chemotherapy. Some had lost their wish to continue treatment, some had failed to improve or keep their performance status, and some had to move on to best supportive care. In a sense, these 13 patients could only enjoy the benefit of reducing unfavorable gastrointestinal symptoms caused by the tumor. But from the quality of life point of view, it may be considered as the right choice for them. Patient performance status and their will may change during the treatment course, and we should be flexible in all circumstances with decision making.

As shown in the survival curve in Fig. 2, we clearly demonstrated that regardless of treatment order, if both reduction surgery and chemotherapy are carried out, we can improve the prognosis of certain stage IV patients, that is, patients diagnosed stage IV with H, P, and CY (redundant included). Therefore, we should not hesitate to carry out both reduction surgery and systemic chemotherapy as long as patient PS is well enough. Moreover, metastatic-site specific treatment strategy was not necessary to achieve this result. Our study excluded M (LYM) from the favorable factor that extended lymph node metastasis is truly an unfavorable factor for stage IV patients.

The current study was based on a small number of data acquired retrospectively at a single institution in Japan. Although the extent of the disease does not permit us to carry out prospective studies, our results may be rationalized and validated in a larger randomized control trial. It is beyond the objective of this study, but clinically, it is very interesting to find out the treatment order or when to convert to surgery after neo-adjuvant chemotherapy. These questions could be clarified by a future randomized control trial.