Annals of Surgical Oncology

, Volume 16, Issue 9, pp 2510–2515

Chemosensitivity and Survival in Gastric Cancer Patients with Microsatellite Instability


    • Department of Surgery and Science, Graduate School of Medical SciencesKyushu University
  • Yoshihiro Kakeji
    • Department of Surgery and Science, Graduate School of Medical SciencesKyushu University
  • Yan Zhao
    • Department of Surgery and Science, Graduate School of Medical SciencesKyushu University
  • Rintaro Yoshida
    • Department of Surgery and Science, Graduate School of Medical SciencesKyushu University
  • Koji Ando
    • Department of Surgery and Science, Graduate School of Medical SciencesKyushu University
  • Takanobu Masuda
    • Department of Surgery and Science, Graduate School of Medical SciencesKyushu University
  • Kippei Ohgaki
    • Department of Surgery and Science, Graduate School of Medical SciencesKyushu University
  • Masaru Morita
    • Department of Surgery and Science, Graduate School of Medical SciencesKyushu University
  • Yoshihiko Maehara
    • Department of Surgery and Science, Graduate School of Medical SciencesKyushu University
Gastrointestinal Oncology

DOI: 10.1245/s10434-009-0580-8

Cite this article as:
Oki, E., Kakeji, Y., Zhao, Y. et al. Ann Surg Oncol (2009) 16: 2510. doi:10.1245/s10434-009-0580-8



Conflicting data exist regarding the relevance of high-frequency microsatellite instability (MSI-H) for predicting the prognosis and benefits of 5-fluorouracil (5-FU)-based chemotherapy. This study investigated the usefulness of MSI as either a prognostic indicator or predictor of distinct clinical attributes regarding the use of adjuvant chemotherapy with 5-FU and its analogues in gastric cancer.

Materials and methods

Data and tumor specimens were collected from 240 gastric cancer patients from 1993 to 2002. Five microsatellite loci were analyzed using a high-intensity microsatellite analysis reported previously. A Cox proportional hazard model was used to compare the clinical data and survival as well as any associations between MSI and 5-FU treatment status of patients with MSI or microsatellite stability (MSS) gastric cancers. A 3-(4,5-dimethyl-2-thiazolyl) -2,5-diphenyl-2H-tetrazolium bromide (MTT) assay was conducted in 168 cases to investigate chemosensitivity to 5-FU.


This analysis identified 22 MSI-H (9.4%), 25 MSI-L (10.7%), and 193 MSS (79.9%) tumors. Gastric cancer with MSI-H tended to have increased likelihood to show higher age, antral location of the tumor, and lymph vessel involvement (P < 0.05). Univariate analyses failed to show any difference between the MSI-H and MSS/MSI-L groups with respect to overall survival. Furthermore, survival after the administration of 5-FU did not correlate with MSI status, and MSI was not associated with 5-FU sensitivity by MTT assay.


The results of this study indicate that MSI status has no clear influence on overall survival or response to 5-FU in gastric cancer.

Gastric cancer is one of the leading causes of cancer death worldwide, although its incidence has decreased in Western and some Asian countries, including Japan.1 Multiple genetic and epigenetic alterations in oncogenes, tumor-suppressor genes, cell-cycle regulators, cell adhesion molecules, DNA repair genes, and genetic instability as well as telomerase activation are implicated in the multistep process of gastric carcinogenesis.24 The specific combinations of these alterations differ in the two histological types of gastric cancer, indicating that intestinal- and diffuse-type carcinomas have distinct carcinogenetic pathways.2 Microsatellite instability (MSI) is a phenomenon which is caused by a mismatch repair gene deficiency and is observed in around 20% of gastric cancer patients. Gastric cancer with high-frequency microsatellite instability (MSI high; MSI-H) represents a well-defined subset of carcinomas, showing distinctive clinicopathological features.1,58 In colon cancer, tumors with MSI-H are characteristic of hereditary nonpolyposis colorectal cancer syndrome (HNPCC) which, in the majority of cases, is associated with early age at onset and which is caused by germline mutations in one of the mismatch repair (MMR) genes.912 In contrast, the MSI-H phenotype in gastric cancer is predominantly caused by epigenetic hypermethylation of the hMLH1 as MMR gene rather than germline mutations in one of the MMR genes.1315 Hypermethylation of the promoter of the hMLH1 gene occurs in more than 30% of human gastric cancer tissues.13,16,17 Thus MSI in gastric cancer is frequent enough not to be ignored. Not only genetic factors but also dietary intake, cigarette smoking, and alcohol consumption may influence the development of hMLH1-methylated gastric cancer.16 MSI-H and microsatellite stable (MSS)/MSI-L gastric cancer differ in some pathological and clinical findings.1,5,18 Many studies found a better prognosis of MSI-H tumors in comparison with MSI-Land MSS tumors with respect to overall and disease-free survival.1,5,19 However, other studies found no influence of MSI on survival.2022 The mechanism by which MSI may influence clinical outcome is unknown in gastric cancer.

Currently, tumor stage is the most important predictor of prognosis for gastric cancer. A recent pivotal Japanese trial showed that systemic adjuvant chemotherapy using an oral 5-FU derivative improved survival time for stage II and III gastric cancer.23 In the same manner as for gastric cancer, adjuvant chemotherapy improved the survival of stage II and III colon cancer. However, some studies found that MSI-H tumors may be resistant to 5-FU.24,25 In contrast, others found a better chemosensitivity in MSI-H tumors.2527 However, these findings remain controversial. So far, no-one has shown any clear results in regard to whether MSI is a prognostic or predictive factor for 5-FU sensitivity in gastric cancer. This study was therefore a retrospective study of unselected patients with gastric cancer, irrespective of age at diagnosis or tumor stage, to identify tumor MSI status and analyze its relationship with overall survival and chemosensitivity.

Patients and Methods

Tissue Samples

Primary gastric carcinoma tissues and corresponding normal mucosa were obtained from 240 Japanese patients who underwent surgery in the Department of Surgery II at Kyushu University Hospital between 1993 and 2000. No patient had received chemotherapy before surgery. Informed consent was obtained from each patient prior to tissue acquisition, and the study was approved by the ethical committee of Kyushu University. All the tumors, examined microscopically by pathologists, were diagnosed as adenocarcinoma and classified according to the criteria of the Japanese Research Society for Gastric Cancer. The cancer tissues and the well-separated normal gastric mucosa obtained from the gastrectomies were immediately cut into two pieces. Half of the specimen was kept in liquid nitrogen for DNA extraction, and the another half was immediately put into medium (Minimal Essential Medium; MEM) for chemosensitivity tests. The genomic DNA was prepared by proteinase K digestion and phenol–chloroform extraction, which was then followed by ethanol precipitation. In all cases, the histopathological type of the tumors was determined to be adenocarcinoma.

Microsatellite Instability Analysis

MSI was analyzed in this study using a DNA sequencer with five microsatellite markers. The oligonucleotide primers were synthesized and then purified by high-performance liquid chromatography (HPLC). The sequences of the primers for polymerase chain reaction (PCR) were: D2S123-5′; 5′-AAACAGGATGCCTGCCTTTA, D2S123-3′; 5′-GGACTTTCCACCTATGGGAC, D5S107-5′; 5′-GGCATCAACTTGAACAGCAT, D5S107-3′; 5′-GATCCACTTTAACCCAAATAC, D10S197-5′; 5′-ACCACTGCACTTCAGGTGAC, D10S197-3′; 5′-GTGATACTGTCCTCAGGTCTCC, D11S904-5′; 5′-ATGACAAGCAATCCTTGAGC, D11S904-3′; 5′-CTGTGTTATATCCCTAAAGTGTGA; D13S175-5′; 5′-TGCATCACCTCACATAGGTTA, D13S175-3′; 5′-TATTGGATACTTGAATCTGCTG. The PCR reactions using genomic DNA were performed using a TAKARA GeneAmp PCR Reagent Kit and were run in a Perkin-Elmer GeneAmp PCR system 9700 (Norwalk, CT). The thermal conditions of the system were as follows: one cycle at 95°C for 4 min; 35 cycles at 95°C for 0.5 min, 55°C for 0.5 min, and 72°C for 0.5 min; and one cycle at 72°C for 10 min. The DNA derived from the cancer tissues were amplified with ROX-labeled 5′ primer and cold 3′ primer, whereas the DNA from the normal tissues was amplified with HEX-labeled 5′ primer and cold 3′ primer. The running condition of the Perkin-Elmer Genetic Analyzer 310 (Norwalk, CT) was described previously.27 The data were processed by the ABI software GeneScan. An alteration in the length of a microsatellite PCR fragment obtained from a cancerous tissue specimen was defined as MSI positivity. The detailed determination method was previously reported.12,28 MSI status was classified as follows: high frequency of microsatellite instability (MSI-H) when MSI was demonstrated in two or more of five analyzed markers, low frequency of microsatellite instability (MSI-L) when MSI was demonstrated in one out of five analyzed markers, and microsatellite stability (MSS) when no positive MSI were detected in any of the loci.

Chemosensitivity Test

One hundred and sixty-eight patient specimens underwent an in vitro chemosensitivity test. The succinate dehydrogenase inhibition (SDI) test was conducted as previously described by Takeuchi et al.29,30 Cell viability was estimated based on succinate dehydrogenase (SD) activity, and was determined using 3-(4,5-dimethyl-2-thiazolyl) -2,5-diphenyl-2H-tetrazolium bromide (MTT).

Statistical Methods

Statistical significance was determined using Student’s t-test and the chi-square test with Yates’s correlation factor, and was also confirmed via a two-tailed Fisher exact test. A difference was considered significant when the P-value was less than 0.05.


Correlations Between MSI Status and Clinicopathological Factors in Gastric Cancer

Screening identified 240 patients with gastric cancers for whom pathological material and follow-up data were available. Table 1 shows 22 MSI-H (9.4%), 25 MSI-L (10.7%), and 193 MSS (79.9%) tumors. In many previous reports, MSI status was categorized into two groups, including MSI-H and MSS/MSI-L, since MSI-L tumors demonstrate very similar clinicopathological features to those of MSS tumors.2426 In our past reports, MSI was classified into type A and type B MSI.12,3134 However, a classification of MSI-L and MSI-H was used in this article simply to discuss the pathological characteristics. Furthermore, in the same way as noted in other previous reports, MSS and MSI-L were categorized in the same group as MSS/MSI-L. MSI-H was associated with higher age (P < 0.05), antral location of the tumor (P < 0.05), and lymph vessel involvement (P < 0.05); however, MSI-H was not associated with the intestinal type according to Lauren classification, depth of invasion or tumor stage (Table 2).
Table 1

Frequency of microsatellite instability of gastric cancer

MSI status

Incidence (%)


193 (79.9%)


25 (10.7%)


22 (9.4%)

Table 2

Clinicopathological features and microsatellite instability of gastric cancer


MSS/MSI-L (n = 218)

MSI-H (n = 22)
























Depth of invasion

  T1, T2




  T3, T4




Histological lymph node metastasis









Lymph vessel involvement

  ly (–)b




  ly (+)c




Vascular involvement



















  I + II




  III + IV




aCase for which the histological diagnosis could not be determined were excluded

bly (–) No lymphatic invasion was observed

cly (+) The lymphatic invasion was observed

d Case for which the location could not be determined were excluded

e The tumor location was expressed as followings; the esophagus (E), upper (U), middle (M), and lower (L)

MSI Status and Prognosis in Gastric Cancer

The median follow-up of all patients was 983 ± 861 days and within the study period, 64 patients (28.2%) died. Adjuvant chemotherapy was performed for 76/240 (31.7%) patients. A Cox regression model for a univariate analysis was used to assess the effect of MSI status on survival. This analysis failed to show any difference between the MSI-H and MSS/MSI-L groups with respect to overall survival (Fig. 1).
Fig. 1

Overall survival in all cases. The bold line is MSS and dotted line is MSI-H/L

MSI Status and Chemosensitivity in Gastric Cancer

The regimens in the 76 cases who received chemotherapy were as follows: tegafur/uracil in 27 cases, 5-FU bolus injection in 20 cases, doxifluridine in 11 cases, 1-hexylcarbamoyl-5-fluorouracil (HCFU) in 8 cases, tegafur in 6 cases, and tegafur/gimeracil/oteracil potassium in 4 cases. They are all derivatives of 5-fluorouracil.

A Cox regression model for a univariate analysis was used to assess the effect of MSI status on survival in the 76 patients who underwent adjuvant chemotherapy. We found a small difference between the patients who received adjuvant chemotherapy and those who did not receive any adjuvant chemotherapy, but it was not significant (Fig. 2a). Survival after administration of 5-FU or its derivatives did not correlate with MSI status (Fig. 2b). When the chemotherapeutic sensibilities of these patients were studied in an MTT assay in 168 patients, MSI was not associated with the 5-FU sensitivity (Fig. 3). The chemosensitivity test was repeated for the other chemotherapeutic agents, such as cisplatin (CDDP), adriamycin, and mitomycin; however, no difference was found among the patients regarding each MSI status group (data not shown).
Fig. 2

a Overall survival in all cases following the use of adjuvant chemotherapy. The bold line is for the patients who received adjuvant chemotherapy and the dotted line is for the patients who did not received adjuvant chemotherapy. b Overall survival following the use of adjuvant chemotherapy. The bold line is for MSS and dotted line is for MSI-H/L
Fig. 3

In vitro chemosensitivity test using MTT assay. When the chemotherapeutic sensibilities of these patients were studied in an MTT assay, MSI was not associated with 5-FU sensitivity. Cell viability was estimated based on succinate dehydrogenase (SD) activity, and was determined using 3-(4,5-dimethyl-2-thiazolyl) -2,5-diphenyl-2H tetrazolium bromide (MTT)


Gastric cancer (GC) remains a leading cause of cancer mortality worldwide. Although the association between clinicopathological factors and MSI has been well investigated in colon cancer, their proportion and association with clinicopathological factors remain to be proven in gastric cancer. It is unclear whether MSI is a prognostic factor or a predictive factor in gastric cancer. The most important clinical findings concerning microsatellite instabilities is their effect on the prognosis of patients and chemosensitivity to 5-FU for adjuvant chemotherapy in colon cancer. There is improved survival in patients with non-microsatellite instability-high (MSI-L/N) tumors after 5-fluorouracil-based chemotherapy that does not extend to patients with microsatellite instability-high tumors (MSI-H).24 Benatti et al. also reported that type of genomic instability could influence the prognosis of CRC, in particular in stages II and III, and 5-fluorouracil-based chemotherapy does not improve survival among MSI-H patients. Therefore, MSI is regarded as a predictive factor in colon cancer. However, the survival difference was influenced by the patients of hereditary nonpolyposis (HNPCC) colorectal cancer who were diagnosed at a younger age and less advanced in comparison with their sporadic MSI-H counterparts.35 Controversy remains regarding the use of microsatellite instability for predicting the prognosis or benefit of 5-fluorouracil-based chemotherapy in patients.

Microsatellite instability (MSI) is a consequence of the inactivation of a mismatch repair gene and is observed in approximately 10–30% of gastric cancer cases. The clinical significance of high levels of MSI (MSI-H) in gastric cancer has been reported.12,28,31,33,34,36 The purpose of this study was to investigate the prognostic significance of high levels of MSI (MSI-H) in gastric cancer patients, and to investigate the predictive significance in relation to 5-fluorouracil-based chemotherapy. Tumors resected from 240 patients with sporadic gastric cancer were examined using a unique fluorescent technique, high-resolution fluorescent microsatellite analysis (HRFMA), to determine the frequencies of the MSI (MSI-H) in gastric cancer.31,32 In this system, several devices were prepared to improve reproducibility of the polymerase chain reaction products, migration accuracy of electrophoresis, and characteristics of the detection system. HRFMA provides precise and objective analyses of microsatellite alterations.31,33 In previous reports, gastric cancer with MSI-H had a significantly higher frequency of antral location and intestinal subtype, but a lower incidence of lymph node metastasis in comparison with gastric cancer with MSI-L or MSS.1,7,37,38 The MSI-H gastric cancers in the current study were also significantly associated with older age and antral location, but not intestinal subtype. Gastric cancers with and without high-frequency microsatellite instability (MSI-H) represent distinctive pathways of carcinogenesis.39 A high frequency of MSI was reported in intestinal metaplasia and this suggests that intestinal metaplasia and well-differentiated adenocarcinoma in the stomach may have the same molecular backgrounds.37 In the analysis of Japanese patients, the high incidence of the diffuse type of gastric cancer may result in the low frequency of MSI-H. As a result, the significant association between MSI-H and intestinal-type cancer might not be clearly proven. In the current analysis, MSI-H was not significant but slightly more frequently found in the intestinal type of gastric cancer than in the diffuse type of cancer. Moreover, previous reports considered the low frequency of lymph node metastasis to be a reason for good prognosis of MSI-H patients. However, the occurrence of lymph node metastasis is not so infrequent in MSI-H tumors, and a rather large degree of lymph vessel invasion was also observed in our cases. These phenomena may therefore have affected the results regarding the prognosis.

The current analysis showed no clear influence of MSI status on overall survival. All of the patients in the current study underwent a D2 lymphadenectomy according to the Japanese Guidelines for Diagnosis and Treatment for Carcinoma of Stomach, and the prognosis was better than that described in a previous report, in spite of the high rate of lymph node metastasis. Even if there are lymph node metastases or lymph vessel involvement in MSI-H patients, a D2 lymphadenectomy may overcome the negative impact of the prognosis for the lymph node metastasis. Accordingly, no difference was found in the prognosis between MSI-H and MSI-L/N tumors. At least, MSI cannot be regarded as a prognostic factor in our results.

Multimodal treatment protocols are increasingly employed to improve the survival of patients with locally advanced adenocarcinomas of the upper gastrointestinal tract; however, only 30–40% of patients respond to 5-FU and cisplatin-based neoadjuvant chemotherapy.40 Recently, S-1, an oral derivative of 5-fluorouracil, has proven to be an effective adjuvant treatment for East Asian patients who have undergone a D2 dissection for locally advanced gastric cancer.23,41 Adjuvant chemotherapy was done on advanced-stage cancer and the patients treated with adjuvant chemotherapy showed poorer prognosis (Fig. 2a). Many studies have evaluated the role of high levels of microsatellite instability (MSI) as a predictor of the response to chemotherapy in colorectal cancer. Adjuvant chemotherapy improves the overall survival among patients with microsatellite-stable tumors or tumors exhibiting low-frequency microsatellite instability.25 In the current gastric cancer cases, MSI-H phenotype did not predict any benefit of adjuvant chemotherapy on overall survival (OS). However, no difference could be detected in our study since adjuvant chemotherapy had only been administered to 76 cases. Therefore, sensitivity to 5-FU was investigated with an MTT assay using tissue samples. The assay showed that MSI phenotype has no correlation with chemosensitivity in an in vitro sensitivity test. Therefore MSI phenotype was not a predictive factor for adjuvant chemotherapy of gastric cancer.

In conclusion, in a series of 240 patients with gastric cancer, MSI-H did not demonstrate any particular benefit as either a prognostic factor or a predictive factor for the response to chemotherapy in comparison with patients with MSS/MSI-L tumors. Due to the conflicting results of the available data and the lack of any prospectively performed randomized studies, it is therefore considered to be premature to use MSI status routinely to determine optimal individual patient management in gastric cancer cases.

Copyright information

© Society of Surgical Oncology 2009