Molecular Biomarker Study in a Randomised Phase III Trial of Irinotecan Plus S-1 versus S-1 for Advanced Gastric Cancer (GC0301/TOP-002).

AIMS: Gastric cancer is a common and heterogeneous disease; however, global standard and biomarkers for selecting chemotherapy regimens have not been established. This study was designed retrospectively to identify molecular biomarkers for irinotecan plus S-1 (IRI-S) and S-1 therapy from subset analyses in GC0301/TOP-002, a randomised phase III trial for advanced gastric cancer. MATERIALS AND METHODS: Paraffin-embedded primary tumour specimens were collected from 126 of 326 randomised patients in GC0301/TOP-002. The mRNA was measured for thymidylate synthase, dihydropyrimidine dehydrogenase, topoisomerase I, excision repair cross-complementing gene 1 (ERCC1) and thymidine phosphorylase; categorised into low and high to analyse their association with efficacy end points. RESULTS: There was no significant difference in each mRNA between S-1 and IRI-S groups, whereas there were differences among some clinical characteristics. Multivariate analyses for overall survival showed that mRNA levels were not correlated with prognosis. By comparison, between IRI-S and S-1 arms, low thymidylate synthase, low ERCC1 and high thymidine phosphorylase were associated with better prognosis for IRI-S versus S-1 (hazard ratio = 0.653, 0.702 and 0.709, respectively; P < 0.15 for each interaction). CONCLUSION: Low thymidylate synthase, low ERCC1 and high thymidine phosphorylase are candidates for predictive biomarkers for first-line treatment in advanced gastric cancer by IRI-S. Further study is warranted to confirm these results in other clinical trials and cohort studies.

Neoadjuvant chemotherapy with S-1 and cisplatin followed by D2 gastrectomy with para-aortic lymph node dissection for gastric cancer with extensive lymph node metastasis.

BACKGROUND: Locally advanced gastric cancer with extensive regional and/or para-aortic lymph node (PAN) metastases is typically unresectable and associated with poor outcomes. This study investigated the safety and efficacy of S-1 plus cisplatin followed by extended surgery with PAN dissection for gastric cancer with extensive lymph node metastasis. METHODS: Patients with gastric cancer with bulky lymph node metastasis along the coeliac artery and its branches and/or PAN metastasis received two or three 28-day cycles of S-1 plus cisplatin, followed by gastrectomy with D2 plus PAN dissection. The primary endpoint was the percentage of complete resections with clear margins in the primary tumour (R0 resection). A target sample size of 50 with one-sided α of 0.105 and ß of approximately 0.2 corresponded to an expected R0 rate of 65 per cent and a threshold of 50 per cent. RESULTS: Between February 2005 and June 2007, 53 patients were enrolled, of whom 51 were eligible. The R0 resection rate was 82 per cent. Clinical and pathological response rates were 65 and 51 per cent respectively. The 3- and 5-year overall survival rates were 59 and 53 per cent respectively. During chemotherapy, grade 3/4 neutropenia occurred in 19 per cent and grade 3/4 non-haematological adverse events in 15.4 per cent. The incidence of grade 3/4 adverse events related to surgery was 12 per cent. There were no reoperations or treatment-related deaths. CONCLUSION: For locally advanced gastric cancer with extensive lymph node metastasis, 4-weekly S-1 plus cisplatin followed by surgery including PAN dissection was safe and effective for some patients. Further investigation of this treatment strategy is warranted.

KRAS and BRAF mutations are rare and related to DNA mismatch repair deficiency in gastric cancer from the East and the West: results from a large international multicentre study.

BACKGROUND: Inhibitors of the epidermal growth factor (EGFR) signaling pathway have a major role in the treatment of KRAS wild-type colorectal cancer patients. The EGFR pathway has been shown to be activated in gastric cancer (GC). However, published data on KRAS and BRAF mutation status is limited in GC and has not been compared between GC from different geographic regions. METHODS: The prevalence of KRAS and BRAF mutations was established in 712 GC: 278 GC from the United Kingdom, 230 GC from Japan and 204 GC from Singapore. The relationship between KRAS/BRAF mutation status, DNA mismatch repair (MMR) status, clinicopathological variables and overall survival was analysed. RESULTS: Overall, 30 (4.2%) GC carried a KRAS mutation. In total, 5.8% of the UK GC, 4% of Japan GC and 1.5% of Singapore GC were KRAS mutant. KRAS mutant GC had fewer lymph node metastases in the UK cohort (P=0.005) and were more frequent in elderly patients in the Japan cohort (P=0.034). KRAS mutations were more frequent in MMR-deficient GC in the UK and the Japanese cohort (P<0.05). A BRAF mutation was only detected in a single Japanese GC. CONCLUSIONS: This large multicentre study demonstrated that KRAS mutations and DNA MMR deficiency have a role in a small subgroup of GC irrespective of country of origin, suggesting that this subgroup of GC may have developed along a common pathway. Further studies need to establish whether concomitant mutations or amplifications of other EGFR signalling pathway genes may contribute to the activation of this pathway in GC.

Phase II study of oxaliplatin plus S-1 as first-line treatment for advanced gastric cancer (G-SOX study).

BACKGROUND: The efficacy and safety of oxaliplatin combined with S-1 (SOX regimen) for unresectable advanced or recurrent gastric cancer were investigated. PATIENTS AND METHODS: Oxaliplatin was administered i.v. (100 mg/m(2)) on day 1, while S-1 was administered orally (80 mg/m(2)/day, b.i.d.) for 14 days followed by a 7-day rest. This schedule was repeated every 3 weeks. RESULTS: Among 55 patients enrolled, one patient received oxaliplatin for the other study, and three patients were considered unsuitable against the inclusion criteria. Accordingly, 51 patients were assessable for efficacy. The response rate was 59%, and the disease control rate was 84%. The median progression-free survival time was 6.5 months, the 1-year survival rate was 71%, and the median survival time was 16.5 months. In 54 patients assessed for safety, the major grade 3/4 toxic effects were neutropenia (22%), thrombocytopenia (13%), anemia (9%), anorexia (6%), fatigue (6%), and sensory neuropathy (4%). CONCLUSION: These findings indicate that SOX regimen with oxaliplatin at a dose of 100 mg/m(2) is feasible and shows promising efficacy against advanced gastric cancer.

Efficacy of adjuvant chemotherapy using oral fluorinated pyrimidines for curatively resected gastric cancer: a meta-analysis of centrally randomized controlled clinical trials in Japan.

Adjuvant chemotherapy for gastric cancer has been extensively explored in Japan since the 1950s, and a combination of oral fluorinated pyrimidines (o-FP) and mitomycin C (MMC) has been mainly utilized for adjuvant chemotherapy. However, there is no sufficient evidence for the efficacy of adjuvant therapy. Therefore, we assessed the efficacy of o-FPs over surgery alone (control) by means of a meta-analysis of Japanese centrally randomized controlled clinical trials conducted between 1980 and 2005. For inclusion in this study, studies had to compare adjuvant chemotherapy for curatively resected gastric cancer with surgery alone, mainly targeting o-FP, and central randomization designed to comply with contemporary standards for clinical trials in Japan. For the 4 trials that met the eligibility criteria, the estimated hazard ratio was 0.73 (95%CI=0.60-0.89). Our findings show that in Japan adjuvant chemotherapy using o-FP for long-term maintenance therapy appears to be effective for gastric cancer patients after curative resection.

[Significance of radical recurrent tumor resection for recurrent gastric cancer as assessed by prognosis].

To evaluate the relationship between radical surgery of recurrent tumor and prognosis in cases of recurrent gastric cancer, we analyzed data on 202 patients with relapsed gastric cancer, focusing on surgical recurrent tumor removal. In our series, 18 of the 202 patients underwent radical recurrent tumor resection. Resected tumors were located in the ovarium (n = 4), colorectum (n = 3), liver (n = 3), lymph node (n = 2), locoregional stoma (n = 2), and peritoneum, adrenal gland, brain, and lung (n = 1 each). No surgery-related mortality occurred. One patient remains alive over 5 years after hepatectomy without recurrence, and 17 died within 3 years: 7 patients from primary recurrence and 10 from multiple modes of recurrence. Median survival after recurrence (MSTAR) in the 18 radical surgery patients was 14 months, against 5 months in those treated palliatively (p = 0.0001). MSTAR for the ovary and the liver were 30 months and 15 months in the radical surgery cases, and 2.5 months for the ovary and 5 months for the liver in the palliative cases. Significant differences were thus seen between radical and palliative cases in the ovary (p = 0.010) and in the liver (p = 0.036). Median survival after gastrectomy was 45 months in the radical surgery cases, and 28 months in the palliative cases (p = 0.024). In postoperative gastric cancer follow-up, early detection of recurrence and radical surgery may well benefit patients with relapse, especially in the liver and ovary, in terms of survival.

Should scirrhous gastric carcinoma be treated surgically? Clinical experiences with 233 cases and a retrospective analysis of prognosticators.

BACKGROUND/AIMS: The prognosis of patients with scirrhous gastric carcinoma has been poorest. METHODOLOGY: To clarify the role of surgical treatment, 233 patients with a primary scirrhous gastric carcinoma were retrospectively analyzed. RESULTS: Of the 233 patients, 182 underwent surgical resection, while the other 51 did not. The median survival time of those with unresectable tumors was 88.0 +/- 15.3 days and that of those who underwent resection was 380.0 +/- 41.8 days. In the 182 patients who underwent resection, multivariate analysis revealed four significant factors; lymphatic invasion, serosal invasion, curability, and lymph node dissection. Of these, curability was the most significant. The median survival time of patients whose tumor were curatively resected was 727.0 +/- 116.3 days, significantly longer than 272 +/- 34.9 days for those whose resection ended noncuratively. In 65 patients whose tumor was curatively resected, subset analyses of factors by multivariate analyses revealed an absence of serosal invasion as the single significant prognosticator. The 5-year survival rate was 55.6% in patients with scirrhous cancer without serosal invasion. CONCLUSIONS: For patients with scirrhous gastric carcinoma, palliative resection should not be attempted for poor outcome. However, if curative resection seems feasible, radical surgery would be justified, especially for tumors without serosal exposure.

[Significance of laparoscopy for response assessment of chemotherapy in gastric cancer].

UNLABELLED: We performed laparoscopy before and after chemotherapy in two patients with relapsed and advanced gastric cancer, whose major metastatic sites had been diagnosed as being in the peritoneum. A change in tumor responses when assessed by laparoscopy was found. Case 1: A 63-year-old man presented with an umbilical metastasis and suspected peritoneal metastases after gastrectomy. Laparoscopy revealed peritoneal metastases before chemotherapy. After one course of chemotherapy the umbilical tumor disappeared (CR). Laparoscopy after two courses of chemotherapy revealed increasing peritoneal metastases (PD). The overall response was PD. Case 2: A 67-year-old woman was referred to our hospital with a diagnosis of type 4 gastric cancer. Staging laparoscopy revealed massive lymph node metastases and the patient was positive in peritoneal washing cytology. After four courses of chemotherapy, the primary tumor and the metastatic lymph nodes had decreased in size (PR). In contrast, laparoscopy revealed increasing peritoneal metastases (PD). The overall response was PD. CONCLUSION: In patients with peritoneal and other modes of metastasis, tumor response to chemotherapy may be misjudged by conventional imaging alone. Intraperitoneal examination by laparoscopy provides accurate information, including the tumor response to chemotherapy.

Intratumoral concentrations of tissue inhibitor of matrix metalloproteinase 1 in patients with gastric carcinoma a new biomartker for invasion and its impact on survival.

BACKGROUND: Previously, the authors clarified that the plasma concentration of tissue inhibitor of matrix metalloproteinase 1 (TIMP-1) in patients with gastric carcinoma was a significant predictor of tumor invasiveness and metastasis. METHODS: To further clarify the clinical significance of TIMP-1, the authors used an enzyme-linked immunoassay to assess TIMP-1 protein concentrations in samples of tumor tissue from 86 patients who underwent primary resection for gastric carcinoma. Concentrations in samples of normal gastric mucosa from 73 of these patients also were assessed. RESULTS: Tissue TIMP-1 concentrations were significantly greater in gastric tumors than in normal gastric mucosa and were associated significantly with a variety of pathologic factors, including macroscopic type, depth of tumor invasion in the gastric wall, presence of lymphatic vessel invasion, pattern of tumor infiltration into the surrounding tissue, and disease stage. Significantly greater TIMP-1 concentrations were found in tumors that were exposed to the serosal surface compared with tumors that were limited to the submucosal layer. TIMP-1 protein was significantly greater in tumors with lymphatic vessel invasion, an infiltrative pattern into the surrounding tissue (INF-gamma), and in tumors from patients with Stage III disease. Survival was significantly poorer in patients with TIMP-1 concentrations > or = 10.0 ng/mg total protein. When patients were stratified by disease stage, survival was significantly different in patients with Stage III disease. Multivariate analysis demonstrated that intratumoral concentrations of TIMP-1 were the most significant independent factor for survival. CONCLUSIONS: These findings suggest that the intratumoral concentration of TIMP-1 protein may be a good indicator of tumor aggressiveness and can serve as a significant independent predictor of survival in patients with gastric carcinoma.

Gastrectomized patients are in a state of chronic protein malnutrition analyses of 23 amino acids.

BACKGROUND/AIMS: Malnutrition is one of the major postoperative complications of radical subtotal or total gastrectomy for gastric cancer. This study was conducted to clarify the nutritional consequences of radical gastrectomy with respect to protein metabolism. METHODOLOGY: To evaluate the nutritional status and the abnormalities in protein metabolism in such cases, serum concentrations of 23 amino acids were measured by high performance liquid chromatography in 40 patients who had undergone either subtotal (n = 20) or total (n = 20) gastrectomy more than 6 months prior to this analysis. RESULTS: Serum concentrations of total amino acids and nonessential amino acids were the same between gastrectomized patients and healthy controls (n = 50). However, concentrations of essential amino acids, essential amino acid/nonessential amino acid and branched-chain amino acid/total amino acid ratios were significantly lower in patient groups than in normal controls. Each essential amino acid was decreased and concentrations of glutamate and citrulline were increased in both patient groups compared with controls. The major differences between patients with subtotal and total gastrectomies included an increased ornithine and a decreased arginine concentration in patients with subtotal gastrectomy. CONCLUSIONS: These changes suggest that malabsorption of protein from the intestinal tract causes persistent proteolysis in the skeletal muscle for long periods of time after surgery in these patients and that changes in ornithine and citrulline levels may reflect more severe alterations in those with total gastrectomy.

Expression of facilitative glucose transporter 1 mRNA in colon cancer was not regulated by k-ras.

The expression of facilitative glucose transporter isoforms in colon adenocarcinoma and the possible role of k-ras in inducing GLUT (glucose transporter) mRNA were studied. RT-PCR demonstrated GLUT2 and GLUT3 expression in 100% of the ten normal colon mucosa samples but detected no GLUT1 mRNA. By contrast, GLUT1 mRNA was detected in all 20 (100%) colon cancer samples examined. GLUT4 mRNA was not detected in either normal mucosa or colon cancer tissues. Semiquantitative PCR demonstrated equal amounts of GLUT2 and GLUT3 mRNA in both normal mucosa and colon cancer samples. A point mutation in codon 12 of k-ras was detected in only six of the 20 (30%) colon cancer samples. Thus, a major difference between normal colon epithelia and colon cancer was the acquisition of GLUT1 expression, which was unlikely to have been induced by a point mutation in codon 12 of k-ras.

Suppression of facilitative glucose transporter 1 mRNA can suppress tumor growth.

We attempted to suppress glucose transporter 1 (GLUT1) expression by transfecting MKN45 cells with cDNA for antisense GLUT1. Glucose transport was significantly decreased in cells with antisense GLUT1 compared with wild-type cells or cells with vector alone. Suppression of GLUT1 mRNA resulted in a decreased number of cells in the S phase. This was accompanied by overexpression of p21 protein. Tumorigenicity in the nude mice injected with antisense GLUT1 expressing cells was significantly slower than in those with wild-type MKN45 cells. These results suggest that antisense GLUT1 mRNA inhibits tumor growth through a G(1) arrest and that expression of antisense GLUT1 mRNA via gene therapy can be used as a tool in the treatment of cancer.

Prognostic value of tissue inhibitor of matrix metalloproteinase-1 in plasma of patients with gastric cancer.

Tissue inhibitor of matrix metalloproteinase-1 (TIMP-1) in plasma has been reported to be related to disease progression in patients with gastric cancer. However, the prognostic significance of plasma TIMP-1 concentrations has not been clarified. Concentrations of TIMP-1 protein were measured by enzyme-linked immuno-sorbent assay in plasma samples of 147 preoperative patients who subsequently underwent gastric resection, and prognosis was compared. The cut-off value of plasma TIMP-1 concentrations was defined as 112.5 ng/ml, referring to the TIMP-1 levels in patients with intramucosal gastric cancer. Twenty-nine out of 147 patients had higher plasma TIMP-1 levels than the cut off value. When the patients were divided into those with elevated values and those with normal TIMP-1, such parameters as age, serosal invasion, metastases to lymph nodes, peritoneum, and liver, lymphatic invasion, curability, and stage were significantly different between the two. By univariate analysis of the factors affecting survival, macroscopic type, histology, serosal invasion, metastasis to lymph node, peritoneum, and liver, vessel invasions, curability, and plasma TIMP-1 were significant. However, multivariate analysis revealed that TIMP-1 was the only significant factor. In patients with gastric cancer, plasma TIMP-1 seem to be an independent and most powerful prognosticator for the survival.

Plasma concentrations of VEGF and bFGF in patients with gastric carcinoma.

We examined plasma levels of vascular endothelial growth factor (VEGF) and basic fibroblast growth factor (bFGF) in 54 patients with gastric carcinoma. Postoperative survival was significantly poorer in patients with plasma VEGF levels more than 10.0 pg/ml at the time of surgery. By an univariate analysis of the factors affecting survival, serosal invasion, lymph node metastasis, peritoneal dissemination, lymphatic vessel invasion, curability, and VEGF proteins were significant. By a multivariate analysis only VEGF levels and curability remained significant. Patients with recurrent disease, including liver metastasis, had significantly higher plasma VEGF concentrations than those with resectable primary tumors. VEGF, not bFGF, may serve as an independent prognosticator and a sensitive indicator for liver recurrence in patients with gastric carcinoma.

Marked increase of trypsin(ogen) in serum of linitis plastica (gastric cancer, borrmann 4) patients.

Linitis plastica, or Borrmann 4 gastric cancer, shows very poor prognosis, and the reason has not been understood. In the present study, we examined serum levels of trypsin(ogen) in 44 gastric cancer patients, including 17 early gastric cancer, 18 non-Borrmann 4 advanced gastric cancer, and 9 Borrmann 4 gastric cancer, by using the RIA gnost Trypsin kit (Hoechst Japan, Tokyo, Japan), which was expected to detect trypsin-1, trypsin-2, trypsinogen-1, and trypsinogen-2 in sera. The trypsin(ogen) concentration was much higher in the patients with linitis plastica than in the other gross types of gastric cancer. Hypertrypsinemia was identified in approximately 60% of advanced gastric cancer cases. Lymph node involvement, liver metastasis, or poorly differentiated adenocarcinoma is an important factor of hypertrypsinemia. The serum trypsin(ogen) level in linitis plastica patients was 3484.4 +/- 2319.7 ng/ml (mean +/- SD), which was significantly higher not only than that of the early gastric cancer (384.1 +/- 92.1) but also the stage IV gastric cancer patients (578 +/- 440.4), excluding those with linitis plastica. The elevated serum trypsinogen level in linitis plastica patients may be related to the malignant behavior of this type of cancer cell. Serum trypsin(ogen) of linitis plastica shows significantly higher concentrations than do the other types of advanced gastric cancer. Therefore, serum concentration of trypsin(ogen) might be a good marker of gastric cancer of linitis plastica.

Angiogenesis inhibitor, TNP-470, suppresses growth of peritoneal disseminating foci.

BACKGROUND/AIMS: Angiogenesis is critical not only for growth of primary tumors but also for cells established at distant organs. We investigated the effects of angiogenesis inhibitor, TNP-470, on the establishment and growth of intraperitoneally inoculated human gastric cancer cell line, MKN-45, and survival of nude mice with this tumor. METHODOLOGY: Human gastric cancer cell line, MKN-45, were injected into the peritoneal cavity of an ICR nude mouse and a model of peritoneal dissemination was developed. TNP-470 was injected subcutaneously every other day from day 1 until sacrifice or death. The effects of TNP-470 on MKN-45 cells were also examined in vitro. RESULTS: Although the number of disseminated foci was not significantly different, the maximum size was significantly smaller in a TNP-treated group than those of a control. Survival time was significantly longer in a TNP-treated group. TNP-470 demonstrated no growth inhibition of MKN45 cells in vitro. CONCLUSIONS: Those results suggested that anti-angiogenic agent, TNP-470, might be effective in treating peritoneal dissemination of gastric cancer by inhibiting growth of the seeded tumor cells on the peritoneum.

Neoadjuvant chemotherapy with a combination of irinotecan and cisplatin in advanced gastric cancer--a case report.

We report a case of advanced gastric carcinoma successfully treated with a combination of irinotecan and cisplatin as neoadjuvant chemotherapy. The patient, a 78-year-old man, had type 2 gastric cancer, which had metastasized to the paraaortic lymph nodes. He was treated with irinotecan, 70 mg on day 1 and day 15, and cisplatin, 80 mg on day 1. The course was repeated every 4 weeks. Two courses of treatment resulted in a marked reduction of both the primary tumor and lymph nodes. Subsequently, the patient underwent curative surgery consisting of total gastrectomy, splenectomy, and D3 lymph node dissection. No surgical complications were observed. On microscopic examination, only a few tumor cells were detected in the granulation tissues of the resected stomach and in the lymph nodes. This would be the first case to demonstrate the effectiveness and the safety of irinotecan and cisplatin used in the neoadjuvant setting for treatment of advanced gastric carcinoma.

Is gastric carcinoma different between Japan and the United States?

BACKGROUND: Analyses of surgical results for gastric carcinoma often lead to the conclusion that gastric carcinoma occurring in Japan is different from that diagnosed in the U.S. METHODS: To elucidate factors that might explain the differences in surgical results between the two countries, the authors compared data from a cancer center and a university hospital in Japan and a specialist cancer hospital in the U.S (Memorial Sloan-Kettering Cancer Center [MSKCC]). RESULTS: The mean age and body mass index were significantly greater in patients in the U.S. The N category appeared to be determined less accurately at MSKCC compared with the Japanese centers. The occurrence of early gastric carcinoma was not confined to Japanese patients because 20% of U.S. patients who underwent surgery were determined to have early stage disease. However, mucosal (in situ) carcinoma was detected rarely, and the proportion of advanced stage disease was greater in the U.S. Lesions in the upper gastric body, including the gastroesophageal junction, occurred in > 50% of cases at MSKCC but in only 20% of cases at the Japanese centers (P < 0.001). D2 lymph node dissection was possible with low morbidity and minimum mortality (31% and 3%, respectively, at MSKCC). The 5-year survival rates, stratified by tumor location and T category, revealed more similar results between Japan and the U.S. than had been reported previously. The marked difference between Japanese and American institutions only was observed for T1 and T2 tumors occurring in the lower gastric body and for T3 tumors occurring in the middle and upper third of the stomach. CONCLUSIONS: Based on the findings of the current study, it would appear that the more favorable outcome noted for gastric carcinoma patients in Japan primarily is explained by the differences in tumor location, a greater frequency of early stage disease, and more accurate staging compared with gastric carcinoma patients in the U.S. Results of gastric carcinoma treatment comparable to those obtained in Japan can be obtained in Western centers.

[Distribution of lymphocytes to tumor tissue and regional lymph nodes in patients with gastric cancer--the effects of oral administration of OK-432].

We attempted to clarify the accumulation of radiolabeled lymphocytes to tumors and regional lymph nodes in patients with gastric carcinoma. The effects of oral administration of OK-432 were also studied. Five patients with gastric cancer and one who underwent endoscopic mucosal resection, were entered in the study. Prior to the tracer study in 3 patients with gastric cancer, 5 KE of OK-432 was administered for 2 days. Peripheral mononuclear cells were separated and labeled with [111] In-tropolone. After the resection of stomach, tumor tissue, normal gastric mucosa, regional lymph nodes, and non-regional lymph nodes were dissected and radioactivity was measured by a gamma-counter. Accumulation of lymphocytes to the tumor tissue and n1 lymph node station was more than two times that in the normal gastric mucosa and ten times that in non-regional lymph nodes. Accumulation of lymphocytes to the n2 station was strongly enhanced by oral administration of OK-432.

Tissue inhibitor of matrix metalloproteinase-1 in the plasma of patients with gastric carcinoma. A possible marker for serosal invasion and metastasis.

BACKGROUND: Expression of the tissue inhibitor of matrix metalloproteinase-1 (TIMP-1) in tumor tissue from patients with gastric carcinoma has been reported to be related to disease progression. However, to the authors' knowledge the clinical significance of plasma TIMP-1 concentrations in these patients has not been clarified. METHODS: Concentrations of TIMP-1 protein were examined by enzyme-linked immunoadsorbent assay in plasma samples from 149 patients who underwent resection of their primary tumors and from 18 patients with nonresected or recurrent disease. RESULTS: In the 149 patients whose primary tumors were resected, plasma TIMP-1 concentration was associated significantly with a variety of pathologic factors including macroscopic type, depth of invasion, lymph node and peritoneal metastases, vessel invasion, pattern of tumor infiltration into surround ing tissue, and disease stage. Plasma TIMP-1 concentration was significantly higher in patients with serosal invasion, lymph node metastasis, peritoneal dissemination, or liver metastasis than in those without these factors. Neither carcinoembryonic antigen (CEA) nor CA 19-9 concentrations appeared to be related to these measures of disease progression. In the 18 patients with nonresected or recurrent disease, TIMP-1, CEA, and CA 19-9 were similarly sensitive in predicting peritoneal, liver, and lymph node metastases. The combination of these three factors was able to detect 73.3% of patients with peritoneal metastasis, 83.3% of patients with liver metastasis, and 88.9% of patients with disease recurrence. CONCLUSIONS: In patients with gastric carcinoma, plasma concentration of TIMP-1 appears to correlate with both serosal invasion and metastasis.