During the first three years of this grant, the applicant began testing the hypothesis that the outcome for high-risk patients (pts) with local regional gastric cancer can be improved by intensive pre- and postoperative (postop) adjuvant chemotherapy, given both systemically and intraperitoneally (IP). High-risk (T3-T4) pts are identified by endoscopic ultrasound cisplatin-FU postop. The clinical trial has accrued 56 evaluable pts, and is completed. Tolerance to preop therapy has been demonstrated; the curative resection rate of 61 percent in T3/T4 pts compares favorably with that of similar T3/T4 pts seen concurrently who did not receive neoadjuvant therapy (35 percent). Postop chemotherapy has been tolerable, but IP FU-cisplatin has been cumbersome. The preliminary survival distribution curve is encouraging. On the basis of new data available from a random assignment trial in advanced disease, and in preparation for the next national Intergroup trial, the applicant plans to test a new preoperative regimen of cisplatin and fluorouracil followed by operation, followed by intraperitoneal FUdR-leucovorin (a regimen which is less toxic and easier to deliver). As part of this study, a quality of life assessment will be made both for study pts and for non-study pts undergoing surgery alone. This investigational approach will then be proposed as the experimental area of the next national intergroup trial. The applicant has compared EUS predicted tumor stage and thickness with pathologic proven TNM and tumor thickness. In contrast to its accuracy in pts not previously exposed to chemotherapy, EUS appears to be significantly less useful in predicting T and N stage in treated pts. EUS-predicted tumor thickness has only a moderate correlation with pathologically confirmed measurements, suggesting it will, be of limited use in measuring response to chemotherapy. The applicant will now evaluate laparoscopy with laparoscopic guided ultrasound to stage pts and predict outcome, when compared to EUS. Correlative laboratory studies have focused on invasion and metastases and on molecular biological studies in both chemotherapy-treated and untreated pts. Human gastric cancer cell lines have been established and characterized. Mutations of the p53 suppressor gene decreased response to chemotherapy in gastric cancer cell lines. The applicant will now correlate the presence of mutated p53 with clinical outcome in patients receiving pre and postoperative chemotherapy.