4513 Background: Although neoadjuvant combined modality therapy (NACMTX) has become a standard of care in the United States, median overall survival (OS) for patients (PTS) with EA has changed little over the past 25 years. Progression free survival (PFS) and OS after NACMTX depend on extent of primary tumor response. New regimens to increase pathologic complete response (pCR) are needed. Based on phase I data, SWOG designed a phase II trial (S0356) to test OXP with PI5FU and EBRT for PTS with EA. Objectives included pCR rate ≥ 25%, acceptable toxicity (TOX), PFS, OS and exploration of molecular parameters relevant to pCR. Methods: Eligibility: clinical stage II/III EA, ≥ 18 years, Zubrod PS ≤ 2, standard hematologic/non-hematologic values, and tumor < 2 cm into the gastric cardia. OXP 85 mg/m2 was given day (d) 1, 15 and 29; PI5FU 180 mg/m2/d was given d 8-d43. EBRT 180 cGy/d started d 8 x 25 fractions, 5 d/week to total dose 4500 cGy. S was planned 2–4 weeks after NACMTX, with a second cycle of OXP/ PI5FU after S. Central pathology review of surgical specimens was mandated. The trial used a 2-stage design; the trial was halted at 45 PTS to review pCR rate; it reopened to full accrual. Results: 98 PTS enrolled between 9/15/04 and 8/18/08. 6 PTS were ineligible; 2 PTS did not receive therapy (TX). 90 PTS, 84 men (93%), median age 61.7 years, were analyzed. 4 deaths (4.5%) were due to protocol TX; 2 due to NACMTX, 2 to S. 43% and 18% of PTS had grade 3/4 toxicity, respectively: 39% GI, 22% flu-like/fatigue, 17% pulmonary, 16% hematologic, 14% mucositis and 3% neurologic. 77 PTS (86%) underwent S. 30 PTS (33%) had pCR. 9 PTS (10%) had in-situ cancer or T1N0M0. <50% received postoperative CTX. Conclusions: OXP, PI5FU and EBRT for PTS with EA has produced the highest pCR rate reported to date for a cooperative group trial. Significant but manageable non-hematologic TOX was observed. S mortality is acceptable. Future trials built on this platform should plan to complete all TX before S. Tumor molecular profiles (analyses in progress) may predict benefit from this treatment. Data on PFS and OS will follow. [Table: see text]