With a goal of identifying relations between gene expression and response (mucosal or pathological) or survival in esophageal cancer patients (stages II to M receiving oxaliplatin, 5-fluorouracill (5FU) and radiation, we measured in endoscopic primary tumor biopsies from 38 patients, the expression of seven genes (gamma GCS, gamma GT, MRP-2, ERCC-1, XPA, TS and DPD) prior to treatment, 1 week following oxaliplatin alone and at the end of the combined radio-chemotherapy cycle using real time QRT-PCR. A higher pretreatment level of XPA was related to shorter survival with a hazard ratio of 2.43 (90% confidence interval 1.09 to 5.43) using Cox regression modeling. However, multivariate analysis with a Cox model indicated low expression of XPA or TS and combined stages II and III had a higher probability of survival (for XPA: hazard ratio 3.0 and 90% C.I. of 1.3 to 6.9, with adjustment for stage included; for TS: hazard ratio is 1.98 with 90% C.I. of 0.94 to 4.20. The expression of TS, gamma GCS, ERCC-1 and MRP-2 declined from D 1 to the end of the cycle (p < 0.05, sign test). A validation and further understanding of the findings need to be carried out in a larger study with a more homogeneous population of patients.

4014 Background: In a study designed to explore the relationship between expression of TS, γGCS, γGT, ERCC-1 and XPA genes and clinical outcome for 38 pts with PEC, who received OXP (85 mg/m2 day (D)1, D15 and D29), PI 5FU (180 mg/m2 D8-D42 ) and XRT (5040 cGy beginning D8 for 28 fractions) TS appeared to be prognostic for survival (Leichman et al., Proc. ASCO, 22:263, 2003). Because recurrence and deaths have continued, an updated analysis for pts followed through October 2003 is presented including expression of DPD and MRP-2 genes. Methods: Pts had endoscopic biopsies of PEC prior to OXP (D1), prior to 5FU and XRT (D8) and at the end of treatment (ET). Gene expression was assayed typically in 2 specimens at each time point in 29 (ET) to 37 (D1) pts by real time QRT-PCR and expressed relative to β-actin. The mean mRNA levels at each time were logarithmically transformed and using Cox proportional hazards model with time-dependent covariates explored for relationships to survival (S) and progression free survival (PFS). Results: Overall median S (n=38) and PFS (n=38) are 26.9 and 12.1 months (mo) respectively at a median follow-up of 31 and 30 mo for the censored pts. Stage 4 (n=22) versus Stages 2/3 (n=16) showed no differences in S (90% C.I. on hazard ratio: 0.9, 3.8) or PFS (90% C.I. on hazard ratio: 0.9, 3.2). Now, a year after our initial analysis, only D1 mean XPA expression showed relation to S in a model that included stage (XPA 90% C.I. on hazard ratio: 1.3, 6.9). XPA as a time-dependent covariate at D1, D8 and ET for 28 pts with stage included also showed a relation to S (90% C.I. on hazard ratio: 1.0, 7.5). None of the biomarkers on D1 or as time-dependent covariates showed a relation to PFS. Conclusions: In successive exploratory analyses of PEC pts treated with OXP, 5FU and XRT, TS and XPA, gene expressions demonstrated variable time dependent relationships to survival. Expression of these genes, and their association with survival for PEC pts undergoing multimodality are appropriate for further larger studies. (Supported by CA 915490 and CA 16056) Author Disclosure Employment or Leadership Consultant or Advisory Stock Ownership Honoraria Research Funding Expert Testimony Other Remuneration Sanofi-Synthalabo

PURPOSE: To identify a dose and schedule of oxaliplatin (OXP) to be safely administered in combination with protracted-infusion (PI) fluorouracil (5-FU) and external-beam radiation therapy (XRT) for patients with primary esophageal carcinoma (EC). PATIENTS AND METHODS: Eligibility included therapeutically naïve EC patients with clinical disease stages II, III, or IV. Initial doses and schedules for cycle 1 consisted of OXP 85 mg/m 2 on days 1, 15, and 29; PI 5-FU 180 mg/m 2 for 24 hours for 35 days; and XRT 1.8 Gy in 28 fractions starting on day 8. At completion of cycle 1, eligible patients could undergo an operation or begin cycle 2 without XRT. Postoperative patients were eligible for cycle 2. Stage IV patients were allowed three cycles in the absence of disease progression. OXP and 5-FU increases were based on dose-limiting toxicity (DLT) encountered in cohorts of three consecutive patients. RESULTS: Thirty-eight eligible patients received therapy: 22 noninvasively staged as IV and 16 noninvasively staged as II and III. Thirty-six patients completed cycle 1, 29 patients started cycle 2, and 24 patients completed cycle 2. The combined-modality therapy was well tolerated, but DLT prevented OXP and 5-FU escalation. No grade 4 hematologic toxicity was noted. Eleven grade 3 and two grade 4 clinical toxicities were noted in eight patients. After cycle 1, 29 patients (81%) had no cancer in the esophageal mucosa. Thirteen patients underwent an operation with intent to resect the esophagus; five patients (38%) exhibited pathologic complete responses. CONCLUSION: OXP 85 mg/m 2 on days 1, 15, and 29 administered with PI 5-FU and XRT is safe, tolerable, and seems effective against primary EC. The role of OXP in multimodality regimens against EC deserves further evaluation.