Patients with resectable gastric and rectal cancer undergoing CT/RT followed by surgery are at increased risk for anemia, fatigue, and need for red cell transfusions. Anemia and tumor hypoxia may contribute to adverse treatment outcome. A multi-center, placebo-controlled trial of 184 patients was designed to determine if Procrit treatment can maintain Hb (≥13 gm/dL), reduce the need for PRBC Tx and improve treatment outcome. Procrit was administered weekly (40,000 units starting dose) for 16 wks during CT/RT (5 FU or Xeloda + 45 Gy over 5 weeks) followed by surgery. After 60 pts were randomized, the trial was terminated due to increased incidence of TEE. 59 of these pts (53 rectal and 6 gastric) received at least one dose of treatment; 31 in the placebo and 28 in the Procrit group. Baseline median Hb for the study group was 13 gm/dL. Seven (11.9%) pts experienced venous TEE; 1/31 (3.2%) pts treated with placebo compared with 6/28 (21.4%) pts treated with Procrit (p= 0.045). In exploratory analyses including the multivariate modeling, it was observed that the Procrit treatment was a predictor of TEE (p = 0.060), however this was not observed for Hb (p = 0.998). Using a mixed-effects model, Hb was significantly higher in pts treated with Procrit than pts treated with placebo (p = 0.0004) during CT/RT. In addition, Hb was significantly different across weeks (p < 0.0001). During the study period, there was also a significant treatment by time interaction (p = 0.0061) with Hb maintained/increased in Procrit pts while it steadily decreased in the placebo pts. Four (14.3%) pts treated with Procrit compared with 10 (32.3%) pts treated with placebo required red cell transfusions (p = 0.133). Platelets (PLTs) decreased significantly from baseline during CT/RT (at weeks 4) for the placebo pts and for the pts treated with Procrit who did not experience TEE (p < 0.0001), but there was a trend for an increase in PLT in pts who experienced TEE. The tumor response for rectal ca at MDACC site was similar between both treatment groups with 14/22 (63.6%) in each treatment group (p = 0.777). These findings indicate that treatment with Procrit may prevent anemia during CT/RT/surgery, however there is increased risk for TEE in this setting for prevention of anemia. Correlative analysis is ongoing on biomarkers and molecular markers on biopsy specimens to examine the markers predictive of pathological complete response, TEE, and the treatment outcome.