4552 Background: RND and lack of CPR of the primary tumor correlate with poor survival after induction CRT and resection of esophageal cancer. PET response to CRT (SUVmax change and post-induction SUVmax) is used by some clinicians as an indicator of CPR and RND in order to stratify patients after CRT to observation alone vs completion resection. We aimed to investigate the association of PET response with CPR and RND after induction CRT and resection of esophageal cancer. Methods: An IRB-approved retrospective review of an institutional surgical database identified patients who underwent resection of esophageal squamous cell (SCC) and adeno carcinoma (AC) following CRT. The database was locked on Sept 30, 2008. Categorical variables were analyzed by chi square, continuous variables by t-test, and survival by the Kaplan-Meier method. Results: From 1/96 to 3/08, 493 patients were identified, 82% were male. Median age was 62, chemotherapy cisplatin-based in 87%, mean radiation 50 Gy, in-hospital mortality 4.1% and R0 resection rate 88%. Pathology revealed AC in 80%, lack of CPR in 73% and RND in 35%. While in AC patients CPR and lack of RND were both associated with prolonged survival, PET response was not associated with either. In SCC patients, prolonged survival was associated with CPR but not with lack of RND. In SCC, PET response was associated with CPR but not RND. In these patients, reduction in SUVmax <50, 50–75 and >75% was associated with CPR rates of 29, 44 and 85% respectively (p=0.02). Conclusions: These results do not support the use of PET response to justify observation alone after CRT in esophageal AC. With respect to SCC, though exploratory, these provocative results support further study of the use of PET response to predict CPR. [Table: see text] No significant financial relationships to disclose.