Patients with locally advanced esophageal cancer have a dismal prognosis when treated exclusively by surgery. This fact prompted many investigators to apply neoadjuvant treatment strategies in an effort to improve survival. Results from phase Ⅲ randomized trials are encouraging however, they revealed that only patients with major histopathological response will benefit from treatment. Therefore, predictive molecular markers indicating response or non-response to neoadjuvant treatment would be extremely helpful in selecting patients for current and future treatment protocols. In this paper we review the role of the molecular markers ERCC1 (excision repair cross-complementing 1 gene) and c-erbB-2 (synonym:HER2/neu) in predicting response to radiochemotherapy and outcome for patients with locally advanced resectable esophageal cancers (cT2-4, Nx, M0). The results are promising and it appears that we might expect to unequivocally identify with ERCC1 and c-erbB-2 respectively, approximately up to one third of patients who fulfil the criteria for neoadjuvant treatment for locally advanced esophageal cancer but will not benefit from our treatment protocol. Integration of such markers in the clinical setting might prevent a substantial number of patients from expensive, non-effective and potentially harmful therapies, and could lead to a more individualized type of combined multimodality treatment in the near future.