To be highly successful, a radiotherapeutic dose must be sufficiently large to destroy radioresistant tumors, yet avoid injuring the surrounding healthy tissue. However, many patients exhibit high radiosensitivity and may develop radiation-induced early and late side effects. Because the identification of these radiosensitive patients remains largely problematic, general radiotherapy protocols currently limit the dose given, which risks delivering an insufficient dose to a significant number of less sensitive patients. Therefore, one of the main current challenges of radiobiology is to predict a patient's tumor radioresistance and normal tissue radiosensitivity to tailor a personalized treatment to that individual. Although predictive assays exist, none has demonstrated highly significant results that would be useful in a clinical setting. Therefore, proteomics represents a promising approach for identifying new relevant predictive biomarkers. In this review, the authors first explain the main characteristics of tumor radioresistance and normal tissue radiosensitivity. The authors next describe the existing predictive assays. Finally, the proteomics studies performed to date to identify new biomarkers that probably predicts radiotherapy outcomes are discussed.