PURPOSE: Predicting late effects in patients treated with radiation therapy by assessing in vitro radiation-induced CD4 and CD8 T-lymphocyte apoptosis can be useful in individualizing treatment. EXPERIMENTAL DESIGN: In a prospective study, 399 curatively irradiated patients were tested using a rapid assay where fresh blood samples were in vitro irradiated with 8 Gy X-rays. Lymphocytes were collected and prepared for flow cytometric analysis. Apoptosis was assessed by associated condensation of DNA. The incidences of late toxicities were compared for CD4 and CD8 T-lymphocyte apoptoses using receiver-operating characteristic curves and cumulative incidence. RESULTS: No association was found between early toxicity and T-lymphocyte apoptosis. Grade 2 and 3 late toxicities were observed in 31% and 7% of patients, respectively. More radiation-induced T-lymphocyte apoptosis was significantly associated with less grade 2 and 3 late toxicity (Gray's test, P < 0.0001). CD8 (area under the curve = 0.83) was more sensitive and specific than CD4. No grade 3 late toxicity was observed for patients with CD4 and CD8 values greater than 15% and 24%, respectively. The 2-year cumulative incidence for grade 2 or 3 late toxicity was 70%, 32%, and 12% for patients with absolute change in CD8 T-lymphocyte apoptosis of < or =16, 16 to 24, and >24, respectively. CONCLUSIONS: Radiation-induced T-lymphocyte apoptosis can significantly predict differences in late toxicity between individuals. It could be used as a rapid screen for hypersensitive patients to radiotherapy. In future dose escalation studies, patients could be selected using the apoptosis assay.