The goal of this investigation is to characterize the clinical significance of the rebound interval (RI) after neoadjuvant short-course hormonal therapy (HT) and external-beam radiation therapy (RT), during which the prostate-specific antigen (PSA) may rise because of hormone withdrawal prior to full RT efficacy. The charts of 257 consecutive patients with localized prostate cancer who received short-course neoadjuvant HT and RT were reviewed. A piecewise-linear log PSA versus time curve was generated for each patient and averaged over the population to facilitate identification of the RI start and end dates. Existing definitions of biochemical failure--American Society for Therapeutic Radiology and Oncology (ASTRO), Vancouver and Houston--were applied, as were these same definitions modified to exclude failures during the RI. Sensitivity and specificity were analyzed, using no evidence (by digital rectal examination or radiology) of disease failure as the gold standard. The 5-year biochemical survival with different failure definitions were ASTRO versus ASTRO-modified: 81.6% versus 86.7%; Houston versus Houston-modified: 71.4% versus 76.7%; and Vancouver versus Vancouver-modified: 83.5% versus 85.6%. The sensitivity and specificity comparisons were ASTRO versus ASTRO-modified 58.3% versus 33.3%; 91.4% versus 94.3%, Vancouver versus Vancouver-modified: 50% versus 50%; 92.7% versus 95.5%, Houston versus Houston-modified: 100% versus 66.7%; 90.6% versus 92.2%. The RI after HT and RT is likely not merely an artifact of hormone withdrawal but is correlated with ultimate clinical outcome. Excluding RI failures can marginally improve specificity but may possibly have an unacceptable risk of lowering sensitivity. Further work is needed to design and validate definitions of failure, which account for the RI.