BACKGROUND: Local occurrence of prostate cancer (PCa) after external beam radiation (EBRT) may benefit from definitive local therapy. OBJECTIVE: To evaluate the safety and efficacy of salvage high-intenstiy focal ultrasound (HIFU) in local PCa recurrence after EBRT and to determine prognostic factors for optimal patient selection. DESIGN, SETTING, AND PARTICIPANTS: Between 1995 and 2006, patients with a local PCa recurrence after EBRT were retrospectively included. INTERVENTION: All patients received salvage HIFU with the Ablatherm device. MEASUREMENTS: Prognostic factors (pre-EBRT risk group, androgen-deprivation [AD] use, pre-HIFU prostate-specific antigen [PSA], Gleason score and positive biopsy percentage) were studied in univariate and multivariate analyses. Progression was defined as positive biopsy and/or last PSA > nadir+2ng/ml and/or adjuvant therapy introduction. All complications were recorded. RESULTS AND LIMITATIONS: Some 194 HIFU sessions for 167 patients were performed. Local cancer control was achieved with negative biopsy results in 122 (73%) patients. The median PSA nadir was 0.19ng/ml. The mean follow-up period was 18.1 mo (range: 3-121 mo). Seventy-four patients required no hormone therapy. The actuarial 5-yr overall survival rate was 84%. The actuarial 3-yr progression-free survival rate was significantly lower in three circumstances: (1) worsening of the pre-EBRT stage with 53%, 42%, and 25% for low-, intermediate-, and high-risk patients, respectively, (2) increase in the pre-HIFU PSA, and (3) use of AD during PCa management. In multivariate analyses, the risk ratio for intermediate- and high-risk patients were 1.32 and 1.96, respectively. The risk ratio was 2.8 if patients had received AD. No rectal complications were observed. Urinary incontinence accounted for 49.5% of the urinary sphincter implantations required in 11% of patients. This is a retrospective study in which the role of the PSA doubling time and the time until recurrence was not evaluated. CONCLUSIONS: Salvage HIFU is a curative treatment option for local relapse after EBRT with acceptable morbidity. Careful patient selection is imperative depending upon the aforementioned prognostic factors.