Abstract
BACKGROUND: Unfavorable intermediate-risk prostate cancer (UIR-PCa) represents a biologically heterogeneous subgroup with a higher risk of recurrence compared with favorable intermediate-risk disease. Contemporary management frequently includes dose-escalated radiotherapy (RT) combined with short-term androgen deprivation therapy (ADT). Whether additional intraprostatic dose escalation using a simultaneous integrated boost (SIB) provides incremental oncologic benefit in this setting remains uncertain. METHODS: We retrospectively analyzed 194 patients with UIR-PCa treated at three institutions between 2010 and 2023. All patients received image-guided intensity-modulated or volumetric modulated arc RT to 78 Gy with short-term ADT. An MRI-guided intraprostatic SIB (up to 86 Gy) was delivered in 77 patients (39.7%) at clinician discretion. Primary endpoints were biochemical recurrence-free survival (bRFS), distant metastasis-free survival (DMFS), and prostate cancer-specific mortality (PCSM). Multivariable Cox and competing-risk regression models were used to assess predictors of outcome. RESULTS: After a median follow-up of 105 months, 8-year bRFS and DMFS rates for the entire cohort were 93.7% and 95.7%, respectively. Addition of SIB was not associated with improved bRFS (93.5% vs 93.6%, p = 0.36), DMFS (94.6% vs 96.7%, p = 0.15), or PCSM. Percent positive biopsy cores ≥ 50% was the only independent predictor of inferior bRFS on multivariable analysis. Treatment-related toxicity was low in both groups, with no significant differences in late grade ≥ 2 gastrointestinal or genitourinary toxicity. CONCLUSIONS: In patients with UIR-PCa uniformly treated with dose-escalated, image-guided RT and short-term ADT, long-term oncologic outcomes were excellent. The addition of an intraprostatic SIB was safe but did not confer measurable improvement in biochemical or distant disease control. These findings support a selective rather than routine use of focal intraprostatic dose escalation in contemporary UIR-PCa management.