33 Background: We wished to evaluate the incidence and predictors of the use of long-term (2-3 years) vs. shorter-term androgen deprivation therapy (ADT) in radiation-treated men with high-risk prostate cancer. Methods: We identified 302 patients from the Dana-Farber Cancer Institute patient registry diagnosed with high-risk prostate cancer (T3a or PSA > 20 ng/mL or Gleason score 8-10) between 1993 and 2015. We assessed the intended duration of ADT and used multivariable Cox regression to evaluate predictors of receiving shorter-course ADT than recommended by guidelines (< 2 years). Results: The course of ADT intended by physicians increased following the 2009 publication of trials showing the superiority of 2-3 years versus 4-6 months of ADT, with 43.5% intending ≥ 2 years before vs. 61.4% after (p=0.014). Starting in 2010, 49.4% of patients actually received less than 2 years of ADT. The most common reasons for receipt of shorter-course ADT were intolerance of ADT side effects, patient comorbidity/age, the presence of T3a on MRI only as the sole high-risk feature, or participation in a clinical trial. ACE-27 moderate to severe comorbidity (adjusted hazard ratio [AHR]=2.94), Gleason score less than 8 (AHR=5.66), and PSA < 20 ng/mL (AHR=4.19) all predicted receipt of shorter-course ADT (p<0.05 in all cases). Conclusions: In a tertiary-care setting, rates of long-course ADT for high-risk disease have increased since the 2008/2009 trials supporting its use. However, approximately half of patients continued to receive shorter-course ADT, often due to intolerance of side effects, underlying comorbidity, or physician judgment about the aggressiveness of the disease.