Background: Molecular assays are used to improve risk stratification in prostate cancer. This study determined how the 17-gene Genomic Prostate Score (GPS) assay guides treatment decisions in patients with NCCN favorable intermediate-risk prostate cancer. Methods: This retrospective study included 324 patients from 7 urology practices in the United States. All patients had NCCN favorable intermediate-risk prostate cancer, defined as <50% positive biopsy cores and only one of the following risk factors: grade group 2 (Gleason Score 3+4), clinical stage T2b-T2c, or PSA 10-20 ng/mL. All patients also had a GPS assay report dated May 2017 to April 2019. The GPS assay allows risk stratification into post-GPS very low-, low-, favorable intermediate-, unfavorable intermediate-, and high-risk groups, with higher GPS scores representing increased risk. Data were collected from patients' charts/electronic health records. The proportion of patients who selected active surveillance (AS) over immediate definitive treatment and AS, monotherapy, or multimodal therapy was calculated with 95% confidence intervals (CI) using the Clopper-Pearson method. Results: Patients comprised 79% with Gleason Score 3+4, 19% with PSA 10-20 ng/mL, and 2% with clinical stage T2b; median GPS result was 26. Post-GPS risk groups were 0%, 16%, 11%, 57%, and 16% very low-, low-, favorable intermediate-, unfavorable intermediate-, and high-risk groups respectively. Overall, 31% (95% CI 26%, 36%) of patients selected active surveillance (AS), with rate declining as post-GPS risk increased: 57% (95% CI 42%, 71%) selected AS in the post-GPS low-risk, 51% (95% CI 34%, 69%) in the post-GPS favorable intermediate-risk, 26% (95% CI 20%, 33%) in the post-GPS unfavorable intermediate-risk, and 6% (95% CI 1%, 16%) in the post-GPS high-risk group. Post-GPS unfavorable intermediate/high risk-patients were more likely to receive monotherapy than low/favorable intermediate-risk patients (high, 72%; unfavorable intermediate, 64%; favorable intermediate, 37%; and low, 37%); high-risk patients were more likely to receive multimodal therapy (high, 23%; unfavorable intermediate, 10%; favorable intermediate, 11%; and low, 6%). Median follow-up time was 18 months, and one prostate cancer-related metastasis and 2 deaths (neither prostate cancer-related) were reported. Complications (erectile dysfunction, urinary/bowel incontinence) were more common in treated than AS patients. AS persistence was 91% (95% CI 82%, 95%) at 12 months; patients discontinued AS due to disease progression (61%) or patient preference/unknown reasons (39%). Conclusions: The GPS result is associated with selection of AS and treatment intensity in this first examination of genomic classifiers in a cohort of prostate cancer patients with NCCN favorable intermediate-risk.