5576 Background: In 2006 the National Cancer Institute (NCI) issued a rare clinical alert recommending intraperitoneal chemotherapy (IP) to treat ovarian cancer after GOG-172, a randomized clinical trial comparing IP and intravenous chemotherapy (IV), demonstrated a 16-month survival advantage with IP. The aims of this study were to determine: (1) changes in IP utilization over time, (2) factors associated with adoption of IP therapy, and (3) the impact of IP on survival. Methods: Prospective cohort study of women with stage III, optimally-cytoreduced ovarian cancer treated at National Comprehensive Cancer Network (NCCN) institutions. Trends were evaluated in 823 patients between 2003-2012. In 603 patients, diagnosed after 2006, multivariable logistic regression and propensity-score analyses were used to examine factors associated with receipt of IP therapy and associations between treatment modality, adverse events, and overall survival. Results: IP utilization rose from 0% to 33% between 2003-2006, increased to 50% in 2007-2008, and decreased to 43% during 2009-2012. At treatment initiation 43% of patients received modified IP regimens, 29% received the GOG 172 regimen, and 28% received IP on a clinical trial. In adjusted analyses, younger age, fewer comorbid conditions, and clinical trial enrollment were associated with IP therapy (all P≤0.01). NCCN center was independently associated with receipt of IP; the adjusted proportion of patients receiving IP varied by site between 16% and 71% (P<0.001). After propensity-score weighted adjustment, patients receiving IP were more likely to switch to IV due to toxicity (AOR=2.01, 95% CI=1.30-3.12) compared with IV. Adjusted 5-year overall survival rates were 57% for IP vs. 44% for IV (hazard ratio=0.69, 95% CI=0.50-0.94). Conclusions: The use of IP therapy increased significantly over the past decade at NCCN centers. However, recently only a minority of eligible patients have received it. Despite frequent modifications, IP chemotherapy is associated with significantly improved survival in clinical practice. Future studies should examine whether site-specific variations in IP utilization reflect patients’ informed treatment preferences.