205 Background: Review of treatment plans by a second radiation oncologist is an important quality indicator in radiation oncology. Peer review (PR) can improve quality of care in individual patients by detecting clinical and planning issues and recommending plan changes. This study reports the frequency and nature of these changes across all 14 radiation oncology programs (ROPs) in Ontario, Canada. Methods: We identified all peer-reviewed curative treatment plans delivered in Ontario within a 3-month study period between Dec 2013-Nov 2014 using Cancer Care Ontario’s Activity Level Reporting System, where data on treatment intent and date, disease site treated, PR status, timing of PR, and nature of recommended changes were available. Results: There was considerable variation in the proportion of plans peer-reviewed across ROPs (70.2%, range: 40.8-99.2%). Over the study period, 5,561 curative treatment plans were peer-reviewed and 3.3% had changes recommended. Of those, 21.0% had major clinical and re-planning implications. Recommended changes most often involved minor (63.1%) vs major (36.9%) re-planning implications. Highest proportions of changes were recommended for the treatment of the esophagus, uterus, upper limb, cervix, lower limb, H&N, bilateral lung, right supraclavicular nodes, rectum, and spine (5.0%-7.0%). Plans involving the left breast had slightly more changes recommended (3.0% [95%CI:2.0%-4.5%]) vs right breast (2.4% [95%CI:1.5%-3.8%]). Recommendations were more frequently made when PR was conducted pre-radiotherapy (3.8%) vs during (1.4%-2.8%; p = 0.005), however the nature and implementation of changes were not statistically associated with the timing of PR (p = 0.91; p = 0.23, respectively). Proportion of recommended changes to treatment plans was not statistically associated with ROP patient volume (p = 0.08), proportion of plans peer-reviewed (p = 0.36) or academic status (p = 0.75). Conclusions: Significant variation exists in the proportion of recommended changes across all disease sites and ROPs. PR seems effective in detecting treatment plans with important clinical and planning issues; strategies should be developed to optimize its conduct in radiation oncology.