Systematic Implementation of Effective Quality Assurance Processes for the Assessment of Radiation Target Volumes in Head and Neck Cancer.

PURPOSE: Significant heterogeneity exists in clinical quality assurance (QA) practices within radiation oncology departments, with most chart rounds lacking prospective peer-reviewed contour evaluation. This has the potential to significantly affect patient outcomes, particularly for head and neck cancers (HNC) given the large variance in target volume delineation. With this understanding, we incorporated a prospective systematic peer contour-review process into our workflow for all patients with HNC. This study aims to assess the effectiveness of implementing prospective peer-review into practice for our National Cancer Institute-Designated Cancer Center and to report factors associated with contour modifications.

METHODS AND MATERIALS: Starting in November 2020, our department adopted a systematic QA process with real-time metrics, in which contours for all patients with HNC treated with RT were prospectively peer-reviewed and graded. Contours were graded with green (unnecessary), yellow (minor), or red (major) colors based on the degree of peer-recommended modifications. Contours from November 2020 through September 2021 were included for analysis.

RESULTS: 360 contours were included. Contour grades were made up of 89.7% green, 8.9% yellow, and 1.4% red grades. Physicians with >12 months of clinical experience were less likely to have contour changes requested than those with <12 months (8.3% vs. 40.9%; p<.001). Contour grades were significantly associated with physician case load, with physicians presenting more than the median number of 50 cases having significantly less modifications requested than those presenting less than 50 (6.7% vs. 13.3%; p=.013). Physicians working with a resident or fellow were less likely to have contour changes requested than those without a trainee (5.2% vs. 12.6%; p=.039). Frequency of major modification requests significantly decreased over time after adoption of prospective peer-contour review, with no red grades occurring >6 months after adoption.

CONCLUSIONS: This study highlights the importance of prospective peer contour-review implementation into systematic clinical QA processes for HNC. Physician experience proved to be the highest predictor of approved contours. A growth curve was demonstrated, with major modifications declining after prospective contour review implementation. Even within a high-volume academic practice with subspecialist attendings, over 10% of patients had contour changes made as a direct result of prospective peer-review.