Evaluation of a deformable image registration quality assurance tool for head and neck cancer patients

Introduction A challenge in implementing deformable image registration (DIR) in radiation therapy planning is effectively communicating registration accuracy to the radiation oncologist. This study aimed to evaluate the MIM® quality assurance (QA) tool for rating DIR accuracy. Methods Retrospective DIR was performed on CT images for 35 head and neck cancer patients. The QA tool was used to rate DIR accuracy as good, fair or bad. Thirty registered patient images were assessed independently by three RTs and a further five patients assessed by five RTs. Ratings were evaluated by comparison of Hausdorff Distance (HD), Mean Distance to Agreement (MDA), Dice Similarity Coefficients (DSC) and Jacobian determinants for parotid and mandible subregions on the two CTs post‐DIR. Inter‐operator reliability was assessed using Krippendorff's alpha coefficient (KALPA). Rating time and volume measures for each rating were also calculated. Results Quantitative metrics calculated for most anatomical subregions reflected the expected trend by registration accuracy, with good obtaining the most ideal values on average (HD = 7.50 ± 3.18, MDA = 0.64 ± 0.47, DSC = 0.90 ± 0.07, Jacobian = 0.95 ± 0.06). Highest inter‐operator reliability was observed for good ratings and within the parotids (KALPA 0.66–0.93), whilst ratings varied the most in regions of dental artefact. Overall, average rating time was 33 minutes and the least commonly applied rating by volume was fair. Conclusion Results from qualitative and quantitative data, operator rating differences and rating time suggest highlighting only bad regions of DIR accuracy and implementing clinical guidelines and RT training for consistent and efficient use of the QA tool. As deformable image registration (DIR) is becoming increasingly used in clinical practice, this study aimed to evaluate a new QA tool to rate the accuracy of DIR of head and neck cancer patient planning and diagnostic imaging. Regions on deformed images were qualitatively evaluated using good, fair and bad rating levels which were 1) compared to quantitative metrics recommended in the AAPM TG132 Report and 2) compared inter‐operator ratings using Krippendorff's alpha reliability test. Results suggest to only utilise qualitative assessments for only the bad rating level, as well as develop clinical guidelines and training to support the clinical implementation of the QA tool.