Search

This study aimed to determine the relationship between geometric and dosimetric agreement metrics in head and neck (H&N) cancer radiotherapy plans. A total 287 plans were retrospectively analyzed, comparing auto-contoured and clinically used contours using a Dice similarity coefficient (DSC), surface DSC (sDSC), and Hausdorff distance (HD). Organs-at-risk (OARs) with ≥200 cGy dose differences from the clinical contour in terms of Dmax (D0.01cc) and Dmean were further examined against proximity to the planning target volume (PTV). A secondary set of 91 plans from multiple institutions validated these findings. For 4995 contour pairs across 19 OARs, 90% had a DSC, sDSC, and HD of at least 0.75, 0.86, and less than 7.65 mm, respectively. Dosimetrically, the absolute difference between the two contour sets was <200 cGy for 95% of OARs in terms of Dmax and 96% in terms of Dmean. In total, 97% of OARs exhibiting significant dose differences between the clinically edited contour and auto-contour were within 2.5 cm PTV regardless of geometric agreement. There was an approximately linear trend between geometric agreement and identifying at least 200 cGy dose differences, with higher geometric agreement corresponding to a lower fraction of cases being identified. Analysis of the secondary dataset validated these findings. Geometric indices are approximate indicators of contour quality and identify contours exhibiting significant dosimetric discordance. For a small subset of OARs within 2.5 cm of the PTV, geometric agreement metrics can be misleading in terms of contour quality. [ABSTRACT FROM AUTHOR]

Deep learning methods for image segmentation and contouring are gaining prominence as an automated approach for delineating anatomical structures in medical images during radiation treatment planning. These contours are used to guide radiotherapy treatment planning, so it is important that contouring errors are flagged before they are used for planning. This creates a need for effective quality assurance methods to enable the clinical use of automated contours in radiotherapy. We propose a novel method for contour quality assurance that requires only shape features, making it independent of the platform used to obtain the images. Our method uses a random forest classifier to identify low-quality contours. On a dataset of 312 kidney contours, our method achieved a cross-validated area under the curve of 0.937 in identifying unacceptable contours. We applied our method to an unlabeled validation dataset of 36 kidney contours. We flagged 6 contours which were then reviewed by a cervix contour specialist, who found that 4 of the 6 contours contained errors. We used Shapley values to characterize the specific shape features that contributed to each contour being flagged, providing a starting point for characterizing the source of the contouring error. These promising results suggest our method is feasible for quality assurance of automated radiotherapy contours.