Click here to


Are you sure ?

Yes, do it No, cancel

Dosimetric and Geometric Evaluation of Five Commercial Contour Propagation Tools for Online Adaptive Radiotherapy

D Nash1,2*, S Juneja1, A McWilliam2,3, AL Palmer1,4,5, E Vasquez Osorio2,3, (1) Queen Alexandra Hospital, Portsmouth, UK, (2) University of Manchester, Manchester, UK (3) The Christie NHS Foundation Trust, Manchester, UK (4) University of Surrey, Guildford, UK (5) University of Portsmouth, Portsmouth, UK


(Sunday, 7/12/2020)   [Eastern Time (GMT-4)]

Room: AAPM ePoster Library

Purpose: In online adaptive radiotherapy, CBCT contours are required to determine whether a re-plan is indicated. Manual contouring by oncologists is impractical, with automatic propagation of contours from planning CT to CBCT being ideal. The aim of this study was to evaluate whether contours propagated by five commercial algorithms are geometrically similar to those manually drawn by a clinician, with the dosimetric impact of any differences evaluated. This is the first step to using automatic contouring directly in plan adaption.

Methods: Contours for the spinal cord, brainstem and parotids for five H&N patients were propagated to five weekly CBCTs via Pinnacle, ProSoma, RayStation, ADMIRE and Mirada, and independently contoured by an oncologist who reviewed the initial treatment plan. The treatment plan was then recalculated on the shading-corrected CBCTs in RayStation. To assess contour geometrical similarity, the mean distance to agreement (mDTA) was calculated for all propagated contours using the clinician’s contours as reference. To assess dosimetric impact, relevant DVH parameters were extracted: D1cc for the brainstem and spinal cord PRVs, and mean dose for parotids and larynx. These DVH parameters were compared against values derived from the clinician’s contours. Dosimetric differences were tested using Wilcoxon signed-rank tests.

Results: Good geometrical agreement was found for all propagated contours (mean mDTA 2.5±0.9mm), with larger differences for the brainstem (3.1±0.8mm). Good dosimetric agreement was demonstrated for the parotids and brainstem (p>0.2), but not for the cord and the larynx (p<10e-9).

Conclusion: It has been shown that for most organs a good geometrical and dosimetric agreement is reached with any of the evaluated tools. Dosimetric differences, for the larynx and cord, could be explained by proximity to steep dose gradients. Further research is required to determine metrics, considering dose gradients, to assess whether contour adjustments are required for adaptive radiotherapy.


Deformation, Image Fusion, Cone-beam CT


TH- External Beam- Photons: adaptive therapy

Contact Email