Session Item

To demonstrate whether a therapeutic radiographer trained dosimetrist is competent to independently contour clinical target volumes (CTVs) for radiotherapy to the prostate +/- seminal vesicles on CT.  

A retrospective contour comparison of therapeutic radiographer outlined prostate and seminal vesicle contours across 55 datasets using CHHiP guidelines for Group 1 and 2 risk cohorts.  Contours were compared in terms of absolute volume, and similarity using DICE coefficient and mean distance to agreement.  This was completed across a purposive sample of low and intermediate risk prostate cancer patients treated with radiotherapy by 26 different clinical oncologists across the group.  Both prostate and seminal vesicle contours were considered independently.

Prostate contours alone were comparable to clinical oncologist contours across the sample with a mean DICE coefficient of 0.84.  Contouring for seminal vesicles was less accurate with a mean DICE coefficient of 0.63.  Mean distance to agreement across prostate contours was 1.45mm.  Mean distance to agreement across seminal vesicle contours was 1.39mm.  Mean absolute volume difference for prostate was 3.45cc.  Mean absolute volume difference for seminal vesicles was 1.75cc.

Prostate contours were generally comparable to the clinically-delivered clinical oncologist volumes, which is in keeping with this being a fairly stable anatomical structure, and that in radiotherapy planning the prostate itself is contoured in its entirety in all cases.  Confounding factors for this included the potential inclusion of MR imaging by the consultant clinical oncologists, and variations in practice amongst the clinical oncologists.

Seminal vesicle contours had less overlap.  Seminal vesicle contouring varies according to clinical risk factors, and although the therapeutic radiographer contoured according to CHHiP criteria as a chosen baseline, the clinical oncologist contours were not necessarily based on the same set criteria but rather their clinical judgement at the time of outlining the patient informed by clinical practice recommendations. This is a limitation of this study, and further work should aim to be prospective to allow a true “like for like” comparison.  This study potentially opens the door for such work.

This study demonstrates the ability of the therapeutic radiographer to accurately delineate the prostate structure itself on CT anatomy, and therefore suggests a start point for further development work to allow autonomous target volume delineation across the cohort, to include accurate and appropriate seminal vesicle contouring.