Can We Detect A Contour Change Easily?
Name Answer
Santam Detecting contour change is a function of two basic processes: 1. The resolution of the underlying imaging and display system and 2. The person doing the contouring. First coming to the resolution it is well known that the human eye is bad at determining the interface in gray scale image. The display system is a case in point too. First some TPS like the eclipse will compress the imaging data and make the image less "clear" as compared to another TPS. For example image clarity is much superior in the advantage Sim workstation. That said the 2nd component for detection is perhaps more important. Training, knowledge and an ability to visualize a tumor in the mind in three dimensions are important. This has been repeatedly highlighted in studies where contouring becomes more "consistent" with training and experience. Adding information from other imaging modalities is of crucial importance as is integration of clinical data. It is important to remember perhaps that of all the methods of seeing the tumour seeing is the best. In short unless the contrast between the interface is good detecting contour changes remains a challenging task
Andrew I was asking something less difficult! Can you detect when the patient's contour has changed? And I think that now the answer is 'yes, easily' thanks to CBCT. However like all 'new' knowledge, the question is what do you do with it? Because for the last century and the first decade of this century we have largely ignored all the smaller subtle changes.
Indranil Surely it is possible to identify most gross contour changes on CBCT. However, its all not black and white. Contour changes are sometimes very difficult to precisely delineate. In some occasions the contour changes may represent radiation related edema rather than a change in tumor size - e. the glottic region and the post ph wall.
Andrew It may be semantics, but the contour of the patient (i.e., the skin contour) is very accurately seen in CBCT and so can be used. The difficulty comes if the internal contours and, more importantly, volumes are not adapted with the contour change.
Santam We haven't got a cone beam CT so have attempted to approach this problem using the KV CT simulator. Yes the external contours are easy to delineate and even the contours of the soft tissue - fat interface are easily delineated. However one thing I have noticed is the question of radiation induced edema. I will be posting some images shortly and they do blur the planes. Andrew have you come across this in your protocol.
Andrew Yes, I have. In the last 3-4 weeks the margins do become less clear. This is particularly a problem around the GTV margins, however the normal anatomical planes are usually still clear. For instance with base of tongue tumours, I am now voluming the CTV at the junction between the intrinsic and genioglossus/hyoid muscles which remains clear if not very sharp throughout the whole treatment time. But yes, weeks 5+ are definitely more difficult.
Pranshu Was not sure what you meant with "voluming CTV at the junction…".. is it that you limit your CTV at that level… irrespective of GTV volume? To add a comment to your original question, about what do we do with the changes that we detect in contours and also what should be the ideal threshold to re-plan? One method employed here at PMH is to fuse CBCT images to planning CT images and then assess changes in isodoses. Obviously, without using deformation and without actually planning on new images, this just gives an rough estimate of the effect of change in contour on isodose distribution. If there is any major concern of overdose or underdose of an important target, then re-plan may be considered. But yes, I believe, it is still early days in terms of evidence as to what and when re-plan needs to be done. Till that time, we need to rely on our clinical acumen. Also, from what I understand, re-plan if required has to be done on a fresh planning CT scan and can not be done on CBCT which is generated by back-projection in a manner different from diagnostic CT.
Unless otherwise stated, the content of this page is licensed under Creative Commons Attribution-ShareAlike 3.0 License