Is Automated Voluming Accurate
Name Answer
Santam From the kind of body contours that eclipse draws i will say it has a long way to go.. time saver yes accurate no not yet.
Andrew What software have you tried?
Indranil I have very limited experience with this. I have worked with some early in house development on Pinnacle and I think its pretty much sucks for the targets - definitely not bad for some normal structures.
Santam Eclipse / Oncentra / Advantage Sim / Plato - all of them do allow some level automatic segmentation. But the quality of contouring is not good. That said I have read in literature about several approaches to contouring including atlas based segmentation methods which are theoretically showing promise.
Andrew Yes, I agree with the "promise" bit, but a lot of the hype is for the possibility - the lure of the promise - rather than the hopefulness of the early product. Moreover, we have no experience of critiquing volumes with others to give ourselves some hope of being able to critique a piece of software. When our unit does a planning review meeting (each week), I have to continually ask for the GTV and CTV to be turned on so that I can see (never mind critique) the medical decision making drawn on the anatomy. Left alone we look at isodose coverage of PTV. And anyone doing IMRT knows that that is NOT the problem, the problem is where the line is drawn.
Pranshu Just a different thought.. for anyone treating with parallel opposed fields, coverage of PTV is NOT the problem. With IMRT, contrary to what we think, every PTV drawn may not be covered as nicely as we think IMRT could… an example.. Nasopharynx/ Nasal cavities extending to paranasal sinuses and in close proximity to chiasm or optic nerves or brain stem.. no matter what you do with IMRT.. there is some element of PTV dose compromise that needs to be done to maintain normal tissue constraints. So I think that anyone who really knows IMRT knows that coverage of PTV MAY be a problem if normal tissue constraints are to be respected to levels we want. Back to the original point, auto-contours, from the limited exposure I have, can be good when there is an air-tissue or bone-tissue interface.. like Mandible or lung [without any consolidations].. but we have a long way to go before we can confidently auto-contour primary targets.
Andrew 1. Just on a matter of terminology - I keep the term "CONTOUR" for normal tissues, and the term "VOLUME" for target areas.
Andrew 2. the issue of PTV compromise is interesting. What I do is NEVER alter a PTV. If there is overlap between a OAR and a PTV then one or other is sacrificed (I ask the patient what they would prefer). So some of my highly conformal IMRTs have a PTV coverage of only 45Gy because of the juxtaposed optic nerve PRV. I maintain that this is the only way in which we will ever work out whether this compromise is a reasonable thing and just how detrimental it might be. (If you shave the PTV so that coverage is 'perfect', then the failure comes after a perfect plan and we can't learn anything!)
Pranshu Well.. I guess you missed my word "dose" between "PTV" and "compromise". And for that reason, looking at isodose coverage of PTV by itself is not sufficient without supplementing it with information on CTV coverage. In a situation where CTV coverage itself is not being satisfied, then we surely need to have a discussion about "intent of treatment" and decide accordingly. Anyways, that is besides the point raised in this question, so will stop at that.
Unless otherwise stated, the content of this page is licensed under Creative Commons Attribution-ShareAlike 3.0 License