Streamlining dosimetry (i.e. Rapid plan, etc....)

Panacea Financial Contract Review
This forum made possible through the generous support of SDN members, donors, and sponsors. Thank you.
D

dieABRdie

Full Member
5+ Year Member
Joined
Aug 25, 2018
Messages
538
Reaction score
1,235
I am determined to streamline as many processes at our clinic as possible. One of those being treatment planning.

We have structure and plan templates already set up. We are implementing auto-contouring for OARs. Does anyone have any suggestions for streamlining the actual dosimetry? The only one I am familiar with is RapidPlan with Varian. But wondering what others have experience with.
Members don't see this ad.
 
You can use standardized beam arrangements/arcs and optimization objectives per common treatment site, (breast/h&n/lung/prostate/etc). that will do a decent first pass and then use multi criteria optimization/trade off exploration in the newer Eclipses TPS version on predefined PTV optimization/overlapped structures to select a great plan without much additional tweaking.

Here's an article about it Effectiveness of Multi-Criteria Optimization-based Trade-Off exploration in combination with RapidPlan for head & neck radiotherapy planning - Radiation Oncology

You can try MCO on OARs but I found that you get more success on just modifying homogeneity and upper objectives on the PTV.
 
EZFluence from Radformation is the best I know of.

(with the disclaimer being it's the only one I have experience with, auto/AI treatment planning lags behind in terms of software solutions given the complexity)
 
Members don't see this ad :)
Auto oar contouring with radformation is pretty good. Im not really a fan of rapid plan. EZ fluence is good too especially for whole breast tangents.
 
When it comes to autocontouring, Limbus AI is excellent.
 
madchemist89 said:
Auto oar contouring with radformation is pretty good. Im not really a fan of rapid plan. EZ fluence is good too especially for whole breast tangents.
Click to expand...

Agree with radformation for OAR or even elective nodal levels in head and neck and breast (especially in larger patients with nice fat planes). Pelvic LN's OK, but not great IMO.

I haven't used any auto-planning software yet.
 
Palex80 said:
When it comes to autocontouring, Limbus AI is excellent.
Click to expand...
Ugh, we just got a temporary license but went clinical in not ’official‘ ways and so it got pulled so we can properly implement Mim. It’s been frustrating. I just want a few extra hours of my week back lol
 
BobbyHeenan said:
Agree with radformation for OAR or even elective nodal levels in head and neck and breast (especially in larger patients with nice fat planes). Pelvic LN's OK, but not great IMO.

I haven't used any auto-planning software yet.
Click to expand...
Totally agree.

I find it fascinating how much AI struggles with pelvic structures yet does generally well with things like H&N.

For Radformation, I've felt like I got better results manually changing the contouring window to "Pelvic". I believe it defaults to "Abdomen"? Haven't looked in awhile.

Could be my imagination, though.
 
FYI - There is a lot of work on the payor side with regards to AI planning.

I envision a day in the not too distant future where you will have to upload your the dicom files/targets into a payor-based AI platform and have them run your "comparison plans" to compare your IMRT vs. 3D or proton vs. IMRT.
 
BobbyHeenan said:
FYI - There is a lot of work on the payor side with regards to AI planning.

I envision a day in the not too distant future where you will have to upload your the dicom files/targets into a payor-based AI platform and have them run your "comparison plans" to compare your IMRT vs. 3D or proton vs. IMRT.
Click to expand...
That will be thoroughly annoying.
 
There was a cross platform "company" out of STL that was trying to do this exact thing for some insurers. Far as I know, it went nowhere.

Payors don't care about quality. They care about cost. Quality? Would be: did you actually do the thing you said you did ie did you commit fraud? Did the patient actually have cancer? (E.g. horrific like Mata case)

Other than that, lo f'n l.
 
I could see ASTRO "preemptively" introducing some sort of quality comparison tool, doing all the heavy lifting in regards to coding and validation, putting out a white paper, and eviCore/UHC/AIM immediately lapping it up to abuse the rest of us with it.

ASTRO is the monk antagonist in The Da Vinci Code:

1670263586807.png
 
Regarding radformation and pelvic nodes, they are created too big, overlap with bowel often, so I either use them as the PTV or trim them with an internal margin of 1 or 2 mm.
 
IonsAreOurFuture said:
Regarding radformation and pelvic nodes, they are created too big, overlap with bowel often, so I either use them as the PTV or trim them with an internal margin of 1 or 2 mm.
Click to expand...
Yeah, what I'll often do is take the AI-generated bowel bag structure, clean it up, then use that to guide my node CTV creation (as a hard stop).

Sometimes I just delete it and do it manually if it's too wonky, though.
 
You must log in or register to reply here.
Top