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

[dieABRdie]

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.

 

[brushbeamscanning]

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.

 

[elementaryschooleconomics]

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)

 

[madchemist89]

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.

 

[Palex80]

When it comes to autocontouring, Limbus AI is excellent.

 

[BobbyHeenan]

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.

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.

 

[taserlaser]

When it comes to autocontouring, Limbus AI is excellent.

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

 

[elementaryschooleconomics]

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.

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.

 

[BobbyHeenan]

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.

 

[madchemist89]

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.

That will be thoroughly annoying.

 

[Fpg1245]

That will be thoroughly annoying.

Always coming up with new ways to make life even more miserable

 

[BobbyHeenan]

That will be thoroughly annoying.

Your annoyance level is not in their "things we care about" algorithm.

 

[sirspamalot]

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.

 

[elementaryschooleconomics]

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:

 

[IonsAreOurFuture]

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]

Raystation has a feature where you just click a button once to get a VMAT plan. It's really handy to get an IMRT comparison plan quickly for proton appeals.

 

[communitydoc13]

Raystation has a feature where you just click a button once to get a VMAT plan. It's really handy to get an IMRT comparison plan quickly for proton appeals.

I feel like I'm being trolled.

 

[elementaryschooleconomics]

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.

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.

 

[sirspamalot]

I just have a very very good dosimetrist who gets it done quickly and accurately lol.