Learning how to improve planning/dosimetry

[Radonky]

I now realize as an attending, I wish I was better with plan review. We had rockstar dosi so I took this sorely for granted, but now have someone new I work with, and the plans are not meeting where I thought would be possible. I realize I never developed the language/deeper understanding to effectively communicate how to get to a better plan and I want to learn. The new dosimetrist is very good and hard working but not familiar with dose painting, inhomogeniety on SBRT plans, has difficulty pushing the optimizer on key OAR's if it creates insignificant breakup in the 100% prescription isodose line. On an SBRT case did not use arcs appropriately, so liver got roasted but still "met constraints".

When I ask another dosimetrist to take a look at the plan, it improves. So I am bit by bit adding knowledge on how to improve plans but was hoping to gain some tricks on what you guys do.

Is this something that has to be taught on a case by case basis? Are there "go to's" you guys try first? I asked someone in our department to teach me VMAT planning so at least I can get a better understanding beyond what we learned in physics.

---

Members don't see this ad.

 

[TheWallnerus]

I now realize as an attending, I wish I was better with plan review. We had rockstar dosi so I took this sorely for granted, but now have someone new I work with, and the plans are not meeting where I thought would be possible. I realize I never developed the language/deeper understanding to effectively communicate how to get to a better plan and I want to learn. The new dosimetrist is very good and hard working but not familiar with dose painting, inhomogeniety on SBRT plans, has difficulty pushing the optimizer on key OAR's if it creates insignificant breakup in the 100% prescription isodose line. On an SBRT case did not use arcs appropriately, so liver got roasted but still "met constraints".

When I ask another dosimetrist to take a look at the plan, it improves. So I am bit by bit adding knowledge on how to improve plans but was hoping to gain some tricks on what you guys do.

Is this something that has to be taught on a case by case basis? Are there "go to's" you guys try first? I asked someone in our department to teach me VMAT planning so at least I can get a better understanding beyond what we learned in physics.

Just try and sit there with your dosimetrist from A to Z. Every choice they’re making, and there are a lot, ask them “why?” and make sure there’s some logic behind it. If they say “we can’t do that,” make them try. The worse thing that can happen: you botch a plan. (A botched plan inside a computer that never makes it to R&V has zero chance of harm.) Every rad onc, even if you have no dosimetry experience, has latent planning skills. It’s just going to take work, time, and effort to make them un-latent. How do you get to Broadway? Practice, practice, practice.

 

[taserlaser]

When I get plans improved in a way I am not expecting, or something that exceeds what I was hoping for, or any kind of plan success really, I always ask the planner (and sometimes physicist involved) what worked well. Over the years I’ve been able to add to my own personal repertoire of tips and tricks, which has been helpful, especially when working with junior planners and tough situations. I agree with the Wallnerus here - be inquisitive, and practice.

 

[madchemist89]

Youtube has some good videos.

 

[gmsquid]

What tps? Dm me and depending on what I may be able to help.

 

[CaesarRO]

Youtube has some good videos.

Links?

 

[temujim]

I now realize as an attending, I wish I was better with plan review. We had rockstar dosi so I took this sorely for granted, but now have someone new I work with, and the plans are not meeting where I thought would be possible. I realize I never developed the language/deeper understanding to effectively communicate how to get to a better plan and I want to learn. The new dosimetrist is very good and hard working but not familiar with dose painting, inhomogeniety on SBRT plans, has difficulty pushing the optimizer on key OAR's if it creates insignificant breakup in the 100% prescription isodose line. On an SBRT case did not use arcs appropriately, so liver got roasted but still "met constraints".

When I ask another dosimetrist to take a look at the plan, it improves. So I am bit by bit adding knowledge on how to improve plans but was hoping to gain some tricks on what you guys do.

Is this something that has to be taught on a case by case basis? Are there "go to's" you guys try first? I asked someone in our department to teach me VMAT planning so at least I can get a better understanding beyond what we learned in physics.

You will find most dosimetrists don't understand the TPS. Just ask them a few basic questions. Simply reading the TPS user manual will help a lot. But funny enough the physics/dosimetry support hired by TPS developers is of very uneven quality. So Wallnerus is correct, taking some time to learn the ropes and the process and seeing what works.

Elekta/proknow has videos of dosimetrists who won plan challenges explaining how they went about planning. You can also download the datasets from plan challenges and do your own plan and compare to what others were able to achieve.

In plan review one thing I've found to be helpful is looking at the 25% IDL. Easy way to find areas that can be improved in an IMRT plan.

 

[RickyScott]

I really encourage anyone new to this to draw avoidnacw structures .5 to 1 cm away and then try optimize to see what kind of fall off is possible for a given volume and acceptable hot spot. I have some recipes for certain situations

 

[thecarbonionangle]

This is another missed opportunity for ACGME to tighten the requirements and improve the quality of education in this field and indirectly, and not intended , close down the hellpits, of which our field is replete. Many places do not allow even a month of a dosimetry rotation. Mandate 3 months of dosimetry, planning. People can write their own notes and do their own contours. Ridiculous that our field is being destroyed by “leaders”.

 

[Neuronix]

Mandate 3 months of dosimetry, planning. People can write their own notes and do their own contours.

100% agree with this.

If I ran a residency, residents would have at least one month of dosimetry, therapy, and physics year one.

How are you supposed to spend years seeing patients in clinic having little idea how the procedures work that you're recommending for patients?

In most programs you can graduate having never learned how to generate a radiation plan, how to do a QA, or how to set a patient up on the table.

In this way, the entire specialty is failing our trainees in my opinion.

 

[yesmaster]

It seems kind of silly in retrospect that residency programs push more basic science research time, MBA, global health, MPH, medtech incubator. Anything to avoid spending more time in "the basement"/radiation oncology.

 

[temujim]

This is another missed opportunity for ACGME to tighten the requirements and improve the quality of education in this field and indirectly, and not intended , close down the hellpits, of which our field is replete. Many places do not allow even a month of a dosimetry rotation. Mandate 3 months of dosimetry, planning. People can write their own notes and do their own contours. Ridiculous that our field is being destroyed by “leaders”.

It's a travesty that ACGME only requires one full time (PhD level or equivalent) medical physicist per program. If it got bumped it up to 1:1 guessing half of programs would close.

 

[RadRadRad]

I’ve said this before and been mocked. Don’t really care. You 100% can do your own plans. I don’t do them from scratch but if you see something you don’t like, copy plan, hit optimize, change constraints and see what happens . Sometimes you get what you want sometimes dosi did all they good and anatomy just isn’t favorable. Look at beam angles too if not doing vmat
Good luck and have fun. It’s something different to break up the day

 

[Lamount]

I try to give dosimetry full proof planning objectives. I utilize PTVevals and PPTV volumes to engineer hotspots, carve out OARs and control dose falloff. I never give them a catch-22… and, when they cover my internal PPTVs to the specified dose, I can be sure the are pushing as hard as is possible. These planning volumes I create also make plan eval a lot easier. I almost never need a replan

 

[grenz]

I try to give dosimetry full proof planning objectives. I utilize PTVevals and PPTV volumes to engineer hotspots, carve out OARs and control dose falloff. I never give them a catch-22… and, when they cover my internal PPTVs to the specified dose, I can be sure the are pushing as hard as is possible. These planning volumes I create also make plan eval a lot easier. I almost never need a replan

Would love to hear more about this. For spine SBRT, what are the PTVs you give dosimetry and what do you evaluate in the DVH? It’s virtually impossible to meet conventional PTV coverage goals and cord PRV simultaneously with most cases when I contour per the Sahgal trial. I have moved towards cropping the PTV 2mm from cord PRV and just letting them optimize off that.

 

[Lamount]

Would love to hear more about this. For spine SBRT, what are the PTVs you give dosimetry and what do you evaluate in the DVH? It’s virtually impossible to meet conventional PTV coverage goals and cord PRV simultaneously with most cases when I contour per the Sahgal trial. I have moved towards cropping the PTV 2mm from cord PRV and just letting them optimize off that.

I treat mostly lung… so my site specific tricks are focused in the chest. That being said, my general framework is as follows

PTV_rx D95> dose I would give go the PTV if no OAR was at risk… PTV margin is determined by setup uncertainty (obviously)

PTV_rxEval = PTV_rx cropped off OAR by x mm assuming a 4-5% drop off/mm. E.g. if the max cord_PRV dose that you want is 22Gy and the Rx dose is 27 Gy, would crop 4 mm of cord. I will then go through and manually remove any sharp edges/spikes, because RT won’t do that anyway. **this is just an example. I haven’t treated spine in years, so I don’t know if these are reasonable doses or constraints for spine.

PPTV_highdose. This is whatever higher dose I would want to go to some sub volume. This may not be relevant in spine… but I use this a lot in lung. I just make sure to abide by same 4-5%/mm falloff assumptions, as above.

PPTV_lowdose. This will go to the true PTV and is the highest dose I am comfortable going to the entire region of overlap between PTV and OAR. This is important because this will make sure that your dose won’t fall off abruptly right next to your target.

With this, dosimetry will almost always give me a plan that achieves what I want to achieve. To evaluate, I just look at OARs and the D95 of all of my planning volumes… if the D95s are the doses that I want, I know that I am getting as much dose in as I can.

 

[dieABRdie]

I now realize as an attending, I wish I was better with plan review. We had rockstar dosi so I took this sorely for granted, but now have someone new I work with, and the plans are not meeting where I thought would be possible. I realize I never developed the language/deeper understanding to effectively communicate how to get to a better plan and I want to learn. The new dosimetrist is very good and hard working but not familiar with dose painting, inhomogeniety on SBRT plans, has difficulty pushing the optimizer on key OAR's if it creates insignificant breakup in the 100% prescription isodose line. On an SBRT case did not use arcs appropriately, so liver got roasted but still "met constraints".

When I ask another dosimetrist to take a look at the plan, it improves. So I am bit by bit adding knowledge on how to improve plans but was hoping to gain some tricks on what you guys do.

Is this something that has to be taught on a case by case basis? Are there "go to's" you guys try first? I asked someone in our department to teach me VMAT planning so at least I can get a better understanding beyond what we learned in physics.

I feel your pain. I understood enough to know whether a plan could get better, but not really enough to tell them how to fix the problem. Its a really awkward position to be telling a dosimetrist the plan could be improved... but they are telling you "This is as good as I can get it." So do you delay the patient's start date.. and tell them "Um... sorry we don't know what we are doing... ". Or do you sign a plan that you wouldn't give your own family member? If that issue doesn't get addressed you end up creating a lot of acrimony; where the treatment planner thinks the doctor is unreasonable and the doctor thinks the treatment planner is incompetent.

I was very fortunate one of the dosimetrists from our program was willing to spend the evening with me on Zoom whenever I needed help. So you better believe it I know how to do it now. Not once have they told me "This is as good as it gets, doc!" and I didn't proceed to sit down that evening and absolutely eviscerate those constraints. Now the plans are gorgeous... but it was painful.

Maybe the dosimetrist you mentioned might be willing to help? Is the department willing to send them for further training? There are some online resources. Here's one I have utilized;

Planning Tutorials

If you are new to the field, click on the following link: RADIOTHERAPY from START to FINISH 2 & 5 Minute Tutorials Video Series: 2 Minute Tutorials: 3D Lung Parallel Opposed Obliques Planning2 …

dosepedia.com

I haven't done this one but here's another: https://www.mededportal.org/doi/10.15766/mep_2374-8265.9297

One major issue we had is they weren't really using good "opti" structures to push the dose where you want it. You really should generate those up front depending on the site. Before we started doing that.... they were basically chasing dose all over the place.

Good luck

 

[communitydoc13]

Does anyone worry about MU/Gy or limits of segmentation?

This comes up with SBRT lung.

If you don't explicitly draw an internal structure (within the IV for instance) with an objective to make it hot, while constraining your PTV-ITV space, a typical RA optimization will naturally throw some hot spots at the periphery of your ITV. I actually don't know if this is bad, as it seems to me this is the region with the greatest dosimetric uncertainty and concern for underdosing. In a patient with empty lungs, like a bad COPDer, dosimetry at the interface is absolutely bonkers. Often, you will get plans with a Dmax around 117-120, but they will meet all traditional protocolized constraints. Small hotspots may fall out of the ITV at the air interface.

Of course if you force dose to the middle of the ITV, you get a more traditional looking stereotactic plan with a cohesive hotspot and it will appear better in terms of stereotaxy (high dose gradient and hotter at the target). However, the cost is typically a significant increase in segmentation, MU and segmentation across the target.

Is it ever too much? I have a hard time believing the dosimetry at times.

 

[evilbooyaa]

Does anyone worry about MU/Gy or limits of segmentation?

This comes up with SBRT lung.

If you don't explicitly draw an internal structure (within the IV for instance) with an objective to make it hot, while constraining your PTV-ITV space, a typical RA optimization will naturally throw some hot spots at the periphery of your ITV. I actually don't know if this is bad, as it seems to me this is the region with the greatest dosimetric uncertainty and concern for underdosing. In a patient with empty lungs, like a bad COPDer, dosimetry at the interface is absolutely bonkers. Often, you will get plans with a Dmax around 117-120, but they will meet all traditional protocolized constraints. Small hotspots may fall out of the ITV at the air interface.

Of course if you force dose to the middle of the ITV, you get a more traditional looking stereotactic plan with a cohesive hotspot and it will appear better in terms of stereotaxy (high dose gradient and hotter at the target). However, the cost is typically a significant increase in segmentation, MU and segmentation across the target.

Is it ever too much? I have a hard time believing the dosimetry at times.

No. Until someone proves to me a clinical utility of modulation factor (MU/Gy) and WHY it should be limited, I say let it fly until you reach your goals, assuming plan passes QA. I've heard conflicting things whether a higher modulation factor leads to a higher chance of plan failure (or vice-versa) from my physics staff.

 

[communitydoc13]

assuming plan passes QA

My understanding is that our typical QA process provides only the most nominal checks on high gradient plans like SBRT. (you can move your array to get agreement).

It's the compounding uncertainties that bother me. Not that I know the answer. When I look at the beams eye view for a complex shaped, modulated beam lung SBRT plan, I am amazed that we have confidence in dosimetry.

 

[sirspamalot]

Its magic. We can't possibly understand fluence maps. As far as I'm concerned whatever QA physics is doing is alchemy.

 

[grenz]

Does anyone worry about MU/Gy or limits of segmentation?

This comes up with SBRT lung.

If you don't explicitly draw an internal structure (within the IV for instance) with an objective to make it hot, while constraining your PTV-ITV space, a typical RA optimization will naturally throw some hot spots at the periphery of your ITV. I actually don't know if this is bad, as it seems to me this is the region with the greatest dosimetric uncertainty and concern for underdosing. In a patient with empty lungs, like a bad COPDer, dosimetry at the interface is absolutely bonkers. Often, you will get plans with a Dmax around 117-120, but they will meet all traditional protocolized constraints. Small hotspots may fall out of the ITV at the air interface.

Of course if you force dose to the middle of the ITV, you get a more traditional looking stereotactic plan with a cohesive hotspot and it will appear better in terms of stereotaxy (high dose gradient and hotter at the target). However, the cost is typically a significant increase in segmentation, MU and segmentation across the target.

Is it ever too much? I have a hard time believing the dosimetry at times.

Do you try DCA plans for your lung sbrt? It drastically reduces the # of MU, and treatments are typically shorter than VMAT. Unless you’re trying to specifically carve dose out of an OAR, we usually go DCA as the default.

 

[communitydoc13]

Do you try DCA plans for your lung sbrt

I'll bring this up to the team. We have new physics staff with good ideas, but some tension as expectations of complex plans like lung SBRT vary. I like the idea of DCA where applicable.

 

[HolySmokes]

Does anyone worry about MU/Gy or limits of segmentation?

Is it ever too much? I have a hard time believing the dosimetry at times.

I have similar concerns but less so with SBRT since we start with open fields so our MU/cGy doesn't end up terribly high. I have started to have more concerns with this since our TPS upgrade (Pinnacle v16 to v18)... now an anal SIB plan can easily run 1200(!) MU for 180cGy fraction and the treatment time, in vivo accuracy, and integral dose all give me pause when it gets to that high of a ratio.

Anyone have any insight from a non-SBRT perspective?

 

[NotMattSpraker]

I have similar concerns but less so with SBRT since we start with open fields so our MU/cGy doesn't end up terribly high. I have started to have more concerns with this since our TPS upgrade (Pinnacle v16 to v18)... now an anal SIB plan can easily run 1200(!) MU for 180cGy fraction and the treatment time, in vivo accuracy, and integral dose all give me pause when it gets to that high of a ratio.

Anyone have any insight from a non-SBRT perspective?

This is a great question Ive discussed before with my teams. No one has ever been able to tell me why it is bad except by saying things very vague like "leakage" and "neutron contamination".

It does reflect an inefficient and overmodulated plan though. In the Lattice trials the physicists worked hard to keep the ratio close to 4-5, similar to a regular SBRT plan.

The one place I care clinically is for benign treatments such as HO. My clinic loooooves ezFluence and try to use it in all cases. When I saw 3000 MU for an AP/PA 7 Gy fraction, I just asked them to stop and use 1-2 FiF or a wedge. It probably doesnt matter but in lower dose benign but for me its the one place worth the comment.

 

[TheWallnerus]

a typical RA optimization will naturally throw some hot spots at the periphery of your ITV.

If optimizing purely to the target, with no constraints on any OAR, and no upper bound max, the hot spots will be at the center of the target. (If you get too strict on lung avoidance for a target siting within the lung, you can get the hot spots bowing out to target edges.)

I hate saying hot spot here, but I do it for peer pressure sake, but by definition per ICRU hot spots are outside the PTV. What we are talking in these instances is inhomogeneity.

Do you try DCA plans for your lung sbrt

This also is an easy way to get hot spots in the center, and this is essentially rapidarc optimization to target alone as I mentioned above. One can hybridize a plan: do two plans one at full dose RA and full dose DCA, then mix/match the two plans together at different weights if that makes sense. Makes a wonderful quick (graphic) way to fluff up, or fluff down, the inhomogeneities without having to rerun calcs or optimizations. The weights of the different plans can be interactively applied real time in Eclipse.

 

[TheWallnerus]

When I saw 3000 MU for an AP/PA 7 Gy fraction, I just asked them to stop and use 1-2 FiF or a wedge. It probably doesnt matter but in lower dose benign but for me its the one place worth the comment.

If you aren’t looking at the total MUs of a plan, each arc, and each beam, you aren’t being as whole a rad onc as you should be. Especially do “sanity checks” on electron MUs.

(Also tell physics neutron contamination a moot point with 6MV beams, not to do >=10MV IMRT, or both.)

 

[grenz]

I'll bring this up to the team. We have new physics staff with good ideas, but some tension as expectations of complex plans like lung SBRT vary. I like the idea of DCA where applicable.

Encourage them to have an open mind to DCA and give it a few cases to get on a learning curve. My dosi and physics team have come to appreciate its simplicity and robustness as experience has grown.

 

[RadRadRad]

Treatment time for lung sbrt is almost completely dependent on method of respiratory motion management
If you are able to treat free breathing due to limited motion a 9-11 field coplanar plan can be delivered within a minute of time for DCA
Do whatever you like better but notion that you are saving machine time is silly ( and this is something I frequently hear as rationale for arcs )

It’s breath hold or gating that drive machine time …

 

[brushbeamscanning]

In my opinion the more modulated a plan the higher the chance you will cause more heterogeneity and deviation from your theoretical plan due to interplay effects.

An extreme example of this are large PTV volumes on freebreathing treatments using small spot size proton beam pencil beam scanning even if you treat each area 2-3 times per fraction. This also comes into play for tomotherapy units where the target is larger than the field size. All these factors can lead to a much higher chance of delivering an unintended dose cloud especially when things are not static. To be honest it probably all averages out and is not clinically meaningful.

I've become a semi-competent dosimetrist mostly due to wondering how the sausage is made and experimenting when I've gotten push back from dosimetry claiming a plan cannot be further improved.

 

[communitydoc13]

I agree with all of the above. The errors probably average out, even for complicated solutions like very high segmentation or pencil beam. Neutron production is not my concern at 6x, and for most applications, increased scatter dose is also of minimal concern.

From a workflow and uniformity standpoint, I just don't want the team working too hard on complicated, very highly segmented solutions that are not clinically meaningful. This may be a poor bias on my part.

It will be interesting how inputs are set when all of this is done by the computer in entirety. Will these systems incorporate a cost-function that limits segmentation and plan complexity?

Now that we are all pretty much prescribing SBRT to the target periphery, are folks insisting on a minimal hot spot within the GTV to mimic the initial unmodulated SBRT prescriptions that were to target center, or are they treating with solutions that are not that hot in the middle (115-125% hot)?

 

[evilbooyaa]

I agree with all of the above. The errors probably average out, even for complicated solutions like very high segmentation or pencil beam. Neutron production is not my concern at 6x, and for most applications, increased scatter dose is also of minimal concern.

From a workflow and uniformity standpoint, I just don't want the team working too hard on complicated, very highly segmented solutions that are not clinically meaningful. This may be a poor bias on my part.

It will be interesting how inputs are set when all of this is done by the computer in entirety. Will these systems incorporate a cost-function that limits segmentation and plan complexity?

Now that we are all pretty much prescribing SBRT to the target periphery, are folks insisting on a minimal hot spot within the GTV to mimic the initial unmodulated SBRT prescriptions that were to target center, or are they treating with solutions that are not that hot in the middle (115-125% hot)?

I expect an inner margin inside of GTV/ITV to get 120-125% higher than Rx dose. If GK-SRS is Rx to 50% IDL (200% hot spot *AHEM* heterogeneity for wallnerus's benefit) and Linac-SRS/SBRT is 80% IDL, I expect to see at least a 125% max heterogeneity, although I'll allow higher in non-lung cancers. With IMRT/SBRT style planning, it's usually not necessary.

In lung cancers, going much higher given the fact that not each voxel of ITV will be tumor for each part of the breathing cycle is a bit concerning.

 

[RickyScott]

Most of the time there is not much difference between sbrt lung w/ da vs imrt/vmay. Generally absolute best fall off comes w/hot spots around 40%. Prior to vmat era, found same thing when varying the block edge (including trying negative block edge).

 

[temujim]

If you aren’t looking at the total MUs of a plan, each arc, and each beam, you aren’t being as whole a rad onc as you should be. Especially do “sanity checks” on electron MUs.

(Also tell physics neutron contamination a moot point with 6MV beams, not to do >=10MV IMRT, or both.)

Neutrons aren't a big deal. Even at 10X+. Just ask the proton centers!

 

[TheWallnerus]

Neutrons aren't a big deal. Even at 10X+. Just ask the proton centers!

Ha!

Before there was pencil beam scanning, the only way to do IMPT was to insert a big brass dot decimal style milled block in the path of the proton beam

After the patient was irradiated, there was a necessary time delay of several minutes before staff/therapists would go back in the room to “un-table” the patient… because all that stuff was radioactive

The treatment blocks were radioactive for a long time in general, had to be taken to some vault storage shielded area for… months?… before they were safe to be melted down or disposed of

 

[RickyScott]

Ha!

Before there was pencil beam scanning, the only way to do IMPT was to insert a big brass dot decimal style milled block in the path of the proton beam

After the patient was irradiated, there was a necessary time delay of several minutes before staff/therapists would go back in the room to “un-table” the patient… because all that stuff was radioactive

The treatment blocks were radioactive for a long time in general, had to be taken to some vault storage shielded area for… months?… before they were safe to be melted down or disposed of

how many proton centers are still treating this way?

 

[TheWallnerus]

how many proton centers are still treating this way?

More than zero

 

[hookem19]

Links?

https://m.youtube.com/@dosimetryguide

These are the best videos I have found- they have a great intro video that teaches basic principles for using the optimizer and then specific disease site videos.

 

[Radiator20]

In lung cancers, going much higher given the fact that not each voxel of ITV will be tumor for each part of the breathing cycle is a bit concerning.

My approach to this concern is to have dosi also create the minimum intensity projection from the 4D. I delineate the tumor extent on the minIP and use this as the intentional _hotspot volume. Every voxel where tumor appears on minIP is going to be tumor at all points throughout the respiratory cycle, therefore can get as hot as reasonably achievable. For safety, I also double check that this volume is contained within the tumor on the free breathe scan (and, if not, I crop the volume accordingly). If it's a large target even on minIP, and you want to be extra conservative, you could always contract by a couple mm for setup uncertainty, and certainly away from any nearby OARs plus a PRV margin. But I find this is a nice way to put the extra dose in the place it's likely to do the most good.

 

[Radonky]

I expect an inner margin inside of GTV/ITV to get 120-125% higher than Rx dose. If GK-SRS is Rx to 50% IDL (200% hot spot *AHEM* heterogeneity for wallnerus's benefit) and Linac-SRS/SBRT is 80% IDL, I expect to see at least a 125% max heterogeneity, although I'll allow higher in non-lung cancers. With IMRT/SBRT style planning, it's usually not necessary.

In lung cancers, going much higher given the fact that not each voxel of ITV will be tumor for each part of the breathing cycle is a bit concerning.

how are you instructing dosi to get the hotspot? With GK you prescribe to the 50% IDL, with VMAT dosi is telling me they don't plan like that. They are creating a structure within the GTV to push dose into. If I don't tell them I want a hotspot, its usually not there.

 

[TheWallnerus]

how are you instructing dosi to get the hotspot? With GK you prescribe to the 50% IDL, with VMAT dosi is telling me they don't plan like that. They are creating a structure within the GTV to push dose into. If I don't tell them I want a hotspot, its usually not there.

This is one it’s helpful to have your own planning experience I could write several paragraphs of tips and tricks but, ultimately, you are correct. As a twist, they could try planning *without* VMAT. God never decreed that all SBRT had to be VMAT… or even IMRT.

 

[temujim]

how are you instructing dosi to get the hotspot? With GK you prescribe to the 50% IDL, with VMAT dosi is telling me they don't plan like that. They are creating a structure within the GTV to push dose into. If I don't tell them I want a hotspot, its usually not there.

There are two general ways (I am aware of) to drive a hot spot in an IMRT plan. First is to just to set a max dose constraint at whatever hot spot limit you have (110-130% of Rx eg). This leverages your OAR or other constraints to use advantageous fluences resulting in that hot spot. The second is to pair that high max dose constraint with a structure within your target where you have a minimum/mean/gEUD/whatever to ask for a higher than Rx.

 

[RickyScott]

There are two general ways (I am aware of) to drive a hot spot in an IMRT plan. First is to just to set a max dose constraint at whatever hot spot limit you have (110-130% of Rx eg). This leverages your OAR or other constraints to use advantageous fluences resulting in that hot spot. The second is to pair that high max dose constraint with a structure within your target where you have a minimum/mean/gEUD/whatever to ask for a higher than Rx.

I have always used second way.

 

[TheWallnerus]

There are two general ways (I am aware of) to drive a hot spot in an IMRT plan. First is to just to set a max dose constraint at whatever hot spot limit you have (110-130% of Rx eg). This leverages your OAR or other constraints to use advantageous fluences resulting in that hot spot. The second is to pair that high max dose constraint with a structure within your target where you have a minimum/mean/gEUD/whatever to ask for a higher than Rx.

Or not even set a max dose constraint

 

[evilbooyaa]

how are you instructing dosi to get the hotspot? With GK you prescribe to the 50% IDL, with VMAT dosi is telling me they don't plan like that. They are creating a structure within the GTV to push dose into. If I don't tell them I want a hotspot, its usually not there.

I have a discussion with them regarding peak vs plateau planning and what the goals are. It's about letting the hotspot fly free. 80% IDL is primarily for 3D based plans for LINAC-based SRS that has been some.

You can restrict the hotspot and it's not going to break constraints most of the time, but I can guarantee you if you take the same optimization and remove the hotspot limitation that is used for homogenous IMRT planning, you'll get a more conformal plan....

Hotspot is my personal preference but is not mandatory. The priority is to meet constraints. I just know that most patients can achieve more than the absolute bare minimum from whatever your scorecard says.

 

[TheWallnerus]

In doing VMAT or IMRT SRS/SBRR planning you must impress on your dosimetrists the quote from Tom Hardy in Inception

You mustn’t be afraid to dream a little bigger, darling.

 

[dieABRdie]

Or not even set a max dose constraint

I have a discussion with them regarding peak vs plateau planning and what the goals are. It's about letting the hotspot fly free. 80% IDL is primarily for 3D based plans for LINAC-based SRS that has been some.

You can restrict the hotspot and it's not going to break constraints most of the time, but I can guarantee you if you take the same optimization and remove the hotspot limitation that is used for homogenous IMRT planning, you'll get a more conformal plan....

Hotspot is my personal preference but is not mandatory. The priority is to meet constraints. I just know that most patients can achieve more than the absolute bare minimum from whatever your scorecard says.

It took me forever to get them to stop putting an upper limit on the GTV.

For a while I was intentionally putting concentric PTVs at 100% - 0.5cm = 110% - 0.5cm 120%, etc. in order to get them to do it. Then once the plan was ready; I would delete the inner PTVs from the plan. I found I could do that in the treatment planning system without losing the plan... and it would look like an actual SBRT plan.

So I guess you could trick them if they just can't understand it.

 

[Lamount]

I just do an internal margin on PTV (or PTV_eval) of 8 mm prescribed to 130-140%... gets me a great plan every time