Radiation Prescriptions?

[brendav]

Advertisement - Members don't see this ad

Hi -

Does anyone have very clear explanations for the following physics questions? I've asked around in my physics division and can't seem to get consistent answers.

1. Why do we generally prescribe electrons to the 90% isodose (rather than 100%)?

2. Why do we generally prescribe SBRT plans to lower isodose lines like the 70% line? Is it just because of the sharp dose fall off?

3. What is meant my re-normalizing a plan from, for example, the 100% isodose to the 110% isodose? Do the isodose curves ever change? (I played around with this in Eclipsys and they seem to remain constant)

4. What is meant by prescribing to a prescription point, such as for supraclav breast fields? Is it really prescribing to a point or are you actually checking the isodose lines / coverage and so prescribing to the isodose lines?

5. Is feathering the same as a junction shift?

Thanks,
Brenda

 

[FrostyHammer]

I'd like to add one more question: why do books state that "field borders for ____" are such & such. Shouldn't field borders depend on what your contoured volume is, not a fixed anatomical definition?

 

[Burt Radnolds]

Hi -

Does anyone have very clear explanations for the following physics questions? I've asked around in my physics division and can't seem to get consistent answers.

1. Why do we generally prescribe electrons to the 90% isodose (rather than 100%)?

2. Why do we generally prescribe SBRT plans to lower isodose lines like the 70% line? Is it just because of the sharp dose fall off?

3. What is meant my re-normalizing a plan from, for example, the 100% isodose to the 110% isodose? Do the isodose curves ever change? (I played around with this in Eclipsys and they seem to remain constant)

4. What is meant by prescribing to a prescription point, such as for supraclav breast fields? Is it really prescribing to a point or are you actually checking the isodose lines / coverage and so prescribing to the isodose lines?

5. Is feathering the same as a junction shift?

Thanks,
Brenda

1) One specific reason for this is that many dose prescriptions for skin cancers were originally in orthovoltage energies, which have a higher RBE than electrons. By prescribing to the 90% with electrons, you are heating up parts of the plan to 110%, overcoming the decreased RBE.

However, there are many ways to do this, and you could just prescribe a higher dose, use 3D planning to ensure adequate coverage, whatever. It's kind of a stupid rule of thumb in my opinion.

2) You have the right idea. The goal is to find coverage with an isodose line than provides the steepest dose fall off while covering the target, then prescribe the full dose to that isodose line. This will also heat up the center of the plan. The isodose line chosen is dependent on what radiosurgical system you are using, and other planning parameters as well. Also, not everyone plans SBRT this way.

3) No, the general shape of the isodose curves will never change if you are just re-normalizing to a new isodose line. You are just changing the dose to each line, not the shape.

4) Prescribing to a point is kind of a remnant of old school planning that still serves a purpose in modern treatment planning. The idea is that you use a specific field shape, and prescribe the dose to a specific point, knowing that it will give you a plan that looks the way you want it to look. You absolutely should look over the isodose lines to ensure that coverage, hotspots, junction, etc are appropriate. In most situations these days, I would recommend using contours to ensure you are happy with your coverage and dose to critical structures.

5) Yes.

Frosty Hammer) Again, seen #4. Old school planning was based on bony anatomy or anatomical landmarks, which is the basis of the field borders you are seeing mentioned in books. In many situations, these borders may still give you an acceptable plan (whole pelvis, breast, palliation). However, your general rule of thumb should be to use contours and real 3D planning to ensure that you are getting appropriate coverage.

 

[Cancerdancer]

1) One specific reason for this is that many dose prescriptions for skin cancers were originally in orthovoltage energies, which have a higher RBE than electrons. By prescribing to the 90% with electrons, you are heating up parts of the plan to 110%, overcoming the decreased RBE.

However, there are many ways to do this, and you could just prescribe a higher dose, use 3D planning to ensure adequate coverage, whatever. It's kind of a stupid rule of thumb in my opinion.

Prescribing to 100% and just using a higher dose is poor form that can lead to misses.
To understand the answer to this question you need to look at isodose curves for electrons in tissue. The 100% isodose volume is miniscule, happening basically only at Dmax, and not approaching the lateral edges of your field at all due to lateral dose constriction.
If you prescribe to 100%, your surface is only (depending on the energy) in the range of 80% of prescription dose. For a skin cancer you are therefore underdosing any tissue shallower than Dmax, and any tissue deeper than Dmax. You will also be surprised at how little volume laterally is getting the dose you prescribe.
By prescribing to 90% (some even use 80%) you actually have a volume that you can fit to your target. You can use bolus to bring your surface dose up to this prescription dose (and/or reduce depth of your prescription dose).
It's not a stupid "rule of thumb" it's an important concept...there is a constriction at depth and laterally with electrons and prescribing electrons must take this into account. Use your energy, bolus, and margin laterally to craft your plan carefully depending on what isodose line you will choose. I often prescribe to 85%

 

[Burt Radnolds]

Prescribing to 100% and just using a higher dose is poor form that can lead to misses.
To understand the answer to this question you need to look at isodose curves for electrons in tissue. The 100% isodose volume is miniscule, happening basically only at Dmax, and not approaching the lateral edges of your field at all due to lateral dose constriction.
If you prescribe to 100%, your surface is only (depending on the energy) in the range of 80% of prescription dose. For a skin cancer you are therefore underdosing any tissue shallower than Dmax, and any tissue deeper than Dmax. You will also be surprised at how little volume laterally is getting the dose you prescribe.
By prescribing to 90% (some even use 80%) you actually have a volume that you can fit to your target. You can use bolus to bring your surface dose up to this prescription dose (and/or reduce depth of your prescription dose).
It's not a stupid "rule of thumb" it's an important concept...there is a constriction at depth and laterally with electrons and prescribing electrons must take this into account. Use your energy, bolus, and margin laterally to craft your plan carefully depending on what isodose line you will choose. I often prescribe to 85%

My entire point throughout all my responses was to use 3D planning to ensure that you have an appropriate plan that provides the coverage that you desire. I entirely agree with everything you said. Electron plans are dependent on many things, energy, bolus, field size, field shape, surface contour, etc. I can assure you I understand all of those. HOWEVER, there are many ways to get to the same endpoint, specifically regarding prescription dose and chosen isodose line. Even you yourself just said that sometimes you prescribe to the 85% instead of the hallowed 90%.

You can achieve the exact same plans by tweaking the dose per fraction and the isodose level you are prescribing to. My point is to use your BRAIN and not arbitrary rules of thumb. By the way, look at NCCN guidelines (by no means the be all end all, but was still likely written by an expert in the field) and it specifically states that the reasoning behind prescribing to 90% for electrons is to overcome the relative difference in RBE with orthovoltage.

Also, at no point did I say that you should just prescribe to Dmax by default, which seems to be the crux of your criticism.

 

[Cancerdancer]

Just trying to elucidate these points to Brendav so that she didn't take your answer to mean it's ok to give 60 Gy to 100% and not understand the implications of doing so...to the young resident the danger in doing so might not be clear unless spelled out. They too often depend on 3D planning but then don't understand how to prescribe based on isodose/depth curves when treating a simple skin CA that isn't CT sim'd...

 

[Burt Radnolds]

I actually think its a bit of a tricky question, and that the reasons behind it are multifactorial (ensuring proper coverage, RBE difference, etc), hence our discussion on it. I do agree with the reasoning behind your answer and think that its an important concept for a resident to understand. I also think there is validity to my initial answer, but perhaps it was incomplete, and possibly could be misconstrued by someone without a lot of experience with electrons.

 

[FrostyHammer]

Thanks a lot for the answers everyone! I have a few more since you're good at explaining things:

1) In SRS/SBRT plans, why is the hot spot even an issue if it's in the tumor target? It's 20%, 30%, how does it matter if it's not going to a normal piece of tissue?

2) How come some lung cases are done without 4DCT? If you don't use an ITV for all lung cases, aren't you forced to put a huge margin on everything to make sure everything is covered (and that too, you're not sure how much it moves)? Also, since without 4DCT, you sim lung cases at some random unknown point in the breathing cycle, adding margins on top of that is a blind shot b/c you don't know where in the breathing cycle it was at. Let's say you simmed at the peak of inspiration of the cycle, and then you're adding margins above the tumor (which wouldn't be needed if it was in the peak inspiration) but wouldn't add enough margins caudal to the tumor...

3) Same thing really with breast. I get the point of using flash, but what if you sim and catch the portion in end-inspiration. And then you add MLC blocks to block out the heart, really close to the back of the breast (maybe where the tumor bed is), which means half the treatment time you're not hitting the tumor...

4) Why are fiducials used for the prostate so much? Seems like many other areas could benefit that move more than the prostate...

5) I never understand why it's not a big deal if PA beams that go through the table/couch isn't anything of concern. I know the table is thin and made of a special material, but no one takes into account that it can still have a small interaction with the table too? Same with AP beams and a person's clothes or another 2-3 blankets the therapists might add on that. Are we seriously assuming zero interaction with all those pieces of matter?

6) Heterogeneity correction is difficult, especially with all that streaking artifact of teeth (not sure how that happens in terms of physics) and all that scatter with artificial knees, etc. Where do you even begin with that? Often you'll have air, bone, and artifact all mishmoshed in the same CT slice. And your dose measurements depend greatly on density accuracy!

7) Y'all were mentioning "feathering". What's that?

8) I will never understand how treatment of a solitary (or even two) brain mets involves surgical resection. The brain is being hematogenously seeded with tumors and you're subjecting the pt to surgery to remove a met? Isn't this the point of high-dose (and high BED) SRS, to avoid surgery? Because after surgery you end up doing SRS for the microscopic disease to the tumor bed anyway!

9) Just out of curiosity, how would you treat a eyelid basal cell? Hitting the eye would be bad. Just a very thick bolus?

10) Contouring on MRI fusions is common, and I'm aware that the sim CT is the main thing on which treatment plans are done. But what if the MRI fusion isn't precisely lined up? Many times the patient is slightly rolled, not straight, or whatever. But the attending will still freely contour on the MRI w/o paying much attention to the CT (not that you can see much on it anyway, comparatively). Won't that screw up target definition? Can you even use an MRI that isn't perfectly lined up to the sim CT?

Thanks a lot everyone for the much-needed furthering of my education!!

 

[brendav]

Thanks to everyone above for their help! Really great explanations!

I actually think its a bit of a tricky question, and that the reasons behind it are multifactorial (ensuring proper coverage, RBE difference, etc), hence our discussion on it. I do agree with the reasoning behind your answer and think that its an important concept for a resident to understand. I also think there is validity to my initial answer, but perhaps it was incomplete, and possibly could be misconstrued by someone without a lot of experience with electrons.

 

[Gfunk6]

Thanks a lot for the answers everyone! I have a few more since you're good at explaining things:

1) In SRS/SBRT plans, why is the hot spot even an issue if it's in the tumor target? It's 20%, 30%, how does it matter if it's not going to a normal piece of tissue?

Well, with intracranial SRS if you have a 'super hot-spot' you would worry about increasing the risk of tumor necrosis. Less of an issue in lung SBRT but may be in an issue for SBRT of prostate, pancreas and liver where critical structures are sometimes 'in' the PTV (e.g. prostatic urethra).

2) How come some lung cases are done without 4DCT? If you don't use an ITV for all lung cases, aren't you forced to put a huge margin on everything to make sure everything is covered (and that too, you're not sure how much it moves)? Also, since without 4DCT, you sim lung cases at some random unknown point in the breathing cycle, adding margins on top of that is a blind shot b/c you don't know where in the breathing cycle it was at. Let's say you simmed at the peak of inspiration of the cycle, and then you're adding margins above the tumor (which wouldn't be needed if it was in the peak inspiration) but wouldn't add enough margins caudal to the tumor...

Not all places have CT simulation with 4D technology. You can do a 'pseudo-4D CT scan' by fusing a free breathing scan, an inspiratory scan and an expiratory scan. However, appropriate concerns have been raised that patients sometimes perform inspirations/expirations during these scans that do not correspond with the physiological reality. Another way to do 'pseudo-4D' is to perform fluoroscopic imaging of lung cancer during normal breathing to get an idea of its motion.

3) Same thing really with breast. I get the point of using flash, but what if you sim and catch the portion in end-inspiration. And then you add MLC blocks to block out the heart, really close to the back of the breast (maybe where the tumor bed is), which means half the treatment time you're not hitting the tumor...

There's always a yin-yang between target dose prescription and meeting OAR constraints. Most would favor the former in the case of heart dose to breast, but if the resection cavity is nowhere near the area that you are blocking then you can probably have your cake and eat it too.

4) Why are fiducials used for the prostate so much? Seems like many other areas could benefit that move more than the prostate...

$$$. Its an additional procedure. There are plenty of other ways to perform prostate IGRT.

5) I never understand why it's not a big deal if PA beams that go through the table/couch isn't anything of concern. I know the table is thin and made of a special material, but no one takes into account that it can still have a small interaction with the table too? Same with AP beams and a person's clothes or another 2-3 blankets the therapists might add on that. Are we seriously assuming zero interaction with all those pieces of matter?

The interaction is not zero but is clinically insignificant in most cases.

6) Heterogeneity correction is difficult, especially with all that streaking artifact of teeth (not sure how that happens in terms of physics) and all that scatter with artificial knees, etc. Where do you even begin with that? Often you'll have air, bone, and artifact all mishmoshed in the same CT slice. And your dose measurements depend greatly on density accuracy!

Contours are always a problem in these scenarios. Consider MRI or PET fusion. Also, you can try to fuse with a MV CT scan which has less scatter artifact from metal. Dosimetry does their best, but when you have a patient with prostate cancer and bilateral hip implants you just do the best you can.

7) Y'all were mentioning "feathering". What's that?

Most often used in the context of using multiple posterior fields to radiate the entire spine. There will be overlap between the fields and you want to avoid a hotspot (can cause spinal cord myelopathy) or a coldspot (underdose tumor). Therefore you 'feather' the junctions - which entails moving them every few fractions so that areas of dose overlap or underlap will be spread over the entire volume instead of concentrating in one spot.

8) I will never understand how treatment of a solitary (or even two) brain mets involves surgical resection. The brain is being hematogenously seeded with tumors and you're subjecting the pt to surgery to remove a met? Isn't this the point of high-dose (and high BED) SRS, to avoid surgery? Because after surgery you end up doing SRS for the microscopic disease to the tumor bed anyway!

If a patient is having major neurologic symptoms from brain mets and steroids don't do much, then resection is the only way to immediately decompress. SRS and WBRT will take weeks or months. Therefore, sometimes the clinical scenario dictates urgent decompression.

9) Just out of curiosity, how would you treat a eyelid basal cell? Hitting the eye would be bad. Just a very thick bolus?

Bolus over the eye and a wax-coated lead shield under the eyelid to prevent scatter dose to the lens/globe.

10) Contouring on MRI fusions is common, and I'm aware that the sim CT is the main thing on which treatment plans are done. But what if the MRI fusion isn't precisely lined up? Many times the patient is slightly rolled, not straight, or whatever. But the attending will still freely contour on the MRI w/o paying much attention to the CT (not that you can see much on it anyway, comparatively). Won't that screw up target definition? Can you even use an MRI that isn't perfectly lined up to the sim CT

You have a few options including deformable registration (MRI or CT is 'deformed' by software to better match) and/or just worry about fusion in the area of interest (e.g. GTV). My dosimetrists will alert me when such fusions are highly inaccurate due to differences in patient position and/or anatomy and recommend that I 'interpret with care.'

 

[medgator]

$$$. Its an additional procedure. There are plenty of other ways to perform prostate IGRT.



.'

Not sure what other practical ways there are to do that outside of fiducials or daily cone beam CT.