MRI Linac

[RickyScott]

I know this has been addressed, but wanted to hear thoughts on MRI linac. Couldnt much of the benefit can be subsumed by purchase of an MRI Simulator- regular mri with flat table- (1-2$ million) that is readily available for carte blanche MRIs and frequent use of fiducials. The MRI linac makes the claim that you can adaptively avoid the duodenum a bit for pancreatic body tumors, but alternatively you could achieve much of this same benefit with said mri sim and (paying full time dosimetrist to replan quickly based on daily mris in rare cases if you really had to !). Is that really the only benefit to such an expensive machine in a disease site where role of radiation is still controversial? ( 8 million dollar cost, plus very expensive large vault installation and annual service contract)

(This technology brings to mind old saying that If radiation oncologists led the fight on polio, we would have some really cool iron lungs today)

---

Members don't see this ad.

 

[Gfunk6]

I agree. MRI Linacs can be very useful in highly specific clinical scenarios but they are fairly useless as a bread-and-butter machine. Personally, if one already has a versatile, workhorse linac, I would rather get a dedicated ablative platform to complement it than an MRI linac.

Plus you have to work hard to prevent ferromagnetic materials being brought in the vault plus it is loud as hell plus it can't fit in a conventional vault.

 

[evilbooyaa]

UCLA has one and I haven't been overtly impressed by the research coming out of it. Apparently they use it for lung SBRT which to me is a strange utilization pattern for MRI.

 

[OTN]

Worthless, especially considering the $8M price. I have yet to see any data suggesting pancreatic SBRT with an MRI-Linac leads to any better outcomes than that which can be achieved with CBCT-based SBRT, and that's the only clinical scenario I can think where it may be even remotely beneficial.

I know the machine in WashU paid for itself, but that was due to hospital-based billing with replanning on the fly for each fraction, and I can't imagine that would be allowable long-term.

The whole endeavor is one of those "industry figured something out that no one asked for" kind of things.

 

[Mandelin Rain]

Seems like adaptive planning cervical cancer, prostate localization, and intracranial SRS would be obvious uses for MR guidance, but I know nothing about these linacs nor their capabilities.

SBRT lung is a laughable use, however.

 

[Neuronix]

I know the machine in WashU paid for itself, but that was due to hospital-based billing with replanning on the fly for each fraction, and I can't imagine that would be allowable long-term.

The whole endeavor is one of those "industry figured something out that no one asked for" kind of things.

In my opinion, all image guidance technology advances since the MV EPID are either niche academic and limited to a few research institutions OR take off more generally because there's a business model.

CBCT got its own billing code, became profitable, and everyone started using it. Are there indications for which there is no alternative but to use CBCT (i.e. kV or even MV with fiduials)? No. Has CBCT ever been proven in a high quality fashion to improve patient outcomes? Nope. But we all use it because it pays well and you can't argue that it's any worse. More visualization doesn't hurt, and it might be useful.

Same with MRI-linac. More visualization doesn't hurt, and it might be useful. Academic niche sites are demonstrating uses. I'd be surprised if anything ever comes out that absolutely requires MRI guidance for treatment. Pancreas seems the most obvious candidate if anything ever does. Still, people are trying everything we can probably think of at this point. Because frankly, that's one of the benefits of being in academics--trying out the latest toys even if they're not the most profitable. The adaptive re-planning aspect of the device is the only thing that makes the business model even close to work at this point since you can't have a gigantic money losing device sitting around even in most academic departments. There could be some benefit of marketing and "halo effect" with the MRI-linacs as well since they make some cool pictures/videos.

PS: Intracranial SRS is a tough sell since there are no couch kicks.

 

[RickyScott]

Seems like adaptive planning cervical cancer, prostate localization, and intracranial SRS would be obvious uses for MR guidance, but I know nothing about these linacs nor their capabilities.

SBRT lung is a laughable use, however.

But , even if you believe these sites to benefit from mri guidance, couldnt you capture almost all of this benefit with a standalone MRI machine in the department i.e. MRI simulator, you can perform an mri (including 4d/cine mri) right before treatments all day long and replan all the time for a fraction of the cost, buy a pet scanner as well while you are about it. That is really what I am getting at.

40 second pet scan
Human images from the world's first total-body medical scanner unveiled

 

[fiji128]

This is a little off topic but if you follow Viewray VRAY stock price (maker of the MRIdian system and I don't own any rad onc or medical related stocks) their stock price will change pretty dramatically based off how many of these they say they're selling. For the record, 8 new orders worth $49 million in 2018, 6 new orders worth $34 million in 2017 and 4 new orders worth $22 million in 2016. So I guess there are about 18 of these out there at this time. I agree a lot of this tech seems like over kill for terminal diseases like brain mets and unresectable pancreas. But I did see a pretty cool presentation where they were treating an gastric MALT lymphoma with this so as to make sure they were hitting the target with a minimum target volume.

From their website:

Discover MRIdian®
Built upon a patented split-magnet MR design, MRIdian offers a unique unobstructed radiation beam path and optimal source-axis-distance (SAD) to unlock sophisticated beam dosimetry, exceptionally sharp SRS and SBRT-tailored penumbra, and high dose rate beam delivery.

SmartVISION®*
Providing high-definition, diagnostic-quality MR imaging, MRIdian’s SmartVISION was designed specifically to maintain high-fidelity beam delivery, while virtually eliminating the risks of skin toxicities, trapped or distorted dose, and other concerns which may occur when high magnetic fields interact with radiation beams.

SmartADAPT™
With MRIdian’s SmartADAPT, clinicians generate daily MR setup scans in seconds and leverage high-contrast anatomical detail to rapidly reshape dose delivery based on the current position of both the tumor and adjacent critical structures – all while the patient is in the treatment position.

SmartTARGET™
MRIdian’s SmartTARGET visualizes the tumor’s edges and surrounding organ position in real-time using a non-ionizing, streaming video perspective never seen before in radiation oncology. When tumors or organs-at-risk change shape or position, SmartTARGET instantly reacts, automatically controlling beam delivery.

SmartSITE™
Able to fit within almost any existing standard linear accelerator vault, MRIdian’s compact SmartSITE footprint addresses common physical space limitations and avoids the excessive delays, interruptions and costs necessary to remove walls or build custom, large-scale vaults.

Board of Directors

• Daniel Moore, Chairman
• Caley Castelein M.D.

• James F. Dempsey Ph.D.
• Scott Huennekens
• Scott Drake

• Dr. Henry A. McKinnell Jr.
• Aditya Puri
• D. Keith Grossman

• Brian K. Roberts
• Theodore T. Wang, Ph.D.

Scientific Advisory Board
John Bayouth, PhD, University of Wisconsin, Madison
Jeffrey Bradley, MD, Washington University, St. Louis, Missouri
Jürgen Debus, Prof. Dr., NCT Heidelberg, Germany
Carlos A. Perez, MD, Washington University, St. Louis, Missouri
Michael Steinberg, MD, University of California, Los Angeles
Vincenzo Valentini, MD, Universita Cattolica S. Cuore Roma
Hong-Gyun Wu, MD, Seoul National University Hospital, Seoul,South Korea

 

[OTN]

Using such expensive technology to treat gastric MALT seems pretty wasteful to me.

I get what you're saying when it comes to CBCT, but this isn't as big of a leap as "no soft-tissue guidance to CBCT" was. There's a marginal benefit to MRI over CBCT-based imaging in the vast majority of cases, and for the others you should be able to fuse an MRI with a tx planning CT as others have mentioned.

 

[Haybrant]

The support you need for a machine like this is pretty intense. Can’t just pop your standard physics team and hope things will run smoothly.

 

[Neuronix]

Using such expensive technology to treat gastric MALT seems pretty wasteful to me.

I get what you're saying when it comes to CBCT, but this isn't as big of a leap as "no soft-tissue guidance to CBCT" was. There's a marginal benefit to MRI over CBCT-based imaging in the vast majority of cases, and for the others you should be able to fuse an MRI with a tx planning CT as others have mentioned.

CBCT is basically 3D visualization of bones for patient setup. There is good contrast of bone and soft tissue. If the soft tissue target aligns with respect to bone, great. If the soft tissue target doesn't align with the bone, you need fiducials (assuming you can get them). Tumor and other soft tissue have very little difference in density for clear CBCT visualization in most settings.

If there are critical and moveable/deformable OARs in the area (pancreas SBRT), you essentially cross your fingers and even then do a number of complicated steps to try to make it safe which end up still taking an hour on the machine per treatment by the time you're done all your setup and motion management/breath-holds. The logistics of this are very hard for many practices/institutions which has been a major limiting step for pancreas SBRT. Alternatively, I see people out there kind of hacking pancreas and other types of SBRT together in an unsafe way and then patients end up with toxicity. If nothing else, MRI guidance is an SBRT package that takes care of target visualization, OAR avoidance, and motion management.

In any case, with regards to CBCT, you see basically nothing in the CNS on a CBCT and pelvis and abdomen are unreliable and full of artifacts. For lung SBRT it isn't bad in large part because you're looking at air-tumor interface which has a huge difference in density, though motion during treatment remains an issue. Same for prostate and the rectal-prostate interface. Curiously, I've seen practices out there still just aligning pelvic bones because the therapists can't identify the prostate-rectal interface. This is malpractice, but it's happening out there. MRI-guidance is so clear that anyone can setup any site in the body for SBRT with confidence.

This argument about standalone daily MRI or diagnostic CT daily has been going around for 20 years now. It never took off. Why? From a sim standpoint, the patient still has to move from sim to treatment table and there can still be changes in things like bowel and stomach between the MRI and the sim. Remember CT-on-rails? Ok so CT-on-rails helped with this problem, but still never took off. That's because the MRI-RT systems include the whole package. MRI-RT allows imaging during treatment for motion management. The adaptive radiotherapy is built into the workflow and console. This makes it into a package. Yes it's an expensive package, but you're not just paying for an MRI-sim every day. You're paying for rapid adaptive workflow and motion management as well. Trying to do separate MRI with adaptive radiotherapy every day in a standalone fashion is very clunky and time inefficient. The logistics of two different machines have never allowed this to happen in any meaningful way, while now we have 18 or however many ViewRay systems plus the Elekta systems coming that make it work.

 

[scarbrtj]

Think it's a little too soon to be gung-ho or pooh-pooh either way on MRI Linacs. Some say once you've seen it in action, you can't unsee the amount of motion ever again. Of course this is done real time so @RickyScott an MRI sim is not the same thing. Perhaps the majority of failures in lung SBRT are related to motion-caused dose depression at the tumor level? As Rumsfeld said, there are known knowns, known unknowns, and unknown unknowns. No matter how sure we are that an MRI linac is "worthless" or "laughable" today, we may discover unknown unknowns with MRI linacs.

 

[Palex80]

Two vendors currently offer MRI-linacs.

One being Viewray, the other Elekta.

The main difference between the two, as far as I understand, is the quality of the imaging which depends on the strength of the magnetic field. Elektas linac has a 1.5 Tesla MRI, allowing it to deliver imaging which have almost diagnostic quality. The MRidian from Vierway has 0.35 T, which is quite low and thus imaging is not particularly good.

There are sites you can treat, where these machine can be beneficial. But certainly these machines are only good for bigger clinics which have several other linacs and would like to concentrate selected cases on an MRI linac to enhance treatment results. Otherwise the investment simply doesn't pay off, if you are forced to use the MRI linac to treat everything with it.

Possible scenarios from my point of view, when an MRI linac could be beneficial:

lung SBRT - you could use it to turn beam on and off according to breathing, something you cannot do with CBCT and which is certainly more robust than using SGRT. Breath hold techniques can help if you don't have an MRI linac, but not all SBRT-patients are compliant of fit enough to do that.
liver SBRT - you basically see very little with CBCT in liver SBRT (that's my personal experience) other than the liver surface and surrounding bones / chest wall. Even bowel can sometimes become difficult to distinguish sometimes. MRI linacs are certainly superior in terms of what you can actually see and you can use it just like in lung SBRT to turn beam on and off according to breathing.
pancretic/adrenal/kidney SBRT - certainly a new field with less often indications that lung or liver, but yet potentially interesting. Just like in liver I often find it very difficult to actually see something with plain CBCT in these cases. You can skip fiducials if you use an MRI linac.
prostate - for normal prostate EBRT with fiducials I see only little benefit with an MRI linac. However if you want to try out prostate SBRT and MRI linac would help you to minimize intrafraction positioning uncertainties. There are way to combat that with a "normal" linac, for example the Calypso system. But that also means fiducials, which the MRI linac does not need.
cervix/bladder EBRT - those are other potential indications where an MRI linac can help, especially when it comes to adaptive online replanning.

Obviously the MRI linac will be of little help in brain cases, breast cases, any bone tumor/metastasis.

CBCT has come under fire for the "unnecessary" dose it delivers, especially when treating kids or young adults (think of young breast cancer patients getting comprehensive nodal irradiation in breast cancer or mediastinal RT for hodgkins/thymoma). An MRI linac can spare them that dose, certainly the clinical benefit of this is uncertain and we would need decades of follow up to show it.

 

[scarbrtj]

CBCT has come under fire for the "unnecessary" dose it delivers,

Or perhaps it should be lauded for the dose it delivers

Curiously, I've seen practices out there still just aligning pelvic bones because the therapists can't identify the prostate-rectal interface. This is malpractice, but it's happening out there.

FWIW and as an aside, this is 100% not malpractice.


Also... I am old enough to remember cries of "worthless!" when CBCT was first appearing.

 

[medgator]

FWIW and as an aside, this is 100% not malpractice.

I don't think most clinics would consider a kV orthogonal pair to be adequate without fiducials in treating prostate ca, certainly not the same as a daily cbct

 

[Radiator20]

I'd take an MR sim over an MR linac any day. MR linac mostly feels like a solution looking for a problem. I'd say it's appropriate for a research setting... but then I look at the research actually being done on them, and... yeah. Not clear that many people have even figured out how to investigate this technology thoughtfully.

 

[scarbrtj]

I don't think most clinics would consider a kV orthogonal pair to be adequate without fiducials in treating prostate ca, certainly not the same as a daily cbct

Without any daily imaging, PTV margins should be ~1.5cm for prostate CA. A daily setup to bony anatomy in prostate allows for ~7mm PTV margins. A daily setup to soft tissue anatomy allows ~3-5mm margins. Daily setup to fiducials would allow ~2-3mm PTV margins. But since daily setup to bony anatomy is the "biggest bang for the buck" in the entire daily image guidance pantheon, it is certainly not inadequate nor is it malpractice.

 

[medgator]

Without any daily imaging, PTV margins should be ~1.5cm for prostate CA. A daily setup to bony anatomy in prostate allows for ~7mm PTV margins. A daily setup to soft tissue anatomy allows ~3-5mm margins. Daily setup to fiducials would allow ~2-3mm PTV margins. But since daily setup to bony anatomy is the "biggest bang for the buck" in the entire daily image guidance pantheon, it is certainly not inadequate nor is it malpractice.

Yeah but a posterior ptv margin of 1.5 cm might be when a patient gets bad proctitis.

I've actually seen fu patients from old school docs who treated like that. Crazy rates of proctitis long term

 

[Neuronix]

FWIW and as an aside, this is 100% not malpractice.

I should clarify. This depends on dose and margin. I was referring to the doses and margins you'd expect for CBCT/fiducial guided dose escalated prostate RT.

Also... I am old enough to remember cries of "worthless!" when CBCT was first appearing.

This is exactly what I'm referring to. They said the same about CBCT as they're saying about MRI-RT now. I agree that they're quite different technologies. But, everything in science undergoes a lifecycle of...
1. That is impossible.
2. That is worthless.
3. That is obvious and old news.

Right now MRI-RT is in step 2. I remember not long ago when it was step 1. Before the ViewRay, as a physics MD/PhD applying to rad onc residencies talking about how to make MRI-RT devices work I remember being laughed out of several offices...

The main difference between the two, as far as I understand, is the quality of the imaging which depends on the strength of the magnetic field. Elektas linac has a 1.5 Tesla MRI, allowing it to deliver imaging which have almost diagnostic quality. The MRidian from Vierway has 0.35 T, which is quite low and thus imaging is not particularly good.

I try not to take sides in this debate of ViewRay vs. Elekta, especially considering the Elekta device isn't cleared in the USA yet and so we have no experience with it. From a purely imaging standpoint, field strength is only one component of image quality. The images are not 5x better just because the field strength is 5x higher. I'm curious to see how much different the image qualities end up being and how well Elekta can differentiate itself in that regard.

 

[RickyScott]

I dont think MRIi/linac is completely worthless, I just see a very marginal benefit over having mri in department and maybe it should be relegated to a handful of hospitals and cant see possible benefit outside of pancreas. With cone beam, from the onset it was relatively quick and cheap compared to mri, and that is why the ct on rails did not take off, not because of concern over positioning/movement. I see absolutely no need to gait/motion management for stereo lung vs itv in 95% cases.

cone beam will often give you info outside of bony alignment ie you can see bulky lung or head and neck ca shrink and can replan accordingly.

 

[medgator]

I dont think Mri is worthless, I just see a very marginal benefit and maybe it should be relegated to a handful of hospitals .

Sounds like something else we know....

 

[RadOncDoc21]

I rather have a stand alone MRI and an in-house radiologist or at least one who can turn over a study in a day and I’ll be happy.

 

[RickyScott]

Sounds like something else we know....

yes, but this is even more marginal than protons. Plus, If you were going to sink 8 million for machine and construct large bunker, may as well bite the bullet and get financing for small proton unit which you can promote. I dont think mri linac marketing will bring in pts.




"I rather have a stand alone MRI and an in-house radiologist or at least one who can turn over a study in a day and I’ll be happy."-
same here

 

[Palex80]

I think many of you think that the MRI-linac is not better than an MRI-sim or an in-house fastly available MRI, because you may be thinking still in terms of "classic" adaptive RT.
Treat the patient with CBCT --> see differences in CBCT --> take the patient to CT and replan --> irradiate with new plan next day.

This is however not the concept of what MRI linacs are about or what they want to achieve. The technology there is supposed to allow us to quickly (not instantly yet, but quickly) replan while the patient is still on the table and deliver treatment during the same session, based on the individual anatomy at that timepoint.

Additionally MRI linacs are the only machines than can fully visualize what we are treating, the moment we are treating, without using surrogates like fiducials who are being monitored from the outside or the surface of the patient (SGRT).
This is quite a leap in my opinion and when combined with triggered RT to compensate for breathing, this is motion-triggered RT in its pure form (and the way it was supposed to be all along).

 

[RadOncDoc21]

I think many of you think that the MRI-linac is not better than an MRI-sim or an in-house fastly available MRI, because you may be thinking still in terms of "classic" adaptive RT.
Treat the patient with CBCT --> see differences in CBCT --> take the patient to CT and replan --> irradiate with new plan next day.

This is however not the concept of what MRI linacs are about or what they want to achieve. The technology there is supposed to allow us to quickly (not instantly yet, but quickly) replan while the patient is still on the table and deliver treatment during the same session, based on the individual anatomy at that timepoint.

Additionally MRI linacs are the only machines than can fully visualize what we are treating, the moment we are treating, without using surrogates like fiducials who are being monitored from the outside or the surface of the patient (SGRT).
This is quite a leap in my opinion and when combined with triggered RT to compensate for breathing, this is motion-triggered RT in its pure form (and the way it was supposed to be all along).

Maybe as the job market crashes, I will find more time to contour every patient before every single fraction.

 

[Radiator20]

I think many of you think that the MRI-linac is not better than an MRI-sim or an in-house fastly available MRI, because you may be thinking still in terms of "classic" adaptive RT.
Treat the patient with CBCT --> see differences in CBCT --> take the patient to CT and replan --> irradiate with new plan next day.

This is however not the concept of what MRI linacs are about or what they want to achieve. The technology there is supposed to allow us to quickly (not instantly yet, but quickly) replan while the patient is still on the table and deliver treatment during the same session, based on the individual anatomy at that timepoint.

Additionally MRI linacs are the only machines than can fully visualize what we are treating, the moment we are treating, without using surrogates like fiducials who are being monitored from the outside or the surface of the patient (SGRT).
This is quite a leap in my opinion and when combined with triggered RT to compensate for breathing, this is motion-triggered RT in its pure form (and the way it was supposed to be all along).

Palex, I agree that online adaptive planning is an important feature of MR linacs. But I don’t see how that workflow is feasible outside of SBRT. Can’t adaptively replan 30 or 35 fractions.

 

[scarbrtj]

Can’t adaptively replan 30 or 35 fractions

And why not? Are you saying *never can we* or *currently we can't*? Theoretically IMHO you can on-the-fly replan real-time in any fractionation scenario given powerful enough PCs and algorithms. This is ripe for AI of course.

But, everything in science undergoes a lifecycle of...
1. That is impossible.
2. That is worthless.
3. That is obvious and old news.

“A lot of times, people don't know what they want until you show it to them.”
- Steve Jobs

 

[wandering star]

For those who are as interested in this discussion as I am, there's an older thread two years ago that's also worth reading:

MR-guided radiotherapy

I know a very smart guy who, when he first learned about cameras in cell phones 20 years ago, said something like "that's stupid." Cell phone cameras back then were terrible. Today, technology has developed to the point where people leave their fancy cameras at home and just use their smartphone camera. Calling the cell phone camera "stupid" was true back when it could only take 320x240 pixel images, but now they're great.

There are huge technical hurdles in integrating MR with linac, but they're getting better at it. I can see a future in 15 years where many linacs in the US come bundled with an MRI scanner. I can even see a future where the premium MRI linac costs more but it's bought anyway because enough patients feel more comfortable knowing that their tumor is seen every fraction rather than simple alignment to bony anatomy.

However, I also agree that there's not a killer app yet, and currently it's expensive, requires too many compromises, and doesn't make sense except as a research tool. As with the cell phone camera 20 years ago, its time has not yet come.

 

[nkmiami]

On the viewray, is it an imrt plan that is being replanned/adapted, or is it essentially a small 3d field, say bring a leaf or 2 over any bowel that is now nearer to the pancreas?

 

[Neuronix]

It's full IMRT. The plans can look very different if you ask it to.

 

[OTN]

CBCT is basically 3D visualization of bones for patient setup. There is good contrast of bone and soft tissue. If the soft tissue target aligns with respect to bone, great. If the soft tissue target doesn't align with the bone, you need fiducials (assuming you can get them). Tumor and other soft tissue have very little difference in density for clear CBCT visualization in most settings.

If there are critical and moveable/deformable OARs in the area (pancreas SBRT), you essentially cross your fingers and even then do a number of complicated steps to try to make it safe which end up still taking an hour on the machine per treatment by the time you're done all your setup and motion management/breath-holds. The logistics of this are very hard for many practices/institutions which has been a major limiting step for pancreas SBRT. Alternatively, I see people out there kind of hacking pancreas and other types of SBRT together in an unsafe way and then patients end up with toxicity. If nothing else, MRI guidance is an SBRT package that takes care of target visualization, OAR avoidance, and motion management.

In any case, with regards to CBCT, you see basically nothing in the CNS on a CBCT and pelvis and abdomen are unreliable and full of artifacts. For lung SBRT it isn't bad in large part because you're looking at air-tumor interface which has a huge difference in density, though motion during treatment remains an issue. Same for prostate and the rectal-prostate interface. Curiously, I've seen practices out there still just aligning pelvic bones because the therapists can't identify the prostate-rectal interface. This is malpractice, but it's happening out there. MRI-guidance is so clear that anyone can setup any site in the body for SBRT with confidence.

This argument about standalone daily MRI or diagnostic CT daily has been going around for 20 years now. It never took off. Why? From a sim standpoint, the patient still has to move from sim to treatment table and there can still be changes in things like bowel and stomach between the MRI and the sim. Remember CT-on-rails? Ok so CT-on-rails helped with this problem, but still never took off. That's because the MRI-RT systems include the whole package. MRI-RT allows imaging during treatment for motion management. The adaptive radiotherapy is built into the workflow and console. This makes it into a package. Yes it's an expensive package, but you're not just paying for an MRI-sim every day. You're paying for rapid adaptive workflow and motion management as well. Trying to do separate MRI with adaptive radiotherapy every day in a standalone fashion is very clunky and time inefficient. The logistics of two different machines have never allowed this to happen in any meaningful way, while now we have 18 or however many ViewRay systems plus the Elekta systems coming that make it work.

"CBCT is basically 3d visualization of bones for patient setup" - this has not been my experience. I've been doing pancreatic SBRT for some time now, and I feel confident in my ability and my team's ability to identify what we need to identify on CBCT to make sure we're treating accurately. The data we do have re: SBRT for pancreas was collected using CBCT for IGRT, not a MRI-linac, and that data does show good safety and efficacy. I disagree that we're just "crossing our fingers". Our outcomes from both an oncologic and toxicity standpoint have been excellent.

CNS and CBCT? Who cares? Fuse an MRI and get to planning. Zero benefit to an MRI-linac here. Zero benefit in H+N, most sarcomas, all thoracic tumors, all skin cancers, and (I would argue) all gynecologic tumors. Maybe some benefit to liver SBRT, but I would need to see good data showing an improvement in local control before I would get excited about a machine that's more than twice the cost of a standard linac.

Our therapists have no problem finding the prostate-rectal interface, suggesting that it's not CBCT that's the problem, but rather inadequate therapist training.

 

[RickyScott]

I dont see a possible benefit in liver if you place fiducials and use mri fusion. If in ten years, this technology becomes cheaper and smaller, it may become more appealing. Cone beam ct was cheap, quick and accesible from the onset. Calypso never really caught on, but can also be deployed.

I find the linac/pet hybrid more interesting from a research perspective- using pet tracers to target mutltiple tumors.
https://www.reflexion.com/technology/

 

[scarbrtj]

I feel confident in my ability and my team's ability to identify what we need to identify on CBCT to make sure we're treating accurately... Our therapists have no problem finding the prostate-rectal interface, suggesting that it's not CBCT that's the problem, but rather inadequate therapist training.

There is only one way to know this ability. And that is to record across hundreds of setups and hundreds of patients the X, Y, Z shifts recorded. And then look at the mean and SD for all those shifts. If the means are greater than 1mm then there's a relatively big problem, a preparation/systematic problem of some sort (therapist training perhaps). The SD of the data represents the random error. This will arise from a panoply of things, but things different than systematic errors. The data must be recorded pre-shift and post-shift. The post-shift data will be almost purely random error: this error represents the ability of therapists and/or software to match consistently. (It will, not surprisingly, differ from institution to institution.) In reality, everyone should do this 1) at IGRT program inception, per site, and 2) at periodical intervals thereafter. Primarily, the post-shift data determines institutional PTV margins: 2.5*(systematic error)+0.7*(random error) is the PTV margin providing CTV coverage ~95% of the time. Alternatively, one can calc the post-shift (X*X+Y*Y+Z*Z)^0.5 for all shifts and set the PTV at a distance which covers 95% of shifts from 0 millimeters out. Once all this has been done, then a "team's ability" is truly known.

 

[OTN]

I dont see a possible benefit in liver if you place fiducials and use mri fusion. If in ten years, this technology becomes cheaper and smaller, it may become more appealing. Cone beam ct was cheap, quick and accesible from the onset. Calypso never really caught on, but can also be deployed.

I find the linac/pet hybrid more interesting from a research perspective- using pet tracers to target mutltiple tumors.
https://www.reflexion.com/technology/

The RefleXion is cool, I guess, but I still don't see the benefit over PET/CT fusion followed by regular linac-based tx.

 

[nkmiami]

The RefleXion is cool, I guess, but I still don't see the benefit over PET/CT fusion followed by regular linac-based tx.

I am also not sure what the additional benefits over fusion. Saw it it at ASTRO this year. The idea is that the machine detects pet photons from positron annihilation and starts firing back photons at this location (after statistically accounting for positron walk) treating an ITV. Dwight Heron seemed to be indicating that it could treat multiple lesions simultaneously.

 

[OTN]

I am also not sure what the additional benefits over fusion. Saw it it at ASTRO this year. The idea is that the machine detects pet photons from positron annihilation and starts firing back photons at this location (after statistically accounting for positron walk) treating an ITV. Dwight Heron seemed to be indicating that it could treat multiple lesions simultaneously.

The machine will be able to treat oligomets more rapidly than a conventional linac, but I disagree with their statement that it's any different than PET/CT fusion --> regular SBRT. The problem with using PET threshold for targeting is that we really don't know the SUV threshold for malignancy, especially in lung.

 

[Palex80]

CNS and CBCT? Who cares? Fuse an MRI and get to planning. Zero benefit to an MRI-linac here. Zero benefit in H+N, most sarcomas, all thoracic tumors, all skin cancers, and (I would argue) all gynecologic tumors. Maybe some benefit to liver SBRT, but I would need to see good data showing an improvement in local control before I would get excited about a machine that's more than twice the cost of a standard linac.

I can‘t agree.

Lung SBRT and liver SBRT without CBCT means you have to treat an ITV. With an MRI linac you no longer need an ITV. In lung the difference is small interms of toxicity, but I have run into some LLL tumors, where Ihad difficulties getting the full dose in because of dose contraints to the stomach. In liver, it‘s worse in my experience. Some tumors are untreatable with SBRT with aböative doses , when the ITV starts expanding into bowel/stomach and when it comes to setup, liver is a lot trickier than lung. I am talking about fiducials–free treatments.

 

[Palex80]

The machine will be able to treat oligomets more rapidly than a conventional linac, but I disagree with their statement that it's any different than PET/CT fusion --> regular SBRT. The problem with using PET threshold for targeting is that we really don't know the SUV threshold for malignancy, especially in lung.

True.
There may be sone potential for example in head and neck, treating hypoxic areas with a boost. Hypoxic areas can change rapidly within days. But u‘ll need a tracer for that, like F–MISO.

 

[Lamount]

Research applications are certainly interesting. High quality on-board imaging during treatment allows one to study relationship between dose, radiomics, and treatment outcomes... i.e. perhaps patients with or without T2 particular changes during treatment benefit from a boost. We don't know what we don't know

 

[xrt123]

Neuronix,
After a new plan is generated every day does that get billed and covered by insurance?

 

[Neuronix]

I don't know all the details of the billing. But my understanding is yes, you can bill for re-planning a certain number of times.

 

[emt409]

I can‘t agree.

Lung SBRT and liver SBRT without CBCT means you have to treat an ITV. With an MRI linac you no longer need an ITV. In lung the difference is small interms of toxicity, but I have run into some LLL tumors, where Ihad difficulties getting the full dose in because of dose contraints to the stomach. In liver, it‘s worse in my experience. Some tumors are untreatable with SBRT with aböative doses , when the ITV starts expanding into bowel/stomach and when it comes to setup, liver is a lot trickier than lung. I am talking about fiducials–free treatments.

IMHO, if you are a center that can afford MR linac, you should also be at a center where having fiducials placed is not that difficult.

 

[nkmiami]

There are rare situation for lung (more common in liver) where treating with stereo is tricky. I have had one case over my career where the stomach was an issue for lung strereo, but have still been able to use hypofractionation in those liver and lung cases to deliver high bed XRT , sometimes with a breath hold or compression belt.