Anyone here uses Ethos? Halcyon?

[Soapcat]

Truebeam user here - presently happy with the machine in my community practice seeing a variety of cases.
I wanted to ask if anyone here has firsthand experience with Ethos or Halcyon.
Any advantages to these machines as far as efficiency/speed/reimbursement, compared to truebeam?
e.g. How fast do you guys churn out adaptive planning cases (lung, head and neck, etc)? (vs. potentially a lot faster w/ halcyon) Is Halcyon a lot faster than truebeam for prostate cases?

---

Members don't see this ad.

 

[NotMattSpraker]

I have used it but no first hand experience purchasing. It is very fast. Varian has on their website that it can carry 50 patients a day and I have seen it do similar during the pandemic. I have heard people say it can carry the same volume as two Truebeams and Varian kind of implies that on their website.

Adaptive is very slow and the workflow bottleneck is planning and QA, not delivery. If I needed to increase throughput and wanted to start adaptive, it seems like a good choice.

If I wanted to start adaptive but had no throughput issues, Id probably be looking at an MR machine for my new adaptive service.

The reimbursement is the same except for adaptive planning charges.

 

[evilbooyaa]

Ethos/Halcyon is way, WAY faster than TB at VMAT. Only limitations are smaller field size and 6MV-FFF beam energy alone.

Ethos for adaptive is, IMO, suboptimal because of how Varian has made the decision tree to go adaptive one that you make before you look at the IGRT plan (as opposed to MR-Linac where you can start with how it looks and decide whether you 'need' to adapt).

AI auto contours are only available below the diaphragm. Pelvis is generally a fine starting point. Abdomen suboptimal. Adapting things above diaphragm means contouring OARs from scratch.

If your machines are running late because of the VMAT patients, or you have 30-50 patients getting VMAT a day and you want to move the meat faster, Ethos/Halcyon is great.

 

[medgator]

Ethos/Halcyon is way, WAY faster than TB at VMAT. Only limitations are smaller field size and 6MV-FFF beam energy alone.

Ethos for adaptive is, IMO, suboptimal because of how Varian has made the decision tree to go adaptive one that you make before you look at the IGRT plan (as opposed to MR-Linac where you can start with how it looks and decide whether you 'need' to adapt).

AI auto contours are only available below the diaphragm. Pelvis is generally a fine starting point. Abdomen suboptimal. Adapting things above diaphragm means contouring OARs from scratch.

If your machines are running late because of the VMAT patients, or you have 30-50 patients getting VMAT a day and you want to move the meat faster, Ethos/Halcyon is great.

How are sbrt/sabr plans on the halcyon vs TB? No electrons either but less of a use case now unless you're seeing a lot of skin cancer.

Pricing is very competitive compared to a TB, they are truly IMRT/vmat workhorses from what I've seen

 

[deleted1111261]

I really like the concept of it. Seems like it can handle 95% of what walks in through door (except superficial).

Have to do some weird stuff for non-IMRT. I heard you still do VMAT, but through on one static field with very low weight and then bill as 3D

 

[fiji128]

I have a good amount of experience with Halcyon. It's principle benefit is for cases that one would treat with VMAT. So prostate, gyn stuff, lung, head and neck and CNS. Treatment is only with a flatting filter free 6X beam so on beam time tends to be fast. A kv CBCT is required to be done prior to each fraction so as to verify patient position as they are moved into the bore after initial therapist set up. Realistically, you can probably comfortably treat about 4 patient an hour on average. But if you have a ton of prostate patients all lined up so no boards need to be changed out you could probably hit 5 to 6 a hour. Halcyon is decent with SBRT stuff but when planning these keep in mind there are no non coplanar beams available (HA WBRT can't be done on this platform to my knowledge for this reason) so you might not always get the best possible plan. Anything that you may want to treat with higher energy or 3D CRT can be a really pain to deal with as the planning more difficult/challenging. Also, as said above, there are no electrons. Also motion management wasn't available until relatively recently and I haven't seen how well it works. This machine does have the benefit of being relatively cheap at I think about $2 million new (it's made in China and getting parts has at times been an issue for this reason) versus more like $6+ million for a true beam.

 

[medgator]

This machine does have the benefit of being relatively cheap at I think about $2 million new (it's made in China and getting parts has at times been an issue for this reason) versus more like $6+ million for a true beam.

TB isn't anywhere close to that price unless the post pandemic supply chain inflation has done something to things

 

[fiji128]

TB isn't anywhere close to that price unless the post pandemic supply chain inflation has done something to things

It's been about 5 years since I have seen an all in estimate for one so memory could be off but I think with all the software, add ons and technical support service contract that what was being quoted.

 

[elementaryschooleconomics]

Single TrueBeam installed in an existing vault is currently less than $6 million, I would say $4-$5 million is a somewhat generalizable estimate.

But, I'm doing this process right now - there's so many extras and site-specific quirks and incentivized discounts I wouldn't feel super confident about a number...which I'm sure is the point.

Let's go with $3.5-$5.5 million assuming an existing vault and a relatively "base" model.

 

[evilbooyaa]

How are sbrt/sabr plans on the halcyon vs TB? No electrons either but less of a use case now unless you're seeing a lot of skin cancer.

Pricing is very competitive compared to a TB, they are truly IMRT/vmat workhorses from what I've seen

SBRT is fine if you're fine with coplanar delivery only, no 6DoF couch (although this is reportedly in development), and no during treatment visualization exacTrak. At my institution, most SBRT is still done on the TB. We are doing non-spine bone, can do Lung, soft tissue, etc. SBRT. Not routinely doing spine SBRT currently. Prostate SBRT doable but no 6DoF couch makes people less enthusiastic (still doable but even IMO as a halcyon proponent it's suboptimal).

We are not doing SRS on Halcyon as there are no cones and I believe no microMLCs like there is on a STX, Edge, or Tx and no exacTrak.

Additionally, All patients (regardless of treatment modality) gets daily CBCT (obligatory as per machine), no kV x-ray or MV capabilities. So if it's a struggle to get daily CBCT approved then it's gonna be uncompensated work. Fortunately for VMAT patients daily CBCT not an issue. The CBCT is faster than TB as well.

Wouldn't do static IMRT or 3D-CRT on it honestly but workarounds are feasible.

Also the 'traditional' surface monitoring is crap, and the thing they're evaluating to replace it is still in the trial phases so don't feel great about that either although may improve.

 

[Palex80]

We love our Ethos. We mostly do 5-6 adaptive parients in 20-30‘ slots, depending on complexity and run the rest of the day the machine with IGRT cases.

We use the machine mostly for pelvis and palliative stuff (lymph nodes, bone). We have done a few breast cases too (no DIBH). So far, no H&N cases and no lung.

I guess, whether or not this machine is good for you, depends a lot on the size of your practice and your case mix.

 

[evilbooyaa]

We love our Ethos. We mostly do 5-6 adaptive parients in 20-30‘ söots, depending ok complexity and run the rest of the day the machine with IGRT cases.

We use the machine mostly for pelvis and palliative stuff (lymph nodes, bone). We have done a few breast cases too (no DIBH). So far, no H&N cases and no lung.

I guess, whether or not this machine is good for you, depends a lot on the size of your practice and your case mix.

H&N and stage III lung getting not SBRT not getting breath hold or respiratory gating are 2 of the best uses for Ethos... unless the field size is an issue for a stage III lung

Europe strikes again b/c palliative stuff that needs to be 3D in the US would be impossible, and routine breast (unless using VMAT) also suboptimal

 

[Doctorer]

In addition to what's been summarized above, it essentially comes down to being a good second or great third machine, not so great as a single machine. The therapists love the speed though.

The price of the Halcyon when you add in Ethos/bells/whistles/etc has been getting up there. We bought another Truebeam.

 

[Palex80]

H&N and stage III lung getting not SBRT not getting breath hold or respiratory gating are 2 of the best uses for Ethos... unless the field size is an issue for a stage III lung

Our physics does not like Acuros dose calculation in lung tissue. That is the reason why no lung yet.

Europe strikes again b/c palliative stuff that needs to be 3D in the US would be impossible, and routine breast (unless using VMAT) also suboptimal

We do VMAT for breast.

 

[NotMattSpraker]

We love our Ethos. We mostly do 5-6 adaptive parients in 20-30‘ slots, depending on complexity and run the rest of the day the machine with IGRT cases.

We use the machine mostly for pelvis and palliative stuff (lymph nodes, bone). We have done a few breast cases too (no DIBH). So far, no H&N cases and no lung.

I guess, whether or not this machine is good for you, depends a lot on the size of your practice and your case mix.

Do you have adaptives clustered in the morning then? Also, I just have to admit that 20 minutes seems unrealistic for an adaptive treatment. But I've also only been part of one adaptive service line and it probably could have had improved workflows.

 

[Palex80]

Do you have adaptives clustered in the morning then? Also, I just have to admit that 20 minutes seems unrealistic for an adaptive treatment. But I've also only been part of one adaptive service line and it probably could have had improved workflows.

Precisely, all clustered.
We manage a prostate with or without lymphatics within 20 minutes. A lymph node also works well in 20´.
Cervical cancer cases take longer.

We only use IMRT with 7 or 9 beams for adaptive, it saves a few minutes for planning.

A 20‘ slot works out: 2‘ for positioning, 1‘ for CBCT, 6´ for contouring, 4‘ for planning, 1‘ for repeat CBCT, 1‘ to check it, 2‘ for QA, 2‘ beam on time.
And 6‘ for contouring is rather long, some cases are done in half of that time.

A lot depends on your workflow: Do you work with rigid registration of CTV-contours, influencers, contour from scratch?

 

[evilbooyaa]

Do you have adaptives clustered in the morning then? Also, I just have to admit that 20 minutes seems unrealistic for an adaptive treatment. But I've also only been part of one adaptive service line and it probably could have had improved workflows.

20 minutes is doable for prostate only cases or any pelvic case where the patient's anatomy and deformable registration works well on transferring contours

 

[NotMattSpraker]

Precisely, all clustered.
We manage a prostate with or without lymphatics within 20 minutes. A lymph node also works well in 20´.
Cervical cancer cases take longer.

We only use IMRT with 7 or 9 beams for adaptive, it saves a few minutes for planning.

A 20‘ slot works out: 2‘ for positioning, 1‘ for CBCT, 6´ for contouring, 4‘ for planning, 1‘ for repeat CBCT, 1‘ to check it, 2‘ for QA, 2‘ beam on time.
And 6‘ for contouring is rather long, some cases are done in half of that time.

A lot depends on your workflow: Do you work with rigid registration of CTV-contours, influencers, contour from scratch?

Thanks for the detailed description! Totally agree it depends on workflow. You can get a little data from my old institution that is really not related to your report here, but still interesting. Different workflow and it was on Viewray https://www.astro.org/ASTRO/media/ASTRO/Daily Practice/PDFs/COVID-Price-et-al-2(ADRO).pdf

It was mostly contour from scratch (better therapist than MD in a busy clinic) but sometimes influencers. Subjectively I do think Ethos was faster than Viewray, but it wasnt 20 minutes fast for most patients. However, I've never adapted a prostate.

I changed jobs and do not have on table adaptive now, but we are discussing it. Its nice to see your experience to know a range of expectations. I fully believe it will get much faster. VR has taken an interesting approach at overlapping some of those steps you list with people working simultaneously.

I still think if you build a program in 2023 you should not pick a 20 minute adaptive time slot. Id probably pick 40-60 to start.