OpenSkin for neuro procedures - phantom too short?

45 views
Skip to first unread message

benjam...@gmail.com

unread,
Jun 26, 2020, 10:09:07 AM6/26/20
to OpenREM
Hi, 

At Southampton we have just started using OpenREM to capture high skin dose events and have come across an issue for neuro interventions (aneurysm coiling etc). We do a lot of complex neuro intervention here and so regularly need to identify if radiation induced alopecia is likely. These procedures start with imaging in the groin (for catheter insertion) and proceed through the torso, neck, the majority of the dose in the head. The OpenSkin phantom is too short to capture the whole length of the area imaged and the doses in the head therefore 'drop off' resulting in an incorrect peak skin dose (see image below). 

Is this by design (because head is not torso shaped!) and can this be changed? - it would be useful to have some indication of what configuration the overlapping field area is (without having to manually calculate [i.e. guess]). I would imagine for PA direction beams the dose will be pretty accurate too so phantom shape would only really be an issue for lateral beams. 

Thoughts? 

Thanks
Ben

dosemapnohead.png



David Platten

unread,
Jun 26, 2020, 11:27:06 AM6/26/20
to OpenREM
Hi Ben,

There is an open issue in the OpenREM source code to add a "head" to the phantom which may solve your problem. See https://bitbucket.org/openrem/openrem/issues/810/openskin-enhancements-fixes-proposals

Kind regards,

David

COLE, Jonathan (ROYAL FREE LONDON NHS FOUNDATION TRUST)

unread,
Jun 27, 2020, 3:42:14 PM6/27/20
to David Platten, OpenREM

Unfortuantely OpenSkin currently does not have a head included in the phantom. There are a couple of people looking at adding one but I don’t know how far they have got.

 

It’s not a super straightforward job to add a head. The current skin map is effectively a rectangle where each pixel is mapped to a 3d position on a cuboid with two semi cylinders stuck on the sides. To add a head would probably require adding a second 2d map, which then needs all the accompanying interface changes. It’s definitely doable, but I likely do not have time at the moment.

 

David, do you have any thoughts on how difficult amending the display and interface would be if there were _two_ skin maps, a torso and a head?

 

Jonathan

--
You received this message because you are subscribed to the Google Groups "OpenREM" group.
To unsubscribe from this group and stop receiving emails from it, send an email to openrem+u...@googlegroups.com.
To view this discussion on the web, visit https://groups.google.com/d/msgid/openrem/df96822b-add0-4b39-8c02-0b9c0c33e071o%40googlegroups.com.



********************************************************************************************************************

This message may contain confidential information. If you are not the intended recipient please inform the
sender that you have received the message in error before deleting it.
Please do not disclose, copy or distribute information in this e-mail or take any action in relation to its contents. To do so is strictly prohibited and may be unlawful. Thank you for your co-operation.

NHSmail is the secure email and directory service available for all NHS staff in England and Scotland. NHSmail is approved for exchanging patient data and other sensitive information with NHSmail and other accredited email services.

For more information and to find out how you can switch, https://portal.nhs.net/help/joiningnhsmail

David Platten

unread,
Jun 29, 2020, 9:30:58 AM6/29/20
to OpenREM
I havn't looked at my JavaScript code that drives the OpenREM display and interface for quite some time. I think that adding another object for the "head" and wrapping a second skin dose map graphic around it wouldn't be too much work.


On Saturday, 27 June 2020 20:42:14 UTC+1, COLE, Jonathan (ROYAL FREE LONDON NHS FOUNDATION TRUST) wrote:

Unfortuantely OpenSkin currently does not have a head included in the phantom. There are a couple of people looking at adding one but I don’t know how far they have got.

 

It’s not a super straightforward job to add a head. The current skin map is effectively a rectangle where each pixel is mapped to a 3d position on a cuboid with two semi cylinders stuck on the sides. To add a head would probably require adding a second 2d map, which then needs all the accompanying interface changes. It’s definitely doable, but I likely do not have time at the moment.

 

David, do you have any thoughts on how difficult amending the display and interface would be if there were _two_ skin maps, a torso and a head?

 

Jonathan

 

From: ope...@googlegroups.com [mailto:ope...@googlegroups.com] On Behalf Of David Platten
Sent: 26 June 2020 16:27
To: OpenREM
Subject: [openrem] Re: OpenSkin for neuro procedures - phantom too short?

 

Hi Ben,

 

There is an open issue in the OpenREM source code to add a "head" to the phantom which may solve your problem. See https://bitbucket.org/openrem/openrem/issues/810/openskin-enhancements-fixes-proposals

 

Kind regards,

 

David

On Friday, 26 June 2020 15:09:07 UTC+1, benjam...@gmail.com wrote:

Hi, 

 

At Southampton we have just started using OpenREM to capture high skin dose events and have come across an issue for neuro interventions (aneurysm coiling etc). We do a lot of complex neuro intervention here and so regularly need to identify if radiation induced alopecia is likely. These procedures start with imaging in the groin (for catheter insertion) and proceed through the torso, neck, the majority of the dose in the head. The OpenSkin phantom is too short to capture the whole length of the area imaged and the doses in the head therefore 'drop off' resulting in an incorrect peak skin dose (see image below). 

 

Is this by design (because head is not torso shaped!) and can this be changed? - it would be useful to have some indication of what configuration the overlapping field area is (without having to manually calculate [i.e. guess]). I would imagine for PA direction beams the dose will be pretty accurate too so phantom shape would only really be an issue for lateral beams. 

 

Thoughts? 

 

Thanks

Ben

 

dosemapnohead.png

 

 

--
You received this message because you are subscribed to the Google Groups "OpenREM" group.

To unsubscribe from this group and stop receiving emails from it, send an email to ope...@googlegroups.com.

w3ns...@gmail.com

unread,
Jul 2, 2020, 4:14:19 AM7/2/20
to OpenREM
Dear all,

I think the most straight forward solution would be a cylindrical, I plan to fuse this with the current version of openskin in the coming weeks. But if someone else has brigther ideas, I'd love to hear all about it and see it!

Zeego_HFP_Neuro.PNG


Best,
W

Op vrijdag 26 juni 2020 16:09:07 UTC+2 schreef benjam...@gmail.com:

COLE, Jonathan (ROYAL FREE LONDON NHS FOUNDATION TRUST)

unread,
Jul 2, 2020, 7:09:06 AM7/2/20
to w3ns...@gmail.com, OpenREM

That sounds sensible to me.

 

You’ll need to call the skin mapping bit twice I think; once for the torso and once for the head cylinder. As there is existing logic to make the half cylinders for the sides of the torso I hope it won’t be _too_ hard. Need some thought about scaling as the torso currently can scale with height and weight, but heads are a bit different and only really scale with age.

 

I’m really pleased you are working on this! I hope to have a bit more time at some point, but it is crazy at the moment with childcare still mostly closed and trying to provide medical physics services with all the extra restrictions on access, cleaning kit etc.

 

Jon

 

 

From: ope...@googlegroups.com [mailto:ope...@googlegroups.com] On Behalf Of w3ns...@gmail.com
Sent: 02 July 2020 08:45
To: OpenREM
Subject: [openrem] Re: OpenSkin for neuro procedures - phantom too short?

 

Dear all,

--

You received this message because you are subscribed to the Google Groups "OpenREM" group.

To unsubscribe from this group and stop receiving emails from it, send an email to openrem+u...@googlegroups.com.
To view this discussion on the web, visit https://groups.google.com/d/msgid/openrem/1a642916-0653-4693-b33d-414ccbd62afao%40googlegroups.com.

Ben Johnson

unread,
Jul 3, 2020, 4:10:16 PM7/3/20
to w3ns...@gmail.com, OpenREM
This sounds fantastic and just what we need. If we can help in any way please let me know

Thanks
Ben

Sent from my iPhone

On 2 Jul 2020, at 09:14, w3ns...@gmail.com wrote:


Dear all,

I think the most straight forward solution would be a cylindrical, I plan to fuse this with the current version of openskin in the coming weeks. But if someone else has brigther ideas, I'd love to hear all about it and see it!

<Zeego_HFP_Neuro.PNG>


Best,
W

Op vrijdag 26 juni 2020 16:09:07 UTC+2 schreef benjam...@gmail.com:
Hi, 

At Southampton we have just started using OpenREM to capture high skin dose events and have come across an issue for neuro interventions (aneurysm coiling etc). We do a lot of complex neuro intervention here and so regularly need to identify if radiation induced alopecia is likely. These procedures start with imaging in the groin (for catheter insertion) and proceed through the torso, neck, the majority of the dose in the head. The OpenSkin phantom is too short to capture the whole length of the area imaged and the doses in the head therefore 'drop off' resulting in an incorrect peak skin dose (see image below). 

Is this by design (because head is not torso shaped!) and can this be changed? - it would be useful to have some indication of what configuration the overlapping field area is (without having to manually calculate [i.e. guess]). I would imagine for PA direction beams the dose will be pretty accurate too so phantom shape would only really be an issue for lateral beams. 

Thoughts? 

Thanks
Ben

dosemapnohead.png



--
You received this message because you are subscribed to the Google Groups "OpenREM" group.
To unsubscribe from this group and stop receiving emails from it, send an email to openrem+u...@googlegroups.com.
To view this discussion on the web, visit https://groups.google.com/d/msgid/openrem/1a642916-0653-4693-b33d-414ccbd62afao%40googlegroups.com.
<Zeego_HFP_Neuro.PNG>

w3ns...@gmail.com

unread,
Jul 14, 2020, 11:28:18 AM7/14/20
to OpenREM
Hi,
A small update on my side, it currently looks like this: (siemens 180 degrees rotational acquisition)
this version still generates 1 skin map that has some nonsense normal vectors because that area would be outside the anatomy of the phantom.
The cylinder has length 24, circumference of 58, it doesnt scale with patient size.

Coming weeks I will be less responsive but I do intend to work on it (I will be busy renovating my new home).

Best,
W

Op vrijdag 26 juni 2020 16:09:07 UTC+2 schreef benjam...@gmail.com:
Hi, 
Reply all
Reply to author
Forward
0 new messages