SOP for Scanner Upgrades/Changes?

19 views
Skip to first unread message

Raamana, Pradeep Reddy

unread,
Sep 28, 2022, 12:56:10 PM9/28/22
to niQC, Nikolaus Weiskopf, Bandettini, Peter (NIH/NIMH) [E], Todd Constable, Wald, Lawrence L., Liu, Thomas

Dear Quality Champions,

 

Hope you are doing well, and enjoying the last few weeks of the summer 😊

 

I am curious if you are aware of any SOPs or best practices in handling a non-trivial scanner upgrade or change (software or hardware)? Some of the questions I have are 1) how would we go about making sure the upgrade didn’t result in any unwanted effects (noise, artefacts), 2) how do you assert that?, 3) what needs to kept in mind in this process?

 

I’d like to capture any related papers or SOPs from different centers into a useful document we can share to the community – please share anything you have from your lab/center. If you don’t want it to be posted openly, please email me separately. Thanks for considering.

 

If that is seen as a useful resource (because nothing exists in that space?) to develop an SOP and/or proper flowchart or framework for this, I’m happy to work with those interested to make a short paper.

 

Thanks,

Pradeep

 

Pradeep Reddy, RAAMANA, Ph.D.,

Assistant Professor,

Depts of Radiology,

    Biomedical Informatics, and

    Intelligent systems,

University of Pittsburgh School of Medicine.

 

Lab : openmindslab.com

Blog: crossinvalidation.com

Logo Description automatically generated

John Pyles

unread,
Sep 28, 2022, 1:52:00 PM9/28/22
to Raamana, Pradeep Reddy, Nikolaus Weiskopf, Bandettini, Peter (NIH/NIMH) [E], Todd Constable, Wald, Lawrence L., Liu, Thomas, niQC
Hi Pradeep,

Thanks for the email. I’d be both very interested in participating in this and also finding out what other sites do.

I think there is an important distinction between a major hardware change (such as a gradient coil replacement), and an upgrade (either minor or major). The latter case of an upgrade (or sometimes “upgrade”) I think is the trickier one to evaluate. The approaches there to maintaining consistency would probably overlap with procedures for multi-site facility quality assurance where different scanner versions or even entirely different models/manufacturers are being compared. A repair situation seems more straightforward as it is should basically be a “before and after” comparison. But I’d be curious about what other sites do beyond the manufacturer QA procedures. 

I’m happy to share what we’ve done in the past for major hardware replacements. I’d just have to put it together into a document.

Thanks for bringing up this topic. It would be nice to have something standard out there for the community.

Best,
John


__________________________________________________

John Pyles, Ph.D.
Associate Director, CHN
Principal Research Scientist
Center for Human Neuroscience
Department of Psychology
University of Washington
email: jo...@uw.edu  |  phone: 206.552.0107
__________________________________________________



--
You received this message because you are subscribed to the Google Groups "niQC" group.
To unsubscribe from this group and stop receiving emails from it, send an email to niqc+uns...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/niqc/DM6PR04MB67469F56CEACF00EF80CF0F1B8549%40DM6PR04MB6746.namprd04.prod.outlook.com [groups.google.com].

Raamana, Pradeep Reddy

unread,
Sep 28, 2022, 2:26:44 PM9/28/22
to niQC

Forwarding Nik’s reply below as it might not have been delivered to the group.

 

Thanks very much everyone for the replies so far, I will keep listening for others before commenting further.

 

On 2022-09-28, 1:14 PM, "Nikolaus Weiskopf" <weis...@cbs.mpg.de> wrote:

Hi Pradeep and others,

 

this is a tricky question. We usually avoid any such upgrades during

longitudinal studies.

 

If it cannot be avoided at all, we try to apply good experimental

design, particularly randomization across the upgrade time point, to

avoid bias. We would also try to treat it similarly to harmonization

across vendors, sites and scanner types. For example, traveling heads

studies (here not traveling but across time) can be useful (e.g.

using quantitative MRI approaches can generally help with these kind of

issues.

 

Best regards,

 

Nik

> Logo Description automatically generated

>

 

--

Professor Dr Nikolaus Weiskopf

Director

Department of Neurophysics

Max Planck Institute for Human Cognitive and Brain Sciences

Stephanstrasse 1a

04103 Leipzig

Germany

 

 

 

Todd Constable

unread,
Sep 28, 2022, 2:39:41 PM9/28/22
to Raamana, Pradeep Reddy, Todd Constable, niQC
I saw Nik’s comment about avoiding upgrades during longitudinal studies. We have of the order of 100 PIs running studies, many of them longitudinal and overlapping (not synchronized in time) so there is never a good time to change the magnet but it has to be done. So we measure as much as we can pre- and post- and just do it. 


R. Todd Constable, Ph.D.
Professor Radiology and Biomedical Imaging
Neurosurgery
Director MRI Research
Yale University School of Medicine
The Anlyan Center
300 Cedar Street
PO Box 208043
New Haven, CT 06520-8043
Website: http://mri.med.yale.edu


My hours may differ from yours - no need for you to respond off-hours.







-- 
You received this message because you are subscribed to the Google Groups "niQC" group.
To unsubscribe from this group and stop receiving emails from it, send an email to niqc+uns...@googlegroups.com.

John Pyles

unread,
Sep 28, 2022, 3:51:46 PM9/28/22
to niQC, Todd Constable, Bandettini, Peter (NIH/NIMH) [E], Wald, Lawrence L., Raamana, Pradeep Reddy, Nikolaus Weiskopf
One somewhat time sensitive thing to note regarding this topic is that all of us with Siemens systems are going to be forced to upgrade to XA software sometime in the next 12ish months probably. (Hats off to those paving the way for us right now and actively troubleshooting in places like www.snug-discuss.org.) It would be great if we could develop an SOP for the research MRI community that could guide sites through critical checks for compatibility as well as data stability. This could also be a good chance to standardize some QA tests that everyone could run and cross compare before/after upgrade. I’d be happy to be involved with figuring out a place that test data could be stored and shared. 

—John


__________________________________________________

John Pyles, Ph.D.
Associate Director, CHN
Principal Research Scientist
Center for Human Neuroscience
Department of Psychology
University of Washington
email: jo...@uw.edu  |  phone: 206.552.0107
__________________________________________________
On 9/28/22 at 11:39:14 AM, Todd Constable <rtc...@gmail.com> wrote:
I saw Nik’s comment about avoiding upgrades during longitudinal studies. We have of the order of 100 PIs running studies, many of them longitudinal and overlapping (not synchronized in time) so there is never a good time to change the magnet but it has to be done. So we measure as much as we can pre- and post- and just do it. 

R. Todd Constable, Ph.D.
Professor Radiology and Biomedical Imaging
Neurosurgery
Director MRI Research
Yale University School of Medicine
The Anlyan Center
300 Cedar Street
PO Box 208043
New Haven, CT 06520-8043

My hours may differ from yours - no need for you to respond off-hours.

On Sep 28, 2022, at 2:26 PM, Raamana, Pradeep Reddy <RAAM...@pitt.edu> wrote:

Forwarding Nik’s reply below as it might not have been delivered to the group.
 
Thanks very much everyone for the replies so far, I will keep listening for others before commenting further.
 

On 2022-09-28, 1:14 PM, "Nikolaus Weiskopf" <weis...@cbs.mpg.de> wrote:

Hi Pradeep and others,
 
this is a tricky question. We usually avoid any such upgrades during
longitudinal studies.
 
If it cannot be avoided at all, we try to apply good experimental
design, particularly randomization across the upgrade time point, to
avoid bias. We would also try to treat it similarly to harmonization
across vendors, sites and scanner types. For example, traveling heads
studies (here not traveling but across time) can be useful (e.g.
> 
> Logo Description automatically generated
> 
 
-- 
Professor Dr Nikolaus Weiskopf
Director
Department of Neurophysics
Max Planck Institute for Human Cognitive and Brain Sciences
Stephanstrasse 1a
04103 Leipzig
Germany
 

-- 
You received this message because you are subscribed to the Google Groups "niQC" group.
To unsubscribe from this group and stop receiving emails from it, send an email to niqc+uns...@googlegroups.com.

--
You received this message because you are subscribed to the Google Groups "niQC" group.
To unsubscribe from this group and stop receiving emails from it, send an email to niqc+uns...@googlegroups.com.

Oscar Esteban

unread,
Sep 29, 2022, 2:58:44 AM9/29/22
to Raamana, Pradeep Reddy, niQC, Nikolaus Weiskopf, Bandettini, Peter (NIH/NIMH) [E], Todd Constable, Wald, Lawrence L., Liu, Thomas
Hi Pradeep,

> because nothing exists in that space?

At least this exists - https://doi.org/10.1016/j.neuroimage.2021.118172, and it seems to me very close to what you are looking for.

Here in Lausanne we have upgraded one of our Siemens machines from XA20 to XA30 very recently, and we're running a scan-rescan (pre/post upgrade) with some 18 subjects. We plan to openly share the data at some point in the future but are happy to chat with interested parties in sharing before public release.

Cheers,
oe



--
You received this message because you are subscribed to the Google Groups "niQC" group.
To unsubscribe from this group and stop receiving emails from it, send an email to niqc+uns...@googlegroups.com.

Fidel Alfaro Almagro

unread,
Sep 29, 2022, 7:01:04 AM9/29/22
to Oscar Esteban, Raamana, Pradeep Reddy, niQC, Nikolaus Weiskopf, Bandettini, Peter (NIH/NIMH) [E], Todd Constable, Wald, Lawrence L., Liu, Thomas
Hi all,

In the confounds paper on UK Biobank that we wrote a couple of years ago, we studied the variance explained and how correlations (between brain variables and non brain variables) were affected by some hardware and software changes in the scanner. We also included some interactive plots in case you wanted to have a more detailed look.

For example, here you can see how much variance (of the IDPs)  is explained by the variance in variables such as the CMRR version (the multiband software), the scanner service pack version, ramping down events, change of cold heads, head coils, or some more varied scanner events:


The grey violin plots at the right of each set of confounds is a set of random variables with the same size as the confound set and is there just for comparison with a "baseline".

Here you can see some Bland-Altman plots of how the p values of the correlation between BODY and BRAIN variables are affected when unconfounding with the scanner service pack and without unconfounding with it, giving an idea of how these correlations are affected by the changes.


There is a whole battery of other similar interactive plots that you can check here:


I hope this helped.

Cheers

To view this discussion on the web visit https://groups.google.com/d/msgid/niqc/CAAsFcgzX7Djj1k7Li66L644W5kXEd_QvhCnLGe%3D0iSXGo8O8tg%40mail.gmail.com.

Chandana Kodiweera

unread,
Sep 29, 2022, 10:37:00 AM9/29/22
to niQC
Hi All,

FYI:
I have implemented the fbirn qa protocol at DBIC.






Cheers,

Chandana
Reply all
Reply to author
Forward
0 new messages