[Intl_DxMedPhys] Siemens MRI ACR protocol

Tue Nov 11 17:31:30 EST 2025

!-------------------------------------------------------------------|
  This Message Is From an External Sender
  This message came from outside your organization.
|-------------------------------------------------------------------!

This thread highlights a significant flaw in the current accreditation process.  There are valid points for both sides.  

I think this could be an opportunity for AI.  Assist the reviewers, and remove the physics phantom scanning by letting AI evaluate each scanner by actual clinical images.  Is your image quality degrading over time?  AI could tell you that.  Are your head scans able to resolve lesions with at least 30% of the reliability of "accredited" scanners?  AI could do that too.

The hard part would be new scanners.  That would be a place where phantom scanning would be helpful, as imperfect as it is.

---
Thanks,
Gary 

On Tue, 11 Nov 2025 22:06:15 +0000, Bob Kobistek via Intl_dxmedphys_wd_osu_list <intl_dxmedphys_wd_osu_list at lists.osu.edu> wrote:

>See Doug Pfeiffer’s contribution to this thread. I believe he is on
>the accreditation committee. According to him, the only “filter” that
>we can’t use on the phantom images is AI.
>
>My skepticism lies in the fact that, as an ACR phantom reviewer, I see
>the scan options used on the ACR protocols and the site protocols. As
>far as I can remember, I was never instructed to compare options to be
>sure no one was slipping in a filter to cheat.
>
>Also, one can argue that if, for example, the clinical protocol uses a
>256 x 192 matrix, then the ACR phantom protocols will always have
>better high contrast resolution than the clinical scans because they
>use a 256 x 256 matrix.
>
>Early this morning, before Doug responded, I sent an Email to Marcello
>Acosta at the ACR, but he’s out of the office until Monday.
>
>BTW, I do remember the practice of chilling phantoms in the
>refrigerator back in the days of 0.2T and 0.3T magnets. I also
>remember an ACR FAQ stating that this was a “no no.”
>
>
>Robert J. Kobistek, MS, FACR, DABR, MRSE(MRSC™)
>Medical Physicist
>RJK Medical Physics, Inc.
>440-463-7879
>
>From: Hardy, Peter A. <peter.hardy at uky.edu>
>Sent: Tuesday, November 11, 2025 4:39 PM
>To: Bob Bell <bell3660 at sbcglobal.net>; Gretchen Raterman Bell
><gretchen.raterman at gmail.com>; Bob Kobistek <bob at rjkmedphys.com> Cc:
>kenharper14 via Intl_dxmedphys_wd_osu_list
><intl_dxmedphys_wd_osu_list at lists.osu.edu> Subject: RE:
>[Intl_DxMedPhys] Siemens MRI ACR protocol
>
>This is an important discussion.
>I think the flaw in the argument that we should be testing the scanner
>in the same condition as is used clinically is that there is very
>little carry over in the protocol between what is done for the ACR
>measurements and what is done clinically, at least for most
>institutions. Recognizing that, the ACR requires a submission to
>include images using both the ACR AxT1 & AxT2 protocol and the
>equivalent clinical AxT1 & AxT2 scans. If either set of scans passes
>the required resolution, slice thickness, SNR, etc, the scanner passes
>the test. In the clinical scans we are not required to use or not use
>low pass or high pass filters. My own view of removing the in-line
>filter during a test is that it assesses the scanner in its rawest
>form. I don’t see any difference turning that filter off and removing
>any of the filters, such as elliptical filters, available in the
>RESOLUTION scan card or setting the RF profile to NORMAL. We may not
>use RF=NORMAL in our clinical settings but using it for the test gives
>a uniform assessment of the scanner’s performance.
>
>Peter Hardy, PhD, DABMP
>Associate Professor of Radiology, Neuroscience and Biomedical
>Engineering Faculty, Division of Diagnostic & Nuclear Medical Physics
>Department of Radiology
>University of Kentucky College of Medicine
>UK HealthCare
>800 Rose Street, Room HX-307
>Lexington, KY 40536-0293
>Phone: (859) 323-2954
>Email: peter.hardy at uky.edu<mailto:peter.hardy at uky.edu>
>
>
>From: Bob Bell <bell3660 at sbcglobal.net<mailto:bell3660 at sbcglobal.net>>
>Sent: Tuesday, November 11, 2025 4:23 PM
>To: Hardy, Peter A. <peter.hardy at uky.edu<mailto:peter.hardy at uky.edu>>;
>Gretchen Raterman Bell
><gretchen.raterman at gmail.com<mailto:gretchen.raterman at gmail.com>>; Bob
>Kobistek <bob at rjkmedphys.com<mailto:bob at rjkmedphys.com>> Cc:
>kenharper14 via Intl_dxmedphys_wd_osu_list
><intl_dxmedphys_wd_osu_list at lists.osu.edu<mailto:intl_dxmedphys_wd_osu_list at lists.osu.edu>>
>Subject: Re: [Intl_DxMedPhys] Siemens MRI ACR protocol
>
>CAUTION: External Sender
>
>Hi Bob,
>
>I don't recall a direct ACR reference for this policy but do recall
>discussions among the ACR MRI Technical Subcommittee members that no
>changes to a protocol should be allowed that were not also applied to
>normal clinical imaging.  To do so would gut the basis of MRI
>performance testing by allowing technical results that could not be
>accomplished clinically.
>
>Features of the initial ACR MRI phantom were designed to ensure fair
>testing by eliminating ways one could bias results.  For example,
>slice thickness, resolution and bandwidth (indirectly from the
>fat/water inserts) are all assessed simultaneously in the ACR T1 axial
>series.  One can't artificially increase SNR by using a wider slice
>thickness, by reducing resolution via a lower matrix or abnormal
>k-space weighting, or by markedly lowering bandwidth.  I also seem to
>recall ACR noting not to chill the large ACR MRI phantom prior to
>testing in order to improve SNR as some did in the early days of
>low-field systems.
>
>Bob Bell
>760-579-3929
>
>
>
>On Monday, November 10, 2025 at 08:07:02 PM PST, Bob Kobistek
><bob at rjkmedphys.com<mailto:bob at rjkmedphys.com>> wrote:
>
>
>
>Is that documented somewhere on the ACR website?
>
>
>
>
>
>Robert J. Kobistek, MS, FACR, DABR, MRSE(MRSC™)
>
>Medical Physicist
>
>RJK Medical Physics, Inc.
>
>440-463-7879
>
>
>
>From: Intl_dxmedphys_wd_osu_list
><intl_dxmedphys_wd_osu_list-bounces+bob=rjkmedphys.com at lists.osu.edu<mailto:intl_dxmedphys_wd_osu_list-bounces+bob=rjkmedphys.com at lists.osu.edu>>
>On Behalf Of Bob Bell via Intl_dxmedphys_wd_osu_list Sent: Monday,
>November 10, 2025 8:01 PM To: Hardy, Peter A.
><peter.hardy at uky.edu<mailto:peter.hardy at uky.edu>>; Gretchen Raterman
>Bell <gretchen.raterman at gmail.com<mailto:gretchen.raterman at gmail.com>>
>Cc: kenharper14 via Intl_dxmedphys_wd_osu_list
><intl_dxmedphys_wd_osu_list at lists.osu.edu<mailto:intl_dxmedphys_wd_osu_list at lists.osu.edu>>
>Subject: Re: [Intl_DxMedPhys] Siemens MRI ACR protocol
>
>
>
>Hi all, Just a quick comment on this email chain. It's my
>understanding ACR allows any filters or instrument settings routinely
>used for patient studies to be used for phantom studies but does not
>allow filters/settings unique to phantom studies
>
>Hi all,
>
>
>
>Just a quick comment on this email chain.  It's my understanding ACR
>allows any filters or instrument settings routinely used for patient
>studies to be used for phantom studies but does not allow
>filters/settings unique to phantom studies alone.  Employing
>filters/settings that are not used for patient studies defeats the
>purpose of annual testing by artificially altering instrument
>performance from normal clinical practice.  For example, it's okay to
>use prescan normalize on Siemens units to improve IIU as long as it is
>also used for patient studies.  However, altering filters/settings to
>those not common for clinical studies just to improve instrument
>phantom performance is a no no.
>
>
>
>Bob Bell
>
>
>
>On Friday, November 7, 2025 at 06:29:18 AM PST, Gretchen Raterman Bell
>via Intl_dxmedphys_wd_osu_list
><intl_dxmedphys_wd_osu_list at lists.osu.edu<mailto:intl_dxmedphys_wd_osu_list at lists.osu.edu>>
>wrote:
>
>
>
>
>
>For what it's worth, I'll summarize some settings I've gotten from
>both the DXlist and from Siemens that have helped tremendously. I work
>on both Aeras and Solas. To improve signal uniformity: - Use B1 filter
>- Use sum of squares
>
>For what it's worth, I'll summarize some settings I've gotten from
>both the DXlist and from Siemens that have helped tremendously.  I
>work on both Aeras and Solas.
>
>
>
>To improve signal uniformity:
>
>- Use B1 filter
>
>- Use sum of squares (coil combination)
>
>
>
>To improve spatial resolution, particularly Solas:
>
>  *   Resolution tab --> B1 Filter should be ON
>  *   Elliptical Filter should be OFF
>  *   Prescan Normalize should be ON
>  *   Sequence tab --> Gradient Mode should be NORMAL or FAST
>  *   Sequence tab --> RF pulse types, try different SAR settings (try
> this very last, he was unsure about this one)
>  *   If none of this improves the high contrast resolution, he highly
> recommended calling into the  1800 number for remote service:
>
>     *   In-Line Image Filter should be turned OFF
>     *   This is a global setting that only FSE or remote service can
> modify; must be turned back on after acquiring phantom images
>     *   It tends to be an overall smoothing filter
>     *   Particularly affects newer software, which is likely why we
> don’t see it on the Aeras
>
>
>
>FYI, I haven't had to do that last bullet point yet, but good to know
>just in case.
>
>Gretchen R. Bell, M.S., DABR
>Diagnostic Imaging Physicist
>Ochsner Medical Center
>(504)842-8506
>
>
>
>
>
>On Fri, Nov 7, 2025 at 8:12 AM Hardy, Peter A. via
>Intl_dxmedphys_wd_osu_list
><intl_dxmedphys_wd_osu_list at lists.osu.edu<mailto:intl_dxmedphys_wd_osu_list at lists.osu.edu>>
>wrote:
>
>It might be easier to import an .exar file. These are available from
>Siemens on their Phoenix site. ACR Phantom QA Test I’m attaching the
>file for a 1. 5T system. Similar files are available for other systems
>and OS levels. This protocol likely
>
>It might be easier to import an .exar file. These are available from
>Siemens on their Phoenix site.
>
>ACR Phantom QA
>Test<https://urldefense.com/v3/__https:/www.magnetomworld.siemens-healthineers.com/clinical-corner/protocols/mr-in-rt/acr-phantom-qa-test__;!!KGKeukY!0J5pWek2Y1lqkWdNUarukJAaQxlT5NXRKlh0kmDROFkR-peOYmfS1MRNkentpqiJPmhBL1nfqqpvJzIqMmXyItg3OYVYBOFj01ry3nznJwbc$>
>
>
>
>I’m attaching the file for a 1.5T system. Similar files are available
>for other systems and OS levels.
>
>This protocol likely includes the updated axial T2 sequence.
>
>
>
>Peter Hardy, PhD, DABMP
>
>Associate Professor of Radiology, Neuroscience and Biomedical
>Engineering
>
>Faculty, Division of Diagnostic & Nuclear Medical Physics
>
>Department of Radiology
>
>University of Kentucky College of Medicine
>
>UK HealthCare
>
>800 Rose Street, Room HX-307
>
>Lexington, KY 40536-0293
>
>Phone: (859) 323-2954
>
>Email: peter.hardy at uky.edu<mailto:peter.hardy at uky.edu>
>
>
>
>
>
>From: Intl_dxmedphys_wd_osu_list
><intl_dxmedphys_wd_osu_list-bounces+peter.hardy=uky.edu at lists.osu.edu<mailto:uky.edu at lists.osu.edu>>
>On Behalf Of Adam Springer via Intl_dxmedphys_wd_osu_list Sent:
>Friday, November 7, 2025 7:26 AM To: kenharper14 via
>Intl_dxmedphys_wd_osu_list
><intl_dxmedphys_wd_osu_list at lists.osu.edu<mailto:intl_dxmedphys_wd_osu_list at lists.osu.edu>>
>Subject: [Intl_DxMedPhys] Siemens MRI ACR protocol
>
>
>
>CAUTION: External Sender
>
>
>
>Hello, List. I have instructions from Siemens for building the ACR
>protocol on a Magntom Avanto, from 2005. Does anyone have a more
>recent set of those instructions? I've had to ask Siemens to reload
>their ACR protocol at a couple sites recently,
>
>Hello, List.
>
>
>
>I have instructions from Siemens for building the ACR protocol on a
>Magntom Avanto, from 2005. Does anyone have a more recent set of those
>instructions? I've had to ask Siemens to reload their ACR protocol at
>a couple sites recently, and it would be convenient if I could just
>change the settings while I'm on site.
>
>
>
>Happy Friday!
>
>Adam
>
>
>
>
>
>
>
>========================
>Adam C. Springer, MS, DABR®
>Medical Physicist & CEO
>KLS Physics Group, LLC
>124 Killgore Rd
>Ruston, LA 71270
>Office: 318-255-5033
>Mobile: 318-272-8292
>
>[2927183d-095e-413d-96b1-930240ed6257]