[Intl_DxMedPhys] Help with Hologic Dimensions artifact

[Travis Greene]

Thank you to those who responded privately to my initial inquiry. I’d like to provide an update on the detector artifact issue and seek your input on a few critical questions.

Update on Artifact Issue:
To my surprise, replacing the detector assembly resolved the artifact issue, despite the raw images not initially showing the artifact. The engineer noted that a similar issue was resolved by replacing a detector on another mobile unit with intermittent, randomly located artifacts.

The radiologist used the unit at an event today, against my recommendation. I reviewed the clinical images (screenshots attached) and found artifacts visible in nearly all images when viewed at 1:1 resolution. Interestingly, the technologist reported no artifacts were present in the day’s images.  This is likely because the standard preview mode does not display them since it is not shown at 1:1. Further review of historical clinical images revealed that this artifact has been present, though less frequently, for several months. I can conclude this issue predates the recent tube leak and is unrelated to it.

Questions for Discussion:

  1.  Clinical Significance of Artifacts: The artifacts are distinct and unlikely to be mistaken for pathology. However, I find them distracting and believe they may reduce the radiologist’s diagnostic accuracy. Given that the radiologist used the system despite my concerns and did not notice the artifacts in clinical images for months, how would you assess whether these artifacts constitute a system failure? If you disagree with the radiologist’s decision to use the system, how would you address this?


  1.  Post-Repair System Performance: Following a major repair, what is the required timeline for resolving artifacts? The QC manual specifies a 30-day resolution period for issues found during annual evaluations. Does this apply to artifacts identified during a major repair evaluation, or is the system considered down until the issue is fully resolved? I’ve heard conflicting timelines (30 days vs. immediate resolution) but lack definitive, sourced guidance.

Thank you,

Travis

From: Intl_dxmedphys_wd_osu_list <intl_dxmedphys_wd_osu_list-bounces at lists.osu.edu> On Behalf Of Travis Greene via Intl_dxmedphys_wd_osu_list
Sent: Wednesday, July 9, 2025 10:57 AM
To: intl_dxmedphys_wd_osu_list at lists.osu.edu
Subject: [Intl_DxMedPhys] Help with Hologic Dimensions artifact

All, I’m hoping you can help me identify the source of an artifact in a bus with a Hologic Dimensions installed. I went to perform a major repair evaluation following the replacement of the x-ray tube, filter assembly, and collimator. 
All,

I’m hoping you can help me identify the source of an artifact in a bus with a Hologic Dimensions installed.  I went to perform a major repair evaluation following the replacement of the x-ray tube, filter assembly, and collimator.  The x-ray tube cracked and leaked oil all over the tube assembly.  First time I’ve ever heard of that happening on a mammo unit, but I digress.  During my evaluation I came across the attached artifact.  Initially, the artifact was visible only in 2D contact images.  It’s towards the chest wall about halfway from center to the left edge.  It was not visible in 3D or 2D mag images.  The artifact is in a single row of the detector going from chest to anterior direction.  I believe this is the gate direction for this detector.

The FSE was still on site so I asked her to take a look and see if they could map out the bad pixels.  The remote tech support guy pulled up the raw images and the artifact could not be found in the for-processing image.  Subsequent testing had the artifact appear intermittently, but usually in the same location.  Several repetitions of flat field shots yielded 3D projection images that had the artifact intermittently and in multiple location in some instances.  The presentation of the artifact was always the same regardless of where it appeared.  Bright towards the chest wall edge and black towards the anterior edge.  Always in a single row of the image.  We looked at another 2D image where the artifact was present in the image and again, it could not be found in the raw image.  If the artifact is not visible in the raw image (what the detector sees) then the artifact has to be added by the bad pixel correction, the gain correction, or the final image processing; right?

The plan is to replace the detector today.  However, I don’t think this will resolve the problem.  Do you think replacing the detector is the right intervention or do you think the problem is caused by some other component?  My thought is that this is a processing error not a readout error.  I wonder if the old Nvidia1650 graphics card in the system is having issues after being rattled around on the highway for several years.

Lastly, for the first time ever, I failed a Hologic Dimensions for AEC reproducibility for the 3D images.  Oddly, the kVp was switching between 29 and 30 when using the ACR phantom exam card.  I had thought when using this exam card, the kVp was fixed by the software to be 29kVp and could not be changed.  Any idea why the exposure settings might have been changing?  Is this a problem with the software?  In over a decade of evaluating these systems, I have never seen this behavior.  Typically, the 3D AEC phantom exposures are nearly identical for all four of them.

I would appreciate any advice you might have on this situation.

Thank you,

Travis

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.osu.edu/pipermail/intl_dxmedphys_wd_osu_list/attachments/20250709/42da3c80/attachment.html>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: XXXXClinical Artifact 2. S0 I0.Tiff
Type: image/tiff
Size: 1842258 bytes
Desc: XXXXClinical Artifact 2. S0 I0.Tiff
URL: <http://lists.osu.edu/pipermail/intl_dxmedphys_wd_osu_list/attachments/20250709/42da3c80/attachment.tif>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: Clinical Artifact 1. S0 I0.Tiff
Type: image/tiff
Size: 1429044 bytes
Desc: Clinical Artifact 1. S0 I0.Tiff
URL: <http://lists.osu.edu/pipermail/intl_dxmedphys_wd_osu_list/attachments/20250709/42da3c80/attachment-0001.tif>