Phantom Testing: Mammography (Revised 1-26-2024)

Revision History

The facility must submit an image of an ACR-approved phantom for review. The phantom image must be from the current testing cycle (the date the testing packet was issued forward). For initial accreditation, renewal, or a reinstatement, clinical and phantom images from each unit must be taken within 30 days of each other. All images should be clearly dated. If you are repeating a test for a deficiency, only submit images from the deficient test (i.e., all clinical images and/or the phantom).

Two types of phantoms are available: the small ACR Mammography Phantom and the ACR Digital Mammography Phantom. Both simulate a 4.2 cm compressed breast of average density and have a wax insert containing fibers, specks and masses. The facility must purchase the phantom directly from the manufacturer. At this time, the ACR has approved the following manufacturers to sell phantoms for use in the ACR Mammography Accreditation Program:

Exposure Instructions

For each unit tested, follow the instructions below to make an exposure for the relevant phantom. Your supervising radiologist must review and approve the phantom image prior to submission to the ACR. Submit only one image for each unit:

• Screen-Film: Submit one 2D phantom image for each unit

• FFDM: Submit only one 2D DICOM phantom image for each unit taken in 2D mode (i.e.-ACR Conventional-Lorad/Hologic, GE Essential- GE)

• DBT: Submit one 3D best-slice phantom image (DICOM, JPEG, JPG, PNG, GIF, TIFF, BMP). “Best-slice” is defined as the slice where the test objects are best visualized.

All Systems Using the ACR Digital Mammography Phantom and the 2016 ACR Digital Mammography Quality Control Manual
	Position phantom as shown. Chest-wall side of phantom must be completely flush with chest-wall side of image receptor.
	Lower the compression paddle to approximately 5 daN (or 12 pounds) of compression force.
	If applicable, position the AEC sensor so that it is under the center of the wax insert.
	Set imaging mode and technique (AEC mode, target, filter, kVp, density control setting, etc.) currently used for a 4.2-cm compressed breast of average density.
	Make an exposure.
	Record the technical factors on the Test Image Data form.

General Electric (GE) FFDM Units (Except the GE Pristina and SenoClaire DBT) Using the Small ACR Mammography Phantom If you are using the Automatic Optimization of Parameters (AOP) AEC mode (which involves auto-timing with auto kVp and/or auto selection of target and filter) for your clinical examinations, the phantom will not appropriately simulate the attenuation of an average breast due to the nature of the unit’s automatic parameter selection and exposure control. Consequently, you must first use GE’s 4.0 cm thick acrylic plates to find a manual technique to expose the phantom.
	Position the acrylic plates on the image receptor as typically done for the AOP and SNR Check.
	Lower the compression paddle onto the acrylic plates and apply moderate compression force.
	Make an exposure under routine clinical AOP conditions (CNT, STD or DOSE) currently used for a 4.2 cm compressed breast of average density.
	Make a note of the technical factors (AOP mode, kVp, mAs, focal spot size, target and filter). These will be used to set a manual technique to expose the phantom.
	Position phantom as shown: chest wall side of phantom must be completely flush with chest wall side of image receptor.
	Do NOT use an acrylic disk on the phantom.
	Lower the compression paddle until it just touches the phantom.
	Reproduce the kVp, focal spot size, target and filter, and select a manual mAs as close as possible to the one determined by the acrylic plates in the first step. Do not use the technical factors in the GE QC Manual.
	Make an exposure.
	Record the technical factors on the Test Image Data form.

All Other FFDM Systems (Including Computed Radiography & the GE Pristina and SenoClaire DBT) Using the Small ACR Mammography Phantom Follow your manufacturer directions for exposing the phantom with the imaging mode and technique (AEC mode, target, filter, kVp, density control setting, etc.) currently used for a 4.2-cm compressed breast of average density for the modality for which you are applying for accreditation (i.e., if applying for 2D accreditation you must submit a 2D phantom).
	Position phantom as shown. Chest-wall side of phantom should be completely flush with the chest-wall side of image receptor.
	Follow your manufacturer’s directions for your weekly phantom QC for using the acrylic disc. If your manufacturer requires you to use it, place the acrylic disk on the phantom, as for routine QC. Be sure that the disk does not cover any test object locations. If your manufacturer’s QC manual specifies that the acrylic disk not be used, don’t use it.
	Lower the compression paddle until it just touches the phantom.
	If applicable, position the AEC sensor so that it is under the center of the wax insert.
	Make an exposure.
	Record the technical factors on the Test Image Data form.

All Screen-Film Units
	Position the phantom as shown.
	Place the acrylic disk on the phantom as for routine QC. Be sure that the disk does not cover any test object locations.
	Take preliminary images of the phantom to ensure correct positioning of the phantom.
	Lower the compression device until it just touches the phantom.
	Move the AEC detector so that it is under the center of the wax insert.
	Set the clinical technical factors (AEC mode, target, filter, kVp, density control setting, etc.) currently used for a 4.2-cm compressed breast of average density.
	Make an exposure.
	Record the technical factors on the Test Image Data form.

There should be one and only one phantom image submitted for each unit.

At least 2 ACR medical physicist reviewers will score the phantom image. Although the test object counts in order to pass vary with the type of phantom used, the size of the test objects that need to be seen in order to pass are the same (the slightly smaller glass sphere “specks” in the ACR Digital Mammography Phantom have been tested to have the same visibility as the corresponding Al₂O₃ specks in the small phantom).

For the small ACR Mammography Phantom, in order to pass there must be no clinically significant artifacts and the 4 largest fibers, the 3 largest speck groups, and the 3 largest masses must be visualized.

Your manufacturer’s phantom QC requirements may differ from the ACR’s criteria for accreditation. The ACR uses 1 phantom image criteria when reviewing images submitted for accreditation from all manufacturers’ units for FDA-approved mammographic modalities. However, as required by the FDA, the ACR will check the facility’s QC to ensure that they meet their FFDM manufacturer’s QC criteria.

ACR phantom image reviewers will essentially follow the same process outlined in the 1999 ACR Mammography Quality Control Manual in scoring the Small ACR Mammography Phantom for all FDA-approved mammographic modalities. Medical physicists and quality control technologists should follow the same procedures as part of their routine QC.

For the ACR Digital Mammography Phantom, ACR phantom image reviewers will follow the same process outlined in the 2018 ACR Digital Mammography Quality Control Manual. Medical physicists and quality control technologists should follow the same procedures as part of their routine QC. In order to pass there must be no clinically significant artifacts and the 2 largest fibers, the 3 largest speck groups, and the 2 largest masses must be visualized.

Additional Phantom Details to Supplement the 1999 ACR Mammography QC Manual

Subtracting is a penalty against the phantom score to reflect the presence of disturbing artifacts that could interfere with detection or diagnosis on clinical images. Artifacts should be deducted if they are seen anywhere in the area of the wax insert.

For fibers, we deduct from the fiber score for artifactual structures larger than or equal to the last scored fiber, even though they would not be confused with the last scored fiber. If the artifact is at least as apparent with a diameter at least as large as that of the last fiber scored and the artifact has a fiber-like, rather than a mass or speck-like shape, deduct the artifact from the raw score as follows: 

• If the last fiber being scored is seen as a whole fiber and the artifact is seen as essentially a full fiber (at least 75%) then deduct 1.0. 

• If the last fiber being scored is seen as one-half of a fiber and the artifact is seen as at least a half fiber then deduct 0.5 for a fiber-like artifact.

• Do this even if the artifact has a larger diameter than the last fiber scored. The normal rule that the deduction is only from the last fiber scored still applies. For example:

To count a fiber as a whole point if it appears incomplete or broke, the phantom image review instructions say to “Count each fiber as 1 point if the full length of the fiber is visible and the location and orientation are correct. Count the fiber as 0.5 point if not all, but more than half, of the fiber is visible, and its location and orientation are correct.” 

In general, fiber images with breaks mean that the entire length is not visible and they should not be given the full point. However, if there is only 1 break in the fiber and the break is smaller than the width of the fiber, you may give it a full score as long as the full length of the fiber is visible.

Because of the small variability that may occur in the length of the fibers within any phantom, some leeway is allowed. You may score a fiber as a full point if its visible length is within ±2 mm of the length of the largest fiber. If the fiber image is shorter than 2 mm of the length of the largest fiber (but is longer than 50%) you must score it as 0.5 point.

For speck groups, subtract artifactual specks, one for one, from the specks in the last scored group to reduce the score. Subtract if the speck is at least as apparent (i.e., at least as bright as and of equal or greater diameter to the specks in the last scored group). Only subtract from the speck(s) seen in the last scored group (with either a full score or 0.5). Do not subtract for artifacts that are obviously due to dust or emulsion pickoff and are much brighter than a speck would be. The example in the table is given for 1 and for 2 artifactual specks.

For mass scoring, the masses should be viewed without a magnifier. A viewing distance of greater than 18 inches is suggested.