Introduction
The Nuclear Medicine Accreditation Physics Subcommittee has defined scoring criteria for planar (uniformity and resolution) and SPECT (uniformity, spatial resolution, and contrast). These criteria are based on results obtained from a variety of cameras that are operating satisfactorily.
Data must be collected, and images prepared, according to the phantom instructions. If sites do not follow the testing requirements as specified in these instructions, then sites may be required to resubmit images or may be subject to failing the submission. Note: Specialty scanners may not be capable of meeting the image acquisition parameters, and guidance is provided within these instructions. Contact the ACR for additional questions or clarifications.
Breast imaging units should follow the small FOV requirements; for spatial resolution the phantom should be placed on a 45-degree angle to visualize all quadrants (refer to the phantom atlas for image examples).
Common reasons for image rejection:
Planar images acquired with incorrect parameters (e.g., matrix and counts)
Data form does not accurately reflect information for images submitted
Planar images submitted are not acquired as static planar images (Sites should not submit the daily QC results page)
Planar uniformity images with graphic overlays
Screenshot images with the upper threshold not equal to 100% and lower threshold not between 0-5%
Attenuation correction was not applied (required for cameras that rotate 360 degrees; for SPECT CT units, CT attenuations correction is acceptable)
As appropriate, planar and/or SPECT phantom images must be obtained and submitted for review using the phantom that has been approved by the ACR Committee on Nuclear Medicine Accreditation. Please see the article on Modalities with Low Volume or Emergency Use as needed.
Example phantom images can be reviewed in the NM/PET Phantom Image Atlas, available as an attachment in our Testing Overview article
The ACR-approved SPECT phantom is commonly used for quality control in nuclear medicine. For cameras that are used to perform planar and SPECT imaging studies, an ACR-approved phantom should be used for evaluating planar and tomographic image quality. The phantom is a cylinder with an internal height of 18.6 cm and internal diameter of 20.3 cm (flangeless phantom). The lower portion of the phantom contains 6 sets of acrylic rods arranged in a pie-shaped pattern with the following diameters: 4.8, 6.4, 7.9, 9.5, 11.1, and 12.7 mm. The upper section contains six solid spheres with the following diameters: 9.5, 12.7, 15.9, 19.1, 25.4, and 31.8 mm.
The ACR has also approved a Data Spectrum Small SPECT phantom. The phantom is a cylinder with an internal length of 15 cm and an internal diameter of 13.9 cm (flangeless phantom). The lower portion of the phantom contains 6 sets of acrylic rods arranged in a pie-shaped pattern with the following diameters: 4.8, 6.4, 7.9, 9.5, 11.1, and 12.7 mm. The upper section contains six solid spheres with the following diameters: 6.4, 9.5, 12.7, 15.9, 19.1, and 25.4 mm.
Cameras that should use Small SPECT phantom: GE 530c, GE 570c, MyoSPECT, maiCam, C!, P3000, and ClearVision
Optional (can use either the Deluxe or Small ACR phantom): c.cam, CardioMD, D-SPECT, Digirad dedicated cardiac SPECT cameras (e.g., Cardius, 2XPO, 3XPO, X-ACT, 2020TC SPECT, SPECTpak), CorCam, and Ventri
The following are available directly from Data Spectrum of Durham, NC (order forms for the large and small phantoms are available in our Testing Overview: Nuclear Medicine and PET article):
Jaszczak Deluxe Flangeless ECT phantom and PET faceplate (can be used for both SPECT and PET acquisitions)
Jaszczak Deluxe Flangeless ECT phantom (for SPECT only)
Small SPECT Phantom (for SPECT only)
Flangeless PET phantom (for PET only)
PET faceplate made to fit an existing flangeless or flanged Jaszczak Deluxe ECT phantom
One set of Site Scanning Data Forms (attached below), along with all pertinent worksheets, should be completed for each camera system that is undergoing accreditation.
Phantom Requirements by Module
All Modules (Planar Only or SPECT with Planar): All applicants are required to submit a Tc-99m or Co-57 intrinsic or system flood field image, as well as a Tl-201 and/or a Ga-67/In-111 flood field image (if applicable). If both Tl-201 and Ga-67/In-111 are currently used, planar uniformity and planar spatial resolution images are required for both isotopes.
Module 1 (Planar Imaging Only): If a unit performs only planar imaging and is only applying for accreditation in Module 1 - Planar Imaging, the number and complexity of the tests you are asked to perform, evaluate, and submit for review are different than those required for a unit that performs SPECT. You must provide images documenting the planar uniformity and spatial resolution of your camera detector(s).
If accreditation is requested for planar imaging only, intrinsic or system spatial resolution images must be submitted. The phantom for this data is the four-quadrant bar phantom that you should already own and use on a regular basis. The smallest bars must be between 2 and 3 mm. Planar images acquired with the ACR SPECT phantom are not required for applicants that are requesting accreditation for Module 1 only but are acceptable in lieu of the four-quadrant bar phantom. If using the ACR SPECT phantom in lieu of the four-quadrant bar phantom, please refer to the spatial resolution requirements in the Planar and SPECT protocol section.
The Planar Only protocol should be followed for camera systems seeking accreditation for the Planar Module only. Images must be submitted for all detectors on systems with more than one detector.
For each of the tests requiring the SPECT phantom, plan the various acquisitions so that the minimum time is required. Please note that if the additional isotopes include both Ga-67 and In-111, you only need one submission for Ga-67 or In-111. If you are performing less than 25 exams using T1-201, Ga-67, In-111 in a year, you are not required to apply for these isotopes. Please see Units for Emergency Use as needed.
Planar Only Protocol |
Acquisition One: Tc-99m (Preferred) or Co-57 |
Field Uniformity - Each Detector | Spatial Resolution - Each Detector (Four-Quadrant Bar Phantom)Refer to the Planar and SPECT protocol section if using the ACR rod phantom for spatial resolution |
Set the acquisition for Tc-99m or Co-57 and acquire an intrinsic or system uniformity image using a 256 x 256 matrix (with zoom 1x). For large rectangular field cameras (smallest dimension ≥ 32 cm), acquire 10 million counts. For large and small circular detectors and small rectangular field cameras (smallest dimension < 32 cm), acquire 5 million counts. For intrinsic uniformity, the point source must be set at a distance from the detector that is 3-5 times the longest dimension of the detector. For a 50 cm detector, this would be 150-250 cm from the detector. Intrinsic floods acquired using the Siemens point source holder will not be accepted. | Set the appropriate energy and acquire an intrinsic or system spatial resolution pattern image using a 512 x 512 matrix (with zoom 1x). Some scanners may not be capable of acquisition with a 512 x 512 matrix. For these systems, use the largest matrix available on the system. For large rectangular field cameras (smallest dimension ≥ 32 cm), acquire 5 million counts. For large and small circular detectors and small rectangular field cameras (smallest dimension < 32 cm), acquire 3 million counts. |
Acquisition Two: Tl-201 |
Field Uniformity - Each Detector | Spatial Resolution - Each Detector(Four-Quadrant Bar Phantom)Refer to the Planar and SPECT protocol section if using the ACR rod phantom for spatial resolution |
Set the energy windows for Tl-201 that are used clinically and acquire an intrinsic or system uniformity image as described above for Tc-99m. For intrinsic uniformity, the point source must be set at a distance from the detector that is 3-5 times the longest dimension of the detector. For a 50 cm detector, this would be 150-250 cm from the detector. Intrinsic floods acquired using the Siemens point source holder will not be accepted. | Set the energy windows for Tl-201 that are used clinically and acquire an intrinsic or system spatial resolution pattern as described above for Tc-99m. |
Acquisition Three: Ga-67/In-111 |
Field Uniformity - Each Detector | Spatial Resolution - Each Detector(Four-Quadrant Bar Phantom)Refer to the Planar and SPECT protocol section if using the ACR rod phantom for spatial resolution |
Set the energy windows for Ga-67 or In-111 that are used clinically and acquire an intrinsic or system uniformity image as described above for Tc-99m. For intrinsic uniformity, the point source must be set at a distance from the detector that is 3-5 times the longest dimension of the detector. For a 50 cm detector, this would be 150-250 cm from the detector. Intrinsic floods acquired using the Siemens point source holder will not be accepted. | Set the energy windows for Ga-67 or In-111 that are used clinically and acquire an intrinsic or system spatial resolution pattern image as described above for Tc-99m. |
Planar Images for SubmissionSubmit only gray scale images.Use linear mapping of the display with the lower threshold set at 0 (or ≤ 5% of the maximum upper level) and the upper threshold set to 100% or the maximum count.All images should be appropriately labeled with camera head number for systems with more than one detector and isotope if applying for TI-201, Ga-67, or In-111. Make sure the Site Scanning Data Forms are filled out completely and match the acquired data.Images with graphic overlays will not be accepted.DICOM images are preferred.
|
Module 2 and/or 3 (SPECT with Planar): If your facility performs SPECT imaging and is applying for accreditation in SPECT Module or Nuclear Cardiology Module, you must perform, evaluate, and submit both planar and SPECT information for review. Most of the planar and SPECT data are obtained with the ACR-approved phantoms available from Data Spectrum. Submission of accreditation data using any other phantom without ACR permission will disqualify the site. The ACR-approved phantoms are the Data Spectrum Deluxe flanged or flangeless phantom and the Data Spectrum Small flangeless Jaszczak phantom (Small SPECT phantom). The Small SPECT phantom is only approved for use on certain cameras. The ACR-approved phantom was chosen for the evaluation of tomographic systems because it is relatively easy to fill and set up for a SPECT study and can be used to measure planar resolution, and tomographic uniformity, spatial resolution, and contrast. The wide variety of elements for evaluating spatial resolution and contrast are sufficient to enable the user to detect relatively subtle differences in system performance.
The recommended activity for the ACR-approved phantom is 10-20 mCi Tc-99m. The recommended activity for the Small SPECT phantom is 5-15 mCi Tc99m. Lower quantities may be used but the imaging time will be proportionately longer. These activities are recommended to provide a good compromise between count rate limits and acquisition time.
For both phantoms, the spheres must be placed in order of decreasing size and aligned with the rods in decreasing size. The sphere diameters must be the same as listed in the worksheets. The reviewers will use this information to properly score the images.
Acquire and process the best possible images on your system within the guidelines provided. All settings must be documented.
The ACR requires semiannual but strongly recommends quarterly testing of each SPECT system with the ACR-approved phantom described above, in addition to other QC tests.
For systems with more than one detector, planar images must be submitted for each detector. Proceed to the SPECT Imaging section below for the following cameras that acquire only SPECT images: D-SPECT, Digirad dedicated cardiac SPECT cameras (e.g., Cardius, 2XPO, 3XPO, X-ACT, 2020TC SPECT, SPECTpak), Veriton, Starguide, GE 530c, GE 570c, MyoSPECT, Clearvision, and GVI.
Sites performing SPECT imaging are required to submit Tc-99m SPECT phantom images. Please note that sites utilizing Tl-201, Ga-67, and/or In-111 for either planar or SPECT imaging are required to submit both planar uniformity and planar spatial resolution images with these isotopes (see the Planar Only protocol).
Planar and SPECT Protocol for Systems Applying for SPECT Accreditation |
Planar Imaging Protocol |
Planar Field Uniformity - Each Detector | Planar Spatial Resolution - Each Detector(Phantom rods or Bars are acceptable) |
Set the acquisition for Tc-99m or Co-57 and acquire an intrinsic or system uniformity image using a 256 x 256 matrix (with zoom 1x). For large rectangular field cameras (smallest dimension ≥ 32 cm), acquire 10 million counts. For large and small circular detectors and small rectangular field cameras (smallest dimension < 32 cm), acquire 5 million counts. For intrinsic uniformity, the point source must be set at a distance from the detector that is 3-5 times the longest dimension of the detector. For a 50 cm detector, this would be 150-250 cm from the detector. Intrinsic floods acquired using the Siemens point source holder will not be accepted.
| ACR Phantom (Recommended): Fill the phantom with an aqueous solution of Tc-99m, taking care to ensure that it is thoroughly mixed. It is advisable to perform the Tc-99m studies first, and then fill the phantom with Tl-201 or Ga-67/In-111 after the Tc-99m has decayed. Place the phantom on top and in the center of the detector. The rods should be at the flat bottom of the cylinder and closest to the collimator. Use the highest resolution low energy parallel hole collimator that is used clinically. The low-energy high resolution (LEHR) collimator is preferred. If you are using the ACR phantom to acquire the planar resolution for other isotopes, use the collimator that is routinely used for the indicated radionuclide. Set the energy window to the window used for clinical acquisitions. The matrix size should be 256 x 256. The count rate should not exceed 50 kcps. If a 256 x 256 matrix is not available on the system, use the highest matrix available. Acquire a static image using a zoom factor of 1 for 600,000 counts. |
Four-Quadrant Bar: Set the appropriate energy and acquire an intrinsic or system resolution pattern image using a 512 x 512 matrix (with zoom 1x). Some scanners may not be capable of acquisition with a 512 x 512 matrix. For these systems, use the largest matrix available on the system. For large rectangular field cameras (smallest dimension ≥ 32 cm), acquire 5 million counts. For large and small circular detectors and small rectangular field cameras (smallest dimension < 32 cm), acquire 3 million counts. |
Planar Images for SubmissionSubmit only gray scale images.Use linear mapping of the display with the lower threshold set at 0 (or ≤ 5% of the maximum upper level) and the upper threshold set to 100% or the maximum count.All images should be appropriately labeled with camera head number for systems with more than one detector and isotope if applying for TI-201, Ga-67, or In-111. Make sure the Site Scanning Data Forms are filled out completely and match the acquired data. Images with graphic overlays will not be accepted. DICOM images are preferred.
|
SPECT Imaging ProtocolQC Information: Center-of-rotation and flood uniformity calibrations are regularly performed on SPECT systems. Fill in the appropriate data fields based on the most recent calibrations.The ACR phantom used for the SPECT tests must be described on the Site Scanning Data Forms. For more information, please see the ACR’s SPECT Phantom Video.
|
Positioning of ACR Phantom for Tomography:Place the phantom on the patient support (bed, chair, etc.) so that it is positioned lengthwise, corresponding to the primary patient axis. The phantom’s long axis must be parallel to the rotational (z)-axis of the detector system with the phantom level. Phantoms with flanges may require extra support for leveling. The phantom should be positioned in the center of the field of view of the detector(s).
Note: For cameras that are only capable of rotating through 180 degrees and have the detectors in a fixed relative angle of 90 degrees, special caution must be exercised in the orientation of the spheres and rods. You may use your cardiac acquisition protocol. Ensure that the phantom configuration is consistent (largest sphere adjacent to largest rods, second-largest sphere adjacent to second-largest rods, etc.). Position the phantom on the imaging table such that the largest sphere will be in approximately the 10 o'clock position in the reconstructed image. When set up properly, the largest sphere will be in the upper left of the phantom when viewed from the foot of the table. If you are imaging from 0 to 180 degrees, the largest sphere should be at the 12 o'clock position. See the images below for guidance.![](https://s3.amazonaws.com/cdn.freshdesk.com/data/helpdesk/attachments/production/11085878182/original/nUX3BGZBPuviaIfbuhtJLjEN4GKW94Au5Q.png?1651161740)
|
Collimator: Use the highest resolution low-energy parallel hole collimator that is routinely used for clinical studies. |
Acquisition:Select the appropriate isotope and acquire a SPECT study with the window setting, orbit shape, and detector motion (step-and-shoot or continuous motion for rotating SPECT cameras) that is used clinically.Total counts - For SPECT acquisitions the sum of the counts from all images (for all heads on multiple detector systems) should be approximately 32 million counts for the Deluxe Phantom and 20 million counts for the Small SPECT phantom. A sample calculation to determine the number of seconds per image to use when acquiring the SPECT data is provided below. Alternatively, use the table below to determine the number of counts in the first image for cameras that may be set to acquire the first image for a set number of counts and the rest of the images for the same time.Count rate should not exceed 50 kcps.Use a 128 x 128 matrix regardless of which phantom is used. Acquire 120 (or 128) views over 360 degrees for rotating SPECT cameras. Some cardiac SPECT systems cannot acquire a single set of projections over 360 degrees. For these systems, the phantom should be acquired with the maximum number of rotational steps for the standard rotational arc. This is generally 60 or 64 steps over 180 degrees. Cameras that do not rotate should use the angular sampling that is used clinically.A circular or non-circular orbit can be used. If a circular orbit is used, then the radius of rotation should be as close as possible to 20 cm. For either acquisition type, the center of the phantom must be close to the axis-of-rotation so that the distance from the face of the collimator(s) to the phantom does not change significantly during rotation. The Small phantom should be placed at the optimal clinical imaging position for cameras that do not rotate around the patient.For an acquisition matrix of 128 x 128, pixel size should be near 2.7 to 3.3 mm. Some systems will require an acquisition zoom factor (up to a maximum of 1.6) to achieve the appropriate pixel size.If the camera cannot acquire a 128 x 128 matrix, use the matrix closest to 128 x 128.
Sample Calculation for SPECT AcquisitionCalculation of the time per view to acquire approximately 32M counts in the SPECT study: If the system you are using shows the count rate prior to starting the SPECT acquisition, use the procedure given below. If the count rate is not shown, just prior to starting the SPECT study, acquire a planar acquisition for 10 seconds (be sure the imaging table is not between the detector you are using and the phantom) and divide the acquired counts by 10 to get the rate in counts per second. Example:
R = count rate in counts per second for one of the detectors | N = number of views (120 or 128, whichever is available or whichever you want to use if both are available) | T = time for the first view |
![](https://s3.amazonaws.com/cdn.freshdesk.com/data/helpdesk/attachments/production/11037618910/original/D32NuojemO08gaabFxPbZhfhFWPZaaKYRQ.JPG?1541083161)
Assume R = 28,000 counts per second. Assume N = 120 views.![](https://s3.amazonaws.com/cdn.freshdesk.com/data/helpdesk/attachments/production/11037618965/original/_MOzkXlFd5rhvrOz1HJvtzXTZJxAzqOQpw.JPG?1541083223)
Round time up to the next whole number if the fraction of a second is greater than 0.5. Round time down to the next whole number if the fraction of a second is less than 0.5.![](https://s3.amazonaws.com/cdn.freshdesk.com/data/helpdesk/attachments/production/11037619018/original/QoudIWg5rve0pA_lk-VhKzkiNNEggiZjPQ.JPG?1541083295)
If the system has two detectors, the total acquisition time will be approximately one-half of the time for a single detector system. If the system has three detectors, the total acquisition time will be approximately one-third of the time for a single detector system.
Table for the number of counts to acquire when acquiring the first image by counts and the rest by time Total number of Projections - Rotation Arc (in degrees) | Number of Counts in First Image | Deluxe Phantom- 32 M counts | Small Phantom - 20 M counts | 120-360 | 267,000 | 167,000 | 128-360 | 250,000 | 156,000 | 60-180 | 534,000 | 334,000 | 64-180 | 500,000 | 312,000 |
|
Reconstruction:Reconstruction with filtered back projection is recommended using the Butterworth filter that is available on most commercial systems. For the Butterworth filter, suggested parameters are a cutoff of 0.55 and a slope of 6, which is optimal for many systems (for some vendors the cutoff will be close to 0.35); however, parameter values can be adjusted to optimize contrast of spheres and sharpness of smallest visible rods in the transaxial slices. Lower cutoff numbers will result in a smoother image with less noise and less detail. Sites may use the clinical reconstruction method and filter they are currently using with parameters optimized for the ACR phantom. Note that using iterative reconstruction methods may result in the appearance of edge artifacts in the reconstructed images that could result in a lower uniformity score. Review your images carefully for this type of artifact prior to submission. Attenuation correction is required for cameras that rotate 360 degrees. For most cameras, the effective linear attenuation coefficient for Tc-99m is 0.11 to 0.12/cm. Use the uniform part of the phantom to define the boundary for attenuation correction. Please note that on most systems, the attenuation correction is applied as a separate step in the reconstruction programs or a separate program outside of the reconstruction. The operator is usually required to set the boundaries and to apply attenuation correction. Sites submitting data from cameras that can acquire only 180 degrees of data (i.e., cardiac only cameras) are not required to apply attenuation correction to the data.For SPECT CT units, CT attenuation correction is acceptable.The filter and parameters as well as the slice thickness and attenuation coefficient must be documented on the Site Scanning Data Form. Display the first image of the raw SPECT data. Record the total counts in the image on the Site Scanning Data Form. Display all transaxial reconstructed images with a slice thickness 0.6-0.9 cm by summing 2 or 3 slices. For most large FOV cameras, summing 2 slices will generate the appropriate thickness.
|
SPECT Images for SubmissionSubmit the complete set of transaxial slices (0.6 to 0.9 cm thick). Do not submit more than 28 slices on a single page. Image size should be adjusted with the display zoom so that each image occupies about 80% of the image frame.DICOM reconstructed axial slices reframed to the correct slice thickness may be uploaded in lieu of screen captures.Submit a composite image of 3-6 cm thick through the rod section; 2-3 cm thick for the small phantom. Submit only gray scale images.Use linear mapping of the display with the lower threshold set at 0 (or ≤ 5% of the maximum upper level) and the upper threshold set to 100 % or the maximum count.All images should be appropriately labeled. Slice number(s) or position location should be included on secondary capture images that are submitted. Make sure Site Scanning Data Forms are filled out completely and match the data submitted. For multi-detector cameras, separate SPECT data sets for each detector should not be submitted.Only the transaxial images should be submitted. Do not submit sagittal or coronal images. Do not submit secondary capture images that display the three orthogonal planes (transaxial, sagittal and coronal) on one page.
|
Evaluation of the Images Submitted
Uniformity, noise, and the presence of artifacts are evaluated qualitatively by inspection of planar images and reconstructed tomographic sections. Planar spatial resolution is judged by identifying the smallest bars in a resolution phantom that can be visualized on cameras used for planar imaging only. For tomographic systems, the smallest “cold” rods in the ACR-approved phantom are used. Collimator type (e.g., general purpose, high resolution) is considered in scoring the spatial resolution. The reconstructed transaxial slices containing the "cold" spheres are used to visually evaluate image contrast based on identifying the smallest “cold” sphere seen at high contrast. As with the evaluation of spatial resolution, collimator type is considered when the score for contrast is determined.
Revision History for this Article |
Date | Section | Description of Revision(s) |
12-12-19 | All | Article created; FAQs incorporated; No criteria changes |
9-23-2020 |
| Added link to Phantom Image Atlas |
11-30-2020 |
| Added link to NM Equipment Evaluation Summary Form |
2-4-2021 |
| Added Clearvision GVI to list of cameras that are unable to acquire planar images |
5-6-2022 | All | Clarification of instructions (no new requirements) |
6-22-2022 | Planar and SPECT Protocol for Systems Applying for SPECT Accreditation | Phantom rods and Bars are acceptable for Planar Spatial Resolution - each detectorACR Phantom (Recommended) |
8-29-2022 | Module 2 and/or 3 (SPECT with Planar) | Removed exception for Digirad dedicated cardiac cameras |
11-9-2022 |
| Grammatical edit |
11-18-22 | Introduction | Attached Phantom Order and Site Scanning Data forms |
12-6-22 | Introduction | Moved Phantom Order Forms to Testing Overview: NM/PET article |
4-20-2023 |
| Edited 2020TC SPECT and added SPECTpak |
6-21-2023 | Introduction | Added notice at the beginning regarding a change effective January 1, 2024. |
| SPECT acquisition | Added bullet instruction for cameras that cannot acquire a 128 x 128 matrix |
| SPECT Images for Submission | Clarified 3rd bullet for composite images |
1-4-2024 | Introduction | Added clarification for breast imaging units |
|
| Added 7th bullet for common reasons for image rejection |
| Planar Only Module and Planar Imaging Protocol | Changed longest to smallest |
| Reconstruction | Added bullet for SPECT CT units |
|
| Updated attachments |
1-5-2024 | Planar Only Module and Planar Imaging Protocol
| Updated longest and smallest language |
5-1-2024 | Introduction | Removed notice of January 1, 2024 effective changes |
| Module 1 (Panar Imaging Only) | Added instructions for ACR SPECT phantom use and language on emergency use |
| Planar Only Protocol | Added instructions for ACR rod phantom use |
9-4-2024 | Introduction | Added link to the Modalities with Low Volume or Emergency Use article |
| Planar Only Protocol | Changed longest dimension to smallest dimension for large rectangular field cameras under both field uniformity and spatial resolution |
| Planar and SPECT Protocol for Systems Applying for SPECT Accreditation | Changed longest dimension to smallest dimension for large rectangular field cameras for the planar field uniformity and planar spatial resolution |