Revision History
Introduction
The Nuclear Medicine Accreditation Physics Subcommittee has defined scoring criteria for PET phantoms. 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: Some older Cardiac ONLY 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.
Common reasons for image rejection:
Non-DICOM slices submitted.
1 cm thick phantom slices not submitted.
Data form does not accurately reflect information for images submitted (i.e., SUV values do not match those in the ROI/SUV display image).
Not including ROIs with SUV measurements for the PET phantom submission.
When viewing images in the Nil viewer prior to submission, if the SUV numbers are not displaying on the DICOM slices (i.e., if an asterisk or CON is showing after drawing an ROI using the view tool Circle), then the images are in the wrong format.
Phantom images must be obtained and submitted for review using the phantom that has been approved by the ACR Committee on Nuclear Medicine Accreditation and may be purchased from Data Spectrum (phantom order forms are available in our Testing Overview: Nuclear Medicine and PET article).
Example phantom images can be reviewed in the NM/PET Phantom Image Atlas available in our Testing Overview: NM/PET article.
A PET Equipment Evaluation Summary Form and QC Review is available in our Quality Control: PET article. Upload this form along with the dose calibrator report.
The PET phantom uses the base of the Jaszczak Deluxe Flangeless ECT phantom with the spheres removed (as described below) and a PET faceplate. The ACR-approved phantom is a cylinder with an internal radius of 10.8 cm. The faceplate has fillable thin-walled cylinders (8, 12, 16, and 25 mm in diameter), two additional 25-mm cylinders, one for air and one for “cold” water, and a Teflon cylinder (the PET phantom faceplate design is being used by the ACR with the permission of Peter D. Esser, Ph.D., Columbia University; the intellectual property rights belong to Dr. Esser and uses other than through the ACR require his permission). The lower portion of the cylinder contains six 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. In addition, for the SPECT/PET version of the phantom, 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 spheres must be removed for PET studies. RODS MUST REMAIN.
PET data must be collected and processed according to the instructions provided in the testing package. The acquisition and processing must be essentially the same as those used for clinical whole-body scans. Despite the use of a specific protocol, it is understood that there may still be some differences even if the data are collected on the same type and model PET unit.
The phantom was chosen for evaluating PET tomographic systems because it is relatively easy to fill and set up for a PET study and can be used to measure tomographic uniformity, spatial resolution, and the detectability of “hot” lesions. The variety of resolution and “hot” components enables the reviewers to see relatively subtle differences in system performance.
Acquire the best possible images on your system with the same protocol as a routine clinical whole-body scan. All settings must be documented. Reconstruct the entire phantom as you would for clinical studies.
Site scanning data forms (available as attachments in our Clinical Image Testing: PET article) should be completed for each PET system that is undergoing accreditation. In addition, for each PET system, the applicant is required to submit a list with the frequency of all PET-related quality control (QC) procedures, including yearly, quarterly, monthly, weekly and daily tests (see the PET Quality Control Summary Form available in our Quality Control: PET article).
The ACR requires semiannual testing and strongly recommends quarterly testing of each PET system with an appropriate ACR-approved PET phantom such as described here, in addition to other tests recommended by the vendor.
Preparation of Phantom
Please read all instructions before preparing the phantom. Confirm that all calibrations and QC procedures are current.
One scan will be acquired with the ACR-approved PET phantom. The phantom scan starts 1 hour after Dose A is measured. During the 1-hour preparation time the other doses are measured, vials are filled, background added to the phantom, and phantom positioned in the gantry.
Equipment Required | |
1 ACR-approved PET phantom | 1 1000 ml bag or bottle of water |
2 tuberculin syringes (for measuring Doses A & B) | 3 large syringes (60 ml) |
Large-bore needles (18 gauge) | Dose: FDG or F-18 are acceptable (Most vendors provide F-18 for calibrations/tests of scanners at a significantly lower price than FDG) |
Clock or timer | Phantom dilution worksheet (attached below) |
The Table of Dilutions to produce the 2.5 concentration ratio (2.5:1 ratio may represent a lesion in the liver) is presented below. From the left column on the table, select the administered FDG whole-body dose for your site. The activities that are used for the initial Dose A and for the phantom itself, Dose B, are calculated from the F-18 activity normally used by the facility for whole body scans. If the activity is between the values, you must use the one with the higher level. For example, if your facility uses a 5 mCi dose for the whole body scan, you must select the 6 mCi row from the Table.
The corresponding phantom Doses A and B are along the same row as the Patient dose. The second 25 mm vial must be filled with water and the third left empty (“air”).
The spreadsheet indicates the activity that must be used for filling the phantom (this will also be referred to as the “background”) and the “hot” cylinders (“lesions”). Care must be taken to ensure that the solutions in the background and 1,000 ml bag (or bottle) are thoroughly mixed.
Be sure to adjust the "zero" and "background" settings on your dose calibrator. Follow the directions below to measure (± 10%) the doses and activate the PET phantom. Please record all information on the Phantom Dilution Worksheet (attached below).
Phantom Dose Chart | ||
Patient Dose | Dose A (mCi) | Dose B (mCi) |
4 mCi | 0.14 | 0.33 |
6 mCi | 0.21 | 0.50 |
8 mCi | 0.28 | 0.66 |
10 mCi | 0.35 | 0.83 |
12 mCi | 0.42 | 0.99 |
14 mCi | 0.49 | 1.15 |
16 mCi | 0.56 | 1.32 |
18 mCi | 0.63 | 1.48 |
20 mCi | 0.70 | 1.65 |
Directions for Activating Phantom and Vials |
Protocol Summary for the Two Required Doses:
|
Data Acquisition and Processing |
PositioningIf the phantom does not have a flange on the top, place the phantom on its side at the end of the imaging table. Carefully align the phantom parallel to the axis of the table (left to right). Use a bubble level to position the phantom in the horizontal plane. A folded 3 x 5 card positioned under the end of the phantom can be used to make it level. If a metal plate is present, move the phantom to a position where the plate does not attenuate the photons.If the phantom has a flange at the top, it should be placed on the table with its flange hanging over the edge. Carefully align the phantom so it is parallel to the axis of the table. If the imaging table has a metal plate at the end, the phantom should be moved to a position where the plate does not attenuate the photons. Material must be placed under the bottom of the phantom, to make it level. To avoid excessive attenuation, the material should only be under the bottom (rod side) near the end of the phantom, not the entire length of the phantom. If the table is continuous, material must also be used to level the phantom. |
AcquisitionUse your facility’s whole-body protocol (zoom =1) with the same settings that are used for routine clinical studies. In order to generate SUV measurements, the facility will need to enter the administered activity (refer to the table in the above section of this article)/assay time/residual and weight. For the weight, enter 70 kg for the phantom. |
ReconstructionReconstruct the entire phantom with the same protocol used for whole body scans including pre- and post-reconstruction filters. Generate 1 cm thick transaxial slices for analysis (if unit is only being accredited for brain or cardiac, use the appropriate protocol for reconstruction). All acquisition and reconstruction parameters should be recorded on the Acquisition and Reconstructions Parameters form attached below. |
Attenuation CorrectionAll images must be corrected for attenuation with the same protocol applied to patient data. Record the parameters on the Acquisition and Reconstruction Parameters form attached below. |
SUV AnalysisPASS/FAIL CRITERIA for SUV Measurements
|
ROI Analysis of 1 cm Transaxial Slice from PET Phantom
Regions-of-interest selection: use the minimum, maximum and mean SUV statistics from these regions for the analysis section. See the Phantom Dilution Worksheet (attached below) for SUV parameter instructions.
A. Select the “best” 1 cm central slice (around the mid-point of the vials) showing “hot” cylinders (A, B, C & D). Example: 
B. Draw a background circular region (ROI) in the center of the slice as shown here (diameter = 6 to 7 cm). Small variations in size or location of this ROI are not important. Next draw a circular ROI just inside the largest hot cylinder found in the slice (the ROI is shown in white over the 25 mm cylinder).
C. Place copies of the smaller circular ROI (as shown in B) over the other visible objects in the phantom slice as shown here. The circular ROIs should be inside the Teflon, water, and air regions. All ROIs should be 25 mm in diameter. Note that all four of the hot cylinders may not be observed; only the visible cylinders require ROIs. Make a screen capture of the final ROIs with SUV values.
Images for Submission
Transaxial Slices for Phantom Scan: Prepare images that show all non-interleaved transaxial slices (1 cm thick) of the phantom using the same matrix size as the clinical protocol. Submitted images must be in DICOM format. (DICOM screen capture is not allowed).
Use linear mapping of the display with the lower threshold set at zero and the upper threshold set to the maximum count (or whatever is satisfactory to produce a good gray scale). When there is linear mapping in an image, an ROI with twice the number of counts will have twice the intensity. Label images appropriately.
ROI Images from Phantom Scan: Prepare an image that shows the “best” 1 cm central slice (around the mid-point of the vials) of the hot cylinders from the phantom scan with the final ROIs image as described above. The ROI images should display the SUV measurements (mean, max, min), or a table included showing the SUV measurements. ROI image(s) with SUV measurements may be in DICOM format or screen capture (jpg, bmp, tif, etc.). See image example below. For more image examples, please refer to the NM/PET Phantom Image Atlas available in our Testing Overview: NM/PET article.
Additional Image Upload Instructions
* The ACR uses the NilRead viewer to view images uploaded for accreditation. The instructions below are in regard to using the NilRead viewer.*
To verify the file type is DICOM, open the phantom image with the NilRead viewer and draw an ROI (see the toolbar at the top of the NilRead viewer as shown in the image below).
If the correct DICOM file type has been uploaded, the ROI values should be similar to the data form entries. If the ROI measurements are not seen (see example below), please contact your vendor applications and/or PACs/IT department for assistance as the images are not in the correct DICOM format.
When verifying slice thickness in the NilRead viewer, click on the DICOM popup tool in the upper left corner of the NilRead toolbar to display the DICOM attributes. Type the word "Thickness" in the "Name" column to find the slice thickness values.
Slice thickness can also be verified in ACRedit Plus after the images have been uploaded (as seen in the image below). The phantom is 20 cm in length. With the 1 cm slice thickness requirement, the number of slices should be 20 (+/-2).
The slice thickness is also displayed on the bottom left of the image set.
Evaluation of the Submitted Images
Uniformity and noise are evaluated qualitatively by inspection of reconstructed tomographic sections. Optimal density ranges should be comparable to those used for clinical images. Spatial resolution is judged by identifying the smallest “cold” rods in the ACR-approved phantom and lesion detectability is determined from the “hot” cylinders using the region-of-interest protocol above. The same protocol is used for the “cold” cylinders that demonstrate the effectiveness of the attenuation and scatter corrections.
 | Previous: Clinical Image Testing: PET | Next: Phantom Criteria |  |