Revision History
Between four and seven examinations per unit are required for accreditation. The exact number of examinations depends on the number of modules for which the unit is used. The facility may select which examinations it will submit for accreditation evaluation based on the information below. For details regarding exam-specific parameters, click on the relevant module in the table.
Examination Choices for MR Accreditation by Module | |||
Head/Neck | Spine | MSK | |
|
|
| |
Body | MRA | Cardiac | Breast |
|
|
|
|
The number of required examinations is based on the number of modules your site selected on the application:
Number of Modules on Application | Number of Examinations Per Module | Total Exams |
1 module | 4 examinations are required | 4* |
2 modules | Select 2 examinations from each module | 4 |
3 modules | Select 1 examination from each module, and select 1 additional examination from any of the 3 modules | 4 |
4 modules | 1 examination from each module | 4 |
5 modules | 1 examination from each module | 5 |
6 modules | 1 examination from each module | 6 |
7 modules | 1 examination from each module | 7 |
*If the unit performs and applies for the cardiac or breast modules only, a total of 2 cardiac exams or 2 MR breast exams with known, enhancing biopsy-proven carcinoma are required
The requirements used by the accreditation reviewers represent a technical baseline for producing acceptable diagnostic examinations. Prior to submission of any images for evaluation, the interpreting physicians and technologists at your facility must review the accreditation criteria detailed here. Although some criteria for MRI examinations are requirements for accreditation, others are only intended as a guide; technique parameters mentioned are only suggestions unless otherwise stated. The sequences required for accreditation submission should not be construed as a complete clinical exam.
The pulse sequences that are used clinically for examinations of different body regions are variable due to the personal preferences of users as well as the capabilities of the different MRI systems. Despite this variability, experienced interpreting physicians are able to agree on what constitutes “acceptable” and “unacceptable” diagnostic exams based on both objective and subjective criteria. The intention of accreditation is to provide guidelines on what constitutes optimal image quality above that which is normally acceptable and to promote the best practice at all times.
The categories for scoring examinations submitted for ACR MR Accreditation are:
A. Pulse Sequence and Image Contrast
B. Anatomic Coverage and Imaging Planes
C. Spatial and Temporal Resolution
D. Artifacts
E. Exam Identification
Category A: Pulse Sequences and Image Contrast
The type of pulse sequence (e.g. conventional SE, multishot RARE or gradient echo) and the precise imaging parameters (e.g. TR, TE, FA, ETL, etc.) are not specified and are left to the discretion of the imaging facility unless otherwise stated.
Our articles on examination-specific parameters detail the sequences considered to be the minimum necessary for a quality examination. IF ANY OF THESE SEQUENCES IS NOT SUBMITTED, THE EXAMINATION WILL FAIL. For accreditation review, the facility should submit only the required sequences and their corresponding reconstructed images. You must submit localizer or scout sequences with cross-reference locations for specific clinical examination as noted in the exam specific testing instructions.
All sequences must demonstrate sufficient Signal to Noise (SNR), and not appear too grainy. If contrast is required, it is very important that patient selection is appropriate for the examination using contrast. Please refer to the ACR Quality and Patient Safety/MR Safety web page for more information on IV contrast safety.
Category B: Anatomic Coverage and Imaging Planes
Proper anatomic coverage and imaging planes are important components of clinical MRI exams. The minimum sets of images required for each examination and the anatomy to be included on those images are listed in our articles on examination-specific parameters. FAILURE TO MEET MINIMUM COVERAGE SPECIFICATIONS WILL RESULT IN FAILURE FOR THAT EXAMINATION.
Category C: Spatial/Temporal Resolution
The spatial resolution necessary for quality MRI images varies by examination and sequence. MRI facilities must use the determinants and formulas listed below to determine the spatial resolution of their clinical MRI examinations. The five determinants of pixel/voxel dimensions in an MRI examination are:
Slice thickness (ST)
Field of view along the phase encode direction (FOVp)
Field of view along the frequency encode direction (FOVf)
Number of phase encoding steps (Np) (This is your phase matrix)
Number of frequency encoding steps (Nf) (This is your frequency or read matrix)
Your images will be scored on acquisition parameters, not interpolated parameters. Use the pixel/voxel dimensions from your scan protocols and the formulas below to calculate your in-plane pixel size in both the phase and frequency directions for all of the sequences you are submitting for accreditation review (see our articles on examination-specific parameters for required sequences). Compare the values calculated in the Clinical Test Image Data forms to the values listed in the articles on examination-specific parameters.
If you are using a rectangular FOV, your phase FOV will be different from your frequency FOV. This may also be true for your matrix. If you are not sure, consult your manufacturer.
Formulas for the Calculation of In-Plane Pixel Size | |
To Calculate... | Formula... |
Pixel size in the phase direction | FOVp/Np (field of view in the phase encoding direction divided by the number of steps in the phase encoding direction) |
Pixel size in the read or frequency direction | FOVf/Nf (field of view in the frequency encoding direction divided by the number of steps in the frequency encoding direction) |
Pixel area (for 2D sequences) | Pixel size in the read or frequency direction times the pixel size in the phase direction |
Voxel volume (for 3D sequences) | Pixel size in the read or frequency direction times the pixel size in the phase direction times the slice thickness |
The determinants of temporal resolution are:
Speed of frames per millisecond
Temporal resolution = msec/frames
For cine images, the number of views per segment (nvs) or segmentation factor also controls acquired temporal resolution
Note that most manufacturers use phase sharing (view sharing techniques) to increase the visual smoothness of the cine movies. Our articles on examination-specific parameters refer to temporal resolution before these view sharing techniques.
With view sharing, images that are acquired every 80 msec can be interpolated, so that the cine display shows a new image every 40 msec (or less). However, each image still contains 80 msec worth of data.
To determine the temporal resolution, use this formula: Temporal resolution (cine) = TR x NVS (Where NVS is the number of views per segment, or segmentation factor and TR is the intrinsic or minimum TR of the pulse sequence). Some manufacturers may not display this TR value. If in doubt, please contact your manufacturer’s application specialist.
Category D: Image Artifacts
Artifacts on any image may interfere with image interpretation. Although some artifacts may be unavoidable on certain images (e.g. susceptibility artifacts near sinuses on T2 weighted brains); others may be indicators of inadequate equipment or lack of preventive maintenance at an MRI facility. The artifacts listed in the table below are among the most common. All of the images should be assessed to determine if any of these artifacts are present, especially if they could potentially compromise the diagnostic value of the images. Your examinations will be reviewed for excessive artifacts that may interfere with image quality.
Common Image Artifacts | |
Aliasing | The image appears wrapped around into itself. This is due to a large body portion included in a too small FOV. |
Parallel Imaging | Mismatches between the anatomy on calibration images and diagnostic images appear as chemical shift, motion, ghosting and misregistration along the phase-encoding direction in the middle of the FOV. |
Truncation (Edge Ringing) | Periodic parallel lines or ringing adjacent to borders or tissue discontinuity, in either the phase and/or frequency encoding directions. This is due to a small matrix. |
Black Boundary (India Ink) | Well-defined black contours outlining regions of MR anatomy, without corresponding anatomical structure. |
Heterogeneous Brightness (Shading) | This is due to RF heterogeneity, improper patient positioning, or metal in the magnet or on the patient. |
Heterogeneous Fat Suppression | Uneven darkening of the fat signal in different portions of the image set. This may be due to either a heterogeneous magnetic field or a heterogeneous RF field. |
Susceptibility | Localized field distortion or non-uniformities produced by differing tissue magnetic susceptibility (especially at air-tissue interfaces). |
Chemical Shift | Occurs along the frequency encoding axis at fat/water soft tissue interfaces as a thin intense band of high signal or low signal. |
Ghosting | Periodic replication of partial copies of images of the original structure along the phase encoding axis due to motion. It includes artifacts from swallowing (C-spine), respiration and peristalsis (L-spine), CSF pulsation (brain and spine), vascular pulsation (brain and knee) and cardiac motion (T-spine). |
Geometric Distortion | Size, orientation or shape is not accurately represented on the image. |
Excessive Filtering | Excessive smoothing using software to reduce apparent noise in the image. Excessive filtering or smoothing obscures true anatomical structure and/or contrast. |
Misregistration of 2D Images | Consecutive 2D images do not line up so some anatomy is skipped and other regions are imaged twice. This can also be a particularly serious problem on 2D time-of-flight MRA MIPs. |
Misregistration of Subtracted Images | On subtracted images, there is incomplete subtraction of the background tissue signal with prominent signal at edges that do not align properly. |
Ringing | Accentuation of edges due to either under sampling of k-space (not enough phase encoding steps) or at the leading edge of the bolus on an enhanced 3D MRA study due to IV contrast being present during acquisition of peripheral k-space but not as much during acquisition of the center of k-space. |
Stair Step (Venetian Blind Effect) | In MRA, a vessel goes obliquely through slices, due to slice thickness and vessel size. Venetian blind occurs on multi slab MRA (typically on reformations and MIPs), when the adjacent slabs are not properly and seamlessly overlapped. |
Reformatting Artifacts | Improper MIP and reformations may give the false appearance of vessel occlusion or stenosis when it is only partially included in the MIP volume. Superimposed vessels may falsely appear stenotic on MIP due to stealing of voxels at the vessel edges. Stair step artifact may occur on oblique reconstruction when the slices are too thick or there is insufficient zero filling. |
ECG Lead Artifacts | The ECG leads used for cardiac gating should not produce excessive artifacts that would interfere with the interpretation of the image. |
RF Leak or "Zipper" Artifact | Linear hyperintensity parallel to the phase encoding direction often caused by unwanted sources of RF signals originating within (e.g., light bulb failure) or outside (e.g. inadequate RF shielding) the scanner room. |
Echo Train Blurring | Image blurring due to excessively long echo spacing and/or echo train length. |
Peripheral signal artifacts | Star artifact: A bright spot close to the image center originates very far from isocenter because FID signal from RF 180 pulse or SAT pulse is not crushed out and aliases back into image center.Annefact artifact: Smeared, bright, ribbon ghosting signals in the phase-encoding direction are uncompensated eddy currents that also originate far from isocenter where gradients are non-linear. |
Other | There are other artifacts that are not as common as those listed above but which may be important. |
Category E: Exam Identification
Patient and technical data must be displayed on the images or available in the DICOM header. Do NOT anonymize images. All patient information will be kept confidential by the ACR as stated in the Practice Site Accreditation Survey Agreement. If the parameters listed below in bold in the table below are not available to the reviewer, that examination will fail.
Exam Identification | |||
Each Exam | Each Sequence | Each Image | |
|
The following labels are not required, but are strongly recommended for each sequence:
|
|
Revision History for this Article | ||
Date | Section | Description of Revision(s) |
12-12-19 | All | Article created; FAQs incorporated; No criteria changes |
1-17-2020 | Exam Identification | Changed "patient and technical data must be displayed on the images or readily accessible in the DICOM header" to patient and technical data must be displayed on the images or available in the DICOM header". |
6-8-2020 | Exam identification | Added Do NOT anonymize images |
9-2-2021 | Category C: Spatial/Temporal Resolution | Updated link to clinical data form |
11-9-2022 | Added Breast module | |
8-4-2023 | Exam choices by module | removed black blood from the cardiac exam |
Category A: Pulse Sequence and Image Contrast | Updated submission of complete exam to submission of required sequences and corresponding reconstructed images | |
Removed requirement of scout/localizer images for all clinical exams. The requirement is now exam specific. | ||
8-10-2023 | Cardiac Module | Updated the name of the exam to Cardiac MR contrast enhanced |
2-2-2024 | First paragraph | Updated to "between four and seven examinations" |
Previous: Testing Overview: MRI | Next: Exam Specific Parameters: Head/Neck |