MRI ACR Phantom Quality Control: A Complete Testing Guide

Introduction

The ACR MRI accreditation phantom is the central tool of a defensible MRI quality control program, supporting seven standardized image-quality tests plus system-level checks that are run weekly by a technologist and comprehensively each year by a qualified MR medical physicist. A complete program ties each measurement to a defined ACR pass criterion and action level, documents the trend over time, and feeds the clinical and phantom images that the American College of Radiology reviews for accreditation. ¹²

Unlike CT or mammography, MRI does not produce ionizing radiation, so there is no NRC, state radiation-machine, or MQSA dose limit driving the QC schedule. Instead, image quality is the deliverable: geometric fidelity, spatial resolution, slice accuracy, uniformity, freedom from ghosting, and the ability to detect low-contrast detail. The ACR phantom turns those abstract qualities into reproducible, numeric measurements that a technologist and a physicist can both perform and trend. ¹³

This guide explains how the ACR MRI Quality Control Program is structured, what each of the seven phantom tests measures, the system-level checks layered on top of them, the pass/fail criteria and action levels, the accreditation submission workflow, and the failure modes that most often trip up a clinical site. DRPS performs annual MRI surveys and supports QC program design as part of its accreditation support and medical physicist consulting services across Florida, Maryland, Virginia, Washington DC, California, and Nevada.

Topic Explanation

What is the ACR MRI Quality Control Program?

The ACR MRI Quality Control Program is a tiered structure that assigns routine image-quality testing to a trained technologist and comprehensive performance evaluation to a qualified MR medical physicist or MR scientist, all anchored to the ACR accreditation phantom. It is not a single annual checkbox. It is a continuous program with weekly technologist QC, periodic quick checks, an annual physicist survey, and a three-year accreditation cycle. ¹²

The program has three roles working together:

• The MRI technologist runs the weekly phantom QC, records the results, performs frequent quick checks (such as central-frequency verification, table positioning, and a visual artifact review), and flags any result that crosses an action level.
• The qualified MR medical physicist or MR scientist establishes the QC program, sets baseline values, performs or oversees the comprehensive annual evaluation, reviews the technologist QC trend, and signs the performance report used for accreditation.
• The supervising physician and facility ensure corrective action is taken when QC fails and that the program is documented for ACR review.

The ACR phantom is the shared reference object for all of this. Because every accredited site images the same standardized phantom with prescribed sequences, results can be compared against fixed acceptance criteria rather than against each scanner's own history alone. This is the same accreditation philosophy described in our overview of ACR accreditation physics requirements.

What is the ACR MRI phantom?

The ACR MRI phantom is a fluid-filled cylinder containing precisely positioned internal structures designed to make each image-quality property measurable. It is supplied in a large version (the original accreditation phantom, roughly 190 mm inner length and 190 mm inner diameter) and a small version intended for small-bore systems, extremity coils, and small field-of-view applications. The structures inside include length and diameter references for geometry, resolution insert arrays, ramps and wedges for slice measurement, a large uniform region for uniformity and ghosting, and low-contrast disk arrays arranged as spokes. ¹³

A basic ACR phantom QC review starts with a few questions:

• Was the phantom positioned and landmarked correctly, with the correct coil and prescribed sequences?
• Do the geometry, resolution, slice, uniformity, ghosting, and low-contrast measurements meet ACR criteria?
• Has anything drifted relative to the established baseline or to the prior weekly results?
• If a test fails, is the cause a phantom setup error, a coil problem, a gradient or B0 issue, or a genuine system fault?

Those questions map directly onto the seven ACR tests and the system-level checks that follow.

How do the ACR MRI tests fit together?

The seven phantom tests address image quality as the end user experiences it, while the system-level checks address the underlying hardware that produces that image quality. The table in the next section summarizes each ACR test, what it measures, a representative pass criterion, and the typical testing frequency.

Key Technical Principles

The seven ACR phantom tests and system checks

ACR test	What it measures	Representative pass criterion	Typical frequency
Geometric accuracy	Spatial/dimensional fidelity (lengths, diameters)	Measured lengths within ±2 mm of true value (large phantom: ±3 mm)	Weekly + annual
High-contrast spatial resolution	Smallest resolvable hole array (in-plane resolution)	Resolve the 1.0 mm hole array (upper-left and lower-right)	Weekly + annual
Slice-thickness accuracy	Accuracy of the prescribed slice thickness	Measured 5.0 mm slice within ±0.7 mm	Weekly + annual
Slice-position accuracy	Position error from the prescribed slice (ramp wedges)	Bar-length difference ≤ 5 mm (≤ 4 mm recommended)	Weekly + annual
Image intensity uniformity (PIU)	Flatness of signal across a uniform region	PIU ≥ 87.5% below 3 T (fail if < 85%); PIU ≥ 82% at 3 T (fail if < 80%)	Weekly + annual
Percent-signal ghosting (PSG)	Ghost artifact level outside the phantom	Ghosting ratio ≤ 0.025 (2.5%); accreditation fail if > 0.030	Weekly + annual
Low-contrast object detectability (LCOD)	Visibility of low-contrast disks (spokes)	Required total spokes by field strength: < 1.5 T ≥ 7; 1.5 T to < 3 T ≥ 30 (T1) / ≥ 25 (T2); 3 T ≥ 37	Weekly + annual
System checks (SNR, central frequency/B0, transmit gain, gradient calibration, RF/coil)	Underlying hardware performance and stability	Stable relative to baseline; within physicist-defined action levels	Annual (physicist); some quick checks more often

The criteria above are representative starting points written to illustrate the structure of the program. The authoritative, field-strength-specific tolerances and action levels are defined in the current ACR MRI Quality Control Manual and the ACR MRI Accreditation Program Requirements, and every site should test against the current published values, not against approximate figures. ¹²

Geometric accuracy

Geometric accuracy verifies that the scanner reproduces true spatial dimensions, which depends on gradient calibration and B0 homogeneity. The technologist measures known internal lengths and diameters in the phantom and compares them to the true values. Geometric distortion outside tolerance usually points to gradient miscalibration or field inhomogeneity, and it matters clinically wherever measurements drive decisions, such as stereotactic planning, oncology measurements, and orthopedic sizing. Automated geometry analysis has been shown to agree with physicist reference measurements within sub-millimeter limits, tighter than typical manual technologist measurements. ⁴

High-contrast spatial resolution

High-contrast spatial resolution is assessed visually using a resolution insert containing arrays of small holes at decreasing spacings. The technologist identifies the smallest hole array in which the rows and columns are fully resolved. This characterizes the in-plane resolution achievable with the prescribed sequence and reflects the combined effect of field of view, matrix, and system performance.

Slice-thickness accuracy

Slice-thickness accuracy uses crossed ramps within the phantom. The displayed length of the signal ramp is proportional to the slice thickness, so the measured ramp lengths are converted into an effective slice thickness and compared to the prescribed value. A commonly used relationship for the two crossed ramps is:

$$\text{Slice thickness}=0.2\times \frac{(\text{top}\times \text{bottom})}{(\text{top}+\text{bottom})}$$

where "top" and "bottom" are the measured signal-ramp lengths and the 0.2 factor reflects the ramp geometry. The prescribed 5.0 mm slice should measure within ±0.7 mm; a larger deviation indicates an RF or gradient slice-selection problem.

Slice-position accuracy

Slice-position accuracy uses paired wedges (ramps) that produce vertical bars whose relative lengths indicate how far the actual slice is displaced from the prescribed location. The difference in bar lengths is read directly as a position error. This test detects landmarking and slice-prescription errors and confirms that the localizer-to-slice geometry is faithful.

Image intensity uniformity (PIU)

Image intensity uniformity quantifies how flat the signal is across a large uniform region of the phantom. The ACR percent integral uniformity (PIU) is computed from the maximum and minimum mean signal in small regions of interest placed within a large region of interest covering most of the uniform area: ¹⁵

$$\text{PIU}=\left(1-\frac{S_{\max }-S_{\min }}{S_{\max }+S_{\min }}\right)\times 100\%$$

A perfectly uniform image gives PIU = 100%. Real images fall below that because of coil sensitivity profiles, B1 inhomogeneity, and gradient nonuniformity. ACR sets a minimum acceptable PIU that depends on field strength. For the large phantom, systems below 3 T should have PIU ≥ 87.5% and fail accreditation if PIU is less than 85%, while 3 T systems should have PIU ≥ 82% and fail if PIU is less than 80%. Surface coils and parallel-imaging reconstructions tend to lower uniformity. Automated PIU measurement has been shown to agree well with manual evaluation across multiple scanners, and validated automated tools can replace manual analysis of ACR phantom images. ⁵¹⁰

Percent-signal ghosting (PSG)

Percent-signal ghosting characterizes the faint replica of the phantom that appears outside its true boundaries, usually along the phase-encode direction, caused by motion, gradient instability, or RF problems. The ACR ghosting ratio compares mean signal in background regions of interest placed in the ghosting direction and the perpendicular direction to the mean signal inside the phantom: ¹⁵

$$\text{Ghosting ratio}=\left|\frac{(\text{top}+\text{bottom})-(\text{left}+\text{right})}{2\times S_{\text{phantom}}}\right|$$

where each bracketed term is the mean of background regions of interest in the corresponding direction and $S_{\text{phantom}}$ is the mean signal inside a large region of interest within the phantom. ACR sets a low ceiling on this ratio: the ghosting ratio should be ≤ 0.025 (2.5%), and images submitted for accreditation fail if the ratio exceeds 0.030 (3.0%). Ghosting that exceeds the limit points to RF instability, eddy currents, or gradient hardware issues, and the measured value is sensitive to exactly where the background regions of interest are placed. ⁵

Low-contrast object detectability (LCOD)

Low-contrast object detectability measures the faintest detail the system can show. The phantom contains slices of disk arrays arranged as ten spokes, each spoke containing three disks of decreasing diameter at a fixed low contrast. The reader counts the number of complete spokes visible across the relevant slices and sums them. ACR requires a minimum total number of visible spokes that depends on field strength, because higher field generally yields more signal and thus more visible spokes, and the limits were raised in mid-2021. Per the current ACR Large and Medium Phantom Test Guidance, the passing total spoke counts are: below 1.5 T, at least 7 spokes on each series (9 recommended); 1.5 T to below 3 T, at least 30 on the T1 series and at least 25 on the T2 series; and 3 T, at least 37 on both series. LCOD is the test most directly tied to SNR and is the most reader-dependent of the seven. ¹

System-level checks: SNR, B0, transmit gain, gradients, and coils

The annual physicist evaluation adds system-level checks that explain why the image-quality tests pass or fail:

• Signal-to-noise ratio (SNR). SNR is the fundamental currency of MRI image quality. A common two-acquisition method computes SNR from the mean signal in a region of interest in the phantom and the standard deviation of the noise estimated from the difference of two identical acquisitions: ⁶⁷

$$\text{SNR}=\sqrt{2}\times \frac{S_{\text{ROI}}}{{\sigma }_{\text{diff}}}$$

where $S_{\text{ROI}}$ is the mean phantom signal and ${\sigma }_{\text{diff}}$ is the standard deviation in a region of interest of the subtraction image. The $\sqrt{2}$ corrects for the added noise variance of subtracting two images. The NEMA MS 1 standard defines the reference SNR methodology. ⁶

• Central frequency and B0 stability. The Larmor (central) frequency tracks the main field. Drift in central frequency signals B0 instability and can degrade fat suppression, spectroscopy, and EPI-based sequences. It is one of the quick checks performed frequently.
• Transmit gain (reference amplitude). The RF transmit calibration sets the flip angle. Drift in transmit gain changes contrast and can subtly alter uniformity and SNR even when nothing else has visibly failed.
• Gradient calibration. Gradient amplitude calibration underlies geometric accuracy and slice fidelity; miscalibration shows up first as geometric distortion or slice-thickness error.
• RF coil and coil-element checks. Per-channel coil SNR and element integrity (per NEMA MS 6 methodology) catch a failing coil element before it produces a clinical artifact. A degraded element often presents as a uniformity or SNR failure on the phantom. ⁸

These hardware checks are why the physicist survey is more than a repeat of the weekly QC: the technologist tests confirm the image is acceptable, while the physicist tests confirm the machine that produced it is healthy and stable.

Clinical Impact

ACR phantom QC is not an administrative ritual; each failing test corresponds to a specific way clinical images can mislead. Geometric distortion corrupts measurements used for surgical and radiation-therapy planning. Poor uniformity can mimic or mask pathology, especially with surface coils, and undermines quantitative and fat-suppressed imaging. Ghosting overlays artifactual structure on anatomy. Low-contrast failures mean subtle lesions that should be visible are not.

Because the ACR phantom is imaged with standardized sequences, a stable QC trend gives real assurance that the clinical protocols sharing that hardware are performing as expected. When weekly QC drifts, it is an early warning that something in the signal chain (coil, gradient, B0, or RF) has changed before patients are affected. A multi-unit imaging center study using the ACR phantom found that uniformity at 3 T and slice-thickness accuracy were among the parameters most likely to fall near or outside acceptance limits, illustrating that even well-run fleets benefit from systematic phantom testing. ³

QC results also support the broader MRI program. A scanner that passes image-quality QC still needs a separate, robust MRI safety program covering zoning, screening, and projectile control. QC and safety are complementary: one protects the image, the other protects the patient and staff in the magnetic environment.

Practical Optimization Tips

A reliable ACR MRI QC program follows the same disciplined workflow each time.

1. Standardize phantom setup

• Use the correct phantom (large vs small) and the prescribed coil.
• Position and landmark the phantom exactly as specified; setup errors are a leading cause of spurious failures.
• Acquire the ACR series and the site clinical series with the established sequences and parameters.

2. Establish and protect baselines

• Have the MR medical physicist set baseline values at acceptance and after major service.
• Re-baseline after coil replacement, gradient service, software upgrades, or a magnet ramp, because a stale baseline can hide a real change.

3. Trend, do not just spot-check

• Record every weekly result and review the trend, not only the latest pass/fail.
• A parameter drifting toward its action level is actionable before it crosses the limit.

4. Place regions of interest consistently

• For PIU, PSG, and SNR, keep region-of-interest size and placement consistent; these measurements are sensitive to placement, and inconsistent ROI positioning can create apparent changes that are really measurement noise. ⁵
• Consider validated automated analysis to reduce operator dependence and improve reproducibility. ⁴⁵

5. Diagnose before re-testing

When a test fails, identify the likely cause rather than simply re-running:

• Uniformity failure → coil element, B1, or parallel-imaging reconstruction issue.
• Geometric or slice-thickness failure → gradient calibration or B0.
• Ghosting failure → RF instability, eddy currents, gradient hardware, or motion.
• Low-contrast failure → low SNR (coil, noise source, or sequence change).

Common pitfalls to avoid

• Treating setup errors as machine failures. A mispositioned phantom or wrong coil fails the test without anything being wrong with the scanner.
• Ignoring the trend. A single in-tolerance reading can still be part of a clear downward drift.
• Inconsistent ROI placement. PIU, PSG, and SNR depend on where regions of interest are drawn. ⁵
• Skipping re-baselining after service. New hardware needs a new reference.
• Relying on weekly QC alone. Weekly tests confirm image quality; the annual physicist survey confirms hardware health and stability.

Regulatory Considerations

MRI QC is governed by accreditation and professional standards rather than radiation-machine regulation, because MRI uses no ionizing radiation. There is no NRC, FDA radiation-emitting-device dose limit, state radiation-machine inspection, or MQSA equivalent driving MRI QC. Instead, the binding framework is accreditation and the facility's own quality and safety standards. ¹²

Key frameworks to reference:

• ACR MRI Quality Control Program / ACR MRI Quality Control Manual — defines the weekly technologist QC, the annual physicist evaluation, the seven phantom tests, and the field-strength-specific acceptance criteria. ¹
• ACR MRI Accreditation Program Requirements — defines the personnel qualifications, the phantom and clinical image submission, and the three-year accreditation cycle. ²
• ACR–AAPM Technical Standard for Diagnostic Medical Physics Performance Monitoring of MRI Equipment — defines the scope and conduct of the qualified physicist's evaluation. ⁹
• NEMA MS standards — provide the underlying measurement methodologies, including NEMA MS 1 for SNR, NEMA MS 3 for image uniformity, and NEMA MS 6 for RF coil SNR. ⁶⁷⁸
• The Joint Commission diagnostic imaging and MRI safety standards — apply to accredited hospitals and impose MRI safety and QC expectations alongside ACR.

Of the states DRPS serves, Florida, Maryland, Virginia, California, and Nevada are NRC Agreement States and Washington DC is regulated directly by the NRC, but those authorities govern radioactive material and X-ray machines, not MRI image quality. MRI accreditation and QC obligations flow from ACR, the accrediting and payer requirements the facility has accepted, and Joint Commission standards where applicable. A facility should confirm which accrediting body and payer rules apply to its scanners and align the QC program accordingly. For the broader accreditation context across modalities, see our guide to ACR accreditation physics requirements.

Frequently Asked Questions (FAQs)

What is the ACR MRI phantom QC program?

It is the standardized MRI quality control structure built around the ACR MRI accreditation phantom. A trained technologist performs weekly QC on the phantom, and a qualified MR medical physicist performs an annual performance evaluation. Together they verify geometric accuracy, resolution, uniformity, ghosting, low-contrast detectability, and system-level performance against ACR criteria.

What are the seven ACR MRI phantom tests?

Geometric accuracy, high-contrast spatial resolution, slice-thickness accuracy, slice-position accuracy, image intensity uniformity (percent integral uniformity), percent-signal ghosting, and low-contrast object detectability. The annual physicist survey adds system-level checks such as SNR, central-frequency/B0 stability, transmit gain, and RF coil performance.

How often is ACR MRI QC performed?

Core image-quality QC on the phantom is performed weekly by a designated technologist, with several quick checks (such as central frequency, table positioning, and visual artifact review) performed more frequently per the program. A qualified MR medical physicist performs a comprehensive annual evaluation, and ACR accreditation runs on a three-year cycle.

What is percent integral uniformity (PIU)?

PIU is the ACR measure of image intensity uniformity. It compares the maximum and minimum mean signal in small regions of interest inside a large uniform region of the phantom. Higher PIU means a flatter, more uniform image. ACR sets minimum PIU action levels that depend on field strength.

What is the most common reason a site fails ACR MRI QC?

Frequent failure modes include image intensity uniformity below the field-strength-specific limit, geometric distortion outside tolerance from gradient miscalibration, ghosting from RF or gradient instability, and low-contrast detectability below the required number of visible spokes. Many failures trace back to coil problems, B0 drift, or gradient calibration.

Who is qualified to perform the annual ACR MRI physicist survey?

ACR requires a qualified MR medical physicist or MR scientist who meets the program's education, board-certification, and continuing-experience criteria. DRPS uses board-certified medical physicists for annual MRI surveys across Florida, Maryland, Virginia, Washington DC, California, and Nevada.

Does MRI QC require a radiation safety review?

MRI does not use ionizing radiation, so it is not regulated under NRC or state radiation-machine rules. However, ACR accreditation, Joint Commission MRI safety standards, and the facility MRI safety program still apply. QC and safety are complementary parts of a defensible MRI program.

Key Takeaways

• The ACR phantom is the backbone of the program. Seven standardized tests turn abstract image-quality properties into reproducible numbers comparable against fixed ACR criteria.
• Two tiers, two purposes. Weekly technologist QC confirms the image is acceptable; the annual physicist survey confirms the hardware producing it is healthy and stable.
• PIU, PSG, and SNR are calculable, not just visual. Each has a defined ACR or NEMA formula, and each is sensitive to how regions of interest are placed.
• Field strength changes the criteria. Minimum PIU and the required low-contrast spoke count depend on field strength, so always test against the current ACR values.
• Failures point to specific hardware. Uniformity to coils/B1, geometry and slice to gradients/B0, ghosting to RF and gradient stability, low contrast to SNR.
• QC is accreditation-driven, not radiation-regulated. MRI QC obligations flow from ACR, payers, and Joint Commission, not from NRC or state radiation-machine rules.

Conclusion

ACR MRI phantom QC works because it makes image quality measurable and repeatable. The seven phantom tests, run weekly by a trained technologist and reviewed annually by a qualified MR medical physicist, give a facility an objective, trendable picture of geometric accuracy, resolution, slice fidelity, uniformity, ghosting, and low-contrast detectability. The system-level checks layered on top, including SNR, B0 stability, transmit gain, gradient calibration, and coil performance, explain why those image-quality results pass or fail and catch hardware drift before it reaches patients.

A strong MRI QC program is not about generating phantom images for a binder. It is about establishing baselines, trending every result, diagnosing failures by likely cause, and re-baselining after service so the data stays meaningful. Facilities that treat ACR phantom QC as a continuous, physicist-supported process are better positioned to pass accreditation, defend their image quality, and protect the diagnostic value of every MRI exam.

How DRPS Can Help

Diagnostic Radiation Physics Services helps MRI facilities build and sustain a defensible ACR quality control program. This includes the annual qualified MR medical physicist evaluation, QC program design and baseline establishment, technologist QC training and review, ACR phantom-image and accreditation-submission support, failure-mode troubleshooting, and coordination of QC with the facility MRI safety program. DRPS uses board-certified medical physicists for MRI surveys and accreditation support.

DRPS supports facilities across our service locations, including Florida, Maryland, Virginia, Washington DC, California, and Nevada. If your scanner is approaching its accreditation renewal, adding a new coil or sequence, or showing drifting QC results, contact us to schedule a survey.

A strong MRI QC program makes the right answer the easy answer: stable, documented image quality that the clinical team and the accrediting body can both trust.

Related Resources

• ACR accreditation physics requirements
• MRI safety program essentials
• SMPTE display monitor QC
• Mammography quality control under MQSA
• Diagnostic ultrasound QC
• Medical physicist consulting
• Accreditation support

References

1. American College of Radiology. ACR MRI Quality Control Manual. Reston, VA: American College of Radiology. acraccreditation.org
2. American College of Radiology. MRI Accreditation Program Requirements. accreditationsupport.acr.org
3. Ihalainen TM, Lönnroth NT, Peltonen JI, et al. MRI quality assurance using the ACR phantom in a multi-unit imaging center. Acta Oncologica. 2011;50(6):966-972. doi:10.3109/0284186X.2011.582515. doi.org
4. Panych LP, Chiou JG, Qin L, Kimbrell VL, Bussolari L, Mulkern RV. On replacing the manual measurement of ACR phantom images performed by MRI technologists with an automated measurement approach. Journal of Magnetic Resonance Imaging. 2016;43(4):843-852. doi:10.1002/jmri.25052. doi.org
5. Epistatou AC, Tsalafoutas IA, Delibasis KK. An automated method for quality control in MRI systems: methods and considerations. Journal of Imaging. 2020;6(10):111. doi:10.3390/jimaging6100111. doi.org
6. National Electrical Manufacturers Association. NEMA Standards Publication MS 1: Determination of Signal-to-Noise Ratio (SNR) in Diagnostic Magnetic Resonance Imaging. Rosslyn, VA: NEMA. nema.org
7. National Electrical Manufacturers Association. NEMA Standards Publication MS 3: Determination of Image Uniformity in Diagnostic Magnetic Resonance Imaging. Rosslyn, VA: NEMA. nema.org
8. National Electrical Manufacturers Association. NEMA Standards Publication MS 6: Determination of Signal-to-Noise Ratio and Image Uniformity for Single-Channel, Non-Volume Coils in Diagnostic MR Imaging. Rosslyn, VA: NEMA. nema.org
9. American College of Radiology, American Association of Physicists in Medicine. ACR–AAPM Technical Standard for Diagnostic Medical Physics Performance Monitoring of Magnetic Resonance Imaging (MRI) Equipment. acr.org
10. Alaya IB, Mars M. Automatic analysis of ACR phantom images in MRI. Current Medical Imaging. 2020;16(7):892-901. doi:10.2174/1573405615666190903145343. doi.org