ACR CT Accreditation Phantom QC

Introduction

Passing the ACR CT accreditation phantom means proving four things at once — CT number accuracy, low-contrast resolution, image uniformity, and high-contrast resolution — at a radiation dose below the ACR reference limits. The phantom compresses those tasks into four stacked modules, and the ACR CT Quality Control Manual defines exactly how each is scored. Understanding what each module measures, and why, turns accreditation from a paperwork exercise into a genuine image-quality check. ¹²

CT accreditation is not optional for most facilities: it is tied to reimbursement and, in many jurisdictions, to the right to operate. But the deeper reason the phantom matters is that its four modules probe the exact image-quality parameters that determine whether a clinical CT is diagnostic — can it render the right Hounsfield units, resolve a subtle low-contrast lesion, present a flat uniform field, and resolve fine high-contrast detail. ⁵

This article walks through the phantom's construction, what each module measures, the pass criteria, the worked contrast-to-noise-ratio calculation at the heart of the low-contrast test, the dose limits, and the practical pitfalls that trip up otherwise-good scanners. DRPS provides CT physics testing and accreditation support across Florida, Maryland, Virginia, Washington DC, California, and Nevada.

Topic Explanation

What the ACR CT phantom is

The ACR CT accreditation phantom is a solid, water-equivalent cylinder containing four distinct testing modules, each dedicated to a different image-quality task. A facility scans the phantom using its own clinical protocols — adult head, adult abdomen, and pediatric protocols — and submits the images to the ACR, where they are scored against the criteria in the ACR CT Quality Control Manual. The phantom permits testing of CT number accuracy, slice thickness, low-contrast resolution, uniformity, in-plane distance accuracy, and high-contrast resolution in a single reproducible object. ⁵

Because the facility scans with clinical protocols rather than a special test technique, the phantom result reflects the image quality patients actually receive. That linkage — real protocols, real dose, scored image quality — is what gives ACR accreditation its teeth. For the broader accreditation picture, see our overview of ACR accreditation physics requirements.

The four modules

The phantom's four modules are scanned together and analyzed separately:

Module 1 — positioning, CT number, slice thickness, distance. Contains five materials of known composition (a bone-equivalent insert, polyethylene, water-equivalent material, acrylic, and air), plus alignment and distance features.
Module 2 — low-contrast resolution. Contains a series of low-contrast cylinders (nominally around 0.6 percent contrast) used to test the scanner's ability to distinguish objects that differ only slightly from background.
Module 3 — uniformity and noise. A uniform region used to measure the difference between center and peripheral CT numbers and the pixel noise.
Module 4 — high-contrast (spatial) resolution. Bar-pattern resolution gauges read out in line pairs per centimeter.

For a facility building an internal QC program around these modules, the same parameters appear in routine physics testing; see CT number and HU calibration QC and CT image quality: MTF and low contrast.

Key Technical Principles

Module 1: CT number accuracy

CT numbers, expressed in Hounsfield units (HU), are the quantitative backbone of CT. They are defined relative to water and air:

HU = 1000 \times \frac{μ _{material} - μ _{water}}{μ _{water} - μ _{air}}

By definition water is 0 HU and air is −1000 HU. The ACR phantom's Module 1 checks that the scanner reproduces the expected CT numbers of its five materials within tolerance. The approximate reference values are:

Material	Approximate reference CT number (HU)
Air	approximately −1000
Polyethylene	approximately −95
Water-equivalent	approximately 0
Acrylic	approximately +120
Bone-equivalent	approximately +955

The water-equivalent material must read near 0 HU within the acceptance window in the current ACR CT Quality Control Manual — on the order of a few HU either side of zero. ¹ Crucially, the measured CT numbers of the other materials are manufacturer- and kVp-dependent: published work on the ACR phantom shows that air and water values differ measurably between scanner makes, and that polyethylene and acrylic CT numbers rise with kVp while the bone value falls. A defensible QC program therefore defines tolerances that account for the scanner make and the kVp used, rather than applying one universal number. ⁶

Module 2: low-contrast resolution and the CNR

Low-contrast detectability is the hardest task in CT because the signal — a small difference in attenuation — competes directly with image noise. The metric is the contrast-to-noise ratio (CNR):

CNR = \frac{∣ C ˉ _{target} - C ˉ _{background} ∣}{σ _{background}}

where $\overset{ˉ}{C}_{target}$ is the mean CT number in a region of interest on a low-contrast rod, $\overset{ˉ}{C}_{background}$ is the mean CT number in a background region of the same module, and $σ_{background}$ is the standard deviation (noise) of the background region. ⁴

The ACR CT Quality Control Manual sets a minimum passing CNR near 1.0 for the adult abdomen protocol, and requires that the low-contrast cylinders be visible at that technique. ¹ In one published application of the ACR phantom, a well-behaved CT simulator produced CNR values ranging from about 1.02 to 2.05 with all four of the 6 mm low-contrast rods visible — comfortably above the minimum. ⁴

A worked CNR example

Suppose the low-contrast rod region of interest reads a mean of 6.4 HU, the background region reads a mean of 0.6 HU, and the background noise (standard deviation) is 4.8 HU. Then:

CNR = \frac{∣ 6.4 - 0.6 ∣}{4.8} = \frac{5.8}{4.8} = 1.21

A CNR of 1.21 passes the adult-abdomen minimum of 1.0, but only just. The example illustrates the central tension of low-contrast imaging: the contrast (5.8 HU) is fixed by the phantom, so the only lever the facility controls is the noise. Lowering dose raises noise, which lowers CNR; the accreditation limit therefore effectively couples image quality to dose. This is exactly the trade-off explored in CT protocol optimization and CT noise and task-based image quality.

Module 3: uniformity and noise

A CT image of a uniform object should have the same mean CT number everywhere. In practice, beam hardening, scatter, and reconstruction imperfections cause the periphery to differ from the center — the classic "cupping" or "capping" artifact. Module 3 quantifies this by comparing a central region of interest against four peripheral regions. The ACR criterion requires the difference between the center and each peripheral region to stay within a small window (on the order of ±5 HU) to pass. ¹ The same region is used to record image noise against the facility's baseline.

Module 4: high-contrast (spatial) resolution

High-contrast resolution measures the finest detail the scanner can resolve when contrast is abundant — the limiting spatial resolution. It is read from bar patterns in line pairs per centimeter (lp/cm). The ACR program requires resolving a minimum number of line pairs at the adult abdomen technique (typically on the order of 5 lp/cm), with higher requirements for high-resolution protocols. ¹ A well-performing CT commonly resolves 6 to 7 lp/cm on standard body protocols. ⁴

Why modern metrics are being added

Traditional CNR is increasingly limited on scanners using iterative and deep-learning reconstruction, because those algorithms make noise texture and low-contrast detectability nonlinear and dose-dependent. The AAPM Task Group 233 report defines task-based metrics — the task transfer function, the noise power spectrum, and a detectability index — that characterize modern CT performance more faithfully than a single CNR value. ⁴ These do not replace the ACR phantom criteria, but they explain why a scanner can pass the phantom and still benefit from a deeper physics evaluation, especially photon-counting and deep-learning-reconstruction systems.

Clinical Impact

The phantom mirrors clinical image quality

Because the facility scans the phantom with its clinical protocols, each module maps to a clinical failure mode. A CT-number error in Module 1 means quantitative tasks — coronary calcium scoring, urinary stone characterization, attenuation-based tissue analysis — are biased. A low-contrast failure in Module 2 means subtle lesions, such as a low-contrast liver metastasis, may be missed. A uniformity failure in Module 3 means CT numbers depend on position in the field, undermining any quantitative measurement. A high-contrast failure in Module 4 means fine anatomic detail, such as small bony or lung structures, is blurred. ⁵

Dose is scored alongside image quality

A scanner that produces beautiful images at excessive dose does not pass. The ACR program applies CTDIvol pass/fail limits, and a facility can fail for exceeding them even with excellent image quality. Recent multi-site data show the opposite problem is far more common: modern scanners routinely operate at CTDIvol well below the ACR reference values while exceeding the minimum CNR, which means most facilities have headroom to lower dose further without failing image quality. ¹¹ The phantom, in other words, is a floor for image quality and a ceiling for dose — and a good physics program works the space between them.

Practical Optimization Tips

1. Scan with the real clinical protocols

The phantom must be scanned with the same protocols used on patients — not a special high-dose technique that flatters the scanner. Using an inflated technique to pass accreditation defeats the purpose and misrepresents clinical image quality.

2. Center the phantom carefully

Miscentering biases CT numbers and uniformity through the bowtie filter and automatic exposure control, exactly as it does for patients. Use the alignment lights and the phantom's own positioning features; a tilted or off-center phantom can fail Module 1 or Module 3 for a reason that has nothing to do with the scanner's true performance. See CT patient centering and image quality.

3. Read CNR and noise together

A borderline CNR is almost always a noise problem, not a contrast problem — the phantom's contrast is fixed. Before raising dose, confirm the reconstruction kernel and slice thickness match the protocol's intent; a sharper kernel or thinner slice raises noise and can drop CNR below 1.0.

4. Match materials and kVp to your tolerances

Because CT numbers of the phantom materials shift with kVp and scanner make, set CT-number action limits per scanner and per kVp, and re-baseline after a tube replacement or a major calibration. ⁶

5. Trend, don't just pass

The value of the phantom is not the single pass/fail — it is the trend. Plot CT number, noise, uniformity, and CNR over time. A slow drift in the water CT number or a rising noise value signals a developing problem (detector drift, tube aging, calibration slip) long before it fails. For spectral and photon-counting systems, extend the QC to the additional CT-number and spectral checks those systems require. ⁷¹⁰

Regulatory Considerations

ACR CT accreditation sits within a layered framework of federal, state, and program requirements. The ACR CT Accreditation Program requires three elements: a clinical image review, a phantom image review, and an annual system performance evaluation (the medical physicist's survey). The phantom scoring uses the ACR CT Quality Control Manual, whose current edition (the 2017 manual) remains in force. ¹²

Key points:

Qualified medical physicist. The ACR program requires an annual performance evaluation by a qualified medical physicist, who also interprets the phantom results and establishes baselines. ²
Dose limits. The program applies CTDIvol pass/fail limits — on the order of 80 mGy (adult head, 16 cm CTDI phantom), 30 mGy (adult abdomen, 32 cm CTDI phantom), and 25 mGy (pediatric abdomen) — as ceilings. ³
Size-specific dose. For body CT, size-specific dose estimates (SSDE) translate CTDIvol into a patient-size-corrected dose using the AAPM Report 204 conversion factors; see size-specific dose estimates in CT. ¹²
State regulation. CT scanners are X-ray-producing machines regulated by the FDA (equipment performance standards) and by state radiation-control programs. In Florida, for example, diagnostic X-ray systems fall under the state's radiation-control rules, distinct from the NRC's authority over radioactive material. A facility must satisfy both its state registration and inspection requirements and its accreditation obligations.

Facilities pursuing or maintaining accreditation should coordinate the phantom testing with the annual survey, dose monitoring, and protocol management through CT physics testing and accreditation support.

Frequently Asked Questions (FAQs)

What is the ACR CT accreditation phantom?

It is a water-equivalent phantom with four testing modules used in the ACR CT Accreditation Program to evaluate CT number accuracy and positioning, low-contrast resolution, image uniformity and noise, and high-contrast (spatial) resolution. A facility scans the phantom with its clinical protocols and submits the images for scoring against the ACR CT Quality Control Manual criteria.

What does each module of the ACR CT phantom test?

Module 1 tests positioning, slice thickness, in-plane distance, and the CT number accuracy of five materials. Module 2 tests low-contrast resolution using low-contrast cylinders. Module 3 tests image uniformity and noise across the field. Module 4 tests high-contrast spatial resolution using bar patterns measured in line pairs per centimeter.

How is low-contrast resolution scored on the ACR CT phantom?

Low-contrast performance is judged by the contrast-to-noise ratio measured between a low-contrast rod and the background, and by whether the low-contrast cylinders are visible at the facility's adult abdomen technique. The ACR CT Quality Control Manual sets a minimum passing contrast-to-noise ratio near 1.0 for the adult abdomen protocol.

What is the CT number accuracy tolerance for water?

The water-equivalent material should read close to 0 HU, within the acceptance window specified in the current ACR CT Quality Control Manual, which is on the order of a few HU either side of zero. The reference CT numbers of the other materials depend on scanner make and kVp, so tolerances must account for that dependence.

What CTDIvol limits does the ACR CT program apply?

The ACR CT Accreditation Program applies CTDIvol pass/fail limits on the order of 80 mGy for the adult head, 30 mGy for the adult abdomen, and 25 mGy for a pediatric abdomen protocol, measured in the appropriate CTDI phantom. These are ceilings; modern scanners typically operate well below them at acceptable image quality.

Who performs ACR CT phantom testing?

Phantom scanning is usually performed by a CT technologist following the ACR instructions, while the annual system performance evaluation and the interpretation of results against ACR criteria are performed by a qualified medical physicist. DRPS provides CT physics testing and accreditation support for facilities pursuing or maintaining ACR accreditation.

How often is the ACR CT phantom scanned?

The phantom is scanned for the initial accreditation submission, at the annual medical physicist survey, and as part of routine periodic QC as defined by the facility's program and the ACR CT Quality Control Manual. Trending the results over time catches slow drift in CT number, noise, and uniformity before it fails.

Key Takeaways

Four modules, four image-quality tasks. The ACR CT phantom scores CT number accuracy, low-contrast resolution, uniformity, and high-contrast resolution together.
Low contrast is a noise problem. CNR is fixed contrast over noise; because contrast is set by the phantom, dose and reconstruction — which control noise — decide the result. The adult abdomen minimum is a CNR near 1.0.
CT numbers are kVp- and scanner-dependent. Set tolerances per scanner and per kVp, and re-baseline after major service.
Uniformity guards quantitation. Center-to-periphery agreement within a few HU keeps CT numbers position-independent.
Dose is a ceiling. CTDIvol limits (around 80, 30, and 25 mGy for adult head, adult abdomen, and pediatric abdomen) are pass/fail — but most scanners have room below them.
Trend the numbers. The single pass matters less than the drift; trending catches problems early.

Conclusion

The ACR CT accreditation phantom is deceptively simple — four modules in a water-equivalent cylinder — but it interrogates the full chain of CT image quality against real clinical protocols and real dose. Its power is that it couples the parameters together: low-contrast performance depends on noise, which depends on dose, which is itself capped; CT-number accuracy underpins every quantitative task; uniformity keeps those numbers position-independent; and high-contrast resolution guards fine detail.

For a facility, treating the phantom as a trending tool rather than a one-time hurdle turns accreditation into ongoing image-quality assurance. The scanner that passes comfortably, at a dose well below the ACR ceiling, with stable trends across every module, is the one delivering diagnostic images its radiologists — and patients — can rely on.

How DRPS Can Help

Diagnostic Radiation Physics Services supports CT facilities with CT physics testing and annual system performance evaluations, ACR phantom scoring and accreditation support, protocol optimization and dose management, CT-number and uniformity baselining, and QC program design — all performed by board-certified medical physicists.

DRPS supports facilities across our service locations, including Florida, Maryland, Virginia, Washington DC, California, Nevada, New York, Pennsylvania, New Jersey, and Delaware.

A strong CT QC program makes passing accreditation the natural result of good day-to-day image quality, not a scramble before the deadline.

Related Resources

References

American College of Radiology. ACR CT Quality Control Manual (2017 edition; current). Reston, VA: ACR. accreditationsupport.acr.org
American College of Radiology. ACR CT Accreditation Program — Requirements. acraccreditation.org
American College of Radiology. ACR CT Accreditation Program — Radiation Dosimetry (CTDIvol reference and pass/fail values). accreditationsupport.acr.org
Samei E, Bakalyar D, Boedeker KL, et al. Performance evaluation of computed tomography systems: Summary of AAPM Task Group 233. Med Phys. 2019;46(11):e735-e756. doi:10.1002/mp.13763. PubMed
Hobson MA, Soisson ET, Davis SD, Parker W. Using the ACR CT accreditation phantom for routine image quality assurance on both CT and CBCT imaging systems in a radiotherapy environment. J Appl Clin Med Phys. 2014;15(4):4835. doi:10.1120/jacmp.v15i4.4835. PubMed
Cropp RJ, Seslija P, Tso D, Thakur Y. Scanner and kVp dependence of measured CT numbers in the ACR CT phantom. J Appl Clin Med Phys. 2013;14(6):4417. doi:10.1120/jacmp.v14i6.4417. PubMed
Duan X, Zhang Y. Establishing quality control action limits for CT number accuracy in spectral images using an American College of Radiology phantom. Med Phys. 2023;50(10):6071-6078. doi:10.1002/mp.16626. PubMed
Fan M, Thayib T, McCollough C, Yu L. Accurate and efficient measurement of channelized Hotelling observer-based low-contrast detectability on the ACR CT accreditation phantom. Med Phys. 2023;50(2):737-749. doi:10.1002/mp.16068. PubMed
Ahmed Z, Ferrero A, Ren L, et al. Establishing a quality assurance program for photon counting detector (PCD) CT: Tips and caveats. J Appl Clin Med Phys. 2023;24(7):e14074. doi:10.1002/acm2.14074. PubMed
Anam C, Amilia R, Naufal A, et al. Performance of a statistical-based automatic contrast-to-noise ratio measurement on images of the ACR CT phantom. J Imaging. 2025;11(6):175. doi:10.3390/jimaging11060175. PubMed
Jordan DW, Smitherman CC, Bell JM, Moloney WE, Petrone TJ. Beyond CT accreditation: systematic evaluation of achievable image quality, radiation dose, and protocol factors in annual physics testing. J Appl Clin Med Phys. 2025;26(12):e70366. doi:10.1002/acm2.70366. PubMed
Boone JM, Strauss KJ, Cody DD, et al. Size-Specific Dose Estimates (SSDE) in Pediatric and Adult Body CT Examinations (AAPM Report No. 204). College Park, MD: AAPM; 2011. aapm.org