CT Daily QC: Water Phantom Constancy Testing

The daily water-phantom scan is the simplest and most consequential quality-control test performed on a CT scanner, and it takes a technologist only a few minutes. By measuring the mean CT number of water, the image noise, and checking for artifacts against an established baseline each morning, a facility detects detector, calibration, and x-ray tube drift before it ever reaches a patient image. ^{1, 2}

Water is the reference material of CT. It is assigned exactly 0 Hounsfield units by definition, it is cheap and reproducible, and a uniform water phantom exposes the three things most likely to drift: the calibration of the CT-number scale, the noise level at a fixed technique, and the presence of ring or streak artifacts. That is why nearly every CT QC program in the world starts the day the same way. ^{1, 6}

Introduction

Modern CT scanners are remarkably stable, which is precisely why a small, systematic daily check is worth the effort: when something does drift, it drifts quietly. A detector channel whose gain is slipping, a tube whose output is falling, or a calibration table that is slightly stale will not announce itself with an obvious artifact on day one. It shows up first as a water CT number that is a few HU off, or a noise value that has crept up at the same technique. Caught on the daily phantom, it is a service call. Missed for weeks, it is a stack of suboptimal patient studies and potentially a repeated exam. ^{2, 3}

The daily water-phantom constancy test is the layer of CT quality control designed to catch exactly this. It is deliberately simple so that it can be done reliably every morning by the technologist, and it is deliberately relative — it compares the scanner to its own established baseline rather than to an absolute reference, which is the job of the annual physics survey. This guide explains what the test measures, the physics behind each metric, the tolerances that matter, and how daily QC fits into the full program that DRPS supports through CT physics testing and accreditation support across Florida, Maryland, Virginia, Washington DC, California, and Nevada.

Topic Explanation

What the daily water-phantom test is

The daily test is a short scan of a uniform, water-equivalent phantom — usually the water or uniformity section of the scanner's own QC phantom, provided by the manufacturer — performed after the tube warm-up sequence and before the first patient of the day. The technologist places a region of interest (ROI) in the center of the uniform section and records two numbers: the mean CT number of the ROI and its standard deviation, which is the image noise. The technologist also reviews the image for artifacts such as rings, streaks, or shading. The three results are logged and compared to the baseline the medical physicist established at acceptance or after the last major service. ^{1, 2}

The ACR CT Quality Control Manual frames this as the continuous, technologist-level layer of the program, and recommends running the constancy scans in both axial and helical modes. ¹ The international standard IEC 61223-2-6 similarly defines constancy tests for CT, built around repeated measurement of CT number, noise, and uniformity relative to baseline values. ⁶

Why water, and why Hounsfield units

CT does not measure density directly; it measures the linear attenuation coefficient of each voxel and then maps it onto the Hounsfield scale. The Hounsfield unit is defined relative to water:

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

By construction, water is 0 HU and air is −1000 HU. This definition is the reason water is the ideal daily test material: a correctly calibrated scanner must read a mean of approximately 0 HU in a water phantom, with no reference value to look up. If the daily water reading drifts away from 0 (or from the facility's established baseline), the CT-number calibration itself has moved, and every tissue measurement derived from it — lung, liver, bone, contrast enhancement — moves with it. ¹ For the companion test that verifies accuracy across many materials, see CT Number (HU) Calibration QC.

Key Technical Principles

The three daily metrics

The daily constancy test is built on three measurements, each sensitive to a different failure mode:

Mean CT number of water — the calibration check. Drift here points to detector gain, reference-channel, or calibration-table problems.
Image noise (standard deviation) — the dose-efficiency and tube-output check. Drift here points to falling tube output, detector sensitivity loss, or an inadvertent reconstruction change, since the reconstruction algorithm strongly shapes noise magnitude and texture. ⁷
Artifact evaluation — the uniformity and detector-integrity check. Rings point to a single bad detector channel; streaks and shading point to calibration or beam-hardening issues.

The noise–dose relationship

Image noise is the most physically informative of the three because it is tied directly to the radiation dose and the imaging chain. For a given reconstruction, the standard deviation of CT numbers in a uniform region scales inversely with the square root of the dose delivered:

σ \propto \frac{1}{D}

Equivalently, comparing two acquisitions of the same phantom:

\frac{σ _{2}}{σ _{1}} = \frac{D _{1}}{D _{2}}

This relationship has two practical consequences for daily QC. First, because the daily test is always run at the same fixed technique, the dose is held constant, so any change in the measured noise reflects a change in the scanner, not the protocol — a rising noise value at fixed mAs is a genuine warning sign. Second, it quantifies the cost of noise reduction: to halve the noise standard deviation you must quadruple the dose. A facility that reacts to a creeping daily noise value by raising clinical mAs is paying a steep dose penalty to paper over a hardware problem that service should fix. ^{2, 3} The same inverse-square-root physics underlies automatic exposure control, which modulates tube current to hold noise constant as patient size changes. ^{4, 5}

Einstein and colleagues illustrated both the sensitivity and the limits of these daily metrics: analyzing three months of daily water-phantom QC across two scanners of the same vendor, they resolved statistically significant differences in mean CT number (0.70 versus 0.41 HU) and noise magnitude (4.85 versus 4.94) between the units — differences far too small to see by eye. Importantly, they also found that standard water-phantom noise evaluation alone could not discern the scanners' clinically relevant behavior; the attenuation-dependent differences that mattered for reading emerged only on higher-attenuation phantoms, underscoring that the daily water-phantom test is a sensitive early-warning layer rather than a complete characterization. ³

Typical tolerances

Daily constancy tolerances are set relative to each scanner's baseline and should follow the current ACR CT Quality Control Manual and the manufacturer's specification, both of which acknowledge that vendor tolerances vary. ¹ As general practice:

Daily test	Metric	Typical action threshold
Water CT number	Mean HU of central ROI	Within a few HU of baseline; the ACR CT accreditation water module is evaluated near 0 ± a small number of HU
Image noise	Standard deviation of ROI	Within the baseline value ± the facility/vendor tolerance
Uniformity	Center vs peripheral mean HU	On the order of ± a few HU across the field
Artifacts	Visual (rings, streaks, shading)	None visible

The exact numbers are less important than the discipline: establish a baseline with the physicist, set an action threshold, log every day, and investigate anything that crosses it rather than normalizing the drift.

Clinical Impact

The clinical value of daily QC is that it protects quantitative CT, follow-up comparison, and dose efficiency — three things radiologists and referring physicians rely on without thinking about them.

Quantitative tasks are the most obvious. A liver lesion characterized by its HU, a coronary calcium score, a lung-nodule density, or a contrast washout measurement is only as trustworthy as the CT-number calibration behind it. If the daily water reading has drifted +6 HU and nobody noticed, every density measurement that day is biased. Catching the drift on the morning phantom protects the number before it is ever reported. ^{1, 3}

Follow-up comparison is the subtler case. Oncology and surveillance imaging compares a lesion today against the same lesion months ago, often on a different scanner. Einstein and colleagues documented that even same-vendor scanners can carry small but real CT-number and noise offsets, and that fully characterizing and harmonizing those differences required testing beyond the water phantom, using higher-attenuation materials; holding each scanner to a tight daily baseline helps keep cross-scanner comparisons honest, but it is one part of that larger harmonization effort. ³

Dose efficiency is the quiet one. Because noise and dose are linked by the inverse-square-root law, a scanner silently losing tube output will produce noisier images at the same dose — and the natural but wrong reaction is to raise technique, delivering more dose to compensate for a fixable hardware problem. ^{2, 4} Daily noise tracking turns that invisible degradation into a dated log entry that justifies a service call instead of a dose increase. Patient centering compounds the same physics: miscentering in the gantry raises both dose and noise, which is why positioning is part of a complete QC mindset. ⁴

Practical Tips

Warm up first, every time. Run the full tube-conditioning and calibration sequence before the daily phantom scan. A cold tube or a skipped air calibration will fail the water test for the wrong reason and waste a service call.
Fix the geometry. Scan the phantom at the same position, centered in the gantry, at the same technique every day. The entire value of a constancy test comes from changing only one variable — the scanner — so hold everything else constant.
Log the numbers, not just pass/fail. Record the actual mean HU and noise value daily and trend them. A value that is "still passing" but has climbed steadily for two weeks is the finding; a binary pass/fail hides it.
Set baselines with the physicist. Baselines and action thresholds should be established by the qualified medical physicist at acceptance and after any major service, calibration, or tube replacement — not carried over blindly. ²
Run both axial and helical constancy where the ACR program recommends it, since artifacts and calibration behavior can differ between modes. ¹
Escalate drift; do not compensate for it. If the water CT number or noise crosses the action threshold, stop and investigate or call service. Raising clinical technique to hide rising noise trades a hardware fix for a dose penalty.
Keep daily QC in its lane. Passing the daily test does not substitute for acceptance testing, the annual physics survey, or accreditation measurements — it is the early-warning layer, not the whole program.

DRPS integrates daily QC review into CT physics testing and medical physics consulting engagements, setting the baselines and reviewing the trends that technologists log.

Regulatory Considerations

CT quality control sits at the intersection of accreditation requirements, state regulations, and professional practice standards — but it is not regulated as byproduct material. A CT scanner is an x-ray machine, so its oversight comes from the FDA equipment performance standard (21 CFR 1020.33) ⁹, the state radiation-control program that registers and inspects the unit, and the accreditation body, rather than from the NRC. ^{1, 8}

The key frameworks for a CT QC program are:

ACR CT Quality Control Manual (2017) — defines the continuous (daily/technologist) and periodic (physicist) QC structure, including the daily water-phantom constancy test, and is the operational backbone of the ACR CT Accreditation Program. ^{1, 8}
AAPM Report No. 233 (Task Group 233) — the 2019 AAPM guidance on performance evaluation of CT systems, updating testing methods for noise (magnitude, texture, nonuniformity), resolution, and automatic exposure control performance. ²
IEC 61223-2-6 — the international constancy-test standard for CT, structured around repeated measurement of CT number, noise, and uniformity against baseline. ⁶
ACR CT Accreditation Program Requirements — the accreditation-level testing and tolerances, including the water-module CT number and noise evaluation, that an accredited facility must meet. ⁸

For facilities pursuing or maintaining accreditation, the daily water-phantom log is one of the first records an ACR reviewer or state inspector asks to see, because it demonstrates an ongoing, documented QC process rather than a once-a-year check. See CT Acceptance Testing and Commissioning for how the baselines originate and CTDI Measurement and CT Dose QC for the dose side of the annual program.

Frequently Asked Questions (FAQs)

What is the CT daily water-phantom QC test?

It is a short routine scan of a uniform water-equivalent phantom performed before clinical imaging each day. The technologist measures the mean CT number of a central water region, the image noise (the standard deviation of that region), and checks for artifacts, then compares the results to an established baseline to detect drift.

What should the CT number of water be?

By definition, water is assigned 0 Hounsfield units, so a correctly calibrated scanner should read a mean CT number near 0 HU in a water phantom. Daily programs typically act when the value drifts more than a few HU from the established baseline; the ACR CT accreditation phantom water module is evaluated against a limit on the order of 0 plus or minus a few HU. Always use the current tolerance in the ACR CT Quality Control Manual and your scanner's specification.

Why is image noise measured every day?

Image noise, quantified as the standard deviation of CT numbers in a uniform region, is a sensitive early indicator of tube output loss, detector gain drift, or reconstruction changes. Because noise scales inversely with the square root of dose, a rising daily noise value at a fixed technique signals that something in the imaging chain has changed before it becomes visible on patient scans.

Who performs CT daily QC, the technologist or the medical physicist?

The daily water-phantom constancy test is a technologist task, performed each morning after tube warm-up. The qualified medical physicist performs the more comprehensive acceptance and annual testing, reviews the daily QC trends, sets the baselines and action thresholds, and investigates failures.

What is the difference between daily constancy QC and the annual physics survey?

Daily constancy QC verifies that the scanner has not drifted from its own established baseline using a few quick, repeatable measurements. The annual medical physicist survey independently verifies absolute performance against reference standards, including CT number accuracy across materials, dose (CTDIvol), spatial resolution, slice width, and low-contrast detectability.

Does passing daily QC mean the scanner is fully compliant?

No. Daily QC confirms day-to-day stability, but it does not replace acceptance testing, the annual physics survey, or accreditation requirements. A scanner can pass daily constancy and still fail an annual test such as CT number accuracy across materials or a dose measurement, so both layers are required.

Key Takeaways

Water is the reference. Because water is defined as 0 HU, the daily water-phantom scan checks the CT-number calibration against a value that needs no lookup.
Three metrics, three failure modes. Mean CT number catches calibration drift, noise catches tube and detector drift, and the artifact check catches detector and uniformity faults.
Noise tracks the hardware. At a fixed daily technique, rising noise reflects the scanner, not the protocol, and the inverse-square-root law means halving noise would cost four times the dose.
It is a relative test. Daily QC compares the scanner to its own baseline; set that baseline and the action thresholds with the medical physicist.
Log and trend, do not just pass/fail. A value creeping toward the threshold is the early warning that a binary result hides.
Daily QC is one layer. It complements, but never replaces, acceptance testing, the annual physics survey, and accreditation measurements.

Conclusion

The daily water-phantom constancy test earns its place in CT quality control by being fast, repeatable, and physically meaningful. It leverages the single most elegant fact about CT — that water is defined as zero — to turn a three-minute morning scan into an early-warning system for calibration, tube, and detector drift. The payoff is quantitative CT measurements you can trust, follow-up comparisons that hold across scanners and across time, and dose efficiency that does not quietly erode.

The technologist who logs the actual numbers every morning, and the medical physicist who sets the baselines and reviews the trends, together form a QC loop that catches problems while they are still service calls rather than recalled studies. Daily QC is not glamorous, but it is where a defensible CT program lives or dies.

How DRPS Can Help

Diagnostic Radiation Physics Services helps imaging facilities build and maintain CT quality-control programs that satisfy accreditation and state requirements while genuinely protecting image quality. This includes establishing daily QC baselines and action thresholds, reviewing technologist QC logs for drift, performing acceptance and annual physics testing, and preparing facilities for ACR CT accreditation through CT physics testing, accreditation support, and medical physics consulting.

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

Related Resources

References

American College of Radiology. CT Quality Control Manual. 2017. acr.org
Samei E, et al. Performance evaluation of computed tomography systems: Summary of AAPM Task Group 233. Medical Physics. 2019;46(11):e735-e756. doi:10.1002/mp.13763. PubMed
Einstein SA, Rong XJ, Jensen CT, Liu X. Quantification and homogenization of image noise between two CT scanner models. Journal of Applied Clinical Medical Physics. 2019;21(1):174-178. doi:10.1002/acm2.12798. PubMed
Kaasalainen T, Palmu K, Reijonen V, Kortesniemi M. Effect of patient centering on patient dose and image noise in chest CT. AJR American Journal of Roentgenology. 2014;203(1):123-130. doi:10.2214/AJR.13.12028. PubMed
Brisse HJ, Madec L, Gaboriaud G, et al. Automatic exposure control in multichannel CT with tube current modulation to achieve a constant level of image noise: experimental assessment on pediatric phantoms. Medical Physics. 2007;34(7):3018-3033. doi:10.1118/1.2746492. PubMed
International Electrotechnical Commission. IEC 61223-2-6: Evaluation and routine testing in medical imaging departments — Part 2-6: Constancy tests — Imaging performance of computed tomography X-ray equipment. iec.ch
Guido G, Polici M, Nacci I, et al. Iterative Reconstruction: State-of-the-Art and Future Perspectives. Journal of Computer Assisted Tomography. 2023;47(2):244-254. doi:10.1097/RCT.0000000000001401. PubMed
American College of Radiology. CT Accreditation Program Requirements. acr.org
U.S. Food and Drug Administration. 21 CFR 1020.33: Computed tomography (CT) equipment. ecfr.gov