Gamma Camera Bar Phantom Resolution QC

Q: What does a four-quadrant bar phantom measure on a gamma camera?

A four-quadrant bar phantom provides a semi-quantitative check of a gamma camera's intrinsic spatial resolution and spatial linearity. It has four sets of parallel lead bars of different widths; the finest set whose bars are still clearly resolved indicates the camera's limiting resolution, while any waviness or distortion of the bar pattern reveals a spatial-linearity problem.

Q: How is intrinsic resolution different from system resolution?

Intrinsic resolution is the detector's own resolution measured without a collimator, using a point source and a bar phantom or slit. System resolution is the combined resolution of the detector plus collimator at a stated distance and is always poorer, because the collimator dominates the overall spatial resolution in clinical imaging.

Q: How does bar-phantom visibility relate to FWHM?

The camera can resolve a bar pattern only when the bar width is not much smaller than the detector's line spread function. A widely used approximation is that the full width at half maximum is roughly 1.7 times the width of the smallest clearly resolved bars, so a camera that just resolves 2.5 mm bars has an intrinsic FWHM of about 4 mm. Quantitative resolution should still be measured from the line spread function per NEMA methodology.

Q: How often should bar-phantom resolution and linearity be checked?

Intrinsic spatial resolution and linearity are typically checked at acceptance, after major service, and on a routine basis such as weekly, with the phantom rotated so both axes and the whole field of view are assessed over time. Daily QC usually focuses on uniformity, with resolution and linearity confirmed weekly and quantified at the annual physics survey.

Q: What causes a gamma camera to fail a linearity or resolution test?

Common causes include drift in photomultiplier tube gains, failure or degradation of the energy and linearity correction maps, a cracked or hydrated sodium iodide crystal, light-guide or PMT faults, and incorrect energy-window settings. Because uniformity, linearity, and resolution are physically linked, a linearity fault often shows up first as non-uniformity.

Q: Does bar-phantom QC replace NEMA acceptance testing?

No. Bar-phantom QC is a fast routine and constancy check. Full quantitative characterization of intrinsic and system spatial resolution, using line sources and the NEMA NU-1 methodology, is performed at acceptance and at the annual physics survey by a qualified medical physicist.

A four-quadrant bar phantom placed on the uncollimated detector and flooded with a point source is the fastest routine check of a gamma camera's two most fundamental spatial performance parameters: intrinsic spatial resolution and spatial linearity. The finest bar quadrant that remains clearly resolved indexes resolution, while any bending or spacing distortion of the bars exposes a linearity fault.¹²³

Bar-phantom testing is deliberately quick and visual, which makes it ideal for weekly constancy monitoring. It does not replace the quantitative line-spread-function measurements of a full NEMA acceptance test, but it catches the drifts and faults that most often degrade clinical images between annual surveys.⁴⁵

Introduction

Spatial resolution and spatial linearity are the geometric foundation of every planar and SPECT image a gamma camera produces. If the detector cannot separate two nearby sources, small lesions blur together; if it cannot map an event to the correct location, straight structures appear curved and count density becomes non-uniform.

The four-quadrant bar phantom has been the workhorse tool for monitoring both parameters for decades because it is simple, robust, and sensitive to exactly the failure modes — photomultiplier drift, correction-map errors, and crystal or light-guide problems — that develop between acceptance tests.¹²⁶

This guide explains what intrinsic spatial resolution and spatial linearity are, how a bar phantom probes them, how bar visibility maps quantitatively onto the line spread function, how resolution and linearity are physically linked to uniformity, the tolerances that matter, a worked example, the clinical stakes, practical tips, the regulatory and accreditation context, and how routine bar-phantom QC fits into a complete performance-monitoring program.

Topic Explanation

Intrinsic resolution, system resolution, and linearity

Three terms must be kept distinct:

Intrinsic spatial resolution is the resolution of the detector alone — crystal, light guide, photomultiplier tubes (PMTs), and positioning electronics — measured without a collimator using a point or line source. It is quantified as the full width at half maximum (FWHM) of the line spread function (LSF).⁴
System spatial resolution is the resolution of the detector plus collimator at a stated source-to-collimator distance. Because the collimator dominates, system resolution is always poorer than intrinsic resolution and degrades with distance. See our guide to gamma camera collimator selection.
Spatial linearity is the geometric fidelity of event positioning: a straight-line source must image as a straight line. Linearity is characterized as absolute linearity (maximum displacement of an imaged line from its true position) and differential linearity (variation in spacing between imaged lines).⁴

A bar phantom probes intrinsic resolution and linearity together, because a set of straight, evenly spaced lead bars tests both whether the fine bars can be resolved (resolution) and whether they appear straight and evenly spaced (linearity).¹⁶

What is a four-quadrant bar phantom?

A four-quadrant bar phantom is a lead sheet (or lead bars embedded in plastic) divided into four quadrants, each containing parallel lead bars and gaps of a different width — for example 2.0, 2.5, 3.0, and 3.5 mm bars in the four quadrants. Placed directly on the uncollimated crystal face and irradiated by a uniform flood field from a point source of technetium-99m (or a cobalt-57 sheet source), the lead bars block photons and the gaps pass them, so the detector "sees" a striped pattern in each quadrant.¹²

The interpretation is direct: the quadrant with the smallest bars whose stripes remain clearly separated indicates the limiting resolution of that region of the detector. Because the bars run in one direction, the phantom is rotated (commonly by 90° each session) so that resolution and linearity are assessed along both the X and Y axes and, over successive sessions, across the whole field of view.³⁶ Historic transmission test patterns, such as the BRH pattern, were developed on the same principle of a precision lead pattern imaged to reveal resolution, uniformity, and distortion simultaneously.⁶

Why a bar phantom is semi-quantitative

Visual reading of a bar phantom is fast but limited: the assumption that "if intrinsic resolution changed, the bar image would change" does not always hold precisely, and different observers may disagree on whether the finest quadrant is "just resolved." Early comparisons of four-quadrant bar-phantom visual scores against LSF-derived FWHM showed only moderate agreement, which is why bar-phantom testing is used for constancy monitoring while quantitative resolution is measured from the LSF at acceptance and annual survey.²⁸ Automated methods that place regions of interest on the bar image and compute a modulation transfer function improve objectivity and reduce interobserver variability, but the reference intrinsic-resolution measurement remains the LSF.⁷

Key Technical Principles

From bar visibility to FWHM

The camera resolves a bar pattern only when the bar width is comparable to or larger than the detector's line spread function. A widely used approximation links the smallest resolvable bar width to the intrinsic FWHM:

FWHM \approx 1.7 \times b_{m i n}

where $b_{m i n}$ is the width of the smallest clearly resolved bars. Equivalently, bars become resolvable at roughly $b_{m i n} \approx 0.6 \times FWHM$ .⁸ For a camera that just resolves the 2.5 mm quadrant:

FWHM \approx 1.7 \times 2.5 mm \approx 4.3 mm

This is consistent with modern single-crystal cameras, whose specified intrinsic FWHM is commonly on the order of 3.5–4.0 mm.⁸ The approximation is useful for a quick sanity check, but the definitive intrinsic-resolution number is the LSF FWHM measured per NEMA NU-1 methodology.⁴

Uniformity, linearity, and resolution are physically linked

A gamma camera positions each event by comparing PMT signals; energy and linearity correction maps then reposition and rescale events to compensate for non-ideal PMT response. When those corrections drift or fail, three things degrade together:

Linearity — events are mislocated, so straight bars appear wavy.
Uniformity — mislocation piles events into some pixels and starves others, creating flood non-uniformity.
Resolution — positioning noise broadens the LSF.

Because of this coupling, flood-field uniformity is the most sensitive daily indicator, and a linearity fault often reveals itself first as non-uniformity.⁴⁵ Integral (whole-field) uniformity is defined per NEMA NU-1 as:

U_{int} = 100 \times \frac{C _{m a x} - C _{m i n}}{C _{m a x} + C _{m i n}}

and differential uniformity is the analogous maximum change over any small (five-pixel) neighborhood.⁹ These are evaluated over the useful field of view (UFOV) and the central field of view (CFOV).

The two intrinsic tests at a glance

Property	Intrinsic spatial resolution	Intrinsic spatial linearity
What it measures	Ability to separate two nearby sources	Geometric fidelity of event positioning
Quantitative metric	FWHM of the line spread function (mm)	Absolute and differential displacement (mm)
Routine tool	Four-quadrant bar phantom (visual)	Same bar phantom (straightness/spacing of bars)
Reference method	LSF via slit/line source, NEMA NU-1	Line-source displacement analysis, NEMA NU-1
Representative modern spec	Intrinsic FWHM ~3.5–4.0 mm	Absolute linearity within ~1 mm
Common failure cause	PMT drift, crystal/light-guide fault	Correction-map error, PMT gain drift

The "representative spec" column reflects typical modern single-crystal cameras and vendor specifications assessed with NEMA methodology; each camera should be evaluated against its own manufacturer specification and the tolerances in the facility's QC program.⁴¹⁰¹¹

Worked example: integral uniformity from flood counts

Suppose a routine intrinsic flood acquisition, after the appropriate smoothing and pixel binning defined in the standard, yields a maximum pixel count $C_{m a x} = 1040$ and a minimum $C_{m i n} = 960$ within the UFOV. The integral uniformity is:

U_{int} = 100 \times \frac{1040 - 960}{1040 + 960} = 100 \times \frac{80}{2000} = 4.0%

A value of 4.0% is within the range typically seen at acceptance for a well-performing camera; published intrinsic-uniformity results report UFOV integral values on the order of a few percent, with acceptance figures around 5% or better and central-field values lower.¹¹ Because uniformity is count-statistics dependent, the flood must contain enough counts (commonly millions of counts over the field) so that statistical noise does not masquerade as a uniformity defect.⁵⁹

Clinical Impact

Resolution and linearity faults do not announce themselves as obvious artifacts — they quietly degrade diagnostic accuracy. A detector that has lost resolution blurs small structures, reducing the detectability of small lesions in bone, parathyroid, and tumor imaging. A linearity fault distorts anatomy and, through the uniformity coupling, can create ring or streak artifacts in SPECT reconstruction that mimic or mask disease.⁴⁵

SPECT is especially unforgiving. Small, stable non-uniformities that are invisible on a planar flood are amplified by the reconstruction into concentric ring artifacts centered on the axis of rotation, which is why SPECT cameras require high-count uniformity correction floods and disciplined resolution and linearity monitoring. For the tomographic QC that builds on these intrinsic checks, see SPECT center-of-rotation QC and quantitative SPECT/CT calibration.

Routine bar-phantom testing is the early-warning system: catching a resolving-power drop or a bar-pattern distortion at the weekly check prevents weeks of subtly degraded studies and the downstream cost of repeat imaging or misdiagnosis.³¹⁰

Practical Optimization Tips

Acquire the bar image correctly

Use a point source at adequate distance. For an intrinsic flood, place the technetium-99m point source at a distance of at least about five times the largest field-of-view dimension so the flood is uniform across the crystal; center it carefully.⁵
Collect enough counts. A resolution/linearity bar image needs sufficient count density to make the fine bars visible above statistical noise; follow the density specified in your QC procedure.⁵⁹
Rotate the phantom. Rotate the bar phantom (for example 90° per session, cycling through orientations) so both axes and all four bar sizes are assessed over the field of view across successive sessions.³⁶
Set the correct energy window. An incorrect or drifted photopeak window degrades apparent resolution and uniformity; confirm the window (and peaking) before interpreting the bar image.⁵

Read the image deliberately

Identify the smallest quadrant whose bars are clearly and completely separated across the quadrant, not just at the edges.
Inspect every quadrant for waviness, kinks, or uneven spacing — these are linearity faults even when resolution looks acceptable.
Compare against the baseline reference image from acceptance testing, not just against memory; a side-by-side comparison is the most sensitive way to detect gradual drift.¹⁰

Escalate the right way

A borderline or failed bar image should trigger a flood-uniformity check and photopeak/energy-window verification before service is called, because a mis-peak or a correction-map issue is a common and correctable cause.⁴⁵
Document baseline values, tolerances, results, and corrective actions so the record supports accreditation review. For the instrument that verifies your source activity, see dose calibrator quality control.

Avoid common bar-phantom errors

Treating the bar phantom as a quantitative resolution measurement. It is a constancy check; quantitative FWHM comes from the LSF.⁷⁸
Reading only one orientation. Resolution and linearity are direction-dependent; a single orientation misses half the information.³
Under-counting the image. Noise can hide fine bars or fake a distortion.⁹
Ignoring the coupling with uniformity. A "resolution" complaint is frequently a linearity/uniformity problem in disguise.⁴

Regulatory Considerations

Gamma camera performance monitoring is governed less by numerical regulatory limits than by accreditation and professional standards, but those standards are enforceable through accreditation and reimbursement. A defensible program aligns routine bar-phantom testing with recognized performance methodology and documents it.

NEMA NU-1 defines the standardized methods for measuring gamma camera performance, including intrinsic and system spatial resolution and spatial linearity; the 2018 edition is the reference methodology for acceptance measurements.⁴
AAPM Report No. 177 provides acceptance-testing and annual-physics-survey recommendations for gamma camera, SPECT, and SPECT/CT systems, including which tests to perform and how often.¹⁰
IAEA guidance — Human Health Series No. 6 and the earlier IAEA-TECDOC-602 — describes acceptance, reference, and routine QC procedures for scintillation cameras, including uniformity, resolution, and linearity.³¹²
Accreditation and oversight. The ACR Nuclear Medicine and PET accreditation programs and The Joint Commission require documented QC, qualified medical physicist oversight, and phantom imaging within tolerance; accreditation frequently conditions Medicare reimbursement. Byproduct material used to test and operate the camera is regulated under 10 CFR Part 35 or the equivalent Agreement State program, with the camera itself operated as part of a licensed nuclear medicine program.

DRPS serves facilities in Florida (Chapter 64E-5), Maryland, Virginia, Washington DC, California, Nevada, Pennsylvania, New York, New Jersey, and Delaware; confirm the specific accreditation and state requirements with the authority having jurisdiction. For the broader survey-program context, see preparing for an NRC inspection.

Frequently Asked Questions (FAQs)

What does a four-quadrant bar phantom measure on a gamma camera?

It provides a semi-quantitative check of intrinsic spatial resolution and spatial linearity. Four sets of parallel lead bars of different widths let the reader identify the finest resolved set (resolution) and detect any bending or uneven spacing of the bars (linearity).¹⁶

How is intrinsic resolution different from system resolution?

Intrinsic resolution is measured on the detector alone, without a collimator. System resolution includes the collimator at a stated distance and is always poorer, because the collimator dominates the combined resolution in clinical imaging.⁴

How does bar-phantom visibility relate to FWHM?

A common approximation is that the intrinsic FWHM is about 1.7 times the width of the smallest clearly resolved bars, so a camera that just resolves 2.5 mm bars has an intrinsic FWHM near 4 mm. The definitive value is measured from the line spread function per NEMA methodology.⁴⁸

How often should bar-phantom resolution and linearity be checked?

Typically at acceptance, after major service, and routinely (often weekly), with the phantom rotated to cover both axes and the whole field over time. Daily QC usually emphasizes uniformity; annual physics surveys quantify resolution and linearity.³¹⁰

What causes a gamma camera to fail a linearity or resolution test?

PMT gain drift, degraded energy or linearity correction maps, a cracked or hydrated sodium iodide crystal, light-guide or PMT faults, and incorrect energy-window settings. Because the parameters are coupled, a linearity fault often appears first as non-uniformity.⁴⁵

Does bar-phantom QC replace NEMA acceptance testing?

No. It is a fast routine check. Full quantitative characterization of intrinsic and system spatial resolution and linearity, using line sources and the NEMA NU-1 methodology, is performed at acceptance and annual survey by a qualified medical physicist.⁴¹⁰

Key Takeaways

The four-quadrant bar phantom is the standard routine check of gamma camera intrinsic spatial resolution and spatial linearity, read visually and quickly.¹³
Intrinsic resolution (detector alone, no collimator) is quantified as the FWHM of the line spread function; system resolution includes the collimator and is always poorer.⁴
Bar visibility maps to FWHM through the approximation $FWHM \approx 1.7 \times b_{m i n}$ , but the definitive number is the measured LSF.⁸
Resolution, linearity, and uniformity are physically coupled; a linearity fault often shows up first as flood non-uniformity, and SPECT amplifies small non-uniformities into ring artifacts.⁴⁵
Bar-phantom QC is a constancy check, not a substitute for NEMA NU-1 acceptance testing and the annual physics survey.⁴¹⁰
Correct technique — adequate source distance, sufficient counts, phantom rotation, correct energy window, and baseline comparison — is what makes the test sensitive.⁵

Conclusion

Routine bar-phantom testing endures because it is fast, robust, and sensitive to the exact faults — PMT drift, correction-map errors, and crystal or light-guide problems — that develop between acceptance tests. Read deliberately and compared against a baseline, it flags resolution loss and linearity distortion before they degrade clinical studies. It works best not in isolation but as one layer of a complete program: daily uniformity, weekly resolution and linearity, and a quantitative NEMA NU-1 annual physics survey performed by a qualified medical physicist. Together, these keep planar and SPECT images geometrically faithful and diagnostically reliable.³⁴¹⁰

How DRPS Can Help

Diagnostic Radiation Physics Services (DRPS) supports nuclear medicine facilities across Florida, Maryland, Virginia, Washington DC, California, Nevada, Pennsylvania, New York, New Jersey, and Delaware with gamma camera and SPECT/CT PET/CT and nuclear medicine physics acceptance testing, NEMA NU-1 performance measurement, annual physics surveys, routine QC program design, accreditation support, and medical physicist consulting performed by board-certified medical physicists.

A strong gamma camera QC program is not a stack of passed bar images. It is a documented, baseline-referenced method that turns a two-minute weekly check into early detection of the faults that would otherwise quietly erode diagnostic accuracy.

Related Resources

References

National Electrical Manufacturers Association. Performance Measurements of Gamma Cameras. NEMA Standards Publication NU 1-2018. Rosslyn, VA: NEMA; 2018. nema.org
O'Connor MK, Oswald WM. The line resolution pattern: a new intrinsic resolution test pattern for nuclear medicine. Journal of Nuclear Medicine. 1988;29(11):1856-1859. PubMed
International Atomic Energy Agency. Quality Assurance for SPECT Systems. IAEA Human Health Series No. 6. Vienna: IAEA; 2009. iaea.org
Halama JR, Graham D, Harkness BA, et al. Acceptance Testing and Annual Physics Survey Recommendations for Gamma Camera, SPECT, and SPECT/CT Systems. AAPM Report No. 177. College Park, MD: American Association of Physicists in Medicine; 2019. aapm.org
Zanzonico P. Routine quality control of clinical nuclear medicine instrumentation: a brief review. Journal of Nuclear Medicine. 2008;49(7):1114-1131. doi:10.2967/jnumed.107.050203. PubMed
Paras P, Hine GJ, Adams R. BRH test pattern for the evaluation of gamma-camera performance. Journal of Nuclear Medicine. 1981;22(5):468-470. PubMed
Tazegul TE, Polemi AM, Snyder A, Snyder C, Collins PG. Automated phantom analysis for gamma cameras and SPECT: a methodology for use in a clinical setting. Journal of Applied Clinical Medical Physics. 2020;21(11):205-214. doi:10.1002/acm2.13057. PubMed
Hander TA, Lancaster JL, McDavid W, Kopp DT. An improved method for rapid objective measurement of gamma camera resolution. Medical Physics. 2000;27(12):2688-2692. doi:10.1118/1.1328083. PubMed
Edam AN, Fornasier MR, De Denaro M, Sulieman A, Alkhorayef M, Bradley DA. Quality control in dual head gamma-cameras: comparison between methods and software used for image analysis. Applied Radiation and Isotopes. 2018;141:288-291. doi:10.1016/j.apradiso.2018.07.027. PubMed
Busemann Sokole E, Płachcínska A, Britten A, et al. Routine quality control recommendations for nuclear medicine instrumentation. European Journal of Nuclear Medicine and Molecular Imaging. 2010;37(3):662-671. doi:10.1007/s00259-009-1347-y. PubMed
Ejeh JE, Adedapo KS, Akinlade BI, Osifo BOA. Gamma camera intrinsic uniformity in an unstable power supply environment. Hellenic Journal of Nuclear Medicine. 2011;14(2):146-148. PubMed
International Atomic Energy Agency. Quality Control of Nuclear Medicine Instruments. IAEA-TECDOC-602. Vienna: IAEA; 1991. iaea.org