SPECT/CT Quality Control: Uniformity, COR & Performance

Q: How is integral uniformity calculated?

Integral uniformity is IU = (max − min) / (max + min) × 100%, computed over the pixel counts in the useful or central field of view of a high-count flood image. Differential uniformity evaluates the worst local rate of change over a small sliding window, capturing sharp regional non-uniformities that the global integral metric can miss.

SPECT/CT quality control is the scheduled program of gamma camera and tomographic performance tests—field uniformity, energy peaking, center-of-rotation calibration, spatial and energy resolution, sensitivity, multi-detector registration, and CT subsystem checks—that confirms a hybrid system is imaging within tolerance. Each test isolates a specific failure mode: uniformity catches detector and electronics drift, center of rotation catches reconstruction misalignment, resolution and energy tests catch detector degradation, and CT co-registration catches the spatial mismatch that corrupts attenuation correction.

In this edition of the PhysicsPulse^TM Series, we walk through the core gamma-camera and SPECT tests as defined by NEMA NU-1, show the worked math behind integral uniformity and energy resolution, lay out a practical daily-to-annual QC cadence, and explain how facilities across Florida, Maryland, Virginia, Washington DC, California, and Nevada can keep a SPECT/CT system accreditation-ready over its full service life.

Introduction

A SPECT/CT scanner is two imaging systems fused into one, and its quality control has to cover both. The gamma camera (one, two, or three scintillation detectors mounted on a rotating gantry) acquires the emission data, while an integrated CT scanner provides attenuation maps and anatomic localization. A defensible QC program tests the gamma-camera planar performance, the tomographic (SPECT) performance, and the CT subsystem, and then confirms that the two modalities are spatially co-registered. ¹

The reason this matters is that SPECT image quality degrades quietly. A camera can pass a casual visual flood check while a center-of-rotation offset is smearing every reconstruction, or a single photomultiplier tube can drift enough to create a regional non-uniformity that becomes a ring artifact in the tomographic image. SPECT instrumentation is more complex than planar imaging and requires careful, structured quality control to ensure optimum performance; the published QC tests exist precisely because the failure modes are not always visible without a phantom, a metric, and a tolerance. ⁹

This guide is written for the technologists who run daily QC, the physicists who perform acceptance and annual testing, and the administrators who own the accreditation obligations. It explains what each test measures, how the key numbers are computed, how often the tests run, and where the tolerances come from. DRPS performs this work as part of its PET/CT and nuclear medicine physics and accreditation support services.

Topic Explanation

SPECT/CT quality control verifies, on a defined schedule and against defined tolerances, that the gamma camera, the tomographic reconstruction chain, and the CT subsystem all perform as specified. The governing performance framework for the gamma-camera side is the NEMA standard NU-1, Performance Measurements of Gamma Cameras, which defines how parameters such as uniformity, spatial resolution, energy resolution, and sensitivity are measured so that results are reproducible and comparable across vendors and over time. ²

For the radionuclides imaged on these systems—Tc-99m, I-123, In-111, Lu-177 SPECT, and others—see Understanding Common Isotopes in PET & Radiopharmaceutical Therapy. A few terms recur throughout and are worth defining first.

Field uniformity (intrinsic vs. extrinsic; integral vs. differential)

Field uniformity describes how evenly a gamma camera responds to a uniform flux of photons across its field of view. It is the single most informative routine test because almost every detector or electronics problem shows up as a non-uniformity first.

Intrinsic uniformity is measured with the collimator removed, using a point source (commonly Tc-99m) placed at five or more field-of-view diameters from the bare crystal so the flux is effectively uniform. It characterizes the crystal, photomultiplier tubes (PMTs), and position/energy electronics.
Extrinsic (system) uniformity is measured with the collimator in place, using a flood source (a Co-57 sheet source or a refillable Tc-99m flood). It characterizes the camera the way it actually images patients, collimator included.
Integral uniformity (IU) is a global metric: the largest fractional difference between the maximum and minimum pixel counts across the useful field of view (UFOV) or central field of view (CFOV).
Differential uniformity (DU) is a local metric: the worst rate of change in counts over a short sliding window (typically five contiguous pixels) in rows and columns, which catches sharp regional non-uniformities that a global metric can average away.

Center of rotation (COR)

Center of rotation is the calibrated relationship between the mechanical axis the detector heads rotate about and the center of the reconstruction matrix. If the projection data are offset from the assumed axis, filtered back-projection or iterative reconstruction places counts in the wrong location, producing characteristic blurring or a ring/tuning-fork artifact. COR calibration measures the offset using a point or line source and corrects it. ¹³

Spatial resolution and energy resolution

Spatial resolution is the camera's ability to distinguish two nearby sources, quantified as the full width at half maximum (FWHM) of a line- or point-spread function. Routine resolution and linearity are often checked visually and semi-quantitatively with a four-quadrant bar phantom. Energy resolution is the precision of the camera's energy measurement, expressed as the percent FWHM of the photopeak, and it governs how well the system rejects scattered photons.

Sensitivity and multi-detector registration

Planar sensitivity is the count rate the camera records per unit activity (e.g., counts per minute per MBq), and SPECT/system sensitivity extends this to the tomographic geometry. On multi-head cameras, multi-detector registration verifies that all detectors map the same physical point to the same image location—essential for artifact-free reconstruction.

Key Technical Principles

The core gamma-camera and SPECT tests each pair a physical parameter with a defined measurement method, a phantom or source, a frequency, and a tolerance. The table below summarizes the standard test set. Treat the tolerances as representative starting points; the binding values come from the manufacturer's specification, NEMA NU-1 methodology, the applicable accreditation program, and your state or NRC requirements. ²⁴

The SPECT/gamma-camera QC test set

Test	Metric	Method / phantom	Typical frequency	Representative tolerance
Energy peaking	Photopeak window centered on the radionuclide	Daily peak check on each isotope used	Daily	Window centered on photopeak (e.g., 140 keV ± ~10% for Tc-99m)
Field uniformity (intrinsic or extrinsic)	Integral & differential uniformity (%)	Point source (intrinsic) or flood source (extrinsic)	Daily	Visual + quantitative within manufacturer limits; degraded flood prompts service
High-count flood uniformity	Integral uniformity (UFOV & CFOV), differential uniformity	High-count flood (e.g., 30+ million counts)	Weekly / monthly	IU on the order of a few percent over UFOV/CFOV
Spatial resolution & linearity	Visual bar resolution; FWHM (acceptance)	Four-quadrant bar phantom (routine); line/point source (acceptance)	Weekly / monthly + annual	Smallest bars resolved per spec; intrinsic FWHM within vendor spec (often a few mm)
Energy resolution	% FWHM of the photopeak	Energy spectrum of Tc-99m or Co-57 source	Acceptance / annual	Often roughly 9–11% FWHM at 140 keV for modern NaI(Tl) cameras
Center of rotation (COR)	COR offset (mm / pixels)	Point or line source, full SPECT orbit	Weekly / monthly	Offset within sub-pixel to ~1 pixel of calibration
Multi-detector registration	Spatial offset between detector heads	Point/line source imaged by all heads	Monthly / annual	Heads register to within ~1 pixel
Planar / SPECT sensitivity	Count rate per unit activity (cpm/MBq)	Calibrated flood or point source, known activity	Annual (trend)	Stable vs. baseline within a small percentage; constancy is the key signal
Tomographic uniformity & contrast	Reconstructed uniformity, contrast, artifacts	SPECT phantom (e.g., Jaszczak-type)	Quarterly / annual	Uniform reconstruction, expected cold-rod/sphere visibility, no rings
CT subsystem QC	CT number accuracy, uniformity, noise, slice	CT QC phantom	Daily/periodic + annual	Water within tolerance; per CT acceptance criteria
SPECT/CT co-registration	Spatial alignment of emission and CT data	Point sources / alignment phantom	Periodic + annual	Misregistration within sub-voxel to a small fraction of a voxel

The cadence column reflects common practice; published surveys confirm that daily uniformity and energy checks are nearly universal, while the more labor-intensive tests—COR, resolution, sensitivity, and registration—are performed at more variable intervals across facilities. ¹ The remedy for that variability is a written QC schedule with documented pass/fail criteria and trend review.

Worked math 1: integral uniformity

Integral uniformity reduces a flood image to one number that captures the worst global non-uniformity. The standard expression is:

IU = \frac{max - min}{max + min} \times 100%

where $max$ and $min$ are the highest and lowest pixel counts in the region of interest (UFOV or CFOV) of a high-count flood, usually after a light smoothing step defined by the measurement method. Suppose a high-count flood yields a maximum pixel count of $max = 10, 450$ and a minimum of $min = 9, 620$ within the CFOV. Then:

IU = \frac{10 , 450 - 9 , 620}{10 , 450 + 9 , 620} \times 100% = \frac{830}{20 , 070} \times 100% \approx 4.1%

So this field is non-uniform at the ~4% level by the integral metric. Whether that passes depends on the camera's specification and the measurement method—modern cameras are often specified to a few percent over the UFOV, but the binding number is the vendor's value evaluated under the NEMA NU-1 method.

The reason high count statistics matter is that uniformity is measured against statistical noise. At low counts, Poisson fluctuations alone inflate the apparent non-uniformity; acquiring tens of millions of counts pushes the statistical floor well below the systematic non-uniformity you are trying to measure. This is why the formal uniformity test is a high-count flood, while the daily flood is a quicker constancy and artifact check.

Worked math 2: energy resolution

Energy resolution tells you how tightly the camera measures photon energy, which governs scatter rejection. It is the photopeak FWHM expressed as a percentage of the photopeak energy:

% Energy Resolution = \frac{FWHM}{E _{photopeak}} \times 100%

For Tc-99m, the photopeak is $E_{photopeak} = 140 keV$ . If the measured photopeak FWHM is $FWHM = 14 keV$ , then:

% Energy Resolution = \frac{14 keV}{140 keV} \times 100% = 10%

A value near 10% at 140 keV is in the range commonly seen for modern NaI(Tl) gamma cameras, but the acceptance value is the camera's published specification. A degrading energy resolution (a widening photopeak) means the camera is accepting a broader band of energies, which lets in more scattered photons and reduces image contrast—so this metric is a sensitive indicator of detector and PMT health.

Worked math 3: how a COR offset degrades resolution

A center-of-rotation error displaces every projection, and that displacement adds directly to the reconstructed blur. Conceptually, the SPECT spatial resolution combines the intrinsic system resolution with any geometric error from COR misalignment. A common first-order way to think about it is to combine the contributions in quadrature:

R_{total} \approx R_{system}^{2} + R_{COR}^{2}

where $R_{system}$ is the system resolution FWHM with perfect alignment and $R_{COR}$ is the additional blur from the COR offset. Take a system whose tomographic FWHM is $R_{system} = 10 mm$ . If an uncorrected COR offset contributes an effective blur of $R_{COR} = 6 mm$ , then:

R_{total} \approx (10)^{2} + (6)^{2} = 136 \approx 11.7 mm

A 6 mm misalignment has degraded resolution from 10 mm to roughly 11.7 mm—about a 17% loss—and in practice a COR error also produces characteristic ring or "tuning-fork" point-source artifacts rather than simple symmetric blur. This is why COR is calibrated and then verified on a schedule: a small mechanical or electronic drift that no one would notice on a planar image can quietly corrupt every tomographic study. It is also worth remembering that COR calibration alone does not guarantee a clean reconstruction—detector gaps and head-tilt errors can produce sinogram artifacts even when the COR test passes. ³

Clinical Impact

SPECT QC failures translate directly into diagnostic errors: false defects, missed lesions, and unreliable quantification. Each test maps to a recognizable artifact.

Non-uniformity → ring artifacts. Because SPECT reconstructs from projections acquired all around the patient, a fixed regional non-uniformity sweeps into a concentric ring in the reconstructed slice. In cardiac perfusion imaging, that ring or a regional sensitivity dip can mimic or mask a true perfusion defect—which is why uniformity tolerances are tighter for SPECT than the eye would demand of a planar image.
COR offset → blurring and artifacts. An uncorrected COR error blurs the reconstruction and can create point-source "tuning-fork" patterns, degrading small-lesion detectability and the accuracy of any quantitative measurement. ³
Energy-resolution drift → scatter and contrast loss. A widening photopeak admits more scatter, washing out contrast and degrading lesion conspicuity.
Sensitivity drift → quantification error. Programs that rely on quantitative SPECT—including Lu-177 post-therapy imaging and other dosimetry workflows—depend on stable, calibrated sensitivity. For how quantitative imaging feeds dosimetry, see Lu-177 Theranostics Dosimetry.
CT misregistration → attenuation-correction artifacts. On SPECT/CT, the CT provides the attenuation map; if the emission and CT data are spatially misaligned—from patient motion or a calibration error—the correction lands on the wrong location and can create false defects, most notoriously in cardiac SPECT/CT. Co-registration QC guards against this.

The patient never sees the QC, but the reading physician sees its consequences. A disciplined QC program is, functionally, a diagnostic-accuracy program.

Practical Optimization Tips

A strong SPECT/CT QC program is built on three habits: acquire enough counts, trend the numbers, and never sign off a study with an open artifact. The following practices come up repeatedly in well-run nuclear medicine departments.

Run daily QC before the first patient. Peak the camera on every radionuclide in use that day, not just Tc-99m, then acquire the daily flood and review it both visually and against the quantitative limits. A flood that "looks fine" can still be drifting—the number is the early-warning system, so investigate any new non-uniformity, edge-packing change, or spot before clinical imaging.
Acquire high-count floods for the formal uniformity test. Use enough counts (commonly tens of millions) so Poisson noise does not masquerade as non-uniformity, and apply the smoothing and region definition specified by the measurement method so your IU and DU values stay comparable over time and against the specification. ²
Calibrate and then verify center of rotation. Run COR calibration per the manufacturer's procedure, then verify it on the routine schedule—these are different steps. A passing COR test does not rule out every geometric problem; head tilt and inter-detector gaps can still create sinogram artifacts, so inspect raw projections and sinograms when an image looks wrong. ³
Use the right phantom for the question. A four-quadrant bar phantom is a fast routine check of resolution and linearity; a line- or point-source measurement is the quantitative FWHM tool at acceptance and annual testing; and a SPECT performance phantom (e.g., a Jaszczak-type with cold rods and spheres) is what accreditation programs expect for tomographic uniformity, contrast, and artifact assessment.
Trend, don't snapshot. Log IU, DU, COR offset, energy resolution, and sensitivity over time. A single in-tolerance value tells you little; a slow creep in IU or a steady widening of the photopeak flags a failing component before it fails an acceptance limit—and published surveys note that trend analysis is exactly the practice facilities most often skip. ¹
Don't forget the CT half. Run the CT subsystem QC (CT number accuracy, uniformity, noise) on its own schedule and verify SPECT/CT co-registration so attenuation correction lands on the right anatomy. The CT-side discipline mirrors diagnostic CT QC; see CT Protocol Optimization for related principles.

Common pitfalls to avoid

Confusing the daily flood with the uniformity test. The daily flood is a constancy/artifact check; the formal NEMA-method high-count flood is the quantitative uniformity test.
Calibrating COR but never verifying it. Drift happens between calibrations.
Trusting a passing test over a bad-looking image. Some artifacts (detector gaps, head tilt) survive in-tolerance QC.
Ignoring the CT subsystem and co-registration. Half of a SPECT/CT system is a CT scanner.
Recording numbers without trending them. The value of QC is in the trajectory, not the snapshot.

Regulatory Considerations

SPECT/CT QC sits at the intersection of accreditation requirements, radioactive-material regulation, and X-ray (CT) regulation—three different authorities for one machine. The QC program should satisfy all three and be documented well enough to defend during an inspection or accreditation survey.

NEMA NU-1, Performance Measurements of Gamma Cameras, is the technical standard that defines how the gamma-camera parameters (uniformity, resolution, energy resolution, sensitivity, count-rate performance) are measured. It is the methodology your acceptance and annual reports should reference so results are reproducible and comparable. ²
The ACR Nuclear Medicine and PET Accreditation Program requires documented QC, qualified medical physicist oversight, and phantom imaging within tolerance for accredited facilities. Accreditation is frequently tied to reimbursement and to state requirements, so QC failures have business consequences, not just technical ones. For the broader physics expectations behind accreditation, see ACR Accreditation Physics Requirements. ⁴
AAPM reports on rotating scintillation camera SPECT (for example, the Task Group reports issued as AAPM Report No. 22 and Report No. 52) provide the foundational acceptance-testing and QC methodology that much of clinical SPECT QC is built on. ⁵
IAEA Human Health Series guidance on quality assurance for SPECT and SPECT/CT systems offers internationally recognized QC protocols and tolerances, useful as a cross-check and for programs that follow IAEA methodology. ⁶
SNMMI procedure standards and guidelines describe expected instrumentation QC practices for nuclear medicine, including gamma-camera and SPECT quality control, and are a common reference for departmental policy. ¹⁰
NRC 10 CFR Part 35 (Medical Use of Byproduct Material) governs the medical use of the radionuclides imaged on these systems, including authorized-user, dose-calibrator, and survey requirements, while 10 CFR Part 20 sets the underlying dose limits. The dose calibrator that measures every administered activity has its own QC obligations; see Dose Calibrator Quality Control. ⁷⁸

Jurisdiction matters because a SPECT/CT is regulated by two regimes at once. The radioactive material (the injected radiopharmaceutical) is byproduct material under the NRC or an Agreement State, while the CT subsystem is an X-ray–producing device regulated by the FDA and the state radiation-control program. Of the states DRPS serves, Florida, Maryland, Virginia, California, and Nevada are NRC Agreement States that license medical use of byproduct material and register X-ray machines under their own rules, while Washington, DC is regulated directly by the NRC for byproduct material. A facility must confirm which authority applies to each half of the system before relying on any single QC framework.

Frequently Asked Questions (FAQs)

What is SPECT/CT quality control?

SPECT/CT quality control is the scheduled program of tests that verify a hybrid gamma camera and CT system is performing within tolerance. It includes field uniformity, energy peaking, center-of-rotation calibration, spatial and energy resolution, sensitivity, multi-detector registration, and CT subsystem checks, performed on daily, weekly, quarterly, and annual cadences.

How often should gamma camera and SPECT QC be performed?

Daily QC typically includes energy-peak and field-uniformity checks; weekly or monthly QC adds high-count uniformity, center-of-rotation, and resolution/linearity with a bar phantom; quarterly and annual QC by a medical physicist add SPECT sensitivity, tomographic resolution, multi-detector registration, and CT co-registration. Confirm exact intervals against the manufacturer, ACR, and your state or NRC requirements.

What is the difference between intrinsic and extrinsic uniformity?

Intrinsic uniformity is measured with the collimator removed, using a point source to expose the bare detector, and characterizes the detector and electronics. Extrinsic (system) uniformity is measured with the collimator in place, using a flood source, and characterizes the camera as it images patients.

What is center of rotation and why does it matter?

Center of rotation (COR) is the alignment between the mechanical axis the detectors rotate around and the center of the reconstruction matrix. A COR offset misplaces projection data and blurs or rings reconstructed images, degrading tomographic spatial resolution, so COR is calibrated and verified on a regular schedule.

How is integral uniformity calculated?

Integral uniformity is $IU = (max - min) / (max + min) \times 100%$ , computed over the pixel counts in the useful or central field of view of a high-count flood image. Differential uniformity evaluates the worst local rate of change over a small sliding window, capturing sharp regional non-uniformities that the global integral metric can miss.

Why does SPECT/CT need CT subsystem QC too?

A SPECT/CT scanner contains a CT scanner used both for attenuation correction and anatomic localization, so it requires its own CT quality control—including CT number accuracy, uniformity, noise, and image co-registration with the SPECT data. Misregistration between the emission and CT data can introduce attenuation-correction artifacts.

Does failing SPECT QC affect accreditation?

Yes. Accreditation programs such as the ACR Nuclear Medicine and PET Accreditation Program require documented QC, qualified medical physicist oversight, and phantom performance within tolerance. Persistent uniformity, resolution, or registration failures can jeopardize accreditation and the reimbursement that depends on it.

Key Takeaways

SPECT/CT QC covers two systems. It tests the gamma-camera planar performance, the tomographic SPECT performance, and the CT subsystem, and then confirms the two modalities are co-registered.
Uniformity is the workhorse test. Integral uniformity, $IU = (max - min) / (max + min) \times 100%$ , must be measured on a high-count flood so Poisson noise does not masquerade as non-uniformity; regional non-uniformities become ring artifacts in SPECT.
Center of rotation must be verified, not just calibrated. An uncorrected COR offset blurs reconstructions and produces characteristic artifacts; combine it with system resolution in quadrature to see why even a small offset matters.
Energy resolution is an early-warning metric. A widening photopeak ( $% ER = FWHM / E_{photopeak} \times 100%$ , e.g., ~10% at 140 keV for Tc-99m) admits more scatter and signals detector or PMT degradation.
Trend the numbers. A single in-tolerance value says little; a slow drift in IU, COR offset, energy resolution, or sensitivity flags a failing component before it fails an acceptance limit.
Anchor the program to standards. Use NEMA NU-1 methodology, ACR accreditation requirements, AAPM and IAEA QC guidance, and the applicable NRC or Agreement State rules.

Conclusion

SPECT/CT quality control is not a box-checking exercise—it is the mechanism that keeps a complex hybrid system producing diagnostically reliable, quantitatively trustworthy images. The tests are designed so that each one isolates a specific, often-invisible failure mode: uniformity for detector and electronics drift, center of rotation for reconstruction misalignment, resolution and energy measurements for detector health, sensitivity for quantitative stability, and CT co-registration for attenuation-correction integrity.

The facilities that get the most from their QC are the ones that acquire enough counts, follow the NEMA NU-1 methodology, trend every metric over time, run the CT subsystem and co-registration checks as seriously as the emission tests, and never release a study with an unexplained artifact. Done well, SPECT/CT QC protects the patient, the reading physician, the accreditation, and the reimbursement that depends on all three.

How DRPS Can Help

Diagnostic Radiation Physics Services provides qualified medical physicist support for nuclear medicine and SPECT/CT programs across Florida, Maryland, Virginia, Washington DC, California, and Nevada. Our PET/CT and nuclear medicine physics services include SPECT/CT acceptance testing, NEMA NU-1–based performance verification (uniformity, spatial and energy resolution, COR, sensitivity, and multi-detector registration), CT subsystem and co-registration QC, annual surveys, and QC-program design and trend review.

We also provide accreditation support for the ACR Nuclear Medicine and PET Accreditation Program, helping facilities assemble documentation, pass phantom requirements, and satisfy the qualified-medical-physicist oversight that accreditation requires. If you are commissioning a new SPECT/CT, troubleshooting ring artifacts or COR drift, or preparing for accreditation, our board-certified medical physicists can help. Contact DRPS or review our service areas to get started.

Related Resources

References

Ramkishore N, Crocker J, Martin R, Yap KS, Brady Z. A survey of gamma camera and SPECT/CT quality control programs across a sample of public hospitals in Australia. Phys Eng Sci Med. 2024;47(3):1153-1166. doi:10.1007/s13246-024-01436-7. doi.org
National Electrical Manufacturers Association. NU 1: Performance Measurements of Gamma Cameras. Rosslyn, VA: NEMA. nema.org
Kheruka SC, Naithani UC, Aggarwal LM, Painuly NK, Maurya AK, Gambhir S. An investigation of a sinogram discontinuity artifact on myocardial perfusion imaging. J Nucl Med Technol. 2012;40(1):25-28. doi:10.2967/jnmt.111.095562. doi.org
American College of Radiology. Nuclear Medicine and PET Accreditation Program. acraccreditation.org
American Association of Physicists in Medicine. Rotating Scintillation Camera SPECT Acceptance Testing and Quality Control (AAPM Report No. 22) and Quantitation of SPECT Performance (AAPM Report No. 52). College Park, MD: AAPM. aapm.org
International Atomic Energy Agency. Quality Assurance for SPECT Systems. IAEA Human Health Series No. 6. Vienna: IAEA; 2009. iaea.org
U.S. Nuclear Regulatory Commission. 10 CFR Part 35: Medical Use of Byproduct Material. ecfr.gov
U.S. Nuclear Regulatory Commission. 10 CFR Part 20: Standards for Protection Against Radiation. ecfr.gov
Groch MW, Erwin WD. Single-photon emission computed tomography in the year 2001: instrumentation and quality control. J Nucl Med Technol. 2001;29(1):12-18. PubMed
Society of Nuclear Medicine and Molecular Imaging. SNMMI Procedure Standards and Practice Guidelines. Reston, VA: SNMMI. snmmi.org
International Atomic Energy Agency. IAEA Human Health Series No. 1: Quality Assurance for PET and PET/CT Systems (companion hybrid-imaging QA guidance). Vienna: IAEA; 2009. iaea.org
Cherry SR, Sorenson JA, Phelps ME. Physics in Nuclear Medicine. 4th ed. Philadelphia, PA: Elsevier Saunders; 2012. elsevierhealth.com

PhysicsPulse^TM Series

Nick Wellnitz

Diagnostic Radiation Physics Services