PET/CT Performance Testing with NEMA NU 2: Sensitivity, NECR, and Resolution

Introduction

NEMA NU 2 is the common language of PET/CT performance — a standardized way to measure what a scanner can actually do, so that vendor claims can be verified and scanners can be compared, accepted, and monitored on an apples-to-apples basis. It defines reproducible measurements of spatial resolution, sensitivity, scatter fraction and count-rate performance (including noise equivalent count rate), accuracy of corrections, image quality, and, in the 2018 edition, time-of-flight resolution and co-registration accuracy.¹

Every PET/CT vendor publishes performance specifications, but those numbers are only meaningful if everyone measures them the same way. NEMA NU 2 provides that shared methodology: specific phantoms, source geometries, activity concentrations, acquisition rules, and analysis algorithms. When a physicist performs NEMA testing during acceptance, the question is not "does this scanner make nice pictures?" but "does this installed system meet the performance the manufacturer promised, measured by a method everyone agrees on?"¹²

This guide walks through the principal NEMA NU 2 measurements, the math behind sensitivity and noise equivalent count rate, what the metrics mean clinically, how modern long axial field-of-view (AFOV) scanners stretch the standard, and how performance testing fits into accreditation and the radioactive-material regulatory framework. DRPS performs this work as part of PET/CT and nuclear medicine physics and medical physics consulting across Florida, Maryland, Virginia, Washington DC, California, and Nevada.

Topic Explanation

What NEMA NU 2 covers

NEMA NU 2 specifies the standardized performance measurements for PET scanners and the conditions under which they must be acquired and analyzed.¹ The core tests, with the quantities they characterize, are summarized below.

Each test isolates one dimension of performance. Together they form a performance "fingerprint" of the scanner that can be checked against vendor specifications and tracked over the equipment's life.¹² For the upstream physics of how these systems form images, see our guide to time-of-flight PET imaging and the overview of common PET and radiopharmaceutical-therapy isotopes.

Why a standard was necessary

Before standardized methods, manufacturers measured sensitivity and resolution under conditions that flattered their own designs, and buyers could not compare systems meaningfully. NEMA NU 2 fixed the phantoms, geometries, and analysis so a sensitivity of, say, a few cps/kBq means the same thing regardless of who built the scanner.¹² This standardization is also what makes acceptance testing a true verification: the physicist reproduces the manufacturer's stated measurement conditions and confirms the result.

Key Technical Principles

Coincidence types and the count-rate problem

A PET scanner records coincidence events — pairs of 511 keV photons detected within a narrow timing window. Not all coincidences are useful:¹³

• True coincidences (T): both photons come from the same annihilation and travel undeviated. These carry correct line-of-response information.
• Scattered coincidences (S): at least one photon Compton-scattered before detection, so the assigned line of response is wrong. Scatter degrades contrast.
• Random coincidences (R): two photons from unrelated annihilations happen to arrive within the timing window. Randoms add noise and grow roughly with the square of activity.

As injected activity rises, trues increase linearly at first, but scatter and randoms grow too — randoms quadratically. Beyond a certain activity, the data quality stops improving and then deteriorates. NEMA NU 2 captures this with the noise equivalent count rate.¹³⁴

Noise equivalent count rate (NECR)

NECR collapses the three coincidence rates into a single effective count rate that reflects the statistical quality of the reconstructed data. The standard form is:¹³⁴

where $T$, $S$, and $R$ are the true, scattered, and random coincidence rates, and $k$ accounts for the method of randoms estimation. When randoms are estimated by a noiseless method, $k=1$; when randoms are estimated from a delayed coincidence window (which itself adds noise), $k=2$.¹³

The structure of the equation explains its behavior. At low activity, $R$ and $S$ are small, so $\mathrm{NECR}\approx T$, rising with activity. At high activity, $R$ dominates the denominator. Because $T$ grows linearly with activity $A$ while $R$ grows roughly as $A^2$:

The result is a curve that rises, reaches a peak NECR at an optimal activity, then falls. The peak NECR and the activity at which it occurs are key performance descriptors, and they directly inform how much activity is useful to administer for a given scanner.³⁴

A worked NECR example

Suppose at a particular activity concentration a scanner records $T=600,000$ cps, $S=200,000$ cps, and $R=400,000$ cps, with randoms estimated from a delayed window ($k=2$). The NECR is:

So although the scanner detected 600,000 true counts per second, the noise contributed by scatter and randoms reduces the effective useful count rate to about 225,000 cps. Repeating this calculation across a decaying source produces the full NECR curve, from which the peak is read.¹³

PET sensitivity

System sensitivity quantifies how efficiently the scanner converts activity into recorded trues. NEMA NU 2 measures it with a thin line source threaded through concentric metal sleeves; the count rate is recorded for each added sleeve and extrapolated to zero attenuating material:¹²

where $R_{\text{trues}}$ is the attenuation-corrected (extrapolated) true coincidence rate and $A$ is the source activity, giving sensitivity in counts per second per becquerel (commonly reported as cps/kBq). Higher sensitivity means more of the emitted signal is captured, enabling lower injected activity, shorter scans, or better image statistics. Long axial field-of-view scanners achieve dramatic sensitivity gains precisely because they intercept solid angle that conventional scanners miss; reported NEMA NU 2-2018 sensitivity for a total-body system is far above that of conventional axial coverage.⁵

Spatial resolution

Spatial resolution is measured from point sources and reported as the full width at half maximum (FWHM) of the reconstructed point-spread function in the radial, tangential, and axial directions at defined radial offsets.¹ FWHM relates to the Gaussian standard deviation $\sigma$ by:

Resolution degrades toward the edge of the field of view because of effects such as photon non-collinearity, detector parallax (depth of interaction), and positron range. Reporting FWHM at multiple radial positions captures this spatial variation, which matters for lesion detectability away from the scanner center.¹²

Clinical Impact

Acceptance testing protects the investment

A PET/CT system is a major capital purchase tied to contractual performance specifications. NEMA NU 2 acceptance testing is how a facility verifies, independently, that the installed scanner actually delivers its promised sensitivity, resolution, count-rate performance, and image quality before final acceptance. A measured shortfall — for example, peak NECR well below specification — is a documented basis for vendor remediation.²⁵

Baselines for ongoing QA

The NEMA results captured at acceptance become the baselines against which routine quality assurance is compared. Drift in sensitivity, a falling NECR peak, or degrading uniformity can signal detector aging, electronics problems, or calibration issues before they visibly harm clinical images. This complements daily and periodic QA such as SPECT/CT quality control and dose calibrator quality control in a comprehensive nuclear medicine QA program.

Quantitative accuracy and SUV

PET is increasingly used quantitatively — standardized uptake values (SUV) drive treatment response assessment and, in theranostics, dosimetry. Quantitative accuracy depends on the accuracy-of-corrections performance that NEMA characterizes at clinically relevant count rates. The same count-rate physics that shapes NECR also governs whether SUVs remain accurate when activity is high. The relationship between NECR and effective image signal-to-noise has been demonstrated directly in total-body PET, reinforcing why count-rate performance is not an abstract bench metric but a determinant of image quality.⁶ For the downstream use of quantitative PET in therapy planning, see Lu-177 theranostics dosimetry and Y-90 radioembolization dosimetry.

Choosing administered activity

Because NECR peaks at a specific activity, NEMA data inform how much radiopharmaceutical is worth injecting. Pushing activity past the NECR peak increases patient dose without improving — and possibly worsening — image quality. The NECR curve is therefore a quantitative bridge between scanner performance and ALARA-conscious dosing.³⁴

Practical Optimization Tips

1. Plan the acquisition geometry carefully

NEMA tests are exacting about source position, phantom fill, and activity. Small errors in line-source placement or fill activity propagate into sensitivity and NECR results. Use calibrated sources, document the dose calibrator measurement, and follow the standard's geometry precisely.¹

2. Match the standard edition and the scanner class

Confirm whether testing is to NEMA NU 2-2007, 2012, or 2018, because methods and reported quantities differ — most importantly the addition of time-of-flight resolution in the 2018 edition. For total-body and long-AFOV systems, anticipate that the standard geometry may need supplementary or adapted measurements to characterize the very long axial coverage fairly.¹⁵

3. Verify against the right specification sheet

Compare measured results to the vendor's NEMA specification for the same configuration (crystal, electronics, software version, reconstruction). A mismatch in configuration is a common reason measured and specified values appear to disagree when the scanner is actually fine.

4. Capture defensible baselines

Record raw data, analysis software versions, and environmental conditions. These baselines are only useful for trending if they are reproducible, so standardize the procedure and store it with the physics report.²

5. Connect NECR to clinical dosing

Translate the NECR peak into practical guidance on administered activity and scan duration for your clinical protocols, in collaboration with the nuclear medicine physician and technologists. This turns a bench measurement into protocol optimization.³⁴

Common pitfalls to avoid

• Comparing across NEMA editions. A 2007 sensitivity and a 2018 sensitivity are not always directly comparable.
• Ignoring randoms estimation method. Whether $k=1$ or $k=2$ materially changes reported NECR.
• Forcing long-AFOV scanners into the standard geometry alone. Supplementary tests are usually needed.
• Treating NECR as the only metric. Resolution, TOF, sensitivity, and image quality all matter.
• Skipping configuration matching. Always compare like-for-like against the vendor spec.
• Letting baselines drift undocumented. Reproducibility is what makes trending meaningful.

Regulatory Considerations

PET/CT performance testing intersects two regulatory worlds: the radioactive-material framework that governs the PET radiopharmaceuticals, and the accreditation and professional-standard framework that governs imaging quality.

• Radioactive material (PET side). Possession and medical use of positron-emitting radiopharmaceuticals such as F-18 fall under 10 CFR Part 35 (medical use of byproduct material) or the equivalent Agreement State program, with dose limits set by 10 CFR Part 20.⁷⁸ Of the jurisdictions DRPS serves, Florida, Maryland, Virginia, California, and Nevada are Agreement States administering their own equivalent rules, while Washington DC is regulated directly by the NRC. The CT subsystem is regulated as a radiation-producing machine under state radiation-control programs.
• Accreditation. PET/CT imaging is commonly accredited through programs such as the ACR PET Accreditation Program, which require evidence of physicist oversight, image quality, and quantitative performance. NEMA-based acceptance and QA data support these submissions.⁹
• Professional standards and guidance. SNMMI, the IAEA Human Health Series, and AAPM provide guidance on PET/CT acceptance testing and ongoing QA that complements the NEMA measurement methods.²¹⁰

Documented NEMA acceptance results, traceable baselines, and a physicist's report are what make a PET/CT performance program defensible during accreditation review and consistent with the broader radiation-safety obligations of a nuclear medicine license. For the licensing side, see our radioactive material license support and the guide to PET/CT shielding calculations.

Frequently Asked Questions (FAQs)

What is NEMA NU 2?

NEMA NU 2 is the National Electrical Manufacturers Association standard that defines standardized measurements of PET scanner performance. It specifies phantoms, source geometries, activity levels, and analysis methods for spatial resolution, sensitivity, scatter fraction and count-rate performance, accuracy of corrections, image quality, and — in the 2018 edition — time-of-flight resolution and co-registration accuracy.

Why does NEMA NU 2 matter for a clinical PET/CT program?

It provides a vendor-neutral way to compare scanners, verify that an installed system meets its published specifications during acceptance testing, and establish baselines for ongoing quality assurance. Without a common standard, performance claims from different manufacturers would not be comparable.

What is noise equivalent count rate (NECR)?

NECR is a single figure of merit that combines true, scattered, and random coincidence rates into an effective count rate that reflects the statistical quality of the data. It is defined so that as scatter and randoms grow with activity, the effective useful count rate eventually peaks and then declines, identifying the activity at which a scanner performs best.

Does a higher peak NECR always mean better images?

Higher peak NECR generally indicates better count-rate capability, but it is not the only factor. Spatial resolution, sensitivity, scatter fraction, time-of-flight resolution, reconstruction, and the clinical activity and uptake conditions all affect image quality. NECR is one important metric among several, not a complete description of a scanner.

How is PET sensitivity measured under NEMA NU 2?

Sensitivity is measured with a line source surrounded by successive metal sleeves of known thickness. The count rate is recorded for each sleeve, and the data are extrapolated to zero attenuating material to obtain the system sensitivity, typically expressed in counts per second per becquerel (cps/kBq).

Did NEMA NU 2-2018 change anything for long axial field-of-view scanners?

Yes. NEMA NU 2-2018 added time-of-flight resolution as a standard measurement and refined methods relevant to modern systems. Total-body and long axial field-of-view scanners often require additional or adapted tests beyond the standard geometry because their very long axial coverage changes how sensitivity and count-rate behavior are characterized.

Who should perform PET/CT NEMA testing?

Acceptance testing and performance verification should be performed or overseen by a qualified or board-certified medical physicist with PET/CT experience. The physicist confirms that measured performance matches vendor specifications and establishes the baselines used for ongoing quality assurance and accreditation.

Key Takeaways

• NEMA NU 2 is the standardized language of PET performance. It fixes phantoms, geometries, and analysis so scanners can be compared and verified.
• The core tests are resolution, sensitivity, scatter/count-rate (NECR), accuracy of corrections, image quality, and — since 2018 — TOF resolution and co-registration.
• NECR collapses trues, scatter, and randoms into one effective count rate that peaks at an optimal activity, directly informing dosing and ALARA.
• Sensitivity (cps/kBq) governs how much signal is captured, and long-AFOV scanners achieve large gains by intercepting more solid angle.
• Acceptance testing verifies vendor claims and sets QA baselines; trending those baselines catches degradation early.
• Performance data underpin quantitative PET (SUV) and accreditation, and the program lives within the 10 CFR Part 35 / Agreement State framework for the radiopharmaceuticals involved.

Conclusion

NEMA NU 2 turns "this scanner performs well" into measurable, comparable, defensible numbers. By standardizing how spatial resolution, sensitivity, scatter fraction, NECR, accuracy of corrections, image quality, and time-of-flight resolution are measured, it lets a facility verify what it bought, baseline what it has, and detect when performance drifts.

For a modern PET/CT program — especially one moving toward quantitative imaging and theranostics — that rigor is not optional. The NECR peak informs dosing, sensitivity informs scan time, and the full performance fingerprint informs whether the scanner is fit for the clinical and quantitative tasks asked of it. A qualified medical physicist, using NEMA NU 2 methods, is what connects those measurements to safe, accurate, accreditation-ready clinical practice.

How DRPS Can Help

Diagnostic Radiation Physics Services performs PET/CT acceptance testing and ongoing performance QA using NEMA NU 2 methods as part of PET/CT and nuclear medicine physics and medical physics consulting. This can include sensitivity, NECR, spatial resolution, image quality, time-of-flight, and co-registration measurements; comparison to vendor specifications; baseline establishment for QA trending; SUV and quantitative accuracy verification; and the documentation needed for accreditation support and radioactive material license compliance.

DRPS supports nuclear medicine and PET/CT facilities across our service locations, including Florida, Maryland, Virginia, Washington DC, California, and Nevada. Strong performance testing is not paperwork — it is how you prove your scanner does what your patients and your quantitative reads depend on.

Related Resources

• Time-of-flight PET imaging
• Common PET & RPT isotopes
• SPECT/CT quality control
• Dose calibrator quality control
• Lu-177 theranostics dosimetry
• PET/CT shielding calculations guide
• PET/CT and nuclear medicine physics
• Accreditation support

References

1. National Electrical Manufacturers Association. NEMA Standards Publication NU 2-2018: Performance Measurements of Positron Emission Tomographs (PET). Rosslyn, VA: NEMA; 2018. nema.org
2. International Atomic Energy Agency. PET/CT Atlas on Quality Control and Image Artefacts. IAEA Human Health Series No. 27. Vienna: IAEA; 2014. iaea.org
3. Yang X, Peng H. The use of noise equivalent count rate and the NEMA phantom for PET image quality evaluation. Physica Medica. 2015;31(2):179-184. doi:10.1016/j.ejmp.2015.01.003. doi.org
4. Strother SC, Casey ME, Hoffman EJ. Measuring PET scanner sensitivity: relating count rates to image signal-to-noise ratios using noise equivalent counts. IEEE Transactions on Nuclear Science. 1990;37(2):783-788. doi:10.1109/23.106715. doi.org
5. Spencer BA, Berg E, Schmall JP, et al. Performance evaluation of the uEXPLORER total-body PET/CT scanner based on NEMA NU 2-2018 with additional tests to characterize PET scanners with a long axial field of view. Journal of Nuclear Medicine. 2021;62(6):861-870. doi:10.2967/jnumed.120.250597. doi.org
6. Leung EK, Abdelhafez YG, Berg E, et al. Relating 18F-FDG image signal-to-noise ratio to time-of-flight noise-equivalent count rate in total-body PET using the uEXPLORER scanner. Physics in Medicine and Biology. 2022;67(12). doi:10.1088/1361-6560/ac72f1. doi.org
7. U.S. Nuclear Regulatory Commission. 10 CFR Part 35: Medical Use of Byproduct Material. nrc.gov
8. U.S. Nuclear Regulatory Commission. 10 CFR Part 20: Standards for Protection Against Radiation. nrc.gov
9. American College of Radiology. PET Accreditation Program Requirements. Reston, VA: ACR. acr.org
10. Society of Nuclear Medicine and Molecular Imaging. SNMMI Procedure Standards and PET/CT acceptance testing guidance. Reston, VA: SNMMI. snmmi.org
11. National Electrical Manufacturers Association. NEMA NU 2-2012: Performance Measurements of Positron Emission Tomographs. Rosslyn, VA: NEMA; 2012. nema.org