Gamma Camera Collimator Selection: Balancing Resolution, Sensitivity, and Septal Penetration

Q: What does a gamma camera collimator do?

A collimator is a perforated absorber, usually lead, mounted between the patient and the detector. It only allows photons traveling in accepted directions to reach the crystal, projecting the radionuclide distribution onto the detector. Because it discards the vast majority of emitted photons, the collimator is the dominant factor limiting both the spatial resolution and the sensitivity of the gamma camera.

Q: When should a high-energy collimator be used instead of a low-energy one?

Collimator energy class is chosen to match the photon energy of the radionuclide. Low-energy collimators are for isotopes like Tc-99m (140 keV). Medium-energy collimators suit isotopes such as In-111 and Ga-67, and high-energy collimators are required for I-131 (364 keV). Using a low-energy collimator with a high-energy isotope causes severe septal penetration and image artifacts.

Q: What is the difference between LEHR and LEAP collimators?

Both are low-energy collimators for Tc-99m imaging. LEHR (low-energy high-resolution) collimators use smaller or longer holes for sharper images at the cost of sensitivity. LEAP, also called low-energy general-purpose, sacrifices some resolution for higher sensitivity, which can be preferable for dynamic or low-count studies.

Q: How is collimator performance verified?

Collimator and overall system performance are characterized with NEMA NU-1 measurements of spatial resolution, sensitivity, and uniformity, and confirmed in routine QC through daily uniformity floods, periodic spatial-resolution and linearity checks, and SPECT center-of-rotation and reconstructed-resolution tests.

A gamma camera collimator is the lead aperture that defines where each detected photon came from, and selecting it is the single most important determinant of nuclear medicine image quality. Every collimator forces a trade-off: the geometry that sharpens spatial resolution simultaneously reduces sensitivity, and the choice must also match the photon energy of the radionuclide to control septal penetration. Choosing the wrong collimator degrades images in ways that reconstruction cannot fully repair. ^{4, 6}

Unlike CT, where resolution is set by detector and reconstruction, a gamma camera's spatial resolution is dominated by the collimator. The detector crystal and electronics contribute an intrinsic resolution, but in most clinical geometries the collimator's geometric resolution is larger and therefore controls the combined system resolution. At the same time, the collimator discards the overwhelming majority of emitted photons — typically allowing only on the order of one in ten thousand to reach the crystal — so it also sets sensitivity. These two performance metrics are physically coupled, and collimator selection is fundamentally about choosing where on the resolution–sensitivity curve a given clinical task should sit. ^{4, 6}

This guide explains how parallel-hole collimators work, the governing resolution and efficiency equations, the septal-penetration rule that ties collimator class to photon energy, a worked example for a Tc-99m low-energy high-resolution (LEHR) collimator, a comparison of collimator classes, and the QC and regulatory framework. DRPS supports gamma camera and SPECT performance evaluation as part of its PET/CT and nuclear medicine physics services across Florida, Maryland, Virginia, Washington DC, California, and Nevada.

Introduction

The collimator exists because gamma rays cannot be focused like visible light. A 140 keV photon from Tc-99m travels in a straight line and is too energetic to refract or reflect usefully, so the only way to form an image is to absorb every photon except those traveling in an accepted direction. A parallel-hole collimator does this with a slab of lead drilled with thousands of parallel holes; photons traveling parallel to the holes pass through, while obliquely traveling photons strike the lead septa and are absorbed. The detector therefore sees a projection of the radionuclide distribution. ⁴

This design is brutally inefficient — it is, by intent, a device that throws away almost all of the signal — and that inefficiency is the price of spatial information. Because resolution and sensitivity both flow from the same hole geometry, the physicist and technologist cannot optimize one without consequences for the other. A collimator chosen for the sharpest possible bone-scan images will produce noisy, count-starved dynamic renal studies; a high-sensitivity collimator chosen for fast first-pass cardiac imaging will blur fine detail. ^{4, 6}

Two additional constraints complete the picture. First, the collimator must match the photon energy of the radionuclide: septa thick enough for 140 keV Tc-99m are hopelessly transparent to 364 keV I-131. Second, the choice interacts with the clinical task — static versus dynamic, planar versus SPECT, count-limited versus resolution-limited. This article ties those constraints together with the underlying physics.

Topic Explanation

What a collimator is and the main designs

A collimator is a removable, perforated absorber, almost always lead, that mounts on the face of the detector and accepts only photons traveling within a narrow angular range. The most common type is the parallel-hole collimator, in which all holes are perpendicular to the detector face and the image is the same size as the object. Other geometries serve special purposes: ⁴

Parallel-hole — the clinical workhorse; 1:1 projection, used for nearly all planar and SPECT imaging.
Pinhole — a single small aperture providing magnification and very high resolution for small organs such as the thyroid, at low sensitivity.
Converging / diverging — magnify or minify the field of view; largely historical on modern large-field cameras.
Fan-beam and cone-beam — converge in one or two dimensions to trade field of view for resolution and sensitivity in dedicated brain or cardiac SPECT.

Within parallel-hole collimators, designs are further classified by energy (low, medium, high) and by purpose (high-resolution, general-purpose, high-sensitivity). The combination — for example "low-energy high-resolution," LEHR — names the specific collimator. ⁴

The resolution–sensitivity trade-off

The central fact of collimator design is that spatial resolution and sensitivity move in opposite directions. Anything that narrows the accepted angular range of photons — smaller hole diameter, longer holes — sharpens resolution but reduces the number of photons accepted, lowering sensitivity. The relationship is not linear; for a parallel-hole collimator, geometric efficiency falls roughly as the square of the geometric resolution, so modest gains in sharpness are expensive in counts. This is why no single collimator is "best" — only best for a task. ^{4, 6}

For background on the radionuclides that drive these energy and count-rate considerations, see our overview of common PET and radiopharmaceutical-therapy isotopes, and for downstream SPECT system testing see SPECT/CT quality control.

Key Technical Principles

Geometric (collimator) resolution

For a parallel-hole collimator, the geometric resolution is set by the hole diameter, the hole length, and the source-to-collimator distance. A widely used approximation for the geometric resolution $R_{co l l}$ (full width at half maximum) is: ⁴

R_{co l l} = d \frac{l _{eff} + b}{l _{eff}}

where $d$ is the hole diameter, $b$ is the source-to-collimator distance, and $l_{eff}$ is the effective hole length, reduced from the physical length $l$ to account for photons that penetrate the ends of the septa:

l_{eff} = l - \frac{2}{μ}

with $μ$ the linear attenuation coefficient of lead at the photon energy. Two consequences are immediate. First, $R_{co l l}$ degrades linearly with distance $b$ : resolution is best when the collimator is as close to the patient as possible, which is why minimizing detector-to-patient gap is a core QC and positioning principle. Second, longer or narrower holes improve resolution. ⁴

The overall system resolution combines the collimator term with the detector's intrinsic resolution $R_{in t}$ in quadrature:

R_{sy s} = R_{in t}^{2} + R_{co l l}^{2}

Because $R_{co l l}$ usually exceeds $R_{in t}$ at clinical distances, the collimator dominates system resolution. ^{4, 6}

Geometric efficiency (sensitivity)

The geometric efficiency $g$ — the fraction of emitted photons (from a source in air) that pass through the collimator — for a parallel-hole collimator is approximately: ⁴

g \approx K^{2} (\frac{d}{l _{eff}})^{2} (\frac{d}{d + t})^{2}

where $t$ is the septal thickness and $K$ is a dimensionless shape constant (about 0.26 for a hexagonal hole array). Note that $g$ is independent of source-to-collimator distance for a parallel-hole collimator — a counterintuitive but important result: moving a planar source farther away blurs the image but does not change total counts (for an extended source). Comparing the two equations shows the trade-off algebraically: improving resolution by shrinking $d$ or lengthening $l_{eff}$ reduces $g$ quadratically. ⁴

Septal penetration sets the energy class

Septa must be thick enough to absorb photons that strike them at oblique angles; otherwise septal penetration produces a blurred background and star-pattern artifacts. The penetration fraction through a septal path of length $w$ is approximately $e^{- μ w}$ . Requiring penetration below about 5% means: ⁴

e^{- μ w} \leq 0.05 \Rightarrow μ w \geq 3

so the shortest path through any septum should be at least about three mean free paths of lead at the photon energy. As photon energy rises, $μ$ for lead falls, so $w$ — and therefore the septal thickness — must increase. This is the entire reason collimators come in energy classes: a septum adequate for 140 keV Tc-99m is far too thin for 364 keV I-131, where $μ$ in lead is several times smaller. The cost of thicker septa is fewer or smaller holes, hence lower sensitivity, which is why high-energy collimators are inherently less efficient. ⁴

Worked example: a Tc-99m LEHR collimator

Consider a representative LEHR collimator with hole diameter $d = 1.5 mm$ , physical hole length $l = 35 mm$ , and septal thickness $t = 0.2 mm$ , imaging Tc-99m at 140 keV. The linear attenuation coefficient of lead at 140 keV is approximately $μ \approx 25 cm^{- 1}$ , so: ¹⁰

l_{eff} = 35 mm - \frac{2}{25 cm ^{- 1}} = 35 mm - 0.8 mm = 34.2 mm

At a clinically realistic source-to-collimator distance $b = 100 mm$ (10 cm), the geometric resolution is:

R_{co l l} = 1.5 \times \frac{34.2 + 100}{34.2} = 1.5 \times 3.92 = 5.9 mm

Combining with a typical intrinsic resolution $R_{in t} = 3.8 mm$ :

R_{sy s} = 3. 8^{2} + 5. 9^{2} = 14.4 + 34.8 = 7.0 mm

The geometric efficiency is:

g \approx (0.26)^{2} (\frac{1.5}{34.2})^{2} (\frac{1.5}{1.7})^{2} \approx 1.0 \times 1 0^{- 4}

So roughly one emitted photon in ten thousand is accepted, and the system resolution at 10 cm is about 7 mm — values typical of clinical LEHR performance. If we instead chose a general-purpose (LEAP) collimator with shorter, wider holes, $R_{sy s}$ would worsen (larger numbers) but $g$ would rise by a factor of several, illustrating the trade-off quantitatively. These numbers are illustrative; actual values depend on the specific collimator geometry and should be confirmed by NEMA NU-1 measurement. ^{1, 4}

Comparison of collimator classes

Collimator class	Typical radionuclide / photon energy	Septal thickness	Relative spatial resolution	Relative sensitivity	Typical clinical use
Low-energy high-resolution (LEHR)	Tc-99m (140 keV)	Thin	Highest (sharpest)	Lower	Bone, parathyroid, most static Tc-99m imaging
Low-energy general-purpose (LEAP/LEGP)	Tc-99m (140 keV)	Thin	Moderate	Higher	Dynamic renal, first-pass, count-limited studies
Low-energy ultra-high-resolution (LEUHR)	Tc-99m (140 keV)	Thin	Very high	Lowest of low-energy	High-detail planar imaging
Medium-energy (ME / MEGP)	In-111 (171/245 keV), Ga-67 (up to 300 keV)	Thicker	Moderate	Moderate	In-111 and Ga-67 imaging
High-energy (HE)	I-131 (364 keV)	Thickest	Lower	Lowest	I-131 therapy imaging, post-therapy scans
Pinhole	Tc-99m, I-123 thyroid	N/A (single aperture)	Very high (magnified)	Very low	Thyroid, small-organ, high-magnification imaging

Energy-class selection is non-negotiable: imaging I-131 with a low-energy collimator produces severe septal penetration that no reconstruction can remove. Within an energy class, the high-resolution versus general-purpose choice is a clinical optimization. ^{4, 6}

Clinical Impact

Collimator choice directly shapes diagnostic quality and study feasibility. For a bone scan, where lesion conspicuity depends on resolving small foci against background, a LEHR collimator is standard because the task is resolution-limited and counts are adequate. For a dynamic renal study, where temporal sampling demands high count rates over short frames, a general-purpose collimator's higher sensitivity may be preferable even at some cost in resolution. The collimator is therefore part of protocol design, not just a hardware default. ^{4, 6}

In SPECT, the stakes are higher because resolution degrades with distance and the reconstruction amplifies noise. The distance dependence of $R_{co l l}$ means that body habitus and orbit radius materially affect reconstructed resolution; keeping detectors close to the patient (non-circular orbits) is a key optimization. Modern resolution-recovery reconstruction can partially compensate for the known, distance-dependent collimator blur, which has enabled cardiac SPECT protocols with reduced dose or time — but it presupposes a correctly matched collimator and accurate modeling. ⁶

The emergence of cadmium-zinc-telluride (CZT) detectors has shifted some of these trade-offs. CZT cardiac cameras pair specialized collimation with improved intrinsic energy and spatial resolution and higher geometric efficiency, allowing reduced acquisition time or administered activity for myocardial perfusion imaging while maintaining diagnostic performance, as reviewed by the EANM cardiovascular committee. The underlying collimator physics still governs the resolution–sensitivity balance; CZT changes the detector side of the equation. ⁷

Mismatched collimators have concrete clinical costs: septal penetration from using a low-energy collimator with In-111 or I-131 creates artifactual background that can mimic or obscure disease, and an overly high-sensitivity collimator can blur small lesions below detectability. These are not subtle effects, and they are largely preventable by correct selection.

Practical Optimization Tips

A defensible collimator-selection practice combines correct energy matching, task-based resolution/sensitivity choices, and disciplined QC.

Match the energy class to the radionuclide

Use low-energy collimators for Tc-99m and I-123/Tl-201 (low-energy emissions); medium-energy for In-111 and Ga-67; high-energy for I-131.
Never image I-131 or other high-energy emitters with a low-energy collimator — septal penetration will dominate the image.
For radionuclides with both low- and high-energy emissions (e.g., I-123 with minor high-energy photons, or In-111), follow manufacturer and procedure-standard recommendations and consider medium-energy collimation to suppress penetration.

Choose resolution vs. sensitivity by task

Resolution-limited, count-adequate tasks (bone, parathyroid, static planar): favor high-resolution (LEHR/LEUHR).
Count-limited or dynamic tasks (renal dynamic, first-pass cardiac, low-activity studies): favor general-purpose or high-sensitivity collimation.
Small organs (thyroid): consider pinhole for magnification and resolution.

Optimize geometry and positioning

Keep the collimator as close to the patient as possible; resolution degrades linearly with distance.
Use non-circular (body-contouring) SPECT orbits to minimize average detector-to-patient distance.
Handle collimators carefully — dropped or dented collimators create uniformity defects that propagate into ring artifacts on SPECT.

Common pitfalls to avoid

Using the default collimator for every study. The on-camera collimator is not always the right one for the radionuclide or task.
Ignoring septal penetration with medium/high-energy isotopes. This is the most common and most damaging mismatch.
Assuming reconstruction will fix a bad choice. Resolution recovery models blur, not lost information or gross penetration.
Neglecting collimator condition. Physical damage causes non-uniformities; verify with uniformity floods.
Comparing studies across collimators without accounting for resolution and sensitivity differences. Quantitative trends require consistent collimation.

Regulatory Considerations

Gamma camera performance, including collimator-dependent resolution and sensitivity, is governed by performance standards (NEMA, IEC), QA guidance (IAEA), professional practice parameters, and — for the radionuclides themselves — NRC or Agreement State materials regulation. ^{1, 2, 3, 8, 9}

Key frameworks:

NEMA NU 1-2018 — the standard method for measuring gamma camera performance, including system spatial resolution, sensitivity, energy resolution, and uniformity, enabling apples-to-apples comparison across systems and collimators. ¹
IEC 61675-2 — international performance and constancy-test standard for SPECT/gamma camera systems. ²
IAEA Human Health Series No. 6 and the IAEA Quality Control Atlas — practical QA protocols for SPECT systems, including acceptance and routine testing where collimator performance is verified. ^{3, 4}
ACR–AAPM–SNMMI technical standards and practice parameters — define qualified-personnel, equipment, and QC expectations for nuclear medicine procedures, often required for accreditation. ⁸
10 CFR Part 35 (Medical Use of Byproduct Material) and 10 CFR Part 20 — govern possession and medical use of the radionuclides imaged, the authorized-user and RSO framework, and occupational and public dose limits. ⁹

Because gamma camera imaging uses byproduct material (Tc-99m, I-131, In-111, Ga-67), the radionuclide side is regulated by the NRC or an Agreement State, not as an X-ray machine. Of the states DRPS serves, Florida, Maryland, Virginia, California, and Nevada are Agreement States licensing medical use under their own radiation-control programs, while Washington, DC is regulated directly by the NRC. Acceptance testing and periodic performance evaluation by a qualified medical physicist are commonly required by accreditation bodies and license conditions; see SPECT/CT quality control and our guide to the radiation safety officer role for how these fit into a nuclear medicine program. ^{8, 9}

Frequently Asked Questions (FAQs)

What does a gamma camera collimator do?

A collimator is a perforated lead absorber mounted between the patient and the detector. It allows only photons traveling in accepted directions to reach the crystal, projecting the radionuclide distribution onto the detector. Because it discards the vast majority of emitted photons, the collimator is the dominant factor limiting both spatial resolution and sensitivity.

Why is there a trade-off between resolution and sensitivity?

Spatial resolution and sensitivity are coupled through the collimator hole geometry. Smaller, longer holes or thicker septa improve resolution but accept fewer photons, lowering sensitivity. For a parallel-hole collimator, geometric efficiency scales roughly with the square of resolution, so you cannot improve one without paying in the other.

When should a high-energy collimator be used instead of a low-energy one?

Collimator energy class matches the radionuclide's photon energy. Low-energy collimators are for Tc-99m (140 keV); medium-energy collimators suit In-111 and Ga-67; high-energy collimators are required for I-131 (364 keV). Using a low-energy collimator with a high-energy isotope causes severe septal penetration and artifacts.

What is septal penetration?

Septal penetration is the passage of photons through the lead septa between holes instead of being absorbed. It produces a low-resolution background and star or streak artifacts and worsens as photon energy rises. Septa are made thick enough that the shortest path through a septum is at least about three mean free paths, keeping penetration below roughly 5%.

What is the difference between LEHR and LEAP collimators?

Both are low-energy collimators for Tc-99m. LEHR (high-resolution) uses smaller or longer holes for sharper images at the cost of sensitivity. LEAP/general-purpose sacrifices some resolution for higher sensitivity, which can be preferable for dynamic or low-count studies.

How is collimator performance verified?

Collimator and system performance are characterized with NEMA NU-1 measurements of resolution, sensitivity, and uniformity, and confirmed in routine QC through daily uniformity floods, periodic resolution and linearity checks, and SPECT center-of-rotation and reconstructed-resolution tests.

Can the wrong collimator be corrected in reconstruction?

Only partly. Resolution-recovery reconstruction can compensate for known collimator blur, but it cannot recover information lost to severe septal penetration or a fundamentally mismatched collimator. Correct selection remains the primary control on image quality.

Key Takeaways

The collimator dominates gamma camera resolution and sensitivity. It is the limiting component, not the detector, at clinical distances.
Resolution and sensitivity are inversely coupled. Geometric efficiency falls roughly as the square of geometric resolution; no collimator is universally best.
Energy class is set by septal penetration. Septa must be at least ~3 mean free paths thick; higher photon energy demands thicker septa and lower sensitivity.
Match the collimator to both radionuclide and task. LEHR for resolution-limited Tc-99m studies; general-purpose for count-limited or dynamic work; high-energy for I-131.
Resolution degrades with distance. Keep detectors close and use body-contouring SPECT orbits.
Verify performance with NEMA NU-1 and routine QC. Reconstruction cannot rescue a mismatched collimator.

Conclusion

Collimator selection is where nuclear medicine physics meets clinical protocol design. The governing equations are simple but unforgiving: geometric resolution improves with smaller, longer holes and shorter source distances, geometric efficiency falls as the square of resolution, and septal thickness must scale with photon energy to suppress penetration. Together they mean that every collimator is a deliberate compromise, and the right compromise depends on the radionuclide and the diagnostic question.

A nuclear medicine program that treats the collimator as a thoughtful, task-specific choice — matched to energy, optimized for the resolution–sensitivity balance the study needs, and verified by NEMA and routine QC — produces images that reconstruction can enhance rather than rescue. Getting the collimator right is the most cost-effective image-quality decision in the entire imaging chain.

How DRPS Can Help

Diagnostic Radiation Physics Services supports nuclear medicine facilities with gamma camera and SPECT acceptance testing, NEMA NU-1 performance evaluation, collimator-specific resolution and sensitivity measurement, uniformity and SPECT reconstruction QC, protocol and collimator-selection review, and the periodic physicist evaluations required for accreditation and license conditions.

DRPS provides PET/CT and nuclear medicine physics and accreditation support across our service locations, including Florida, Maryland, Virginia, Washington DC, California, and Nevada.

A correctly chosen, well-characterized collimator is the foundation of every defensible nuclear medicine image — and the most reliable way to make sure the camera is showing you the patient, not the hardware.

Related Resources

References

National Electrical Manufacturers Association. NEMA Standards Publication NU 1-2018: Performance Measurements of Gamma Cameras. Rosslyn, VA: NEMA; 2018. nema.org
International Electrotechnical Commission. IEC 61675-2: Radionuclide imaging devices — Characteristics and test conditions — Part 2: Single photon emission computed tomographs. Geneva: IEC. iec.ch
International Atomic Energy Agency. Quality Assurance for SPECT Systems. IAEA Human Health Series No. 6. Vienna: IAEA; 2009. iaea.org
Cherry SR, Sorenson JA, Phelps ME. Physics in Nuclear Medicine. 4th ed. Philadelphia, PA: Elsevier Saunders; 2012. elsevier.com
International Atomic Energy Agency. IAEA Quality Control Atlas for Scintillation Camera Systems. Vienna: IAEA; 2003. iaea.org
Weng F, Bagchi S, Zan Y, Huang Q, Seo Y. An energy-optimized collimator design for a CZT-based SPECT camera. Nucl Instrum Methods Phys Res A. 2016;806:330-339. doi:10.1016/j.nima.2015.09.115. doi.org
Agostini D, Marie PY, Ben-Haim S, et al. Performance of cardiac cadmium-zinc-telluride gamma camera imaging in coronary artery disease: a review from the cardiovascular committee of the European Association of Nuclear Medicine (EANM). Eur J Nucl Med Mol Imaging. 2016;43(13):2423-2432. doi:10.1007/s00259-016-3467-5. doi.org
American College of Radiology, American Association of Physicists in Medicine, Society of Nuclear Medicine and Molecular Imaging. ACR–AAPM–SNMMI Technical Standard for Diagnostic Procedures Using Radiopharmaceuticals. Reston, VA: ACR. acr.org
U.S. Nuclear Regulatory Commission. 10 CFR Part 35: Medical Use of Byproduct Material. nrc.gov
National Institute of Standards and Technology. XCOM: Photon Cross Sections Database (X-ray mass attenuation coefficients for lead). nist.gov
Society of Nuclear Medicine and Molecular Imaging. SNMMI Procedure Standards and instrumentation guidance. Reston, VA: SNMMI. snmmi.org