5.2 Collimation Principles & Types

5.2 Collimation Principles & Types

Quick Answer: The collimator acts as the directional geometric lens of the gamma camera, restricting incident gamma rays so that only photons traveling perpendicular (or along specific geometric angles) to the crystal face pass through to form an image. Collimation dictates spatial resolution and count-rate sensitivity, governed by strict mathematical trade-offs between hole length ($l$), hole diameter ($d$), septal thickness ($t$), and source distance ($b$).

Because gamma rays cannot be refracted or focused by optical lenses, planar and SPECT gamma cameras depend entirely on absorptive collimation. Made of lead ($Z=82$) or tungsten ($Z=74$), collimators absorb off-axis photons while allowing parallel or geometrically directed photons to pass through thousands of precisely arrayed channels (bores).

Mathematical Fundamentals of Collimator Performance

Collimator performance is characterized by two competing parameters: Collimator Spatial Resolution ($R_c$) and Geometric Sensitivity ($g$).

1. Collimator Spatial Resolution ($R_c$)

For a parallel-hole collimator, geometric spatial resolution ($R_c$, defined as Full Width at Half Maximum [FWHM] of a line source response) is given by:

$R_c = \frac{d \cdot (l + b)}{l} = d + d \cdot \left(\frac{b}{l}\right)$

Where:

• $d$ = Hole diameter (bore size)
• $l$ = Effective hole length (bore height)
• $b$ = Distance from the front face of the collimator to the target radioactive object

Total System Resolution ($R_{\text{sys}}$): Overall spatial resolution combines intrinsic crystal resolution ($R_{\text{intr}}$) and collimator resolution ($R_c$) in quadrature:

$R_{\text{sys}} = \sqrt{R_{\text{intr}}^2 + R_c^2}$

Clinical Rule: Because $R_c$ degrades linearly as distance $b$ increases, the collimator face must be positioned as close to the patient's body as physically possible. Doubling the distance from $5\ \text{cm}$ to $10\ \text{cm}$ significantly increases $R_c$ (poorer resolution), blurring fine anatomical structures.

2. Geometric Sensitivity ($g$)

Geometric sensitivity represents the fraction of emitted gamma photons that pass through the collimator holes without striking septa. For parallel-hole collimators with round or hexagonal holes:

$g = K^2 \cdot \left( \frac{d^2}{l \cdot (d + t)} \right)^2$

Where:

• $K$ = Geometric constant based on hole array packing (hexagonal vs. square)
• $t$ = Septal thickness

The Distance Independence Paradox

A fundamental property of parallel-hole collimators is that for a uniform, extended source larger than the field of view, geometric sensitivity ($g$) is independent of distance ($b$). As distance increases, the count flux from any single point decreases according to the Inverse Square Law ($1/b^2$), but the area of the source viewed by each hole expands proportionally ($b^2$). The two effects cancel exactly. However, spatial resolution still degrades with distance.

3. The Resolution vs. Sensitivity Trade-Off

• To increase spatial resolution: Make holes longer ($l \uparrow$) or narrower ($d \downarrow$). This restricts the angular acceptance cone. However, sensitivity drops dramatically ($g \propto d^4 / l^2$).
• To increase sensitivity: Make holes shorter ($l \downarrow$) or wider ($d \uparrow$). Count rates increase, but off-axis photons enter, increasing $R_c$ and blurring the image.

Septal Penetration Physics

Septal penetration occurs when high-energy photons pass through lead septa between holes rather than being absorbed. Penetrating photons strike the crystal at incorrect spatial coordinates, causing low-frequency background fog, loss of image contrast, and "starburst" artifacts.

To limit septal penetration ($P$) to less than $5%$, the minimum required septal thickness ($t$) is calculated as:

$t \ge \frac{2 \cdot d \cdot \mu}{l - 2 \cdot \mu}$

Where $\mu$ is the linear attenuation coefficient of the collimator material (lead/tungsten) at the specific gamma energy. Higher energy isotopes (e.g., I-131 at $364\ \text{keV}$) require substantially thicker septa than Tc-99m ($140\ \text{keV}$).

Specialized Collimator Geometries

1. Pinhole Collimator

Features a single conical lead horn terminating in a small interchangeable aperture ($1\ \text{mm}, 3\ \text{mm}, 4\ \text{mm}$ tungsten insert).

• Geometric Magnification: $M = f / b$, where $f$ is focal length (aperture to crystal) and $b$ is source distance. Objects closer than $f$ are magnified ($M > 1$).
• Inversion: Images are inverted along both horizontal and vertical axes (upside down and right-to-left flipped).
• Use: Unmatched spatial resolution for tiny organs (thyroid, parathyroid, pediatric hip joint).
• Limitation: Sensitivity decreases with the square of distance from the aperture.

2. Converging Collimator

Holes are angled inward toward a focal point in front of the camera face.

• Magnification: Image is magnified ($M > 1$) without image inversion.
• Sensitivity: Increases as the target organ approaches the focal point.
• Use: Imaging small structures (e.g., pediatric cardiac/renal) on standard large-field-of-view detectors.

3. Diverging Collimator

Holes are angled outward from a focal point behind the detector.

• Minification: Image is minified ($M < 1$), fitting large anatomical structures (e.g., adult lung perfusion) onto smaller gamma camera detectors.

4. Fan-Beam & Cone-Beam Collimators

Designed for SPECT brain and cardiac imaging. Fan-beam collimators converge in the transverse plane (magnifying the brain across the head) while remaining parallel in the axial plane.