10.2 SPECT and SPECT/CT Systems
Key Takeaways
- SPECT acquires multiple projections around the patient (step-and-shoot or continuous rotation) and reconstructs tomographic slices
- Center-of-rotation (COR) QC is essential—COR error produces ring or circular artifacts and misregistration of activity
- SPECT demands stricter field uniformity than planar imaging because nonuniformity reconstructs into rings and false lesions
- Jaszczak (and similar) phantoms evaluate 3-D uniformity, resolution spheres/rods, and contrast; pixel size calibration ties matrix to real mm
- Attenuation correction may use Chang’s method (calculated map) or CT-based maps on hybrid SPECT/CT; patient motion and COR errors remain major artifact sources
10.2 SPECT and SPECT/CT Systems
Quick Answer: SPECT needs COR QC, excellent uniformity, correct pixel size, and solid orbits. Prefer CT-based attenuation correction on hybrid systems when available; watch for COR rings, motion, and misregistration.
Single-photon emission computed tomography (SPECT) rotates one or more gamma-camera heads about the patient, collecting projection images used to reconstruct transverse, coronal, and sagittal slices. Many modern systems are SPECT/CT hybrids that add a CT for attenuation correction, anatomic localization, and sometimes diagnostic CT protocols.
SPECT Acquisition Modes
| Mode | How it works | Teaching points |
|---|---|---|
| Step-and-shoot | Detector stops at each angle, acquires, then steps to next angle | Classic; less angular blur during count collection |
| Continuous (continuous rotation) | Detector moves while acquiring | Faster exams possible; motion during dwell can slightly blur projections |
Common orbits are circular or body-contour / elliptical noncircular paths that keep detectors close to the patient (better resolution). Dual-head 180° opposed or 90° cardiac configurations shorten acquisition time. Parameters include matrix (e.g., 64×64 or 128×128), angular sampling (e.g., 3°–6° steps), time per stop, and zoom.
Pixel size / calibration: reconstruction assumes a known mm-per-pixel. Periodic pixel calibration (or equivalent factory/service checks) keeps quantitative distances and filter cutoffs meaningful. Wrong zoom or calibration warps anatomy and ROI sizes.
Why SPECT Uniformity Requirements Are Stricter
In planar imaging, mild nonuniformity may look like mottling. In SPECT, the same nonuniformity is seen from many angles and reconstructs into rings or circular artifacts, which can mimic lesions or defects (classic example: ring artifact from a cold PMT). Therefore SPECT-capable cameras need tight daily uniformity QC and up-to-date correction maps before tomography.
Center of Rotation (COR) QC
Center of rotation is the alignment between the mechanical axis of rotation and the electronic/matrix center used in reconstruction.
| Concept | Detail |
|---|---|
| How tested | Point source (or multi-point) imaged through a full orbit; software computes COR offsets |
| Frequency | Per manufacturer—often weekly or after head collision/service (know your SOP) |
| Failure pattern | Ring artifacts, smeared point sources, misaligned activity on opposite projections |
| Action | Recalibrate COR; do not run clinical SPECT with failed COR |
COR errors are a high-yield CNMT topic: if you see concentric rings on reconstructed floods or patient slices without a physical ring source, think COR or uniformity—not the patient’s anatomy.
Three-Dimensional Performance: Jaszczak Phantom
The Jaszczak phantom (cylindrical phantom with cold/hot rods or spheres and a uniform section) evaluates SPECT tomographic uniformity, resolution, and contrast in three dimensions—beyond planar bar phantoms.
| Phantom region | What you assess |
|---|---|
| Uniform cylinder section | 3-D uniformity; look for rings |
| Rod / sector inserts | Tomographic spatial resolution |
| Spheres | Contrast and partial-volume behavior |
Periodic SPECT phantom testing (schedule per accreditation, manufacturer, and department) documents that reconstruction filters, COR, and detectors perform as a system.
Attenuation Correction
Photon attenuation in the body underestimates deep activity (e.g., inferior wall of the heart, deep abdominal structures).
| Method | Idea | Pros / cons |
|---|---|---|
| Chang’s method | Assumes uniform attenuation coefficient (soft tissue μ) and applies a calculated correction, often after first-pass reconstruction | No CT needed; fails where density varies (lungs, bone, arms down vs up) |
| CT-based AC (SPECT/CT) | CT provides patient-specific attenuation map scaled to emission energy | More accurate when co-registered; metal, truncation, and motion can still err |
Hybrid workflow sketch: position patient → scout/CT → SPECT orbit (or protocol order per system) → generate μ-map → reconstruct emission with AC (and scatter correction if used) → fuse display. Co-registration of CT and SPECT is mandatory; patient motion between CT and SPECT misplaces the μ-map and creates false defects or hot spots.
Common SPECT Artifacts
| Artifact | Clue | Mitigation |
|---|---|---|
| COR error rings | Concentric rings, especially on uniform phantom or liver | COR QC / recalibration |
| Nonuniformity rings | Rings tied to flood defects | Fix uniformity; service |
| Patient motion | Blur, misregistration, inconsistent projections (sinogram breaks) | Immobilize, shorter protocols, motion correction if available, repeat |
| Truncation (CT FOV) | Arms or body outside CT FOV on SPECT/CT | Reposition arms; larger FOV CT if available |
| Metal | CT streaks → wrong μ-map → emission bias | Note metal; review non-AC images |
SPECT/CT Practical Points
- Low-dose CT often suffices for attenuation correction and localization; diagnostic CT needs appropriate technique factors, contrast protocols, and often a radiologist’s CT scope of practice.
- Confirm correct patient, laterality, and study type before long dual-modality exams.
- After collisions or service, repeat COR, uniformity, and any manufacturer-required hybrid calibrations before clinical use.
Reconstruction Awareness for Technologists
You may not design filters, but you must recognize that filtered back-projection (FBP) and iterative methods (e.g., OSEM) behave differently with noise and attenuation. Over-smoothing hides small lesions; under-smoothing creates noisy “pseudo-lesions.” Always use protocol-approved reconstruction parameters for quantitative or comparison studies (e.g., serial bone SPECT). Review sinograms or rotating cine projections when available—discontinuities flag motion better than a single static slice.
Scatter: many SPECT protocols apply energy-window or model-based scatter correction together with AC. Turning corrections on/off inconsistently between baseline and follow-up invalidates visual comparison.
Multi-head tips: dual- or triple-head systems need each head’s COR and uniformity within limits; a single bad head can ruin the entire orbit. After a head collision, treat the system as out of service until COR and floods pass.
Memory aid: SPECT quality = Uniformity + COR + Orbit closeness + Registration (for hybrid) → “UCOR.”
Which quality control failure is classically associated with ring artifacts on reconstructed SPECT slices of a uniform phantom?
Compared with planar gamma-camera imaging, SPECT generally requires:
Chang attenuation correction differs from CT-based attenuation correction primarily because Chang’s method: