Quality Control and Artifacts in SPECT

Chapter 17: Quality Control and Artifacts in SPECT

Acceptance Testing

  • Must include all planar gamma camera tests and SPECT-specific acceptance tests.

  • Testing is conducted according to NEMA (National Electrical Manufacturers Association) protocols.

  • In the case of multi-head cameras:

    • Calibration factors must not differ by more than 0.5%.

    • Acceptance test results should not exceed a difference of 10% between camera heads.

SPECT Quality Control

  • SPECT imaging necessitates stringent planar quality control measures.

    • Corrected camera uniformity should be 3% or better.

    • Flood source radioactivity should show less than 1% variation in activity concentration.

    • Regular testing mandates:

    • Pixel size calibration.

    • Collimator hole angulation tests.

    • A reference scan for transmission-based attenuation correction needs to be performed daily.

Center of Rotation

  • The location of a line source in each pair of 180º-opposed 2D projections establishes the axis of rotation, which correlates to the Center of Rotation (COR).

  • Misalignment between the COR and the line source is quantified by the COR analysis program, which leads to correction via an offset applied to clinical studies.

  • If the X-offset exceeds 0.5 pixels, it is recommended to clear stored values and determine new correction values.

Implications of Incorrect COR

  • An incorrect COR frequently results in subtle artifacts in clinical studies that degrade spatial resolution.

  • Misalignment may arise from:

    • Mechanical changes in camera head alignment.

    • Gravity effects on the collimator.

    • A tilted camera head relative to the axis of rotation (AOR).

  • Regular checks on COR should be performed weekly until stability is established.

COR and Misalignment Errors

  • Misregistration due to filtered backprojection can lead to SPECT artifacts.

COR Protocol

  • The center of rotation was analyzed with the GE HEALTHCARE NM 630 S.

  • Parameters and values include:

    1. Collimator: Low energy high resolution.

    2. Matrix size: 256 x 256.

    3. Zoom: 1.00.

    4. Gantry mode: Hmode.

    5. Total angular range: 720 degrees.

    6. Angular step: 6 degrees.

    7. Scan mode: Step and shoot.

    8. Stop on counts: 1000 kcts.

    9. Energy set: Tc99m 140.5 keV ±10%.

    10. Direction: Clockwise (CW).

Quantitative Measures of SPECT Performance

  • Spatial resolution is assessed using tomographic reconstruction of a line source of radioactivity to measure:

    • Full Width at Half Maximum (FWHM) and Full Width at Tenth Maximum (FWTM) in radial, tangential, and axial directions.

  • Uniformity involves scanning a cylindrical phantom with uniform activity concentration, measuring uniformity over a 15 x 15 cm area on transverse slices.

  • Sensitivity is determined by comparing counts per second (cps) to activity in a cylindrical phantom.

Spatial Resolution Measurement

  • Measurement involves two aspects:

    • (a) In-plane resolution:

    • A cross section of the cylindrical phantom with three line sources.

    • Count profile is measured in the radial direction.

    • (b) Axial resolution is assessed along the Z-axis via the count profile in the tangential direction, across slice numbers.

Jaszczak Phantom

  • Description:

    • A 20-cm diameter Plexiglas cylinder with both rod and sphere inserts.

    • Fill phantom with radioactive solution, ensuring proper mixing.

    • Rods function as a 3D "bar phantom."

    • Spheres vary in size, being either cold (solid Lucite) or hot (fillable spheres).

  • Uses:

    • Rods measure resolution.

    • Spheres measure contrast resolution.

    • The section devoid of rods or spheres assesses uniformity.

Jaszczak Phantom - Features

  • The Jaszczak phantom and its successors enable measurements of uniformity, semi-quantitative spatial resolution (by rods), and real-life spatial resolution (by spheres).

    • Image is courtesy of Biodex Medical Systems.

SPECT Artifacts

  • Ray Artifacts:

    • Result from filtered backprojection reconstruction and are particularly noticeable around hot spots.

  • Ring Artifacts:

    • Caused by the rotational, multi-view nature of SPECT imaging.

    • They may reflect abnormalities such as flood non-uniformities, COR errors, and collimator defects.

  • Example:

    • Image revealing a subtle ring artifact, highlighting alternating dark and light rings due to a flood non-uniformity, is provided as an illustration (courtesy of Virginia Mason Medical Center).

Uniformity

  • Minor non-uniformities can be reconstructed across the 3D matrix, causing hot or cold spots that may obstruct accurate diagnosis.

Motion Artifacts

  • These artifacts lead to image blurring resulting from activity being positioned in two different locations in the 3D image matrix.

Reconstruction Error in SPECT

  • If the long axis of the left ventricle (LV) is misidentified on the transaxial or midventricular vertical long-axis slice:

    • The heart's geometry in reconstructed tomographic slices can become distorted.

    • This distortion can alter the apparent regional count density depicted in polar maps, causing artificial perfusion defects.

    • Such errors are often accentuated during quantitative analyses when patient data are compared to normal values, especially affecting basal myocardial regions at the bull's-eye plot periphery.

    • Additionally, the apex, which may exhibit physiological thinning and decreased count density, can be displaced from the polar plot's center, further provoking artifacts.

Truncation Artifacts

  • These artifacts arise during transmission-based attenuation correction when the patient’s entirety is not fully captured in all projections.

  • The imaging system is incapable of correcting for data it cannot visualize.

Summary of SPECT Quality Control

  • SPECT demands substantially higher vigilance than planar imaging.

  • Artifacts can be challenging to detect, even during phantom analyses, and may remain unnoticed on clinical images.

  • The presence of artifacts can be obscured by filter selection.

  • Therefore, meticulous attention to quality control and assurance details is imperative in SPECT operations.

Recommended Quality Control Procedures for SPECT Tomographs

Test

Recommended Frequency

1. Uniformity

Daily

2. Collimator visual inspection

Daily

3. Center of rotation

Weekly-biweekly

4. Spatial resolution (bar phantom linearity)

Weekly-biweekly

5. High-count flood/uniformity correction map

Weekly-biweekly

6. Tomographic spatial resolution

Quarterly-annually

7. Jaszczak phantom

Quarterly-annually

8. Pixel size

Quarterly-annually

9. Tilt-angle check

Quarterly-semiannually

10. Collimator-hole angulation

Quarterly

Additional Acceptance Tests for Tomographic Gamma Camera Systems

  • Evaluation of rotational capabilities and limits.

  • Verification of angle indicators.

  • Checking step-size accuracy.

  • Timing assessments.

  • Radial motion and gantry controls analysis.

  • Examination of safety features (specifically emergency stops and pressure-sensitive collimator covers).

  • Pixel calibration procedures.

  • Assessment of the center of rotation performance.

  • Checking for rotational uniformity stability.

  • Collimator-hole angulation.

  • Patient-contour positioning assessments.

  • Uniformity across tomographic slices.

  • Sensitivity metrics related to system performance using the Jaszczak phantom.

  • The determination of lesion contrast.

SPECT Math

  • Total acquisition time is calculated as: acquisition ext{ }time = number ext{ }of ext{ }projections imes time ext{ }per ext{ }projection

    • Alternatively expressed as:
      acquisition ext{ }time = rac{total ext{ }counts}{counting ext{ }rate}

Example Calculations:

  1. A SPECT study acquires data at 20 seconds per stop for 60 projections.

    • Total acquisition time:
      acquisition ext{ }time = 60 ext{ }projections imes 20 ext{ }seconds = 1200 ext{ }seconds = 20 ext{ }minutes

  2. A SPECT study yields a total of 1 million counts with a counting rate of approximately 48,000 counts per minute (cpm).

    • Total acquisition time:
      acquisition ext{ }time = rac{1,000,000 ext{ }counts}{48,000 ext{ }cpm}
      ightarrow 20.8 ext{ }minutes

  3. Time per projection is calculated as:
    time ext{ }per ext{ }projection = rac{total ext{ }acquisition ext{ }time}{number ext{ }of ext{ }projections}

  4. Given a SPECT study acquired for 25 minutes with 62 stops:

    • time ext{ }per ext{ }projection = rac{25 ext{ }minutes}{62 ext{ }stops}
      ightarrow 24 ext{ }sec/stop

  5. Counts per projection is evaluated as:
    counts ext{ }per ext{ }projection = rac{total ext{ }counts}{number ext{ }of ext{ }projections}

  6. A SPECT study with 3,000,000 counts and 64 projections gives:

    • counts ext{ }per ext{ }projection = rac{3,000,000 ext{ }counts}{64 ext{ }projections}
      ightarrow 46,875 ext{ }counts/projection

SPECT IMAGE REVIEW

  • Importance of evaluating SPECT images for quality assessment and diagnosis.