Physical Address
304 North Cardinal St.
Dorchester Center, MA 02124
IEC TR 61948-2: Nuclear Medicine — Radiation Detection System Testing
Reference: IEC TR 61948-2:2001 — Nuclear medicine instrumentation — Routine tests — Part 2: Scintillation cameras and single photon emission computed tomography (SPECT) systems.
1. Scope and Clinical Significance
IEC TR 61948-2 provides a comprehensive technical framework for routine quality control testing of radiation detection systems used in nuclear medicine, with primary focus on scintillation (Anger) cameras and SPECT systems. Published as a Technical Report rather than a full International Standard, IEC TR 61948-2 offers guidance on test methods that are practical for clinical implementation, recognizing that nuclear medicine departments require efficient yet thorough testing protocols to ensure diagnostic image quality while maintaining patient throughput.
The standard addresses the fundamental challenge in nuclear medicine imaging: detecting and localizing gamma radiation emitted from radiopharmaceuticals administered to patients. The Anger scintillation camera, introduced by Hal Anger in 1958, remains the core technology, using a large-area NaI(Tl) scintillation crystal coupled to an array of photomultiplier tubes (PMTs). Position computation relies on the centroid method, where the relative signals from adjacent PMTs determine the interaction coordinates. IEC TR 61948-2 specifies tests for each critical subsystem: the detector head (crystal, light guide, PMT array), the position computation electronics, the energy discrimination circuitry, and the collimator.
Critical Context: IEC TR 61948-2 complements the NEMA NU 1 standard (Performance Measurements of Gamma Cameras). While NEMA NU 1 focuses on acceptance testing at installation time, IEC TR 61948-2 emphasizes routine constancy tests performed daily, weekly, and monthly to detect performance degradation between full NEMA evaluations.
2. Key Test Parameters and Methodology
2.1 Intrinsic Uniformity and Spatial Distortion
Flood-field uniformity is the most fundamental daily quality control test. The standard specifies that a uniform flux of gamma radiation (typically using Tc-99m, 140 keV) shall be applied to the detector without a collimator, using a point source positioned at a distance of at least 5 times the detector field of view (typically 2-3 meters). The resulting flood image should appear uniformly bright across the entire field of view. IEC TR 61948-2 defines two uniformity metrics: integral uniformity (the maximum percent deviation across the entire useful field of view) and differential uniformity (the maximum percent deviation over a 5-pixel distance, which detects localized PMT failures).
Spatial distortion or linearity is assessed using a phantom containing parallel slit openings arranged in a grid pattern. The standard requires measurement of the maximum deviation of the imaged slit positions from their known physical positions. Distortion exceeding 1-2 mm at the detector surface indicates PMT gain drift or crystal damage. Modern digital cameras apply real-time correction maps (linearity correction matrices) that compensate for inherent distortion, and the standard specifies that these corrections must be disabled or their effect documented during acceptance testing.
| Test Parameter | Frequency | Acceptance Criterion | Test Source | Measurement Time |
|---|---|---|---|---|
| Intrinsic Flood Uniformity | Daily | Integral UFOV ≤ 5%, CFOV ≤ 4% | Tc-99m, 10-15 mCi | 2-5 minutes |
| Spatial Resolution (intrinsic) | Quarterly | FWHM ≤ 4.0 mm (UFOV) | Tc-99m, slit phantom | 15-30 minutes |
| Energy Resolution | Monthly | FWHM ≤ 10% at 140 keV | Tc-99m point source | 5 minutes |
| Count Rate Performance | Annually | 20% loss at ≥ 60 kcps | Tc-99m, decaying source | 30 minutes |
| Multiple Window Registration | Quarterly | Offset ≤ 1.0 mm | Dual-isotope phantom | 10 minutes |
| Collimator Hole Angulation | Annually | Deviation ≤ 0.5° from axis | Point source, 1.5 m distance | 20 minutes |
2.2 SPECT-Specific Performance Tests
For SPECT systems, IEC TR 61948-2 introduces additional tests that address the rotational nature of tomographic acquisition. Center-of-rotation (COR) alignment is critical: any offset between the mechanical rotation axis and the electronic center of the detector causes ring artifacts and resolution degradation in reconstructed images. The standard specifies a COR test using a point source positioned at the center of rotation, acquired over 360 degrees at 5-10 degree intervals. The measured COR offset must be within 0.5 pixels (typically 1.5-2.0 mm) for clinical use.
SPECT uniformity requirements are more stringent than planar uniformity because the reconstruction process amplifies non-uniformities, producing ring artifacts that can mimic or obscure clinical findings. The standard requires that the tomographic uniformity, measured from a reconstructed uniform cylinder phantom, shall have a coefficient of variation (COV) of less than 5% in the reconstructed slices. This typically requires intrinsic flood uniformity better than 3% and accurate center-of-rotation correction within 0.3 pixels.
Engineering Insight: The most frequent cause of SPECT ring artifacts is PMT gain drift over time. The standard recommends daily PMT gain stabilization using the detector’s built-in calibration source (typically a low-activity Co-57 or Ba-133 source). A gain drift of just 1-2% can produce visible ring artifacts. Modern systems use automatic digital stabilization that adjusts PMT high voltage or gain settings during the daily uniformity flood, maintaining stability within 0.5% over months of operation.
2.3 Energy Resolution and Scatter Rejection
Energy resolution directly impacts image quality by determining the effectiveness of Compton scatter rejection. IEC TR 61948-2 specifies measurement of the full width at half maximum (FWHM) of the Tc-99m photopeak at 140 keV, using a calibrated multi-channel analyzer. For modern gamma cameras with NaI(Tl) detectors, energy resolution ranges from 9.0% to 10.5% FWHM. The standard emphasizes that energy resolution should be measured at a count rate below 10 kcps to avoid pulse pile-up effects that artificially broaden the photopeak. The energy window setting (typically 20% centered on the photopeak, i.e., 126-154 keV for Tc-99m) involves a trade-off: a narrower window improves scatter rejection but reduces sensitivity, while a wider window increases counts at the expense of contrast degradation.
3. Engineering Design Insights and Practical Applications
The test methodologies defined in IEC TR 61948-2 reflect deep engineering trade-offs in detector design. The Anger position computation algorithm, which calculates event coordinates using weighted summation of PMT signals, imposes fundamental limitations on spatial resolution at high count rates. At count rates exceeding 50 kcps, pulse pile-up causes mispositioning of events (event “bursts” are assigned incorrect coordinates by the centroid computation), degrading both spatial resolution and uniformity. This effect, known as “count rate paralysis,” is particularly relevant for modern applications such as dynamic cardiac imaging where instantaneous count rates can exceed 100 kcps during the first-pass bolus phase.
Collimator selection represents another critical engineering decision. The standard specifies tests for parallel-hole, converging, diverging, and pinhole collimators, each with different sensitivity-resolution trade-offs. A low-energy high-resolution (LEHR) collimator typically provides 7.5 mm FWHM resolution at 10 cm distance with a sensitivity of approximately 180 cps/MBq, while a low-energy general-purpose (LEGP) collimator achieves 10 mm resolution with approximately 250 cps/MBq sensitivity. The choice directly affects the detectability of small lesions; the Rose criterion (signal-to-noise ratio > 3-5 for reliable detection) can be used with measured system performance parameters to predict lesion detectability.
Calibration Alert: The energy window calibration must be verified daily. A shift of even 1-2% in the photopeak position (caused by PMT gain drift) results in asymmetric energy window placement, reducing sensitivity and potentially introducing image artifacts. Most clinical protocols require that the photopeak position be within 1% of its calibrated value before patient imaging begins.
4. Frequently Asked Questions
Q1: What is the difference between intrinsic and extrinsic uniformity measurement?
Intrinsic uniformity is measured without a collimator, using a point source at a distance (typically > 5 detector diameters). Extrinsic uniformity is measured with the collimator attached and a uniform flood phantom placed directly on the collimator surface. Intrinsic measurements characterize the detector electronics and crystal performance, while extrinsic measurements additionally characterize collimator defects (e.g., damaged or clogged septa). IEC TR 61948-2 requires intrinsic daily and extrinsic monthly measurements.
Q2: Why is Tc-99m the preferred test source for routine QC?
Tc-99m decays by isomeric transition emitting 140 keV gamma radiation, which is near-ideal for NaI(Tl) scintillation detection (maximum interaction probability ~85% for a 3/8-inch crystal). Its 6.01-hour half-life is short enough to minimize radiation exposure but long enough to perform a complete day’s QC. The 140 keV energy is also representative of most clinical SPECT radiopharmaceuticals. For daily constancy checks without the need for a new source each day, some protocols use longer-lived sources such as Co-57 (122 keV, 271-day half-life) or Ba-133 (356 keV, 10.5-year half-life).
Q3: How do solid-state digital detectors affect the test protocols?
Modern digital gamma cameras using solid-state detectors (e.g., cadmium-zinc-telluride, CZT) eliminate the PMT array and centroid computation entirely. For these systems, uniformity testing remains required but spatial linearity testing is simplified because CZT detectors have intrinsic pixel-to-pixel correspondence. The energy resolution of CZT detectors (5-6% FWHM at 140 keV) is substantially better than NaI(Tl), enabling narrower energy windows and improved scatter rejection. IEC TR 61948-2 test protocols are generally applicable but acceptance criteria for spatial resolution and energy resolution differ significantly for solid-state systems.
Q4: What is the acceptable count rate for daily flood uniformity testing?
The standard recommends acquiring at least 10 million counts for the daily uniformity flood image to ensure adequate statistical precision for detecting small PMT gain variations. The count rate should be maintained below 30 kcps to avoid rate-dependent uniformity degradation. With a typical Tc-99m source of 370 MBq (10 mCi) at 2.5 meters, the expected count rate is approximately 15-25 kcps, requiring 6-10 minutes of acquisition time for a 10-million-count image.
