3.13 Quality Assurance and Equipment Operation

Quality Assurance and Equipment Operation in MRI

In magnetic resonance imaging (MRI), maintaining image quality, scanner performance, and patient safety requires a structured Quality Assurance (QA) program. While QA refers to the overall administrative and scheduling framework, Quality Control (QC) represents the actual physical testing of the equipment's performance. The American College of Radiology (ACR) has established standardized QC guidelines and phantom designs that form the benchmark for MRI accreditation and regulatory compliance.

---

The ACR MRI Accreditation Phantom

The primary tool used for MRI Quality Control is the ACR MRI Accreditation Phantom.

• Construction: It is a hollow, clear plastic cylinder (outer diameter 203 mm, length 165 mm; inner diameter 190 mm, length 148 mm).
• Phantom Filling: The phantom is filled with a solution of nickel chloride ($NiCl_2$) and sodium chloride ($NaCl$).
  • Nickel chloride is used to shorten the T1 and T2 relaxation times of the water protons, matching the relaxation characteristics of human tissues to allow realistic scan parameters.
  • Sodium chloride provides electrical conductivity, loading the RF receiver coil in a manner similar to a human patient.
• Internal Structures: The phantom contains various grids, wedges, ramps, and disks designed to test specific aspects of scanner performance across 11 defined axial slices.

---

ACR Phantom Quality Control Parameters

Technologists must perform and record seven core QC tests using the ACR phantom, comparing results against established action limits.

1. Geometric Accuracy

This test measures the spatial dimensions of the phantom in the three orthogonal planes (sagittal, coronal, and axial) to check for gradient calibration errors.

• Method: Measurements are taken on the sagittal localizer (phantom length, nominal 148 mm) and axial Slice 5 (vertical, horizontal, and diagonal diameters, nominal 190 mm).
• Action Limit: Measured dimensions must be within $\pm$ 2.0 mm of the known dimensions (e.g., 188.0 mm to 192.0 mm). Errors indicate gradient coil calibration drift.

2. High-Contrast Spatial Resolution

This parameter measures the scanner's ability to resolve small, closely spaced structures, which is determined by voxel size and system stability.

• Method: Tested on Slice 1, which contains three pairs of resolution inserts (upper-left, upper-right, and lower-right grids) consisting of small holes. The holes are arranged in rows of 1.1 mm, 1.0 mm, and 0.9 mm diameters.
• Action Limit: The technologist must resolve at least the 1.0 mm holes in both the horizontal and vertical directions.

3. Slice Thickness Accuracy

This test measures the accuracy of the selected slice thickness, which depends on the transmitter RF bandwidth and the amplitude of the slice-select gradient.

• Method: Tested on Slice 1, which contains two thin crossed ramps sloping at a 10:1 ratio. The length of the ramps is measured, and slice thickness is calculated using the formula: $\text{Slice Thickness} = 0.1 \times \frac{\text{Top} \times \text{Bottom}}{\text{Top} + \text{Bottom}}$
• Action Limit: For a nominal slice thickness of 5.0 mm, the measured thickness must be within $\pm$ 0.7 mm (4.3 mm to 5.7 mm).

4. Slice Position Accuracy

This test ensures that the scanner is planning and executing slices at the exact spatial coordinates requested.

• Method: Tested on Slice 1 and Slice 11, which contain crossed wedges (bars) at the top and bottom. The difference in length between the two bar segments is measured.
• Action Limit: The difference must not exceed $\pm$ 2.5 mm. A failure indicates table positioning errors or gradient offsets.

5. Image Intensity Uniformity

This parameter measures the uniformity of the signal intensity across a homogeneous region of the phantom.

• Method: Tested on Slice 7, which is a uniform open slice. The Percent Integral Uniformity (PIU) is calculated by placing a large region of interest (ROI) and identifying the maximum and minimum pixel values: $\text{PIU} = 100 \times \left(1 - \frac{\text{High} - \text{Low}}{\text{High} + \text{Low}}\right)$
• Action Limit: For systems operating at field strengths below 3T (e.g., 1.5T), the PIU must be $\ge$ 87.5% (typically $\ge$ 90% when using a dedicated head coil). For 3T systems, the action limit is typically $\ge$ 82% due to increased dielectric effects.

6. Percent Signal Ghosting

Ghosting is a phase-encoding artifact caused by scanner instability, patient motion, or physical table vibration.

• Method: Tested on Slice 7. Small ROIs are placed outside the phantom in the phase-encoding direction (top and bottom) and frequency-encoding direction (left and right). The ghosting ratio is calculated: $\text{Ghosting Ratio} = \frac{|(\text{Mean}_{\text{Top}} + \text{Mean}_{\text{Bottom}}) - (\text{Mean}_{\text{Left}} + \text{Mean}_{\text{Right}})|}{2 \times \text{Mean}_{\text{Center}}}$
• Action Limit: The ghosting ratio must be $\le$ 0.025 (or 2.5%).

7. Low-Contrast Object Detectability (LCD)

This test evaluates the scanner's sensitivity to structures with small contrast differences, which is heavily influenced by SNR and system noise.

• Method: Slices 8 through 11 contain circular disks of varying diameters (7.0 mm down to 1.5 mm) arranged in spokes. The technologist counts the number of visible spokes across all slices.
• Action Limit: On a 1.5T system, a minimum of 9 spokes must be visible (or a total score of 37 visible disks across the slices).

---

System Calibration and Tuning Parameters

In addition to phantom imaging, the system performing daily self-calibration generates key operating metrics that must be monitored.

Center Frequency

The center frequency (or transmitter frequency) is the exact frequency at which the RF transmitter operates to excite hydrogen protons. It must match the Larmor frequency ($f_0$) of water protons at the magnet's isocenter.

• Values: Approximately 63.86 MHz at 1.5T and 127.7 MHz at 3T.
• Monitoring: The center frequency is measured daily. The daily variation should not change by more than 1 part per million (ppm), which corresponds to roughly 64 Hz on a 1.5T system. Sudden shifts indicate magnet drift, cryogen boil-off, or room temperature fluctuations.

Transmitter Gain (Transmitter Attenuation)

Transmitter gain represents the RF power (voltage) required to produce a precise 90-degree or 180-degree flip angle in the patient's tissues.

• Units: Expressed in decibels (dB) or transmitter voltage.
• Monitoring: If the transmitter gain suddenly increases (typically by more than $\pm$ 3 dB or a 10% change in voltage), it means the system requires more power to achieve the same flip angle. This indicates a potential RF transmit coil failure, RF amplifier degradation, or incorrect phantom positioning/loading.

---

Daily Quality Control Logs and Environmental Monitoring

A comprehensive QA program requires daily logging of both scanner metrics and environmental conditions to ensure optimal scanner operation.

Environmental Specifications

The MRI scan room is a highly sensitive environment. Daily logs must track:

• Scan Room Temperature: Must be maintained between 18°C and 22°C (65°F to 72°F). High temperatures can cause RF power amplifiers to overheat and drift.
• Scan Room Humidity: Must be maintained between 30% and 60%. Low humidity increases the risk of electrostatic discharge (which causes image noise and artifacts), while high humidity can cause condensation on electrical circuits and degrade RF coil insulation.

Equipment and Safety Checks

Before scanning the first patient of the day, the technologist must verify:

1. Cryogen Level: Monitor the liquid helium level. A level below 50% must be reported to prevent a quench.
2. Emergency Systems: Verify the functionality of the patient intercom, the emergency alert squeeze bulb, and the emergency exhaust/oxygen sensor.
3. Visual Coil Inspections: Inspect all RF coils and cables daily for physical damage. Cracks in coil housings or frayed cable insulation can allow patient contact with bare copper, creating an electrical loop that leads to severe RF burns.