Image Quality and Acquisition Modes in Interventional Radiology

Characteristics of Interventional Radiology Systems

• IR units are significantly more complex and expensive than general fluoroscopy units, typically costing around $1,250,000$.
• Modern IR systems are exclusively digital. There are no analog or image intensified (II) systems used in current IR suites.
• The monitors used in IR are larger and offer higher detail, ensuring the monitor is not a limiting factor for perceiving the high-quality data these systems output.
• IR systems employ digital flat-panel detectors rather than the large drums associated with image intensifiers. This means there is no vacuum tube where electrons are minimized or accelerated.

Digital Detectors and Image Conversion

• Digital detectors in IR function differently than image intensifiers. In an II system, X-rays are converted to light, and the intensifier acts as a brightness amplifier. In digital systems, there is no separate amplification step using light in that manner.
• There are two primary types of digital detectors:
  • Photoconductor: Naturally converts X-rays directly into electrons, which are then stored.
  • Scintillator: Converts X-rays to light and then to electrons. Despite the light stage, there is no brightness amplification step or CCD/camera involved.
• Because there is no camera or CCD, the image is formed by reading out electrons stored directly in the detector elements, similar to the process in digital radiography or mammography.
• IR detectors often feature a large field of view (FOV), with input surfaces commonly reaching $40\,cm$. This allows for visualization of larger anatomical areas, though operators can use collimation or digital zoom if a smaller area is required.
• Dynamic Range: Digital systems have a large, linear dynamic range (a straight-line characteristic curve). Unlike image intensifiers, which require radiation to hit a "sweet spot" for optimal brightness and contrast, digital detectors can distinguish signal differences across very low, medium, and high radiation levels.

System Configurations and Magnification

• Single Plane vs. Biplane:
  • Most IR rooms are single-plane units.
  • Biplane units consist of two C-arms, usually positioned at $90^{\circ}$ to one another. These are primarily used in neuro-interventional procedures.
  • The advantage of biplane units is the ability to acquire both AP and lateral views with a single contrast injection, reducing the total contrast load and procedure time.
• Magnification Modes:
  • Digital systems do not have a traditional "Mag Mode" as found in image intensifier systems.
  • Instead, they use Digital Mag or Selectable Field of View.
  • In digital systems, spatial resolution changes differently as a function of FOV compared to II systems.

Spatial Resolution and Pixel Binning

• In image intensified systems, the spatial resolution improves every time the magnification mode is increased.
• In flat-panel digital systems, spatial resolution stays constant until a certain FOV threshold is reached. For example, moving from a $40\,cm$ FOV to a $30\,cm$ FOV may improve resolution, but further zooming might not change the physical spatial resolution of the image sensor further.
• Pixel Binning:
  • To manage large FOVs, the system assumes the operator does not need fine-detail resolution and employs "binning."
  • This process combines groups of pixels (e.g., taking 4 pixels and treating them as 1 large pixel).
  • Pros: Binning reduces image noise (quantum mottle) because the signal from multiple pixels is averaged. It also reduces the amount of data that must be transferred, which is useful at high pulse rates.
  • Cons: Binning results in a loss of spatial resolution.
• When the operator zooms in past a certain point, the system "unbins" the pixels, using every individual detector element to provide the highest possible spatial resolution.
• Perceptual Resolution: Even if the image's inherent spatial resolution doesn't change when zooming further, digital magnification makes small objects larger on the monitor, making them easier for the human eye and brain to interpret.

Dose Management in IR

• Copper Filtration:
  • IR systems utilize significant copper filtration, ranging from $0\,mm$ up to $0.9\,mm$.
  • Copper is highly effective at attenuating low-energy photons that contribute to skin dose without contributing to the image.
  • Adding just $0.2\,mm$ of copper can reduce skin dose by up to $40\%$. Some settings can result in a $70\%$ savings in skin dose.
• Dose in Magnification:
  • In II systems, dose increases in Mag Mode to compensate for the loss of minification gain and maintain brightness.
  • In flat-panel systems, there is no brightness to maintain. However, vendors still program the systems to increase dose when zooming to satisfy radiologist preferences for reduced noise in magnified views. This is a noise-saving measure rather than a brightness requirement.

Comparison of Digital and Image Intensified Systems

• Artifacts Absent in Digital Systems:
  • Flare: No optics or light bouncing within the system.
  • Vignetting: There is no curved input surface characteristic of an II; the flat panel is uniform.
  • Distortions (S-Distortion): Digital detectors are not affected by magnetic fields in the way electrons moving through a vacuum in an II are. There is also no mapping from a curved to a flat surface.
• Shared and Unique Artifacts:
  • Saturation: Can occur in both, but digital systems use Lookup Tables (LUTs) to automatically map high-end radiation values to different gray levels, correcting the appearance.
  • Electronic Noise: Flat-panel systems are noisier than II systems at extremely low doses because the signal becomes comparable in size to the internal electronic noise of the detector component.
  • Standard Digital Artifacts: IR detectors are subject to dead pixels (dead detector elements) and require flat-field calibrations.
• Temporal Resolution:
  • Image quality differs between stationary and moving objects. Motion can cause image lag or ghosting artifacts, where outlines of previous frames are superimposed on the current one.

Digital Subtraction Angiography (DSA)

• Mechanism: DSA involves taking an initial image (the mask) and then subtracting it from "live" fluoroscopy images during contrast injection using logarithmic subtraction.
• Objective: To remove background structures like bone and tissue, leaving only the contrast-filled vessels visible.
• Dose Characteristics:
  • DSA is a very high-dose acquisition mode. One minute of DSA at $4\,fps$ (at $70\,kV$) is equivalent to approximately $30\text{ minutes}$ of standard fluoroscopy.
  • Typical DSA dose is $10$ to $30\text{ times}$ higher than standard fluoroscopy per unit of time.
  • Skin damage can occur in as little as $5\text{ minutes}$ of continuous DSA.
  • Noise increases during subtraction (noise adds in quadrature), necessitating higher dose to maintain diagnostic quality.
• Motion in DSA: If the patient or table moves, the mask and live images do not align, causing misregistration artifacts (black and white edges). This requires remasking (taking a new mask) or using digital registration image processing to shift the images.

Cone Beam CT (CBCT)

• Also referred to as "spins," this involves the C-arm rotating around the patient ($4$ to $20\text{ second}$ acquisition) to produce a 3D reconstruction.
• Image Quality:
  • Spatial Resolution: Very high, often better than conventional CT. It features isotropic voxels (typically $0.3$ to $0.5\,mm$).
  • Contrast Resolution: Very poor. It does not produce quantitative Hounsfield units due to scatter and the heel effect.
  • It is excellent for bone and contrast-enhanced structures but unsuitable for subtle soft-tissue visualization.
• Artifacts: Prone to aliasing or streak artifacts because it takes fewer angles than a standard CT scanner.
• Occupational Dose: Higher than a single fluoro frame but still low (typically $< 0.1\,mGy$ per spin).

Automatic Exposure Rate Control (AERC)

• Modern digital systems use AERC (or ADRC - Automatic Dose Rate Control) instead of ABC (Automatic Brightness Control) because there is no light to measure.
• Goal: To maintain a constant dose at the detector to ensure constant image noise.
• Modulated Factors: AERC adjusts $kV$, $mA$, pulse width, and filtration based on patient thickness.
• Trends as Patient Thickness Increases:
  • Total dose rate increases.
  • $mA$ and $kV$ generally increase.
  • Copper Filtration Decreases: While counterintuitive, filtration is removed for very thick patients to allow more mid-to-high energy photons to penetrate the patient, even though this significantly increases skin dose from unfiltered low-energy photons.

Skin Dose and Safety Regulations

• Peak Skin Dose: Measured at the entrance surface of the patient. This correlates most directly with skin injury risk.
• Erythema Threshold: Occurs at approximately $2$ to $3\,Gy$.
• Necrosis Threshold: Occurs at approximately $15\,Gy$.
• Sentinel Event (Joint Commission): Updated in 2022. It is no longer defined strictly by a $15\,Gy$ threshold but rather by any fluoroscopy resulting in permanent tissue injury when optimization or practice parameters were not followed.
• Follow-up Guidelines (SIR recommendations):
  • Follow-up required if Cumulative Air Kerma ($CAK$) exceeds $5\,Gy$.
  • Follow-up required if Peak Skin Dose exceeds $3\,Gy$.
  • Follow-up required if fluoroscopy time exceeds $60\text{ minutes}$.
• Notification Protocols: Operators should be notified when cumulative dose reaches $3\,Gy$, and every $1\,Gy$ increment thereafter.

Dose Metrics on the Monitor

• Interventional Reference Point (IRP): Also called the Patient Exposure Reference Point ($PERP$). Located at the isocenter of the C-arm moved $15\,cm$ toward the X-ray tube.
• Cumulative Air Kerma (CAK): The estimated dose at the IRP. It is a surrogate for skin dose but typically overestimates it (actual peak skin dose is often about $60\%$ of the reported $CAK$). It does not account for backscatter or the spreading of the beam.
• Kerma Area Product (KAP): Also called Dose Area Product ($DAP$). This measures the total energy delivered to the patient and is the best surrogate for cancer risk and effective dose.

Typical Procedure Doses

• Typical Fluoro: Measured in tens of $mGy/min$.
• DSA: Approximately $10\,mGy/sec$.
• Nephrostomy: Typically low skin dose ($< 2\,Gy$) and effective dose ($\approx 6\,mSv$).
• Transarterial Chemoembolization (TACE): Can exceed $5\,Gy$ skin dose.
• Anatomical Variations in Effective Dose:
  • Body procedures: Tens of $mSv$.
  • Head procedures: Units of $mSv$.
  • Extremity procedures: Less than $1\,mSv$.

Questions & Discussion

• Question 1: What system component found in an image intensified system is not present in a flat panel IR system?
  • Answer: Camera (or CCD). Flat panels use stored electrons read out directly.
• Question 2: Pixel binning degrades what image property?
  • Answer: Spatial resolution. Binned pixels are larger and lose fine detail.
• Question 3: What frame rate is typically used in DSA?
  • Answer: $4\text{ frames per second}$. Some systems allow up to $8\,fps$, but $4$ is standard.
• Question 4: What artifact is found only in flat panel systems and not in image intensified systems?
  • Answer: Dead pixels (dead detector elements). Vignetting and S-distortion are II-specific.
• Question 5: Compared to conventional CT, what image property is better in cone beam CT?
  • Answer: Spatial resolution (due to smaller detector elements and focal spots).
• Question 6: What does automatic exposure rate control in a flat panel system attempt to hold constant?
  • Answer: Noise (by keeping detector dose constant).
• Question 7: What dose metric is the best surrogate for peak skin dose?
  • Answer: Cumulative Air Kerma ($CAK$).
• Question 8: What is the minimum focus to skin distance (SSD) for an IR suite?
  • Answer: $38\,cm$. (Recall: Portable units are $30\,cm$). Spacer cones are used to maintain this distance and prevent skin burns.