Medical Imaging Science: Computerized Tomography (CT) Scanners
Medical Imaging Science - Specialised Imaging I (HMI 3113): Computerized Tomography (CT) Scanners
Principle of Radiography & Tomography in Medical Imaging
In both radiography and tomography, X-rays pass through the patient and are absorbed differently by body tissues:
Bone (denser) absorbs more X-rays than less dense soft tissue.
Soft tissue absorbs more X-rays than less dense air.
Differential Absorption: The variance in X-ray absorption by body tissues is contained within the X-ray beam that passes through the patient and is recorded on the image receptor (IR).
Limitations of Radiography
Superimposition of All Structures on the IR:
Difficult, sometimes impossible, to distinguish particular details, especially when structures differ slightly in density.
This problem persists even with multiple views, making it hard to discriminate the exact location of structures like tumors.
Difficulty Distinguishing Objects:
It is challenging to distinguish between a homogeneous object of non-uniform thickness and a heterogeneous object.
Conventional Tomography
Method to Overcome Superimposition: Conventional tomography was developed to overcome the problem of superimposition inherent in traditional radiography.
Geometric Tomography: This is the most common method of conventional tomography, differentiated from Computerized Tomography.
Mechanism:
The X-ray tube and the image receptor (IR) are moved simultaneously in opposite directions.
The goal is to blur out unwanted sections (structures above and below the plane of interest) while keeping the desired layer or section in sharp focus.
Limitations of Conventional Tomography
Persistent Image Blurring: Blurring cannot be completely removed like in radiography.
Degradation of Image Contrast: Caused by the presence of scattered radiation, due to the open geometry of the X-ray beam.
Inability to Demonstrate Slight Contrast Differences: Fails to adequately demonstrate subtle differences in subject contrast (e.g., soft tissue), hindering the imaging of very small tissue contrast variations.
Introduction to Computed Tomography (CT)
Medical Imaging Method: CT is a medical imaging technique that utilizes tomography.
Etymology: The word "tomography" is derived from the Greek words "tomos" (slice) and "graphein" (to write).
Principles:
A large series of two-dimensional X-ray images (referred to as "slices" or "cuts") of the inside of an object are taken around a single axis of rotation.
Digital geometry processing is then employed to generate three-dimensional images of the object from these collected slices.
Attenuation in CT
Definition: Attenuation refers to the reduction of the X-ray beam intensity as it passes through material, due to a combination of absorption and scatter.
Measurement: The degree of attenuation is quantified by measuring and comparing the incident X-ray intensity (before passing through the patient) and the transmitted X-ray intensity (after passing through the patient).
Density Relationship:
More dense material leads to less transmitted X-rays (higher attenuation).
Less dense material leads to more transmitted X-rays (lower attenuation).
CT Scan & Slices
Image Output: A CT scan produces axial sections, cuts, or slices of the body.
Definition of a Scan: A scan consists of multiple X-ray attenuation measurements taken around the periphery of an object.
Definition of a Slice/Cut: The cross-sectional portion of the body that is scanned to produce a CT image.
Each slice has a specific width and, therefore, a volume.
The width of the slice is determined by the width of the X-ray beam.
History of Computerized Tomography
Inventor of the First Commercial Scanner: Sir Godfrey Hounsfield, working at EMI Central Research Laboratories in Hayes, United Kingdom, invented the first commercially viable CT scanner using X-rays.
Hounsfield conceived his idea in and it was publicly announced in .
Independent Inventor: Allan McLeod Cormack of Tufts University in Massachusetts independently developed a similar process.
Nobel Prize: Both Hounsfield and Cormack shared the Nobel Prize in Medicine for their contributions to CT.
CT Scanner Components
CT scanners consist of three main components:
Gantry Assembly: The rotating framework of the scanner.
The Computer: Processes and reconstructs image data.
The Operator's Console: Used by the technologist to control the scanner and manipulate images.
Gantry Assembly Details
Moveable Frame: The gantry is a movable frame that houses critical components:
The X-ray tube (including collimators and filters).
Detectors.
Data Acquisition System (DAS).
Rotational components (slip ring systems, gantry angulation motors, positioning laser lights).
All associated electronics.
High Voltage Generator:
Some generators are mounted inside the gantry wall, while newer scanner designs may locate the generator outside the gantry.
Must be a three-phase generator.
Must be capable of producing an instant energy power surge to facilitate rapid X-ray production.
Slip Ring Technology:
Eliminated the need for cables connecting components to a stationary power source.
Allows for continuous rotation of the gantry components, enabling helical/spiral scanning and faster acquisition times.
CT X-ray Tube
Exposure Factors & Heat Capacity:
CT procedures utilize large exposure factors (high mAs and kVp values) and short exposure times.
They produce X-radiation either continuously or in short millisecond pulses.
The constant buildup of heat due to these high-power exposures can rapidly decrease tube life. Tubes will suspend operation until they cool down.
Therefore, CT X-ray tubes must be designed with a very high heat capacity.
Modern CT systems use X-ray tubes with a heat capacity of approximately .
A high heat dissipation rate is also critical.
Cooling Systems: Many CT X-ray tubes employ a combination of oil and air cooling systems to efficiently eliminate heat and maintain continuous operational capabilities.
Anode Design:
CT X-ray tube anodes have a large diameter and a graphite backing.
This design allows the anode to absorb and dissipate large amounts of heat effectively.
Focal Spot Size:
Most X-ray tubes feature more than one focal spot size.
A small focal spot increases image detail but concentrates heat onto a smaller portion of the anode, generating more heat.
Filament Size: CT tubes utilize a bigger filament than conventional radiography X-ray tubes, which increases the size of the effective focal spot.
Anode Angle:
Decreasing the anode (or target) angle effectively decreases the size of the effective focal spot.
Conventional radiography tubes typically have an anode angle between and .
CT tubes employ significantly smaller target angles, approximately between and .
This decreased anode angle also helps to mitigate some effects of the heel effect.
Resolution Compensation: CT systems can compensate for any potential loss of resolution resulting from the use of larger focal spot sizes by:
Employing resolution enhancement algorithms (e.g., bone or sharp algorithms).
Decreasing the section thickness.
Exposures in CT
High kVp: Needed for penetration of thick, dense anatomy (e.g., C7/T1 junction).
Lower kVp: May be used for converting adult to pediatric scans or for scout scans.
High mAs: Needed for dense or thick body regions.
Using too low mAs will result in image degradation.
Using unnecessarily high mAs will increase heat units in the X-ray tube and decrease the tube's lifespan.
Filtration in CT
Purpose: To enhance the resolution of a particular anatomical region of interest and shape the beam intensity.
Mechanism: Inherent tube filtration and additional filters (made of aluminum or Teflon) are utilized to filter out low-energy photons that contribute to the production of scatter radiation and do not effectively penetrate the patient.
Types:
Flat filtration.
Beam shaping filter (e.g., "bow-tie" filter).
Beam Uniformity: Heavy filtration of the X-ray beam results in a more uniform beam.
A more uniform beam leads to more accurate attenuation values or CT numbers for the scanned anatomical region.
Beam Shaping (Bow-Tie) Filter:
Ensures a more constant signal reaches all detectors.
Makes the beam hardness at the detectors more constant.
Absorbs low-energy X-rays, and its shape ensures a more intense beam at the patient's center (where anatomy is thicker), compensating for higher attenuation and resulting in a constant signal at the detectors.
Collimators
Collimators are crucial for reducing patient dose and improving image quality by minimizing scatter radiation.
Pre-patient Collimator:
Its design is influenced by the focal spot size.
Mounted on the tube housing itself.
Creates a more parallel X-ray beam.
Directly reduces patient dose by limiting the X-ray field.
Pre-detector Collimator:
Restricts the field of view for the detectors.
Reduces scatter radiation from reaching the detectors.
Its aperture width helps determine the slice thickness.
Detectors in CT Systems
Function: An electronic device that, when struck by an X-ray photon, produces light or electrical energy.
Evolution of Detector Types:
Early Detectors: Scintillators (e.g., NaCl).
Gas Detectors.
Modern Detectors: Ceramic Scintillators or solid-state detectors.
Detector Requirements
For optimal performance, CT detectors must meet several critical requirements:
High Detection/Quantum Efficiency: Must be extremely high over the entire radiation spectrum to maximize signal capture.
Fast Response: Rapid signal conversion and decay to allow for quick scan times.
Large Dynamic Range: A wide ratio of the largest to smallest detectable signal is necessary to capture a broad range of tissue densities.
Linearity: The detector signal must be proportional to the X-ray intensity.
Insensitivity to Variations: Must be insensitive to temperature and voltage fluctuations for stable performance.
Low Cost: For economic viability.
Small in Size: To enable dense packing in detector arrays.
Reliable: For consistent and long-term operation.
Minimal Radiation Drift and Afterglow: To ensure accuracy and prevent residual signal from previous exposures.
Excellent Scintillation Light Yield (for scintillators): If applicable, the light output should be strong and efficient.
Data Acquisition System (DAS)
Signal Amplification: Once the detector generates a weak analog or electrical signal, it is directed to the Data Acquisition System (DAS). The DAS amplifies this weak signal for further analysis.
In some modern CT scanning systems, signal amplification occurs within the detector itself.
Location: The DAS is typically located in the gantry, immediately after or above the detector system.
Analog-to-Digital Conversion (ADC): Before the projection or raw data (which is in analog form) goes to the computer, it must be converted into digital information. An analog-to-digital converter, an essential component of the DAS, performs this task.
Transfer to Array Processor: The digital signal is then transferred to an array processor.
Array Processor (Specialized High-Speed Computer):
Designed to execute mathematical algorithms for the purpose of image reconstruction.
Solves statistical information using algorithmic calculations essential for mathematical reconstruction of a CT image.
Image/Reconstruction Generator: Recently, processors capable of computing CT reconstruction mathematics faster than traditional array processors have been utilized, especially for CT fluoroscopy. These are sometimes referred to as image or reconstruction generators.
Gantry Assembly – Ancillary Systems
Lasers/High Intensity Lights:
Included within or mounted on the gantry.
Serve as anatomical positioning guides, referencing the center of the axial, coronal, and sagittal planes for patient alignment.
Table and Gantry Controls:
Controls for precise table movement.
Laser localizer for accurate patient positioning.
Gantry display showing table position and gantry tilt.
Communication system between the operator and patient.
CT Couch/Table
Support & Material:
Must be strong enough to safely support patients of various sizes and weights.
Made from materials with a low atomic number to ensure minimal X-ray beam attenuation, preventing artifacts.
Attachments: Accommodates various attachments for different scanning procedures (e.g., coronal scanning).
Motor Drive: Features an accurate motor drive for precise indexing of the patient into the gantry.
Safety Devices: Includes safety features such as automatic shut-off mechanisms.
The Computer
Unique Subsystem: A highly specialized subsystem of the CT scanner.
Processing Power: Capable of performing over equations simultaneously, instantaneously, and continuously.
Cost: Accounts for over one-third of the total cost of the scanner.
Environmental Sensitivity: Sensitive to heat and humidity, requiring controlled operating environments.
Data Flow:
Receives signals in analog form and converts them to binary digits using an analog-to-digital converter.
Stores the digital signal (raw data) during the scan.
Reconstructs the image after the scan is complete.
Images may be reconstructed immediately or stored in memory for later reconstruction.
Evolution of CT Scanner Generations
First Generation Scanner (1973)
Beam & Detector: Utilized a pencil beam of X-rays with one or two detectors.
Scan Mode: Operated on a translate-rotate mode.
The X-ray tube and detector system would translate linearly across the field of view (FOV), acquiring data (e.g., parallel rays across a ).
Then, the entire system would rotate by .
This process would repeat until the tube and detector rotated a total of (in intervals for projections).
Scan Time: Required about minutes to complete one rotation.
Total Rays Measured: Achieved ray measurements.
Second Generation Scanner
Key Difference: The primary difference from the first generation was the use of a fan beam of radiation and multiple detectors.
Scan Mode: Employed a multiple detectors, translation-rotation approach with a large fan-beam.
Efficiency: This allowed a series of views to be acquired during each translation, leading to significantly shorter scan times.
Scan Time: The shortest scan time achieved with a second-generation scanner was seconds per slice, which was times faster than the first-generation system.
Third Generation Scanner (Mid-1970s)
Scan Mode: Evolved to a rotate/rotate geometry. Both the X-ray tube and the detector array rotated simultaneously in a full arc (e.g., clockwise then counterclockwise).
Detectors & Beam Coverage:
Substantially increased number of detectors (up to and later over detectors) arranged in a curvilinear array.
The angle of the fan beam was increased so that the detector array formed an arc wide enough to interrogate the entire patient cross-section without translation.
Scan Time:
Elimination of translational motion drastically reduced scan times.
Early third-generation scanners could deliver scan times shorter than seconds.
Newer systems can achieve scan times as low as seconds per slice.
Fourth Generation Scanner
Detectors: In this generation, the detectors are removed from the rotating gantry and are placed in a stationary ring around the patient.
Tube Motion: The X-ray tube still rotates around the patient.
Detector Count: Requires many more detectors (e.g., ; modern systems use around individual detectors). The number of views is equal to the number of detectors.
Sensitivity to Scatter: The stationary detector design requires a larger acceptance angle for radiation, making it more sensitive to scattered radiation compared to third-generation geometry.
Slip Ring Technology (Early 1990s):
Incorporated into both third- and fourth-generation scanners.
Mechanism: A slip ring is a circular electrical contact with sliding brushes that allows the X-ray tube to rotate continually without the need for cables that would twist and break.
Impact:
Eliminated the inertial limitations (the need to reverse direction) at the end of each slice acquisition.
Allowed the rotating gantry to rotate continuously throughout the entire patient examination.
Enabled greater rotational velocities and, consequently, even shorter scan times.
Fifth Generation - Helical or Spiral Scanning
Continuous Rotation and Table Movement: Recent innovations in slip-ring technology allowed the X-ray tube to rotate continuously in the same direction (overcoming inter-scan delays) while the patient table simultaneously advances continuously through the gantry.
Helical Path: The combination of continuous tube rotation and continuous patient translation results in the X-ray tube traveling in a helical or spiral path around the patient.
Typical table velocities range from .
A complete rotation may be achieved in seconds.
Advantages of Helical CT:
Reduction of Patient Motion: Permits faster scanning of larger volumes, reducing motion artifacts, especially in areas like the heart or other moving organs.
Improved Contrast Media Usage: A greater volume of the patient can be scanned during the passage of contrast media, which can permit a reduction in the total volume of contrast needed.
Enhanced 3D Scans: The continuity of data along the axis of the patient (absence of gaps between scans) significantly improves the quality of three-dimensional reconstructions.
Faster Scans: Generally faster than axial slices.
Single Breath-Hold Studies: Facilitates studies that require patients to hold their breath, minimizing motion artifacts.
Lowers Tube Heating: More efficient scanning reduces overall tube heat load.
Increased Patient Throughput: Allows more patients to be scanned per day.
Thin Slices: Makes many thin slices possible, improving z-axis resolution.
Multi-Detector CT (MDCT) / Multi-Slice CT
Evolution from Helical: Represents the next generation of CT scanners, building upon helical principles.
Detector Configuration: These machines incorporate multiple rows of detector rings (e.g., , or more rows).
Increased Coverage: By using multiple detector rows, MDCT scanners can acquire multiple slices per tube rotation simultaneously. This significantly increases the area of the patient that can be covered by the X-ray beam in a given time.
Cone Beam: Uses a cone X-ray beam rather than a fan beam to cover the multiple detector rows.
Detector Switching: Allows for selection of slice width by combining the output of adjacent detector elements (e.g., , slices).
Benefits of Multi-Slice:
Faster: The same scan can be completed in a shorter time.
Thinner: Thinner slices provide excellent z-axis (longitudinal) resolution.
Further: Wider collimations enable larger scan volumes to be acquired in less time.
Pitch in Helical/Spiral Scanning
Definition: The pitch factor controls the helical motion. It is defined as the distance in millimeters that the table moves during one complete rotation of the X-ray tube, divided by the slice thickness (in millimeters).
Pitch : If the table moves exactly the same distance as the slice thickness during one tube rotation, the pitch equals . This results in contiguous data acquisition with no gaps or overlap.
Pitch > 1 (High Pitch): The table moves further than the slice thickness per revolution.
The resulting image is more blurred, and the helix is stretched.
The displayed image may appear to have a slightly thicker slice.
Pitch < 1 (Low Pitch): The table moves less than the slice thickness per revolution (e.g., overlapping scans).
The image is sharper due to increased data sampling.
Multi-Slice Pitch Formula Adjustment:
Where "Total Active Detector Width" is the X-ray beam collimation across all detector rows.
CT Development: Dual Energy CT Scan (DECT) / "Spectral Imaging"
Limitation of Projection Radiography: Projection radiography overlays a three-dimensional patient volume onto a two-dimensional image plane, leading to superimposition (e.g., bony structures hiding soft tissue lesions).
Dual-Energy X-ray Technique: Combines two radiographs acquired at two distinct X-ray energies (i.e., two different kVp settings).
Benefit: Allows the acquisition of information regarding both tissue density and atomic number. This provides insights into material composition or significantly improves image contrast.
Conventional (Single Energy) CT (SECT): Uses a single polychromatic X-ray beam (ranging from ) emitted from a single source and received by a single detector. Image contrast primarily depends on the differential attenuation of various materials.
Principle of DECT: Based on the differential absorption of X-ray energy (linear attenuation coefficient) at variable kVp settings.
Typically, two energy levels are used (e.g., and ) to acquire images.
These images are then processed to generate additional data sets, providing information beyond what conventional CT can offer (e.g., material decomposition, virtual non-contrast images).
This can be achieved by rapidly switching kVp on a single X-ray tube or by using two X-ray sources operating at different kVp levels.
The Basics of CT Image Formation
Pixel (Picture Element): A two-dimensional square represented by a shade of gray in the image.
Voxel (Volume Element): A three-dimensional volume of tissue corresponding to a shade of gray.
It is the result of a computer averaging the attenuation coefficients across a small volume of material.
Provides depth information in the image.
Each voxel is typically about on a side and thick, depending on the depth and width of the scanning X-ray beam (slice thickness).
Attenuation Measurement: The intensity of the X-ray beam along a given path is mathematically represented as the sum of attenuation coefficients ($ ext{μ}$) through the tissues.
Linear Attenuation Coefficient ($ ext{μ}$): Describes the fraction of an X-ray or gamma ray beam that is absorbed or scattered per unit thickness of the absorber.
Image Reconstruction Workflow:
The measured attenuation along each X-ray path generates a specific shade of gray and an associated numerical value (CT number).
The X-ray tube and detector assembly then change angles, and the process of data acquisition repeats.
Images are reconstructed using algorithms like back projection, which essentially trace backwards along the X-rays' forward paths to reconstruct the image and calculate the absorption due to localized regions.
This is a mathematically intensive process but is efficiently handled by computers.
Detectors measure the transmitted X-rays, and the computer uses the known incident X-ray intensity to calculate the attenuation.
Back Projection Reconstruction Method
Summation Method / Linear Superposition:
Starts with an empty image matrix.
The measured value (attenuation) from each X-ray ray in all radial views is added to each pixel along a line through the image matrix that corresponds to the ray's path.
The basic premise is that any attenuation of the X-ray beam is assumed to have occurred uniformly along the entire ray path.
How it Works: The attenuation coefficient for a particular point (pixel) in the image is built up from the contributions of all X-ray projections that pass through that point. Each of these points (pixels) becomes an element of the final image.
Reversing Acquisition: The back projection method effectively builds up the CT image in the computer by reversing the acquisition steps:
During acquisition, attenuation information along a known path of the narrow X-ray beam is integrated by a detector.
During reconstruction, the $ ext{μ}$ value inferred for each ray is "smeared" or projected back along this same path in the image of the patient.
Visualizing Reconstruction: From every position of the X-ray tube as it rotates around the object:
A dark line is projected back for regions with low X-ray attenuation.
A white line is projected back for regions with high X-ray attenuation.
The darkness (or brightness) of the line is scaled according to the measured X-ray attenuation.
Filtered Back Projection
Role of "Filter": The term "filter" here does not refer to a metallic filter (like aluminum or copper) placed in the X-ray beam. Instead, it refers to a convolution filter (also known as a kernel).
Mathematical Manipulation: A convolution filter is a mathematical manipulation of the raw data or projection data designed to change the appearance of the image.
Application: It is a mathematical process applied to an image projection before the back projection step.
Impact: Filtered back projection removes the inherent blurring seen in simple back projection, resulting in a more mathematically exact and sharp reconstruction of the image.
Reconstruction Algorithms: Convolution filters are also commonly referred to as reconstruction algorithms. Filtered back projection is the most commonly used algorithm for CT systems.
Basic Mathematical Concepts: Fourier Transform
Definition: The Fourier Transform is a fundamental mathematical technique used for converting data from the time domain to the frequency domain, and vice versa.
Image Processing Tool: It is an important image processing tool used to decompose an image into its constituent sine and cosine components.
Domain Conversion:
The input image is in the spatial domain (representing pixel values based on location).
The output of the transformation represents the image in the Fourier or frequency domain.
Frequency Domain Interpretation: In the Fourier domain image, each point represents a particular frequency (rate of change in intensity) contained within the original spatial domain image.
Hounsfield Units (HU) or CT Numbers
Definition: The CT number $CT(x,y)$ in each pixel $(x,y)$ of the image is calculated using the formula: Where:
$ ext{μ}(x,y)$ is the linear attenuation coefficient of the voxel at position $(x,y)$.
$ ext{μ}_{ ext{water}}$ is the linear attenuation coefficient of water.
Scaling: The value of is a scaling factor set by the CT manufacturer.
Interpretation: CT numbers (or Hounsfield units) represent the percent difference between the X-ray attenuation coefficient for a given voxel and that of water, scaled by a factor of .
Reference Points:
Water is assigned a CT number of .
Numbers can be positive or negative, depending on the absorption coefficient relative to water.
Typical Range of CT Numbers:
Air: Approximately .
Soft tissues: Range from approximately to .
Dense bone and areas filled with contrast agents: Can range up to (or higher depending on manufacturer scale, e.g., to ).
Shade of Gray Assignment: These numerical values are used to assign a specific shade of gray to each pixel, forming the final CT image.
The Operator's Console
The operator's console consists of two main parts:
The Scan Console:
Technologist-Controlled: This part provides the technologist with direct control over the scanning parameters.
Controls Include:
Technical factors (e.g., kVp, mAs).
Selection of slice thickness.
Number of scans to be acquired.
Angle of the gantry.
Table movement.
Initiating the scan.
Recording patient data.
Setting the Scan Field of View (SFOV) – determined by the body part and patient size.
Selecting the number of detectors to be used.
Setting the Display Field of View (DFOV) – which influences magnification.
The Display Console:
Post-Scan Manipulation: Used to manipulate data and images after the scan has been acquired.
Tools Include:
Grids: For viewing images more easily.
Reverse Display: Allows changing left to right, top to bottom, or inverting black to white.
Magnify Images.
Annotate Images: Adding text or markers.
Histograms: Bar graphs used to contrast and compare Hounsfield units within an image.
Cursor: Used to precisely locate specific points, measure distances, and calculate a Region of Interest (ROI).
Reformatting/Reconstructions:
Requires Raw Data: This process requires the original raw data for reconstruction.
Used to change the Display Field of View (DFOV).
Can reconstruct images to better visualize specific tissues, such as bone or soft tissue, using different algorithms or window settings.
Field of View (FOV)
Concept: Similar to collimation in a different plane, it determines the size of the image displayed on the screen.
Radiographer's Role: Radiographers should select the appropriate FOV to suit the patient's anatomy and diagnostic needs.
Types:
Scan Field of View (SFOV):
This is the area from which raw data is acquired by the scanner.
It is usually set to encompass the entire anatomical part being scanned (e.g., head, body, large body).
Different diameters of SFOV can be selected depending on the anatomy being imaged.
Display Field of View (DFOV):
This is the diameter of the reconstructed image that is displayed.
It is typically employed as a post-processing tool, allowing the technologist to display only a specific part of the larger SFOV.
The DFOV can be equal to or less than the SFOV.
Changing DFOV: The DFOV may be changed to a smaller size after scanning to magnify and better visualize observed pathology or anatomy of interest without a loss of image quality. This is a "re-reconstruction" using a subset of the raw data.
Relationship to Pixel Size:
The formula linking DFOV, matrix size, and pixel size (for square pixels and matrix) is:
Decreasing the DFOV will decrease the pixel size when the matrix size is kept constant.
As the pixel size increases (meaning more patient information is