Medical Image Analysis: X-ray Imaging

This lecture provides an introductory overview of medical imaging systems, focusing on X-ray
imaging. It covers the fundamental principles of X-ray production, interaction with tissue, image
acquisition, and introduces both X-ray projection radiography and X-ray Computed Tomography
(CT).

1. Introduction to Medical Imaging Systems

• Course Scope: This course provides an understanding of what images produced by
medical imaging systems actually represent.
• Imaging Systems Covered:
o X-ray imaging:
§ X-ray Projection Radiography (e.g., standard X-rays, chest X-rays).
§ X-ray Computer Tomography (CT).
o Magnetic Resonance Imaging (MRI).
o Ultrasound Imaging Systems.
o Radionuclide Imaging.
• What will NOT be covered in depth: Deep dives into image reconstruction
techniques. While essential for CT, MRI, and radionuclide imaging, the focus will be on
image acquisition and representation. We will note the general formulation required for
image acquisition.

2. X-ray Imaging: Projection Radiography

X-ray radiography is one of the oldest and most common diagnostic imaging modalities. Most
people are familiar with it, having likely undergone X-rays for fractures or respiratory
complaints.

2.1. Fundamentals of X-rays

• Discovery: X-rays were discovered in 1895 by Wilhelm Conrad Röntgen through

experiments with cathode ray tubes (Crookes tubes).
• Nature: X-rays are a form of electromagnetic radiation.
o They possess higher energy and occupy the high-frequency range of the
electromagnetic (EM) spectrum.
o Energy Range:
§ Visible light: 1 to 10 electron Volts (eV).
§ Medical X-rays: Typically tens of kilovolts (kV), translating to 10^3 to
10^5 eV. For example, CT machines use 180 to 150 kV.
• Safety Concerns: Due to their high energy, X-rays can cause damage at the cellular
level, necessitating strict safety procedures to prevent overexposure.
2.2. X-ray Production in a Modern X-ray Tube

(Insert diagram of X-ray tube here)

A modern X-ray tube is a vacuum tube with two main components:

1. Cathode:
o A filament of wire.
o Emits electrons via thermionic emission (heating the filament by passing current
through it).
2. Anode Target (Plate):
o Typically coated with tungsten or other high-atomic-number (high-Z)
materials.
o Electrons emitted from the cathode are accelerated by a potential
difference across the tube to hit a spot on the anode target.
o This acceleration and impact cause electrons to penetrate the anode, emitting
energy in the form of X-rays.
o Spot Size Impact: The size of the spot where electrons hit the anode directly
affects the resolution of the final image. A larger spot size is needed for higher
flux, which can impact resolution (e.g., 0.2 to 0.5 millimeters or larger for clinical
X-rays).

• Mechanisms of X-ray Emission: Primarily two types:

1. Characteristic X-rays
2. Bremsstrahlung (Breaking Radiation)

2.2.1. X-ray Spectrum

(Insert diagram of X-ray spectrum here)

• The X-ray output from a typical tube is an entire range of energies (polychromatic),
not monochromatic.
• The spectrum is plotted as the number of X-ray photons emitted (y-
axis) against energy (x-axis), essentially a histogram of photon counts per energy range.
• This spectrum helps visualize the contributions of the two primary emission types.
• Maximum Energy: The accelerating potential difference (kVp) across the tube
determines the maximum energy of the emitted electrons and, consequently, the
maximum energy in the X-ray spectrum.

2.2.2. Characteristic X-ray Emission

• Mechanism: Occurs when an incident electron collides with and knocks out a K-shell
electron (innermost shell) from an atom of the anode target material, ionizing the atom.
o The vacancy in the K-shell is then filled by an electron from a higher energy
shell (L, M, or N shell).
o This transition from a higher to a lower energy shell results in a loss of
energy, which is emitted as an X-ray photon.

2
 • "Characteristic" Name: The energy difference between these shells (e.g., L to K, M to
K) is unique for each element. Therefore, the emitted X-ray photons have specific,
discrete energies, producing sharp peaks in the X-ray spectrum that are characteristic
of the anode material (e.g., molybdenum, tungsten).

2.2.3. Bremsstrahlung (Breaking Radiation)

• Mechanism: This is the most dominant contribution to the X-ray spectrum. It involves
the interaction of incident electrons with the nucleus of the anode atom.
o As an electron approaches the positively charged nucleus, it is decelerated and
its trajectory changes due to the attractive electrostatic force.
o According to physics, when a charged particle is decelerated or accelerated,
it loses energy in the form of radiation.
• Spectrum Shape: This process produces the continuous, broad portions of the X-ray
spectrum.

2.2.4. Energy Transfer

• Both characteristic and Bremsstrahlung radiation fall under radiative transfer (emission
of X-rays).
• Another process that occurs is collision transfer, where electrons collide and are
absorbed by the material, largely emitting energy as heat. This is why X-ray tubes need
cooling.

2.3. X-ray Interactions with Tissue and Image Formation

When X-rays pass through the human body, their interaction with tissue is key to image
formation.

• What is Imaged: X-ray radiographs primarily image the summation of attenuation

coefficients along the path of the X-ray beam.
• Linear Attenuation Coefficient (μ):
o Definition: Mu (μ) is a linear attenuation coefficient with units of inverse length
(e.g., cm⁻¹).
o Function: It's a complex function of both the X-ray photon energy (E) and the
atomic number (Z) of the material. This makes it a material property.
o Role: Describes how X-ray photons are absorbed or scattered within the human
body, providing a mathematical model for these interactions.
• Basic Attenuation Relationship (Beer-Lambert Law):
o For a homogeneous slab of material of thickness x (or d), with incident
intensity I₀, the exiting intensity Iis given by: I = I₀ * e^(-µx)
o For a non-homogeneous object (like the human body) where µ changes location
to location, the relationship becomes an integral over the path s: I = I₀ * e^(-
∫µ(s)ds)
o For a real (polychromatic) X-ray source, which has a spectrum S(E'), the
relationship involves integrating over all energies: I = ∫ S(E') * e^(-∫µ(s,
E')ds) dE'

3
 • Radiograph as a "Shadow": In projection radiography, the image is a linear
superposition of all the attenuation coefficients along the path. This means different
anatomical structures are superimposed, much like a shadow, which can lead to limited
differentiation.
o Contrast: Good contrast depends on the energy range. If energy is too high,
photons may pass through without interaction, providing no information.

2.4. Origin of the Linear Attenuation Coefficient (μ)

The attenuation coefficient originates from two primary interactions of X-rays with tissue,
especially relevant for medical imaging:

1. Compton Scattering:
o Impact: Has a negative impact on image quality, leading to loss of contrast.
o Mechanism: Incident X-ray photons interact with outer-shell electrons, causing
the X-ray photon to be deflected from its initial path. These deflected photons
may hit the detector at an incorrect location, blurring the image.
o Dependence: Primarily dependent on electron density (atomic number) and is
relatively constant across energies at higher energies.
2. Photoelectric Absorption:
o Impact: Desirable impact for imaging, contributing to high-quality
images and good contrast.
o Mechanism: The incident X-ray photon is completely absorbed by an inner-
shell electron (e.g., K-edge), leading to the ejection of that electron (ionization).
o Dependence: Strongly depends on the atomic number (Z) of the material
and diminishes with increasing photon energy. At very high energies,
photoelectric absorption does not occur.

• Other Interactions:
o Rayleigh scattering and Pair production also occur but are generally not
significant for medical X-ray imaging.

2.5. X-ray Detectors for Projection Radiography

Detectors capture the attenuated X-ray beam to form an image.

1. Screen-Film Combination / Storage Phosphors (CR):

o Mechanism: A light-sensitive film is sandwiched between two scintillators
(phosphors).
o Scintillator Role: Scintillators convert X-ray energy into visible light. This light
then exposes the film.
o Types:
§ Directly exposed film: Can be digitized immediately.
§ Photostimulable Phosphors (PSPs): The image is stored latently. It
requires re-excitation with a laser to emit light, which is then recorded
and digitized. This allows for easy digitization.
2. Flat Panel Detectors (DR):

4
 o Modern Approach: These are fully digital systems where the detector directly
converts X-rays into a digital array of images.
o Components: Can include Charge-Coupled Device (CCD) cameras, Thin-Film
Transistors (TFTs), or CMOS sensors.
o CMOS Example: Similar to smartphone cameras, but coated with a scintillator
material. X-rays hit the scintillator, emit light, and the CMOS array records this
light signal.

2.6. Representative X-ray Images

(Insert representative chest X-ray images here)

• Appearance: Typical X-ray images (e.g., chest X-rays) show superimposed anatomical
structures.
• Clinical Applications: Used for diagnosing various conditions, including respiratory
problems and lung-related diseases like COVID-19.

3. X-ray Computed Tomography (CT)

CT is another very common diagnostic imaging system, offering significant advantages over
projection radiography.

3.1. Key Difference from Projection Radiography

• Image Output: CT produces cross-sectional (slice) images of the body, unlike the
superimposed 2D images of projection radiography.
o Analogy: Imagine "slicing the patient" into planes perpendicular to the body's
length (z-axis) and viewing these slices.
• Image Acquisition: While the acquired signals are still projections, CT uses
a mathematical inversion process(like filtered back projection) to convert these
projections into cross-sectional or volumetric images.

3.2. CT System Components and Acquisition Process

(Insert diagram of X-ray CT machine cross-section here)

• X-ray Tubes and Detectors: Principles of X-ray generation and detection are the same
as in projection radiography, but the specific tubes and detectors are often more robust
and higher-energy rated due to continuous operation.
• Scanning Configuration:
o Instead of a single projection, CT involves taking projections from many
different angles around the patient.
o Collimation: The X-ray beam from the tube is typically collimated into a thin
fan beam.

5
 o The X-ray tube and detector array rotate around the patient (e.g., taking a
projection every half-degree).
o Historical Context: Prior efforts like axial tomography and linear tomography
also aimed to acquire cross-sections.

(Insert diagram of parallel beam and fan beam projections here)

• Projection Measurement:
o Fan Beam Projection: The X-ray source is treated as a point, emitting a fan-
shaped beam that covers the entire object. A detector array measures the
attenuated beam.
o Data Collection: The intensity of the measured X-ray beam is plotted as a
function of position along the detector array, forming a projection.
• Sinogram:
o A sinogram is a 2D image created by stacking all the projections acquired at
different angles (theta). It represents the projection data as a function
of theta (scan angle) and t (distance along the detector array).

3.3. X-ray CT Detectors

• Scintillator Crystals with Photodiodes:

o Mechanism: Scintillator crystals convert X-rays into visible light, which is then
received by a photodiode array.
o Photodiode Role: Photodiodes convert visible light into an electric current (the
signal).
o Efficiency: Scintillators are designed for high efficiency to capture most X-ray
photons and convert them to light.
o Speed: Acquisition needs to be very fast, requiring high-end analog-to-digital
converters.
• Photon Counting Detectors (Emerging Technology):
o Advancement: Unlike traditional energy-integrating detectors (which sum up
light from all X-ray energies), photon-counting detectors can differentiate and
count photons based on their energy.
o Materials: Use materials like cadmium telluride and cadmium zinc
telluride for energy thresholding and counting.
o Benefit: Since the attenuation coefficient is energy-dependent, obtaining photon
counts at different energies allows for reconstructing images with different
kinds of contrast, enabling better differentiation between various tissue types
(e.g., distinguishing cancers more clearly). This is particularly useful as much of
the human body is water, making density differences subtle except for bone.

3.4. CT Image Reconstruction

• Core Process: The acquired projection data (sinogram) is processed using mathematical
techniques to generate cross-sectional images.
• Common Algorithm:
o Filtered Back Projection: This is a typical algorithm.

6
 § Steps:
1. Take the raw projection data (the log of the ratio of incident to
measured intensity: log(Nin / Nd) or log(Iin / Iout). This
raw data actually represents the integral of µds along the ray path.
2. Filter this data using a digital filter (e.g., Fast Fourier Transform
filter) to remove noise and prepare for summation. This is a crucial
step derived from reconstruction theory.
3. Back Project (smear): For each point (or pixel) in the filtered
projection, its value is "smeared" (assigned) across the
corresponding pixels in the reconstructed image grid.
4. Summation: This smearing and summation is repeated for every
projection angle, ultimately building up the map of the
attenuation coefficient (μ) for that cross-section.
• Other Techniques: Iterative reconstruction techniques and deep learning-based
techniques are also used.
• Hounsfield Units (HU):
o Normalization: The reconstructed µ values are normalized to a standard scale
called Hounsfield Units (HU).
o Formula: HU = (µ_reconstructed - µ_water_average_energy) /
µ_water_average_energy * 1000. (Note: The source also mentions µ_r in the
denominator which is typically 0, implying µ_water as the reference point for
normalization.)
o Scale:
§ Typically ranges from -1000 to 3095 HU.
§ Often offset by 1000 to range from 0 to 4095 HU, which corresponds to
a 12-bit number.
§ Meaningful Values: Unlike MRI or ultrasound, CT Hounsfield Units
have absolute meaning.
§ Water is defined as 0 HU (or around 0-100 due to noise).
§ Bone has very high HU values (e.g., >2000), indicating high
density/atomic number.
§ This allows for reliable characterization of tissue density.

3.5. Representative CT Images and Clinical Applications

(Insert representative CT images (brain, heart) here)

• Appearance: CT images display cross-sectional maps of X-ray attenuation

coefficients.
• Contrast Agents: For certain scans, contrast agents (e.g., iodine, a high-Z material) are
injected. These agents absorb X-rays preferentially, causing tissues that take up the agent
(like blood vessels or tumors) to appear very bright (higher pixel values, higher HU).
• Clinical Applications: Used for detailed imaging of various body parts like
the brain and heart. They provide much better differentiation of anatomical
structures compared to projection X-rays.

7
4. Key Takeaways
• X-ray imaging utilizes electromagnetic radiation in the high-energy spectrum.
• X-rays are produced when high-energy electrons strike a high-Z anode target, leading
to Bremsstrahlung (continuous spectrum) and Characteristic X-ray (discrete
peaks) emissions.
• Image formation relies on the Linear Attenuation Coefficient (μ), which describes
how X-rays are absorbed or scattered by tissue, primarily due to photoelectric
absorption (good contrast) and Compton scattering (image degradation).
• Projection Radiography creates 2D superimposed "shadows" of anatomical structures
by summing attenuation along the beam path.
• X-ray Computed Tomography (CT) generates cross-sectional images (slices) by
acquiring multiple projections from different angles and using reconstruction
algorithms (e.g., filtered back projection).
• CT image intensity is quantified by Hounsfield Units (HU), which provide an absolute
measure of tissue density/attenuation.
• Advancements in detectors, like photon-counting detectors, promise improved contrast
and tissue differentiation by discriminating X-ray energies.

5. Glossary of New Terms

• Thermionic Emission: The process by which electrons are emitted from a heated metal
surface (e.g., cathode filament).
• Anode Target: The positively charged component in an X-ray tube that electrons strike
to produce X-rays.
• High-Z Material: Materials with a high atomic number (Z), such as tungsten, used in X-
ray tube anodes.
• Characteristic X-rays: X-rays emitted at specific, discrete energies unique to the target
material, forming sharp peaks in the X-ray spectrum.
• Bremsstrahlung: German for "breaking radiation"; X-rays produced when electrons are
decelerated by the electric field of an atomic nucleus, forming a continuous spectrum.
• Linear Attenuation Coefficient (μ): A measure of how much X-ray radiation is
attenuated (absorbed or scattered) per unit thickness of a material.
• Compton Scattering: An interaction where an X-ray photon loses some energy and is
deflected after colliding with an outer-shell electron.
• Photoelectric Absorption: An interaction where an X-ray photon is completely absorbed
by an inner-shell electron, leading to the electron's ejection.
• Scintillator: A material that converts X-ray energy into visible light.
• Photostimulable Phosphors (PSPs): Detector materials that store a latent image after X-
ray exposure, which can then be stimulated by a laser to release light for digitization.
• Flat Panel Detectors: Fully digital X-ray detectors that directly convert X-rays into a
digital image array.
• Cross-Sectional Image: An image representing a slice through an object, typically
obtained by CT.
• Projection: A 2D measurement of X-ray attenuation along specific paths through an
object, taken from a particular angle.

8
 • Sinogram: A 2D image formed by stacking all the X-ray projections acquired at different
angles, used in CT reconstruction.
• Filtered Back Projection: A common mathematical algorithm used in CT to reconstruct
cross-sectional images from multiple projections.
• Hounsfield Units (HU): A quantitative scale used in CT to represent tissue density based
on X-ray attenuation, with water defined as 0 HU.
• Energy Integrating Detectors: Traditional X-ray detectors that sum up the energy of all
incident photons without differentiating their individual energies.
• Photon Counting Detectors: Advanced X-ray detectors capable of counting individual
X-ray photons and differentiating them by energy.

6. Thought-Provoking Questions / Further Research

• Image Reconstruction: How do "iterative reconstruction techniques" and "deep
learning-based techniques" for CT reconstruction differ from filtered back projection, and
what are their advantages/disadvantages in terms of image quality and computation time?
• X-ray Safety: What specific safety procedures are built into X-ray imaging systems to
protect patients and medical staff from overexposure?
• Polychromatic X-rays: Given that real X-ray sources are polychromatic, how does this
affect image contrast and reconstruction accuracy compared to an idealized
monochromatic beam?
• Attenuation Coefficient: Research the actual mathematical relationships and graphs for
how Compton scattering and photoelectric absorption depend on X-ray energy and
atomic number (Z).
• Detector Advancements: Explore in more detail the technological challenges and
clinical benefits of implementing photon-counting detectors on a wide scale in medical
imaging.

7. Potential Exam/Interview Questions

• Describe the components and working principle of a modern X-ray tube. How does the
spot size on the anode affect image resolution?
• Explain the two primary mechanisms of X-ray emission: characteristic X-rays and
Bremsstrahlung. How are they represented in an X-ray spectrum?
• Define the linear attenuation coefficient (μ). Discuss its dependence on X-ray energy and
atomic number, and explain how photoelectric absorption and Compton scattering
contribute to it.
• Compare and contrast X-ray projection radiography and X-ray Computed Tomography
(CT) in terms of:
o The type of image produced.
o The image acquisition process.
o How anatomical structures are represented.
• Outline the general steps involved in CT image reconstruction using filtered back
projection. What are Hounsfield Units, and why are they important in CT?
• Discuss the role of X-ray detectors in both projection radiography and CT. What are the
advantages of newer technologies like flat panel detectors and photon-counting detectors?

9
MRI Physics: Comprehensive Notes
1. Introduction to Magnetic Resonance Imaging (MRI)
This class delves into the physics of MRI to understand how images are produced and the
mechanisms of contrast. Future discussions will cover image acquisition techniques and
hardware.

Key Takeaways:

• MRI leverages fundamental physics principles to generate detailed images of internal

body structures.
• Understanding the origin of magnetization and its interaction with magnetic fields is
crucial for comprehending MRI.

2. Microscopic Origin of Magnetization

At the most fundamental level, the MRI signal originates from properties of atomic nuclei within
the sample (e.g., human body tissue).

2.1. Spin Angular Momentum

• Every nuclei possesses both charge and an intrinsic property called spin angular
momentum.
• Spin angular momentum (denoted by φ) has no counterpart in classical physics and
originates from quantum mechanics.
• Nuclei with an odd atomic number or mass number possess this spin angular momentum.
• This spin property gives rise to magnetic properties and allows the nuclei to interact
with an external magnetic field.

2.2. Magnetic Moment and Gyromagnetic Ratio

• Each microscopic magnetic field has an associated magnetic moment vector (µ).
• This magnetic moment primarily comes from the spin and is defined as: µ = γ × φ
o Here, γ (gamma) is known as the gyromagnetic ratio.
o The gyromagnetic ratio can be measured for different types of nuclei.
o Its units are typically radians per second per Tesla.
o A related notation, γ̄ (gamma bar), is sometimes used, which is γ / (2π).

Key Takeaways:

• The fundamental source of MRI signal is the spin angular momentum of nuclei,
particularly hydrogen protons in the body's water content.

10
 • This spin gives rise to a magnetic moment, which is key for interaction with external
magnetic fields.

3. Macroscopic Magnetization
While individual nuclei have magnetic moments, it's the collective behavior that creates a
measurable signal.

3.1. Spins in the Absence of an External Magnetic Field

• In the human body, hydrogen atoms (mostly protons) are abundant due to water content
and are the overwhelming contributors to the spin signal.
• In the absence of an external magnetic field, individual spins are randomly oriented.
o Analogy: Think of these spins like compass needles pointing in random
directions.
• This random orientation causes the individual magnetic moments to cancel each other
out, resulting in zero net bulk magnetization or macroscopic magnetization.

3.2. Spins in the Presence of a Static Magnetic Field (B₀)

• When a static external magnetic field (B₀) is applied (typically along the z-axis), there
is a preferred orientation of spins along the direction of the applied field.
o From a quantum mechanics perspective, spins align either along or opposite to the
field direction.
• This preferred orientation leads to a net magnetization (M), which is the vector sum of
individual nuclear magnetic moments in the sample.
o If a sample is left in a static B₀ field long enough, this net magnetization, often
denoted as M₀, develops.

3.3. Factors Influencing M₀

• The induced net magnetization (M₀) depends on:

o The strength of the applied static magnetic field (B₀).
o The proton density (PD) of the sample. (Proton density refers to the density of
hydrogen atom protons).
• The ultimate MR signal produced depends on M₀, which in turn depends on B₀ and
proton density.
• This induced magnetization is not static; it is a function of space (r) and time (t),
meaning M(r, t).

3.4. Bulk Angular Momentum

• Similar to individual nuclei having spin angular momentum, the entire sample has a bulk
angular momentum (J).
• This bulk angular momentum is related to the sample's magnetization by the
expression: M = γJ
o Again, γ is the gyromagnetic ratio.

11
Key Takeaways:

• A static magnetic field (B₀) is essential to induce a net macroscopic magnetization

(M₀) by aligning the previously randomly oriented spins.
• The strength of this induced magnetization is proportional to the B₀ field strength and
the proton density of the tissue.

4. Dynamics of Magnetization
Once a static magnetic field is applied, the magnetization doesn't just align; it undergoes a
specific motion.

4.1. Torque and Precession

• The presence of the static magnetic field induces a torque on the magnetization (M).
• The rate of change of angular momentum (torque, dJ/dt) is given by the cross product of
the magnetization and the applied static magnetic field (B): dJ/dt = M × B
o Analogy: This is similar to a current-carrying loop experiencing a torque in a
magnetic field, a concept from basic electromagnetism.
• Since M = γJ and B is typically aligned with the z-axis (B₀), solving the differential
equations for the magnetization vector (Mₓ, Mᵧ, M₂) reveals a sinusoidal dependence
with respect to time for the Mₓ and Mᵧ components.
• This sinusoidal motion indicates a precession of the magnetization vector around the
direction of the applied static magnetic field (the z-axis).
o Analogy: This motion is similar to a spinning top or gyroscope precessing in a
gravitational field, where the axis of the top is the direction of M and the z-axis is
the direction of the static magnetic field.
o (Insert diagram of a precessing magnetization vector around the z-axis, similar to
a spinning top, with B0 along z)

4.2. Larmor Frequency

• The frequency of this precession is called the Larmor frequency (ω₀ or ν₀).
• It is directly proportional to the gyromagnetic ratio and the strength of the static magnetic
field: ω₀ = γB₀ (radians per second) ν₀ = γB₀ / (2π) (Hertz)
• The Larmor frequency for MR signals typically falls within the radio frequency (RF)
range (megahertz).

4.3. Sources of B₀ Fluctuations and Isochromates

• While the applied B₀ field in an MRI scanner is designed to be homogeneous, local

variations can occur due to:
1. Hardware inhomogeneities.
2. Intrinsic properties of the sample:
§ Magnetic susceptibility.
§ Chemical shift.

12
 • These local variations (ΔB) mean that the B₀ field is effectively a function of spatial
position, B₀(r).
• Consequently, the Larmor frequency (ν₀) will also vary slightly locally, which is a
phenomenon exploited for imaging and obtaining contrast.
• Isochromates are groups of nuclei in a spin system that have the same Larmor
frequency.
o For example, hydrogen nuclei in water will form one isochromate, while those in
fat might form another, having slightly different Larmor frequencies due to their
chemical surroundings.

Key Takeaways:

• Under a static magnetic field, the net magnetization precesses around the field direction
at a characteristic Larmor frequency.
• Local variations in the magnetic field influence the Larmor frequency, a principle that
can be used for spatial encoding and contrast.

5. MRI Signal Origin and Measurement

The precessing magnetization is not directly measured; its effects are observed.

5.1. Transverse and Longitudinal Magnetization

• The magnetization vector (M) can be conceptually divided into two components:
o Longitudinal component (M_z): This component is along the direction of the
static magnetic field (z-axis).
o Transverse component (M_xy): This is the in-plane component, perpendicular
to the static magnetic field, comprising Mₓ and Mᵧ. It is often written in complex
form as M_xy = Mₓ + jMᵧ.
• In the initial setup with only a static B₀ field, the individual spins preferentially align
along the z-direction, leading to a net M_z. The transverse components generally cancel
out.

5.2. Inducing the Measurable Signal

• The MR signal (the origin of the image) comes from the rapidly rotating M_xy
component.
• If a coil of wire (receiver coil) is placed outside the sample, this rotating
magnetization induces a current (or voltage/EMF) in the coil. This induced current is
the NMR signal (or MR signal).
• The frequencies of this signal are in the radio frequency (megahertz) range.

5.3. Principle of Reciprocity and Signal Strength

• Directly measuring the net magnetization by integrating flux across a coil is problematic
for imaging, as it only yields a single quantity for the entire sample.

13
 • The principle of reciprocity allows the induced voltage to be rewritten as a volume
integral of the magnetization M. This is crucial for obtaining spatial localization,
allowing measurement of magnetization as a function of position, M(r, t).
• The magnitude of the induced voltage (signal) in the coil is proportional to:
o Larmor frequency (ν₀).
o Voxel volume (dV): An infinitesimal volume over which the magnetic field or
magnetization is constant.
o Magnitude of the induced magnetization (M₀).
o Sine of the tip angle (sin α).
• To get a good signal, the tip angle (α) should be maximized, ideally π/2 (90 degrees),
which means the entire magnetization is flipped into the transverse plane.

Key Takeaways:

• The transverse magnetization (M_xy) is the only component that produces a

measurable signal.
• This signal is an induced current in a receiver coil, driven by the precession of M_xy.
• The strength of the measured signal is influenced by the Larmor frequency, voxel
volume, initial magnetization M₀, and critically, the tip angle (α).

6. Manipulating Magnetization for Signal Generation

To generate a strong MR signal, the longitudinal magnetization (M_z) must be tipped into the
transverse plane (M_xy).

6.1. The Rotating Frame of Reference

• The M_xy component processes rapidly at the Larmor frequency (ν₀).

• To simplify analysis, a rotating frame of reference is often used. If the coordinate
system itself rotates at the Larmor frequency, the precessing magnetization appears
stationary in this new frame, simplifying calculations.

6.2. Applying the RF Signal (B₁ Field)

• The tipping of magnetization into the transverse plane is achieved by applying a radio
frequency (RF) signal, which generates an oscillating magnetic field (B₁ field).
• The B₁ field is typically applied perpendicular to the static B₀ field (e.g., if B₀ is along z,
B₁ might be along x).
• For effective tipping, the B₁ field must oscillate at the Larmor frequency (ν₀) of the
nuclei being imaged. This is the principle of nuclear magnetic resonance.
o If the B₁ field were applied at an arbitrary frequency, the much stronger B₀ field
would dominate, preventing effective tipping.
o Analogy: This is similar to timing your pushes on a swing to achieve a higher
displacement. The B₁ field must be in resonance with the precessing spins to
effectively push them.
• The B₁ field is not static like B₀; it is applied only for a specific duration.

14
 • The tip angle (α) achieved depends on the strength of the B₁ field and the duration of its
application: α = ∫(γB₁(t) dt)
• In the laboratory frame of reference, the magnetization spirals into the transverse plane
when the B₁ field is applied. In the rotating frame of reference, it simply appears to tip.

Key Takeaways:

• An RF (B₁) pulse oscillating at the Larmor frequency is applied perpendicular to B₀

to tip the longitudinal magnetization into the transverse plane.
• This resonance phenomenon is crucial for effective signal generation.
• The tip angle is controlled by the duration and strength of the RF pulse.

7. Relaxation Processes and Signal Decay

Once the B₁ field is turned off after tipping, the induced signal does not last indefinitely due to
natural physical processes.

• Ideally, if left undisturbed, the tipped magnetization would precess forever. However,
physical loss processes dampen this motion, causing the in-plane (transverse)
magnetization to decay.
• There are two main types of relaxation, characterized by different time
constants: T1 and T2. These are tissue-dependent properties.

7.1. Transverse Relaxation (T₂) / Spin-Spin Relaxation

• This process causes the decay of the transverse magnetization (M_xy).

• It occurs due to inhomogeneities in the local magnetic field (B₀) caused by
the neighboring spins themselves.
• These perturbations lead to dephasing of the individual spins.
o When spins dephase, they process at slightly different frequencies, causing their
individual transverse magnetic moments to cancel each other out, leading to a
rapid decay of the net M_xy signal.
• This dephasing results in a rapidly decaying signal called Free Induction Decay (FID).
• The time constant characterizing this decay is T₂.
o T₂ for tissues ranges from approximately 25 to 250 milliseconds.
o (Insert diagram showing spins in phase after 90-degree pulse, then gradually
dephasing over time, leading to FID signal decay)

7.2. Longitudinal Relaxation (T₁) / Spin-Lattice Relaxation

• This process describes the reappearance or recovery of the longitudinal

magnetization (M_z).
• After tipping M_z into the transverse plane (e.g., a 90-degree pulse makes M_z go to
zero), the M_z component slowly starts to reappear over time.
• The rate at which M_xy decays to zero (T₂) and the rate at which M_z reappears (T₁)
are different.
• The time constant characterizing this recovery is T₁.

15
 o T₁ for tissues ranges from approximately 250 to 1500 milliseconds.
o T₁ is always higher than T₂.

7.3. Block Equations

• The evolution of magnetization in the presence of B₀, B₁, and accounting for relaxation
processes (T₁ and T₂) is described by the Block Equations.
• These are differential equations that show how Mₓ, Mᵧ, and M_z evolve.
• They are fundamental for understanding signal behavior and designing different MRI
pulse sequences.

Key Takeaways:

• The MR signal in the transverse plane decays due to T₂ (spin-spin relaxation), which
describes spin dephasing.
• The longitudinal magnetization recovers due to T₁ (spin-lattice relaxation).
• These T₁ and T₂ relaxation times are crucial tissue-dependent properties that
underpin MRI contrast.

8. MRI Contrast Mechanisms

The appearance of an MRI image (contrast) is influenced by three main factors. MRI offers high
flexibility in generating different types of contrast.

8.1. Factors Contributing to Contrast

1. T₁ Relaxation Time: The rate of longitudinal magnetization recovery.

2. T₂ Relaxation Time: The rate of transverse magnetization decay.
3. Proton Density (PD): The concentration of hydrogen protons in the tissue.

8.2. Pulse Sequences

• The type and ordering of RF excitations and their relative timing determine which of
these three factors contributes more to the image contrast. This is managed by a pulse
sequence.
• Two critical timing parameters in a pulse sequence are:
o Repetition Time (TR): The time interval between two successive RF excitation
pulses (e.g., 90-degree pulses).
o Time to Echo (TE): The time from the application of the RF excitation pulse
(e.g., 90-degree pulse) to the measurement of the MR signal (echo).
• Only the transverse magnetization provides the measurable signal; thus, a larger
transverse component leads to a higher signal.
• For contrast, the measured signal must be different in different tissue types.

8.3. Types of Weighted Imaging

By manipulating TR and TE, different types of contrast can be achieved:

16
8.3.1. Proton Density (PD) Weighted Imaging

• Goal: To make the image contrast primarily dependent on the density of protons in the
tissue.
• Settings:
o Long TR: Allows for full or near-full recovery of longitudinal magnetization
(M_z) in all tissues before the next pulse. This minimizes the influence of T₁
differences.
o Short TE: The signal is measured quickly after the excitation pulse, before
significant T₂ decay occurs. This minimizes the influence of T₂ differences.
• Result: Tissues with higher proton density will appear brighter.

8.3.2. T₂-Weighted Imaging

• Goal: To make the image contrast primarily dependent on the T₂ relaxation times of the
tissues.
• Settings:
o Long TR: Reduces T₁ effects by allowing full M_z recovery, similar to PD
weighting.
o Medium to Long TE: The TE is selected to allow differences in T₂ decay to
become apparent. Tissues with longer T₂ will retain their signal for longer and
appear brighter, while those with shorter T₂ will decay faster and appear darker.
• Example: For a tissue with T₂ of 20ms and another with 5ms, measuring at 5ms wouldn't
show much difference. But waiting longer, say 20ms, would show a significant difference
as the 5ms tissue's signal would have largely decayed.

8.3.3. T₁-Weighted Imaging

• Goal: To make the image contrast primarily dependent on the T₁ relaxation times of the
tissues.
• Settings:
o Short TR: This setting exploits differences in T₁ recovery. Tissues with shorter
T₁ will recover their longitudinal magnetization more quickly and will therefore
have more M_z available to be tipped by the next RF pulse, resulting in a stronger
signal and appearing brighter.
o Short TE: Minimizes T₂ effects, similar to PD weighting, ensuring that T₁
differences are the dominant factor.

Key Takeaways:

• Pulse sequences, specifically the timing parameters TR and TE, are manipulated to
create different types of image contrast (T1-weighted, T2-weighted, Proton Density-
weighted).
• These contrasts arise because T1, T2, and proton density vary between different
biological tissues.

9. Example Pulse Sequence: Spin Echo

17
The spin echo sequence is a fundamental technique to generate an MR signal.

• (Insert flow diagram of Spin Echo Sequence: 90-degree pulse -> Dephasing -> 180-
degree pulse -> Rephasing/Echo -> Decay)
• Steps:
1. 90-degree RF Pulse: An initial RF pulse (B₁ field) is applied, typically tipping
the magnetization by 90 degrees into the transverse (xy) plane.
2. Dephasing: Immediately after the 90-degree pulse, the individual spins begin to
dephase due to local magnetic field inhomogeneities (T₂ and scanner
inhomogeneities), causing the M_xy signal to decay (FID).
3. 180-degree RF Pulse: At a time TE/2 after the 90-degree pulse, a 180-degree
RF pulse is applied.
§ This pulse flips the dephased magnetization 180 degrees.
§ It effectively reverses the dephasing process, causing the spins to rephase.
4. Echo Formation: As the spins rephase, they temporarily align again, producing a
new, measurable signal called an echo at time TE (which is TE/2 after the 180-
degree pulse, or TE after the initial 90-degree pulse).
5. Subsequent Decay: After forming the echo, the spins will again begin to
dephase, and the signal will decay.

Key Takeaways:

• The spin echo sequence uses a 90-degree and a 180-degree RF pulse to generate a
stronger, more controlled signal (the echo) by rephasing spins.
• This method effectively counteracts some of the dephasing effects that lead to FID.

Glossary of Key Terms

• Spin Angular Momentum (φ): An intrinsic quantum mechanical property of atomic
nuclei that gives rise to magnetic properties.
• Magnetic Moment (µ): A vector quantity representing the magnetic strength and
orientation of a nucleus, derived from its spin angular momentum.
• Gyromagnetic Ratio (γ): The ratio of a particle's magnetic moment to its angular
momentum; a constant for a given nucleus.
• Static Magnetic Field (B₀): The strong, constant external magnetic field applied in an
MRI scanner, typically along the z-axis.
• Net Macroscopic Magnetization (M or M₀): The bulk magnetization of a sample
resulting from the preferred alignment of individual nuclear magnetic moments in an
external B₀ field.
• Larmor Frequency (ω₀ or ν₀): The characteristic frequency at which nuclear spins
precess around the direction of the static magnetic field.
• Isochromates: Groups of nuclei that precess at the same Larmor frequency, often due to
similar chemical environments.
• Longitudinal Magnetization (M_z): The component of the net magnetization aligned
with the static magnetic field (z-axis).

18
 • Transverse Magnetization (M_xy): The component of the net magnetization
perpendicular to the static magnetic field (in the xy-plane). This is the only measurable
component.
• Radio Frequency (RF) Signal (B₁ Field): An oscillating magnetic field applied
perpendicular to B₀, used to tip the magnetization into the transverse plane by inducing
resonance.
• Tip Angle (α): The angle by which the net magnetization vector is rotated from its
equilibrium (longitudinal) position into the transverse plane by an RF pulse.
• T₁ Relaxation (Longitudinal Relaxation / Spin-Lattice Relaxation): The process by
which the longitudinal magnetization (M_z) recovers towards its equilibrium value after
being tipped.
• T₂ Relaxation (Transverse Relaxation / Spin-Spin Relaxation): The process by which
the transverse magnetization (M_xy) decays due to dephasing caused by local magnetic
field inhomogeneities from neighboring spins.
• Free Induction Decay (FID): The rapidly decaying MR signal observed immediately
after an RF excitation pulse, due to transverse relaxation.
• Proton Density (PD): The concentration of hydrogen nuclei (protons) in a given tissue.
• Pulse Sequence: A specific sequence of RF pulses and magnetic field gradients (not
detailed in this excerpt but implied) used to generate and collect MR signals, determining
the image contrast.
• Repetition Time (TR): The time interval between successive RF excitation pulses in a
pulse sequence.
• Time to Echo (TE): The time interval between the application of an RF excitation pulse
and the measurement of the echo signal.
• Spin Echo: A specific pulse sequence that uses a 90-degree RF pulse followed by a 180-
degree RF pulse to rephase spins and create a measurable echo signal.

Thought-Provoking Questions & Exam/Interview Questions

1. Conceptual Understanding: Explain why individual nuclear spins do not produce a net
signal in the absence of an external magnetic field, but they do so in its presence. What
macroscopic property arises from this?
2. Larmor Frequency: How does the Larmor frequency relate to the static magnetic field
strength (B₀)? Why is this relationship crucial for MRI?
3. Role of RF Pulse: Describe the purpose of the RF (B₁) pulse in MRI. Why must it be
applied at the Larmor frequency?
4. Signal Generation: What specific component of magnetization produces the measurable
MR signal? How is this signal detected by the MRI scanner?
5. Relaxation Explained: Differentiate between T₁ and T₂ relaxation in terms of the
magnetization component affected, the underlying physical processes, and their typical
time scales in biological tissues. Why is T₁ always longer than T₂?
6. Contrast Mechanisms: You are asked to obtain an MRI image that highlights areas with
high water content (long T₂). Which weighting would you choose (T₁W, T₂W, or PDW)
and what TR and TE settings would you use to achieve this? Explain your reasoning.

19
 7. Spin Echo Sequence: Describe the sequence of RF pulses in a basic spin echo sequence.
What is the role of the 180-degree pulse, and what is the outcome?
8. Image Interpretation: If you see a bright lesion on a T₁-weighted image and a dark
lesion on a T₂-weighted image, what might this suggest about the T₁ and T₂ relaxation
times of the tissue in the lesion compared to surrounding healthy tissue? (Note: This is an
interpretive question that builds on the provided definitions).
9. Further Research: How do the concepts of magnetic susceptibility and chemical shift
lead to localized differences in Larmor frequency, and how are these exploited in more
advanced MRI techniques beyond basic T1/T2 weighting?

Here are your notes on Magnetic Resonance Image Acquisition, based on the provided transcript:

Magnetic Resonance Image Acquisition

This class provides an understanding of how Magnetic Resonance (MR) images are acquired,
focusing on the hardware involved and the Pulse Sequences used. It also touches upon contrast
mechanisms.

I. MR Imaging System Hardware (Instrumentation)

The MR imaging system consists of several key components that work together to acquire
images. We will explore these superficially, focusing on their functions and challenges.

A. The Magnet

The magnet is the main and most expensive component of an MR scanner.

• Type: Typically a cylindrical superconducting magnet.

• Construction:
o It contains a superconducting wire, usually niobium titanium wire, immersed
in liquid helium.
o Liquid helium is held at 4 Kelvin, while the wire itself is superconducting at 9
Kelvin.
o The helium is contained within a cryostat, which is composed of a vacuum and
liquid nitrogen.
o Maintenance: While liquid helium refilling is not frequent, liquid nitrogen
refilling is required.
• Function: Once energized, the superconducting wire allows current to flow continuously
with minimal resistance loss, setting up a static magnetic field.
• Magnetic Field Properties:
o The static magnetic field points along the axis of the core of the magnet.

20
 o Field Strengths: Clinical magnets typically range from 0.5 Tesla to 3 Tesla.
Modern clinical magnets are often 1.5 Tesla or 3 Tesla.
o Higher Field Magnets: 9 Tesla magnets exist and are used in small animal
imaging, biomedical research, and general research.
o Challenge: A significant challenge is maintaining the uniformity of the
magnetic field across the entire cross-section of the bore.
• Cost Reduction Efforts: Research is ongoing to use permanent magnets for more
economical and portable MR systems, especially for imaging peripheral organs, which
can significantly reduce system cost.

(Insert diagram of magnet cross-section here, showing superconducting wire, liquid

helium, liquid nitrogen, and cryostat.)

Key Takeaways: Magnet

• The magnet is central to MRI, generating the strong, static magnetic field.
• Superconductivity is crucial for maintaining the field without constant power input.
• Maintaining field uniformity is a key engineering challenge.
• Higher field strengths enable specialized research, while clinical systems balance field
strength with practical considerations like cost and size.

B. Gradient Coils

Gradient coils are current-carrying wires that fit inside the bore of the main magnet.

• Configuration: There are three coils, each producing magnetic field

gradients orthogonal to each other.
• Function: They do not change the direction of the static magnetic field (which still
points along the z-axis). Instead, they add an x, y, and z dependency to the magnetic
field strength, meaning the magnitude of the field varies linearly with position.
• Types of Coils:
o Saddle coils: Produce x and y gradients.
o Two opposing coils (wound around circumference): Produce the z gradient.
• Amplitude:
o The gradient amplitude is limited by the very high currents (100 to 200 Amperes)
flowing through the coils.
o Maximum gradient amplitude: 0 to 6 Gauss per centimeter (10 to 60 milliTesla
per meter). This is significantly smaller than the main static magnetic field (e.g.,
3 Tesla).
• Switching Times:
o Gradient switching is not instantaneous. The time it takes for a gradient to turn
on and reach its maximum value is known as the slew rate.
• Hardware Challenges & Limitations:
o Eddy Currents: The constant and rapid switching of high currents induces Eddy
currents in the metallic components of the magnet housing. These distort the
gradient field, requiring shielding.

21
 o Distortion: The introduction of these current-carrying wires, wound in specific
shapes (like saddle or circular coils), can further distort the magnetic field.
o Self-inductance: The self-inductance of the coils prevents rapid switching of the
gradients.
o Field of View: Making coils smaller reduces the field of view.
o Biological Hazard (Patient Safety): Rapid switching of currents can
induce Eddy currents in the patient, potentially leading to nerve damage.
There's a biological limitation on the maximum switching speed.

(Insert diagram of gradient coils here, showing saddle coils for X/Y and opposing circular
coils for Z.)

Key Takeaways: Gradient Coils

• Gradients are crucial for spatial encoding by making the magnetic field position-
dependent.
• They only change the magnitude of the field, not its direction.
• Rapid switching introduces significant engineering and patient safety challenges (Eddy
currents, nerve stimulation).

C. Radiofrequency (RF) Coils

RF coils are responsible for both exciting the spins (producing the B1 field) and measuring the
signals emitted by the spins.

• Dual Role: They act as both transmit and receiver coils.

• Function:
o Transmission: The transmitted RF pulse causes the spins to precess.
o Reception: They measure the induced currents generated by the precessing spins
(the echo signal).
• Configurations:
o Volume coils: Surround the patient.
o Surface coils: Placed in close proximity to the patient.
• Design Challenges: Designing these coils to generate the correct RF field and shape is a
significant area of research.
• Current Amplitudes: Transmission and reception require very different current
amplitudes, as the emitted current from spins is much smaller than the current needed to
generate the B1 field.

Key Takeaways: RF Coils

• RF coils are the interface between the scanner and the patient's spins, enabling both
excitation and signal detection.

22
 • Their design is critical for efficient and accurate imaging.

D. Electronics and Imaging Console

• Imaging Console: Similar to other diagnostic devices (e.g., CT scanners), the MR system
has a scanning console.
o Interface: It serves as the interface with the MR hardware.
o Control Functions: Allows selection of imaging planes, setting ECG gating (for
cardiac motion), and respiratory gating (for breathing motion).
• Reconstruction Engine: The console is connected to a reconstruction engine (which can
be a processor, embedded system, or workstation).
o Function: This engine processes the acquired data to reconstruct images.
o Speed: Capable of reconstructing 10 to 50 images per second, enabling real-
time imaging.
o Clinical Application: For example, in cardiac MRI, specialized hardware and
protocols allow imaging the entire volume of the heart in a single cardiac cycle,
acquiring several volumes very quickly.

Key Takeaways: Console & Electronics

• The console provides user control and manages critical clinical parameters like
physiological gating.
• The reconstruction engine enables real-time image generation, vital for dynamic studies
like cardiac imaging.

II. Pulse Sequences for Image Acquisition

After understanding the principles of MR signals, we now look at how these phenomena are used
to acquire images in the context of Magnetic Resonance Imaging. This involves Pulse
Sequences.

A. Fundamental Steps of Image Acquisition

The acquisition of an MR image fundamentally involves four main steps:

1. Slice Selection Gradient

2. Frequency Encoding Gradient (also known as the readout gradient)
3. Phase Encoding Gradient
4. Image Reconstruction

These steps localize the MR signal in x, y, and z positions. The signal received from the excited
slice is an integration or summation of signals from all spins within that slice.

B. MR Signal Model

23
The signal received by the coils, denoted as S(t), is a function of the transverse magnetization
(Mxy) that has been flipped into the plane. This Mxy varies due to factors like local magnetic
field variations, proton density, and T1/T2 relaxation times. The general signal model is
represented as an integral over the spatial distribution of these spin properties within the slice.

When demodulated, the received signal is essentially the integral of the spin density (f(x,y))
within the particular slice. The goal of spatial encoding is to extract x, y, and z information from
this signal.

C. 1. Slice Selection Gradient

• Purpose: To isolate a specific cross-section (slice) of tissue along one dimension (e.g.,
the z-axis).
• Mechanism: A gradient along the z-axis (Gz) is applied simultaneously with the RF
excitation pulse.
• Effect on Larmor Frequency: The gradient makes the Larmor precession frequency a
function of position along the z-axis: Nu (z) = γ * (B₀ + Gz * z)
o Where Nu(z) is the precession frequency at position z, γ (gamma) is the
gyromagnetic ratio, B₀ is the static magnetic field, and Gz is the strength of the z-
gradient.
• Slice Excitation: Since the RF excitation pulse has a specific center frequency and a
defined bandwidth, it will only excite spins whose Larmor frequencies match this range.
Because the frequency is now position-dependent, this effectively excites only a thin
section of tissue at a specific z-location.
• Controlling Slice Thickness and Position:
o Slice Thickness (Δz): The thickness of the excited slice is inversely
proportional to the gradient strength (Gz) for a given RF pulse bandwidth
(ΔNu): Δz = ΔNu / (γ * Gz)
§ A stronger gradient (sharp slope) results in a thinner slice.
§ A weaker gradient (shallow slope) results in a thicker slice.
o Slice Position: Can be determined by the center frequency of the RF pulse
relative to B₀ and the gradient.
• RF Waveform Design: Ideally, to excite a rectangular slice in the frequency domain, the
RF pulse in the time domain should be a sinc function.
• Refocusing Gradient: After the main z-gradient and RF pulse, a negative z-gradient is
applied (often for half the duration of the initial Gz).
o Purpose: To rephase spins that have dephased due to slight differences in
precession frequencies within the finite thickness of the selected slice. This
ensures that the FID signal received is coherent.
• Timing: The time t=0 in pulse sequence diagrams typically coincides with the peak of
the RF pulse. The FID signal (Free Induction Decay) is obtained after refocusing,
representing the sum of signals from the entire slab of tissue.

(Insert diagram of a typical pulse sequence diagram for slice selection, showing RF pulse
and Z-gradient with a refocusing lobe.)

24
Key Takeaways: Slice Selection

• Spatial encoding in MRI begins with slice selection using a z-gradient and a specific RF
pulse.
• The Larmor frequency's dependence on position (due to the gradient) is fundamental
here.
• Slice thickness can be precisely controlled by adjusting the gradient strength.
• Refocusing gradients are essential for coherent signal acquisition.

D. 2. Frequency Encoding Gradient (Readout Gradient)

• Purpose: To spatially encode the MR signal along one in-plane direction (e.g., the x-
axis) within the selected slice. This direction is often called the readout direction.
• Direction: Typically orthogonal to the slice selection gradient (e.g., if slice selection is
along z, readout is along x).
• Mechanism: An additional gradient along the x-axis (Gx) is turned on during the
acquisition of the FID signal.
• Effect on Larmor Frequency: Similar to slice selection, this gradient makes the Larmor
frequency dependent on the x-position: Nu (x) = γ * (B₀ + Gx * x)
• Signal Interpretation as Fourier Transform: The signal received (S(t)) can be
mathematically interpreted as the Fourier transform of the spin density (f(x,y)) within
the excited slice.
o The frequency variable U (or kx) in the Fourier domain is directly related to the
gradient and time: U = γ * Gx * t.
o In this step, the other spatial frequency variable V (or ky) is implicitly set to zero.
• k-space: The Fourier domain is referred to as k-space (where U = kx and V = ky).
• Acquisition Strategy: MRI acquisition involves scanning k-space by changing the
gradients to acquire different "spatial frequencies" (U and V values). Once k-space is
sufficiently sampled, an inverse Fourier transform is performed to reconstruct the
image.

(Insert diagram of a typical pulse sequence diagram showing slice selection, followed by the
readout gradient during FID acquisition.)

Key Takeaways: Frequency Encoding

• Frequency encoding maps spatial position to signal frequency.

• It's the first step in "filling" k-space by acquiring data along one dimension (e.g., U-axis
where V=0).
• The acquired signal is essentially a Fourier transform of the object's spin density.

E. 3. Phase Encoding Gradient

• Purpose: To spatially encode the MR signal along the second in-plane direction (e.g.,
the y-axis), allowing for full coverage of k-space.

25
 • Traversing k-space: There are different ways to traverse k-space:
o Polar Scanning (less common): Involves simultaneously changing the strengths
of both x and y gradients (Gx and Gy) to sample k-space radially. The angle of
sampling (theta) is determined by the ratio of Gy to Gx. This protocol requires
repeating the pulse sequence with different gradient strengths to cover various
angles.
o Phase Encoding (more common):
§ Mechanism: A gradient along the y-axis (Gy) is applied for a fixed
duration (Tp) after slice selection and its refocusing gradient,
but before the readout gradient is turned on for data acquisition.
§ Effect: Crucially, applying the phase encoding gradient and then
switching it off imparts a position-dependent phase change to the
precessing spins. This phase difference is preserved even after the
gradient is removed.
§ When the readout gradient (Gx) is then applied during signal acquisition,
there is now both a position-dependent frequency (from Gx) and a
position-dependent phase (from Gy) contributing to the signal.
§ Phase Accumulation: The phase accumulated (phi) is a function of
position along y: phi (y) = γ * Gy * Tp * y
§ Signal Model in k-space: In the Fourier domain (k-space), the phase
encoding gradient controls the V (or ky) variable: V = γ * Gy * Tp.
§ Scanning k-space: To sample the entire k-space, the pulse sequence is
repeated multiple times, each time with a different strength of the
phase encoding gradient (Gy). This changes the 'V' value for each
repetition, effectively stepping through different lines in k-space. The
frequency encoding gradient then samples along the 'U' dimension for that
particular 'V' line.

(Insert diagram of a typical pulse sequence diagram for phase encoding, showing slice
selection, phase encoding gradient (Gy) before readout gradient (Gx), and ADC
acquisition.)

Key Takeaways: Phase Encoding

• Phase encoding provides the second dimension of in-plane localization by imparting

position-dependent phase shifts.
• It requires repeating the pulse sequence multiple times with varying gradient strengths.
• Together, frequency and phase encoding allow sampling of the entire 2D k-space.

F. Timing Parameters in Pulse Sequences

While not extensively detailed, specific timings are crucial for MR image acquisition protocols:

• Time to Echo (TE): This is the time at which the echo signal is measured. It typically
corresponds to the middle of the measurement time window.

26
 • Time to Repetition (TR): This is the time gap between successive repetitions of a
pulse sequence. Since k-space is built up by repeating sequences with different gradient
settings (especially for phase encoding), TR determines the total scan time.

Concept Map/Flow Diagram: MR Image Acquisition Pipeline

Main Magnet (Static B0 Field)

↓
Gradient Coils (Gradients for Spatial Encoding)
↓
RF Coils (Excitation B1 Field & Signal Reception)
↓
Pulse Sequence Control (Console)
↓
[STEP 1: SLICE SELECTION]
- Apply Z-gradient (Gz) + RF Pulse
- Localizes signal to a thin Z-slice (freq. depends on Z)
- (Optional) Apply Z-refocusing gradient for phase coherence
↓
[STEP 2: PHASE ENCODING]
- Apply Y-gradient (Gy) for duration Tp (then switch off)
- Imparts position-dependent phase (preserved)
- (Repeat for different Gy strengths to fill k-space lines)
↓
[STEP 3: FREQUENCY ENCODING (READOUT)]
- Apply X-gradient (Gx) *during* signal acquisition
- Maps X-position to signal frequency
- ADC samples time-domain signal (corresponds to U/kx in k-space)
↓
Acquired Raw Data (F(U,V) in K-Space)
↓
Reconstruction Engine (Performs Inverse Fourier Transform)
↓
Final MR Image (f(x,y,z))

III. Deepening Understanding & Further Research

Here are some questions and areas for further exploration based on the lecture content:

• Thought-Provoking Questions:
o How does the homogeneity challenge of the main magnet (due to manufacturing
imperfections or external fields) relate to the concept of shimming in MRI?
o Given the biological limitations of gradient switching, how do modern MRI
systems achieve faster acquisition times (e.g., cardiac MRI) without
compromising patient safety?
o If the FID signal is difficult to measure, how are "echoes" generated and what are
the different types of echoes (e.g., spin echo, gradient echo)? (This was briefly
mentioned but not explained).

27
 o The lecture mentioned "contrast mechanisms". How do T1 and T2 relaxation
times (also mentioned briefly in f(x,y)) contribute to image contrast, and how are
pulse sequences designed to enhance specific types of contrast?
• Areas for Further Research/Clarification:
o Detailed Physics of Superconductivity: How does the critical temperature and
current density affect magnet design and performance?
o Eddy Current Compensation: What specific shielding techniques and hardware
designs are used to mitigate Eddy currents?
o RF Coil Design: Explore different RF coil geometries (e.g., birdcage, phased
array coils) and their advantages/disadvantages for specific anatomical regions or
applications.
o k-space Trajectories: Beyond polar scanning and Cartesian phase-
encode/frequency-encode, what other k-space sampling strategies exist (e.g.,
spiral, radial) and what are their trade-offs in terms of speed, image quality, and
artifacts?
o Hardware-Software Interaction: How do the scanning console and
reconstruction engine "communicate" and coordinate the complex timing and
gradient control required for advanced pulse sequences?
• Potential Exam/Interview Questions:
o Describe the main components of an MRI scanner and briefly explain the function
of each.
o Explain how a superconducting magnet generates a static magnetic field in an
MRI system. What are the associated challenges?
o What is the primary role of gradient coils in MRI? How do they achieve spatial
localization, and what are their limitations?
o Compare and contrast frequency encoding and phase encoding in MRI. Why are
both necessary for 2D image acquisition?
o Explain the concept of k-space in MRI. How is data acquired to fill k-space, and
what role does the Fourier transform play in image reconstruction?
o Discuss the significance of TE and TR in MRI pulse sequences.

IV. Glossary of New Terms

• ADC (Analog-to-Digital Converter): A device that converts analog signals (like the raw
MR signal) into digital data for processing.
• B0 Field: The strong, static magnetic field generated by the main magnet in an MRI
scanner.
• B1 Field: The radiofrequency (RF) magnetic field generated by the RF coils, used to
excite hydrogen protons.
• Cryostat: A specialized container designed to maintain very low temperatures, typically
used to house the liquid helium and liquid nitrogen in a superconducting magnet.
• Demodulate: To separate the modulating signal from a modulated carrier wave. In MRI,
it refers to removing the high-frequency oscillation (Nu0) from the acquired signal to get
the baseband signal.

28
 • Eddy Currents: Unwanted circulating electrical currents induced in conductive materials
(like magnet housing or patient tissues) by changing magnetic fields (e.g., rapidly
switching gradients).
• ECG Gating: A technique used in MRI to synchronize image acquisition with the heart's
electrical activity (electrocardiogram), reducing motion artifacts in cardiac imaging.
• Echo: The refocused MR signal measured by the RF coils after excitation and application
of appropriate gradients. It is typically what is measured, as opposed to the initial FID
signal.
• FID (Free Induction Decay): The initial MR signal generated by precessing spins
immediately after an RF pulse. It rapidly decays due to dephasing.
• Field of View (FOV): The spatial extent of the region being imaged.
• Fourier Transform: A mathematical operation that converts a signal from its original
domain (e.g., time or spatial position) to a representation in the frequency domain.
Essential for MRI image reconstruction from k-space data.
• Frequency Encoding Gradient (Readout Gradient): A magnetic field gradient applied
during signal acquisition that makes the Larmor frequency position-dependent along one
in-plane direction, allowing spatial encoding.
• Gating: Techniques (e.g., ECG, respiratory) used to synchronize image acquisition with
physiological motions to minimize artifacts.
• Gauss per centimeter (Gauss/cm): A unit used to measure magnetic field gradients.
• Gradient Coils: Coils that generate spatially varying magnetic fields (gradients) in x, y,
and z directions, used for spatial encoding of the MR signal.
• k-space: The raw data matrix acquired in MRI, representing the spatial frequency domain
of the image. The image is reconstructed by performing an inverse Fourier transform on
k-space data.
• Kelvin (K): An absolute temperature scale used to describe very low temperatures
required for superconductivity.
• Larmor Precession Frequency: The frequency at which a proton (or other nuclear spin)
precesses in a magnetic field. It is directly proportional to the strength of the magnetic
field.
• MilliTesla per meter (mT/m): Another unit used to measure magnetic field gradients.
• Niobium Titanium (NbTi): A superconducting alloy commonly used for the wires in
superconducting magnets.
• Nuclear Magnetic Resonance (NMR): The fundamental phenomenon underlying MRI,
involving the absorption and emission of electromagnetic energy by atomic nuclei in a
magnetic field.
• Phase Encoding Gradient: A magnetic field gradient applied briefly before signal
acquisition to impart a position-dependent phase shift to spins along an in-plane
direction, crucial for filling k-space.
• Pulse Sequences: A precisely timed series of radiofrequency (RF) pulses and magnetic
field gradients that determine how MR signals are generated and spatial information is
encoded for image formation.
• Radiofrequency (RF) Coils: Coils that transmit RF pulses to excite spins and receive the
MR signals emitted by the spins.
• Rect Function: A mathematical function that is 1 within a specified interval and 0
otherwise, often used to describe ideal frequency content of RF pulses.
• Refocusing Gradient: A gradient applied to rephase spins that have dephased due to
magnetic field inhomogeneities or differences in Larmor frequency within a slice.

29
 • Respiratory Gating: A technique used in MRI to synchronize image acquisition with the
patient's breathing, reducing motion artifacts in chest and abdominal imaging.
• Saddle Coils: A specific configuration of gradient coils used to generate magnetic field
gradients in the x and y directions.
• Sinc Function: A mathematical function (sin(x)/x) that is the Fourier transform of an
ideal rectangular pulse. Often used to describe the ideal shape of an RF pulse in the time
domain for slice selection.
• Slice Selection Gradient: A magnetic field gradient applied simultaneously with an RF
pulse to excite a specific thin cross-section of tissue.
• Slew Rate: The rate of change of the magnetic field gradient strength over time.
• Spin Precession: The rotational motion of the magnetic moment of a proton about the
direction of the external magnetic field.
• Superconducting Magnet: A magnet made of superconducting wires that, when cooled
below a critical temperature, can conduct electricity with virtually no resistance, creating
a persistent and strong magnetic field.
• Tesla (T): The SI unit of magnetic field strength. MRI scanners are typically rated in
Tesla.
• Time to Echo (TE): The time from the start of the RF excitation pulse to the peak of the
echo signal.
• Time to Repetition (TR): The time between successive identical RF pulses in a pulse
sequence. It influences image contrast and total scan time.
• Transverse Magnetization (Mxy): The component of the net magnetization vector that
is perpendicular to the main magnetic field (B0). It is the source of the detectable MR
signal.
• Uniformity of Field: The consistency of the magnetic field strength across the imaging
volume, critical for accurate image formation.

Here are your detailed notes on Ultrasound Imaging, based on the provided transcript:

Ultrasound Imaging: Fundamentals and

Applications
This document provides a brief introduction to ultrasound imaging, explaining what the images
produced by these systems convey.

1. Introduction to Ultrasound Systems

Ultrasound imaging systems consist of a piezoelectric transducer and associated circuitry.
Unlike projection radiography and computed tomography (CT), ultrasound imaging primarily
involves signal acquisition and display rather than complex image reconstruction algorithms.

30
Most of the system's complexity lies in the hardware responsible for creating, directing, and
displaying the ultrasonic pulses.

2. Basic Principles and Contrast Mechanism

2.1 The Transducer: Transmit and Receive

The transducer is a critical component that performs a dual role: it

both transmits and receives ultrasound energy.

• When transmitting, the transducer sends mechanical pulses (longitudinal pulses) into the
human body. This is achieved by pulsing a piezoelectric crystal with an electrical signal,
causing it to vibrate rapidly and produce ultrasonic waves.
• When receiving, the transducer measures the ultrasound waves that
are reflected and scattered back from anatomical structures within the body. The
piezoelectric crystal's property allows it to generate an electric field when it experiences
mechanical pressure from these returning waves.

A gel is typically applied to the patient's skin before the scan to prevent significant reflection of
the generated pulses at the skin surface due to acoustic impedance mismatch.

2.2 Contrast Generation

The contrast in ultrasound images is generated by the differences in reflection and scattering
properties between various tissues in the body. This mechanism differs from other imaging
modalities. The image pixels generally indicate the reflectivity of the tissue at any given point.

(Insert flow diagram of Ultrasound Signal Pathway: Electrical Impulse -> Transducer
(Piezoelectric Crystal) -> Mechanical Ultrasound Pulse -> Propagation through Tissue ->
Reflection/Scattering at Tissue Interfaces -> Reflected/Scattered Waves -> Transducer
(Receives) -> Electrical Signal -> Signal Processing -> Image Display)

3. What is Ultrasound? Frequencies and Applications

3.1 Definition and Frequency Ranges

Ultrasound refers to sound frequencies above 20 kilohertz (kHz). These frequencies are beyond
the human audible range (which is limited to around 20 kHz or 20,000 Hertz). For medical
ultrasound, the frequency range is specifically between 1 to 10 megahertz (MHz). The physical
principles governing ultrasound propagation are the same as those for audible sound propagation.

3.2 Key Applications of Ultrasound Imaging

Ultrasound is widely used in medicine due to its non-ionizing and non-invasive nature.

31
 • Echocardiography: Real-time imaging of the beating heart, including fetal heart
monitoring, to detect abnormalities.
• Doppler Imaging: A specialized technique that measures velocities, such as blood flow.
• Fetal Ultrasound / Obstetrics: A common and safe method for prenatal imaging.
• Image-Guided Surgery: Used to monitor procedures during long surgeries, avoiding
continuous exposure to ionizing radiation.

Ultrasound offers good resolution depth and its range of applications is continually growing.

Thought-Provoking Question:

• Given its non-ionizing nature, why might ultrasound be preferred over CT or X-rays for
certain clinical applications, particularly in obstetrics?

4. Ultrasound Signal Acquisition

4.1 Time and Depth Correlation

When the ultrasound waves propagate through the body, they are reflected and scattered back to
the transducer. The acquired signal is a function of time. Crucially, in ultrasound, time
translates directly to depth.

The relationship between echo time and depth is given by the formula: Echo Time = 2z /
c Where:

• z = depth at which the reflection occurred

• c = speed of ultrasound in tissue, typically 1500 meters per second

This means that by measuring the time it takes for an echo to return, the system can calculate the
depth from which the signal originated, giving positional meaning to the image coordinates.

4.2 Imaging Coverage: 2D and 3D

• A typical 2D ultrasound image is formed from an "imaging cone".

• Volumetric (3D) coverage can be achieved by moving the transducer.

4.3 Longitudinal Wave Propagation

Ultrasound waves are longitudinal waves. They propagate through tissue via a process
of compression and relaxationof small volumes of tissue.

• When a tissue volume is compressed, it tries to return to its original position. Due
to inertia, it overshoots, compressing the neighboring volume, and this is how the wave
propagates.
• This creates regions of high pressure and low pressure.

32
 • The wave can be described in terms of the speed of particles in the medium or in terms
of acoustic pressure. These descriptions satisfy the wave equation, which is used to
model the signal.

4.4 Resolution Considerations

The lateral extent of coverage for an ultrasound beam varies with depth because the ultrasound
beam spreads over time, similar to a flashlight beam. This spreading leads to a loss of lateral
resolution as the depth increases.

Potential Exam/Interview Question:

• Explain the principle behind determining the depth of anatomical structures in an

ultrasound image. What is the typical speed of sound in tissue, and why is this value
important?

5. The Piezoelectric Transducer in Detail

5.1 Piezoelectric Effect

The core of the transducer is the piezoelectric crystal. These crystals exhibit a unique property:

• When an electric field is applied, they produce an ultrasound wave (mechanical pulses)
by vibrating rapidly.
• Conversely, if they experience mechanical pressure (strain or displacement), they
generate an electric field(inducing an electric potential).

This dual property allows the same transducer to both transmit and receive ultrasound waves. A
common piezoelectric material used is Lead Zirconate Titanate (PZT), which is favored for its
ease of manufacturing into various sizes and shapes.

5.2 Pulse Generation and Probe Assembly

An ultrasound pulse is generated by applying a very short, sharp electrical impulse to the
piezoelectric crystal, causing it to oscillate rapidly and produce the ultrasonic waves. The
complete transducer assembly, encased for practical use, is called an ultrasound probe.

5.3 Theoretical Modeling

For theoretical analysis, a single-element ultrasound probe can be modeled as a vibrating

plate or an oscillating circular membrane. This model helps in understanding signal
characteristics like its form and attenuation.

5.4 Types of Ultrasound Probes for Field of View

There are two main types of ultrasound probes for covering a field of view:

33
 • Mechanical Scanner: A single transducer element mechanically rocks back and forth to
acquire the entire 2D structure.
• Electronic Scanner (Array): This type uses multiple transducer elements arranged in
an array. These elements are fired in specific patterns.

5.5 Focusing and Steering via Wave Interference

A key advantage of electronic scanners is their ability to control the ultrasound beam using wave
interference.

• Since ultrasound is a wave, individual waves emanating from different points or elements
can interfere constructively or destructively.
• By precisely timing the pulses (introducing phase shifts) between the multiple elements
in an array, the waves can be made to interfere constructively at a specific depth or along
a particular direction.
• This sophisticated timing allows for focusing the ultrasound beam at desired depths
and steering it along specific directions, which is essential for modern ultrasound
scanners.

6. Modes of Ultrasound Scanning

Different modes of scanning allow for various types of image acquisition and display:

• A-mode (Amplitude Mode):

o The transducer faces along a specific direction (e.g., z-axis).
o It pulses and records echoes over time.
o Since time corresponds to depth, the signal's amplitude can be plotted against
depth.
o (Insert placeholder for an A-mode scan diagram: a simple plot of signal amplitude
vs. depth)
• B-mode (Brightness Mode):
o This is the most common mode for creating 2D images.
o It's formed by translating (mechanically or electronically) multiple A-mode scans
into a plain image.
o The brightness of each point on the display corresponds to the amplitude of the
reflected echo at that spatial location.
o (Insert placeholder for a B-mode scan diagram: a typical 2D ultrasound image
showing anatomical structures)
• M-mode (Motion Mode):
o Measures a single line scan over a period of time.
o Allows following the movement of structures (e.g., heart valve motion) over time
at a fixed location.
o Each line on the display represents a different time point.
o (Insert placeholder for an M-mode scan diagram: a plot showing movement of a
structure over time)
• 3D Scan:

34
 o Achieved by moving an imaging plane or rotating a 1D array to obtain volumetric
coverage.

Connection to Computer Vision/Deep Learning:

• The raw signal acquisition (A-mode) and subsequent processing to form B-mode images
involve signal processing concepts. In modern deep learning, models could potentially
learn to interpret raw A-mode data or even synthesize B-mode images directly from raw
sensor data, or enhance existing B-mode images.
• The "texture" observed in ultrasound images is a significant feature for diagnosis. Deep
learning models often excel at learning complex textural patterns for classification or
segmentation tasks.

7. Understanding Ultrasound Image Pixels: Reflectivity and

Impedance
7.1 Acoustic Impedance

The key physical quantity influencing reflections at tissue interfaces is acoustic impedance (Z).

• It represents the resistance a material offers to the transference of ultrasonic waves,

analogous to electrical resistance in Ohm's Law.
• Formula for Acoustic Impedance: Z = ρ * c Where:
o ρ (rho) = density of the material
o c = velocity of ultrasound in the material

Reflections occur at interfaces between materials (e.g., different organs) that have different
densities and thus different acoustic impedances.

7.2 Reflectivity and Image Interpretation

The reflectivity (or the fraction of the incident signal that is reflected) at an interface depends on
the difference in acoustic impedances between the two materials (Z2 - Z1).

• Low Signal-to-Noise Ratio (SNR): Ultrasound systems typically have a very low
SNR because a significant portion of the ultrasound signal is attenuated or lost as it
travels through the tissue.
• Relative Intensities: Unlike CT or X-ray radiography where pixel values represent
absolute physical properties (like attenuation coefficients), intensities in ultrasound
images are relative.
o A brighter region indicates that the tissue there is more reflective than a darker
region.
o You cannot directly infer the absolute density or attenuation coefficient from a
single pixel value.

35
 • Importance of Texture: Due to the relative nature of intensities, texture plays a crucial
role in ultrasound diagnosis. Different tissues exhibit distinct textural patterns, which are
interpreted by clinicians to identify pathologies.
• Speckles: Ultrasound images often contain speckles, which are bright spots or granular
patterns. These arise from the complex nature of ultrasound propagation and
measurement, specifically the coherent interference of scattered waves.

7.3 Sources of Ultrasound Signal

The two primary sources contributing to the received ultrasound signal are:

1. Reflection: Occurs from interfaces between structures that have different acoustic
properties (and thus different acoustic impedances).
2. Scattering: Occurs from microscopic inhomogeneities within the tissue that are smaller
than the ultrasound wavelength. While relatively weaker than surface reflections,
scattering is still a significant source of the overall signal.

8. Attenuation and Resolution Trade-offs

8.1 Signal Attenuation

The ultrasound pulse undergoes attenuation (loss of intensity) as it travels into and back out of
the tissue.

• Frequency Dependence: Attenuation increases with frequency. Higher frequency

waves lose more energy faster than lower frequency waves.

8.2 Resolution vs. Depth Trade-off

This frequency-dependent attenuation leads to a critical trade-off in ultrasound imaging:

• High Frequencies (e.g., 7-10 MHz):

o Pros: Provide better resolution.
o Cons: Experience much higher attenuation, limiting their ability to penetrate
deeply into the body.
o Use: Ideal for imaging superficial structures (e.g., thyroid, breast).
• Low Frequencies (e.g., 1-3 MHz):
o Pros: Experience less attenuation, allowing for deeper penetration.
o Cons: Result in lower resolution.
o Use: Essential for imaging deeper structures (e.g., liver, kidneys, fetal ultrasound).

Therefore, choosing the appropriate ultrasound frequency for a scan involves a compromise
between the desired image resolution and the necessary penetration depth.

Flag for Further Research:

36
 • The lecture mentions "mathematical models for ultrasound propagation" and "general
expression for an A scan". Further research into these models (e.g., acoustic wave
equation solutions, impulse response models) would deepen understanding of signal form
and resolution degradation.

Glossary of Key Terms

• A-mode (Amplitude Mode): A basic ultrasound display mode that plots echo amplitude
against depth.
• Acoustic Impedance (Z): A physical property of a medium that describes its resistance
to the propagation of sound waves (Z = ρ * c). Differences in acoustic impedance at
tissue interfaces cause reflections.
• Attenuation: The reduction in intensity of an ultrasound wave as it travels through tissue,
due to absorption and scattering.
• B-mode (Brightness Mode): The most common 2D ultrasound display mode, where
echoes are represented as dots of varying brightness, forming an anatomical image.
• Doppler Imaging: An ultrasound technique that uses the Doppler effect to measure the
velocity and direction of blood flow or tissue motion.
• Echocardiography: The specialized use of ultrasound to image the heart, often in real-
time.
• Longitudinal Wave: A type of wave where the oscillations of the medium particles are
parallel to the direction of wave propagation (e.g., sound waves).
• M-mode (Motion Mode): An ultrasound display mode that visualizes the movement of
structures over time along a single scan line.
• Piezoelectric Crystal: A material that produces an electric potential when subjected to
mechanical stress, and conversely, undergoes mechanical deformation when an electric
field is applied. Essential for ultrasound transducers.
• PZT (Lead Zirconate Titanate): A commonly used piezoelectric ceramic material in
ultrasound transducers.
• Reflectivity: The property of a tissue or interface to reflect ultrasound waves, which
dictates image contrast in ultrasound.
• Speckles: Granular or bright spots in ultrasound images, resulting from the coherent
interference of scattered ultrasound waves.
• Transducer: The component in an ultrasound system that converts electrical energy into
ultrasound waves (transmit) and ultrasound echoes back into electrical signals (receive).
• Ultrasound: Sound waves with frequencies higher than the upper audible limit of human
hearing (typically >20 kHz; medical ultrasound 1-10 MHz).

Key Takeaways
• Ultrasound imaging generates images based on the reflection and scattering of high-
frequency sound waves from internal body structures.

37
 • The piezoelectric transducer is the core component, responsible for both generating and
detecting ultrasound waves.
• Acoustic impedance differences between tissues are the fundamental source of contrast
in ultrasound images.
• Ultrasound images display relative reflectivity rather than absolute tissue properties,
making texture a crucial diagnostic feature.
• A key trade-off exists between frequency and penetration depth: higher frequencies
offer better resolution but shallower penetration, while lower frequencies penetrate
deeper but yield lower resolution.
• Modern ultrasound systems utilize electronic arrays and wave interference
principles to precisely focus and steer the ultrasound beam.
• Ultrasound is a non-ionizing and non-invasive modality, making it particularly safe for
applications like obstetrics and real-time monitoring during surgery.

Here are your detailed notes on Radionuclide Imaging (Nuclear Medicine), structured for clarity
and comprehensive understanding:

Radionuclide Imaging (Nuclear Medicine) -

Lecture Notes
1. Introduction to Radionuclide Imaging
• Radionuclide imaging, also known as nuclear medicine, represents the final category of
radiologic imaging systems.
• It encompasses three main imaging modalities:
o Planar Scintigraphy
o Single Photon Emission Computed Tomography (SPECT)
o Positron Emission Tomography (PET)

2. Physics of Nuclear Medicine

• Core Principle: This set of imaging modalities involves introducing a radioactive
substance (referred to as a tracer molecule or radiotracer) into the body.
• Functional Imaging: Unlike CT, MRI, or ultrasound, which primarily focus on
anatomical structures and different contrast mechanisms, radionuclide imaging is
a functional imaging modality.
o It's concerned with physiological imaging, where the tracer actively interacts
with the body's biology and physiology.
o The goal is to detect variations in concentrations of the radioactive tracer,
which can indicate an abnormality or disease.

38
 o Example: Radioactive isotopes of iodine are readily absorbed by the thyroid
gland. Their uptake can indicate how well the thyroid glands are functioning.

2.1. Key Differences from Other Imaging Modalities

• Internal Source: A major distinction is that the radiation source is inside the body.
o In contrast, other imaging systems (MRI with radio frequency, X-ray, ultrasound)
use external sources of energy deposition.
o The administered radiotracer accumulates within the body and emits radiation,
which is then detected externally.

2.2. What is Measured?

• The preferred choice of emitted radiation for imaging is gamma rays.

o Gamma rays are high-energy X-ray photons.
o The gamma radiation emitted is proportional to the concentration of the
tracer.
• Instead of absolute concentration, what is precisely measured is the activity
concentration.
o This is derived from the mean photon count of the gamma radiation.
o The activity concentration, in turn, indicates the actual concentration of the
radiotracer at a specific tissue location.
• Measurements are taken as a function of both position and time.

2.3. Radioactive Decay: The Source of Radiation

• Definition: Radioactive decay is the rearrangement of inherently unstable atomic nuclei

to a lower, more stable energy state.
• Mass Defect & Binding Energy: This transition corresponds to a greater mass defect,
which translates to a higher binding energy per nucleon for the more stable "daughter
atom".
o Mass Defect: The difference between the sum of individual masses of protons
and neutrons and the actual measured mass of the nucleus.
o Parent Atom: The initial unstable atom that undergoes decay.
o Daughter Atom: The more stable atom formed after decay.
• Disintegration: A radioactive atom undergoing radioactive decay is referred to
as disintegration.
• Energy Release: Radioactive decay is typically accompanied by the release of energy,
often in the form of gamma radiation, which is then measured.

2.4. Types of Radioactive Decay Used in Nuclear Medicine

Two common types of radioactive decay are utilized in radionuclide imaging:

1. Positron Emission (Beta Plus Decay)

o Mechanism: A proton transmutes into a neutron and a positron within the
atom.

39
 § The mass number (A) does not change, but the atomic number (Z)
reduces because a proton is lost.
o Positron: The antiparticle of an electron, having the same mass but an opposite
charge.
o Annihilation Coincidence Detection (ACD): When an emitted positron
combines with a nearby free electron (usually after traveling a few millimeters in
tissue), they annihilate each other.
§ This annihilation event results in the release of two 511 keV gamma
photons.
§ Crucially, these two photons travel in exactly opposite directions (180
degrees apart) from the point of emission.
§ This simultaneous emission of photons in opposite directions is
fundamental to Positron Emission Tomography (PET).
o Radiotracer Example: Fluorine disintegrating into oxygen is an example of
positron emission.
§ A widely used PET radiotracer is Fluorodeoxyglucose (FDG).
§ In FDG, a radioactive fluorine isotope replaces an oxygen atom in
glucose.
§ Clinical Application: Tumors often have a higher metabolic rate than
normal tissue, leading them to uptake more glucose (and thus more
FDG). By measuring the accumulation of radioactivity, one can detect and
localize tumors.
2. Gamma Decay (Isomeric Transition)
o Mechanism: The radionuclide is inherently unstable but does not undergo
transmutation (change in atomic/mass number). Instead, it simply transitions to
a lower energy state by emitting gamma rays.
o Energy Range: In diagnostic imaging, these gamma rays typically have energies
between 100 and 600 keV.
o Application: This type of decay is primarily used in Single Photon Emission
Computed Tomography (SPECT).
o Radiotracer Example: Technetium (Tc) is a widely used isotope for SPECT
imaging.
§ It decays by emitting gamma ray photons with an energy of about 140
keV.
§ Clinical Applications: Technetium is commonly used for bone
imaging and cardiac imaging. For instance, Technetium MDP is
absorbed by bones and provides information about bone metabolism.

3. Data Acquisition and Imaging Systems

3.1. General Principles of Data Acquisition

• Small Quantity Injection: Only a very small quantity of radiotracer is injected into
the patient to minimize potential damage from radioactivity.

40
 • Accumulation: The tracer is hypothesized to accumulate preferentially in regions
corresponding to pathology.
• Detection: The emitted gamma rays (or high-energy photons in the case of PET) exiting
the body are detected and counted.

3.2. The Anger Camera (Gamma Camera)

• Inventor: Named after the person who invented it, Anger.

• Purpose: The Anger camera (or gamma camera) is the primary device used for
detecting the emitted gamma rays in planar scintigraphy and SPECT.
• Components:
1. Collimator: Located at the front of the camera, typically made of lead, with
parallel holes.
§ Function: Ensures that gamma rays hitting the scintillator come from
a specific direction, effectively localizing the measurement to a particular
region of tissue. It blocks scattered radiation and radiation from other
angles.
§ Helps to disregard Compton scattering (scattered photons tend to have
slightly lower energy and can be filtered out).
2. Scintillator Crystal: A very large crystal positioned behind the collimator.
§ Function: Traps incident high-energy gamma rays and converts their
energy into visible light.
§ Mechanism: This conversion occurs through photoelectric absorption:
the scintillator absorbs the gamma rays, producing electrons that travel
through the crystal and release energy as light photons through
interactions with other electrons.
3. Photomultiplier Tubes (PMTs) or Photo Diodes: An array of detectors
connected to the scintillator crystal.
§ Function: Measure the visible light output from the scintillator.
• Imaging Chain Components:
o Positioning Logic: Determines precisely where in the scintillator a particular
gamma ray hit.
o Pulse Height Analyser: Used to reject scattered radiation (e.g., Compton
scattered photons, which have lower energy than primary gamma rays).
o Together, these components help estimate the x,y position of each detected event.

3.3. Attenuation Challenges in Radionuclide Imaging

• Unlike X-rays, where attenuation is primarily linear, radionuclide imaging faces two
types of attenuation:
1. 1/r² Attenuation: Since radiation is emitted in all directions from the source
within the body, the flux of radiation falls off with the square of the distance from
the source.
2. Exponential Attenuation (e⁻ᶠᵘ*depth): As with X-rays, emitted photons are also
attenuated as they travel through tissue.
• Depth Dependence: A critical difference from X-ray imaging (e.g., CT) is that
projections acquired 180 degrees apart are not the same.

41
 o If a radioactive compound is deep within the body, measurements taken from one
side of the patient will be significantly different from those taken from the
opposite side.
o This is because both the depth of the source within the tissue and the distance
to the camera change as the camera rotates, affecting both types of attenuation.
These depth and distance dependencies must be accounted for in reconstruction.

4. Specific Radionuclide Imaging Modalities

4.1. Planar Scintigraphy

• Description: The simplest form of nuclear medicine imaging, very similar to projection
X-ray.
• Data Acquisition: The detector (Anger camera) is translated across the patient's body,
recording the amount of gamma ray photons at each point.
o This provides a projection of the underlying activity onto the camera.
o Each pixel in the resulting image corresponds to a projection along a line.
• Reconstruction: Does not require reconstruction.
• Image Interpretation: The gray value measured is proportional to the attenuated activity
along that line.
o Needs correction for depth and distance from the camera due to attenuation
effects.
• Functional Information: Despite similarities to X-ray projections, planar scintigraphy
measures activity concentration, thus providing functional information.
o Example: Technetium MDP imaging provides insights into bone metabolism.
• Depth Resolution: Depth resolution is difficult to achieve with planar scintigraphy.
o The signal recorded is restricted to a small cylinder of tissue, but differentiating
depth within that cylinder is challenging.

4.2. Single Photon Emission Computed Tomography (SPECT)

• Acronym: Single Photon Emission Computed Tomography.

• Principle: Similar to CT in its acquisition method, SPECT requires tomographic
reconstruction.
• Data Acquisition: An Anger camera capable of rotation is used.
o The camera rotates around the patient's body (along the Z-axis, with the patient
lying down).
o It acquires a series of projection images at various angles, similar to how CT
acquires X-ray projections.
o These projections record emitted gamma radiation along lines projecting onto the
camera face.
• Challenges: As discussed, projections acquired 180 degrees apart are not the
same due to depth-dependent attenuation and variations in distance to the camera. These
factors must be modeled for reconstruction.
• Tracers: Typically uses gamma ray emitters (e.g., Technetium).
• Output: After reconstruction, SPECT produces cross-sections of radioactivity
concentrations.

42
 • Clinical Applications: Most SPECT imaging is used for whole body bone imaging and
cardiac imaging.

(Insert diagram of SPECT imaging geometry with a rotating Anger camera around a patient here)

4.3. Positron Emission Tomography (PET)

• Principle: Utilizes positron emission emitters and relies on annihilation coincidence

detection (ACD).
• Data Acquisition:
o When an emitted positron from the radiotracer annihilates with an electron, two
511 keV gamma photons are emitted in opposite directions.
o Detection: Two detectors (or Anger cameras) positioned 180 degrees apart are
used to simultaneously measure these two photons.
o Coincidence Window: A detection is considered "simultaneous" if both photons
arrive within a very small time window (typically 2 to 20 nanoseconds).
o This coincidence detection indicates that the annihilation event occurred along
the straight line joining the two detecting pixels.
• Reconstruction: Similar to SPECT and X-ray CT, the detectors are rotated to collect
annihilation data at different angles.
o This allows for tomographic reconstruction to generate cross-sections of the
radiotracer activity concentration.
o The problem can be cast as a form of Radon transform for reconstruction.
• Attenuation: Similar to X-ray photons, the 511 keV photons are also attenuated
exponentially as they travel through tissue, and this must be accounted for.

(Insert diagram of PET imaging geometry with two opposing detectors detecting coincidence
events here)

5. Image Interpretation and Hybrid Systems

• Raw Radionuclide Images: Images from radionuclide modalities (Planar Scintigraphy,
SPECT, PET) primarily show areas of radiotracer accumulation, appearing as "blobs" or
hot spots.
o They provide functional information (e.g., metabolic activity, blood flow) but
inherently lack detailed anatomical information.
• Superimposition/Fusion: To enhance interpretation, radionuclide images are
often superimposed or fused with anatomical images obtained from other modalities.
o This is why hybrid scanners like PET-CT and increasingly PET-MRI are
common.
o The PET/SPECT scanner provides the functional imaging data (after
radiotracer injection), while the CT/MRI scanner provides
the structural/anatomical information.
o Experienced nuclear medicine radiologists interpret these fused images to
correlate functional abnormalities with anatomical locations.

43
6. Glossary of Key Terms
• Radionuclide Imaging / Nuclear Medicine: Medical imaging techniques that use
radioactive substances (tracers) introduced into the body to visualize physiological
processes.
• Functional Imaging: Imaging modality that focuses on depicting physiological
processes and variations in concentration/activity, rather than just anatomical structure.
• Tracer / Radiotracer: A molecule carrying an unstable isotope (radionuclide)
administered to a patient, designed to interact with specific biological or physiological
processes.
• Gamma Rays: High-energy electromagnetic radiation (photons) emitted during
radioactive decay or annihilation, preferred for imaging in nuclear medicine.
• Scintillator Crystal: A material that absorbs high-energy photons (like gamma rays) and
re-emits their energy as visible light.
• Photomultiplier Tube (PMT) / Photo Diode: Devices that detect and amplify the visible
light emitted by a scintillator, converting it into an electrical signal.
• Mass Defect: The difference between the sum of the masses of an atom's individual
protons and neutrons and the actual measured mass of its nucleus.
• Binding Energy: The energy required to disassemble an atom's nucleus into its
constituent protons and neutrons; related to mass defect.
• Parent Atom: An unstable atom that undergoes radioactive decay.
• Daughter Atom: The more stable atom formed as a result of radioactive decay.
• Radioactive Decay: The process by which an unstable atomic nucleus loses energy by
emitting radiation, transforming into a more stable state.
• Disintegration: The event of a radioactive atom undergoing radioactive decay.
• Positron Emission (Beta Plus Decay): A type of radioactive decay where a proton in the
nucleus is converted into a neutron, a positron, and a neutrino.
• Annihilation Coincidence Detection (ACD): The detection of two gamma photons
simultaneously emitted at 180 degrees from each other, resulting from a positron-electron
annihilation event, primarily used in PET.
• Fluorodeoxyglucose (FDG): A common PET radiotracer, where radioactive fluorine
replaces oxygen in glucose; used to assess metabolic activity, particularly in cancer
detection.
• Technetium (Tc): A widely used gamma-emitting isotope in SPECT imaging for bone,
cardiac, and other studies.
• Anger Camera / Gamma Camera: The primary detection device used in planar
scintigraphy and SPECT, consisting of a collimator, scintillator crystal, and PMTs.
• Collimator: A lead-based component in front of the scintillator that uses parallel holes to
ensure detected gamma rays originate from a specific direction, improving spatial
localization.
• Pulse Height Analyser: An electronic circuit that filters out detected events based on
their energy, used to reject scattered radiation (e.g., Compton scatter).
• Activity Concentration: The measured quantity in radionuclide imaging, representing
the amount of radioactivity per unit volume of tissue, which is proportional to the tracer
concentration.

44
 • Planar Scintigraphy: A simple nuclear medicine technique that produces a 2D
projection image of radiotracer distribution without tomographic reconstruction.
• Single Photon Emission Computed Tomography (SPECT): A tomographic nuclear
medicine technique that uses a rotating gamma camera to acquire multiple 2D projections
of gamma-emitting tracers, which are then reconstructed into 3D cross-sectional images.
• Positron Emission Tomography (PET): A tomographic nuclear medicine technique that
detects pairs of gamma rays emitted indirectly by positron-emitting tracers (via
annihilation) to create 3D images of functional processes.

7. Key Takeaways and Connections

• Functional vs. Anatomical: Radionuclide imaging fundamentally differs from
modalities like CT and MRI by providing functional/physiological information rather
than primarily anatomical structure.
• Internal Source: The unique characteristic of radionuclide imaging is the internal
administration of the radiation source (radiotracer).
• Radioactive Decay Drives Imaging: The specific type of radioactive decay (positron
emission for PET, gamma decay for SPECT) determines the imaging modality and the
properties of the emitted photons.
• Attenuation is Complex: Radionuclide imaging faces unique challenges with depth-
dependent and distance-dependent attenuation (1/r² and exponential), requiring
sophisticated reconstruction algorithms, especially for SPECT and PET.
• Hybrid Systems are Standard: Due to their lack of inherent anatomical detail,
radionuclide images are often fused with CT or MRI scans to provide anatomical
context for the functional information. This is a crucial concept in medical image
analysis, as it often involves multi-modal image registration and fusion techniques.

(Insert concept map: Start with "Radionuclide Imaging" branching into "Physics (Tracer, Decay,
Attenuation)", "Modalities (Planar, SPECT, PET)", and "Imaging System (Anger Camera)".
Then elaborate on each branch, showing relationships like "Tracer -> Positron Emission -> PET"
or "Tracer -> Gamma Decay -> SPECT", and "Anger Camera -> Collimator, Scintillator,
PMTs").

8. Questions for Deeper Understanding & Exam Prep

• Thought-Provoking Questions:
o Given the internal source, how might motion artifacts in radionuclide imaging
differ from those in X-ray CT?
o Discuss the ethical considerations and safety protocols for handling and
administering radiotracers, particularly concerning the "very small quantity"
mentioned.

45
 o How might advancements in detector technology (e.g., solid-state detectors)
impact the design and capabilities of Anger cameras and PET scanners?
o Beyond the examples provided, what other physiological processes could
potentially be targeted with novel radiotracers for medical diagnosis?
• Further Research/Clarification:
o Investigate the specific mathematical models used for attenuation correction in
SPECT and PET reconstruction algorithms.
o Explore how image reconstruction algorithms (e.g., filtered back-projection,
iterative reconstruction) are adapted for SPECT and PET, considering the unique
attenuation challenges.
o Research the differences in spatial resolution and sensitivity between planar
scintigraphy, SPECT, and PET.
• Potential Exam/Interview Questions:
o "Compare and contrast radionuclide imaging with X-ray CT, focusing on their
fundamental physics, what they measure, and their clinical applications."
o "Explain the principle of Positron Emission Tomography (PET), detailing the role
of positron emission, annihilation, and coincidence detection."
o "Describe the function of a collimator in an Anger camera and explain why it is
essential for single-photon imaging modalities like SPECT."
o "Discuss the two main types of attenuation that affect radionuclide imaging and
explain why projections acquired 180 degrees apart are not the same in SPECT."
o "What is the significance of 'functional imaging' in nuclear medicine, and how
does it complement anatomical imaging modalities like CT or MRI?"

46