Utrasound Physics

2 – Instrumentation, Doppler Physics and

Safety
Keith Willson
Clinical Engineering
Royal Brompton and Harefield
NHS Trust
 Signal Processing and Machine Controls

• Imaging modes
• Pre and Post Processing
• Grey Scale, Dynamic Range, Logarithmic Compression
• Transmitted Power and Receiver Gain
• Time Gain Compensation (TGC)
• Reject
• Spatial and Temporal Smoothing
Signal from one scan line
Maximising transmission
matching layer

thin layer between the piezoelectric elements and the skin

“accoustic matching”

reduces reflection → less attenuation and more energy transmitted

matching layer + gel

Additional intermediate “accoustic matching”

reduces reflection → less attenuation and more energy transmitted

At the transducer – one scan line

noise

mV
mV

Excitation pulse

10-100V

time ~ms
This has to be turned into an image line and displayed
with a dynamic range of around 30 dB – How?

Clipping
Depth gain compensation
Dynamic range compression
Reject
Tine gain (or Depth gain) compensation
Changing the gain of the receiver along each line with time compensates for losses
in signal strength as the ultrasound comes from greater depths.

reflection

scatter

gain

time
Non-linear
(logarithmic) amplification

out
“Reject” removes
Small echoes amplified more
very small signals
- Scatter and weak echoes brought entirely – these are
up to same order as strong just noise
reflections; very strong reflections threshold
reduced in
Demodulation
The high frequency oscillating pulses
received are turned into a more slowly-
varying “envelope” which can be used
to place an image on the screen.

The image uses a “grey scale” where

the brightness of a point on the screen
is related to the amplitude of an echo

However the relationship between the

echo amplitude and the brightness can
be changed by the user, to change the
brightness and contrast of the image to
highlight desired areas of the grey
scale.
M Mode image – the image consists of a single line only, swept with time
to show how the positions of the reflectors change
B-Mode Image This signal forms one line of a 2-D image.
The scan line line is placed in image memory with coordinates given by
the scan format
 Scanner Architecture
Video Display
The ultrasonogra ph 11

Beam Forming

Control
Processors

Post-processing
Pre-processing

A-D Converter Image Memory

Signal Processing
Swept gain
(TGC)
100dB

0dB

At Compression Post Display

transducer processing

Threshold Compression Gamma

Electronic Focussing
Transmit focusing
• Cannot continually change transmit pulse
once sent – not easy to modify
• Can however optimise focal zone of transmit pulse
for a given depth
• Best to make separate images, each with different
focal zone placement, and generate montage of
best bits from each
Pre processing and post processing
Pre-processing takes place before the echos are stored in
memory and cannot be undone.

Post-processing occurs after storage and can be varied at will

Image Formation showing compression
Harmonic Imaging
When a high amplitude ultrasound disturbance passes through an elastic
medium it travels faster during the higher density compression phase than the
lower density rarefaction phase causing harmonic distortions.

Progressively stronger harmonic component with distance travelled.

PRO: reduction in artifacts, improved signal-to-noise ratio and slight

improvement in lateral resolution.

CON: reduced axial resolution due to longer initial pulse length

 RESOLUTION
IMAGE MEMORY-where storage of digitized information contained in the pulse
waveforms occurs
each part of the image memory called a pixel (picture element)
must have sufficient bits (binary digits) as possible to enable various shades of grey to
be visualised
Must have sufficient size to store all pixels for high resolution images
Must have sufficient capacity (number of images that can be stored), speed (time
required to write/record and read/retrieve images)

NB these considerations also apply to long-term storage systems outside the scanner..

[email protected]
Doppler ultrasound and blood/tissue
velocity estimation
The origins and processing of the Doppler
ultrasound signal and how it is used to
provide velocity estimation
Doppler Modes

• Continuous Wave Doppler (CW)

• Pulsed Wave Spectral Doppler (PW)
• Colour Doppler
• (TDI)
• Power Doppler
Doppler Effect
• Perceived shift in the frequency emitted by a source due to relative
motion between the source and an observer
• Caused by changing distance between source and observer changing
time of travel of wave between them.
i.e. changing path length causes successive wave fronts to arrive at
the observer sooner than in stationary case if source is moving
towards observer (or later if moving away).
 The Doppler Effect
When ultrasound interacts with a moving object (i.e. red blood cells) the
reflected frequency changes. If the cells are traveling towards the
transducer the ultrasound wave is ”squashed” ↓λ and ↑f giving a positive
Doppler shift. If RBC’s are traveling away the wave is “stretched” →↑λ and
↓f
The received Doppler shift is the velocity component towards the
observer – if the angle is unknown the frequency shift cannot be
corrected to represent the actual velocity

V cos θ
TDCR
θ

V
When the angle is zero the cosine tends to 1. This gives the maximum
received Doppler shift. In the heart we often measure along the direction of
blood flow e.g. in four chamber view. Errors in angle measurement affect the
cosine much more as the angle becomes larger, so it is good to keep the angle
as small as practicable. At around 50 degrees the error in cosine is equal to
the error in the angle.
 f
The Doppler Shift c
c Scattered

Incident
f+ Df

q v q v

v
Df = 2 f cosq
The received Doppler shift Δf is proportional to the ratio of c
the velocity v of the scatterer to the velocity c of
ultrasound.
There is a factor of two because the change in position of
the scatterer results in a change in distance in both
transmitted and received paths.
Magnitude of the Doppler shift - example
Say vblood = 0.5 ms-1
cblood = 1.5 x 103 ms-1
θ = 0; Cosθ = 1
f = 3 x 106 Hz

Then Df = 2 f Cosθ v/c

= 2 x 3 x 106 x 1 x 0.5 / 1.5 x 103
= 2 x 103 Hz
Continuous wave
Oscillator

f
f
Demodulator
f +Df
f+Df

2f+Df, Df

Low pass
filter and
amplifier

Df
Continuous wave Doppler

• No distance (depth) discrimination

• Doppler shift produced by all scatterers anywhere within the
ultrasound “beam”
• So, prone to interference from unwanted vessels and moving
structures
• However no “aliasing”
Pulsed Doppler
Pulsed Doppler
• Send out short samples (bursts) of a continuous wave signal at regular
intervals
• Make the receiver sensitive only to echoes arriving between certain
times after transmission i.e. from within a certain range of distances
from the transducer
• Add ‘Sample & Hold’ to capture & maintain output between pulses
• Each ultrasound pulse is a sample of how much the wave has shifted
i.e effectively how much the frequency has changed.
 The spectrogram
The Fast Fourier Transform is used to mathematically process the Doppler signal to extract the amplitudes of all frequencies
present in a short time segment.

Frequency shift towards probe

Time bins, containing short

term FFTs

time

Spectral
Frequency shift away from probe
amplitude
Frequency

Amplitude Time
Pulsed Doppler Sampling

1 0.6 0.3

-1 -0.6 -0.3
Aliasing Occurs when sampling rate is too low compared with
frequencies present in the Doppler shift signal
The Nyquist Limit (Aliasing)
The maximum Doppler shift (Δfmax) able to be displayed without
aliasing.
Determined by the sampling rate (PRF).

Nyquist Limit: Δfmax = PRF

2

Recall that: PRFmax =

c
Δfmax = c
2D
2x2D
High PRF
Transducer sends out an additional
pulse before the original pulse has
returned.
In effect it doubles the PRF and
therefore doubles the Nyquist
limit.
The disadvantage is that the exact
origin of the Doppler shift is not
known.
Potential for range ambiguity artifact
(“depth confusion”)
Signal amplitude compared with soft tissue
interfaces
Signal levels about 40db below those from soft tissue
interfaces.
So “Doppler” signal requires much more amplification
than “image” signal – so can pick up unintended
moving tissue in sample volume
So we use a high pass filter to reject these high
amplitude low velocity echos from the Doppler
 Fluid Dynamics

• Velocity at a Stenosis
• Volume Flow
• Flow Profiles and their associated mean velocities
• Spectral Doppler representation of Flow Profiles
• Transmitted Power and Receiver Gain
• Time Gain Compensation (TGC)
• Reject
• Spatial and Temporal Smoothing
Velocity at stenoses
Velocity at stenoses
V1 V2 V1

A1 A2 A1
d2
d1

A1 V1 = A2 V2 (conservation)
Hence V2 /V1 = A1/A2 = d12/d22
Velocity at stenoses
(in vascular examination e.g. Carotid)
• Diameter ↓ 30% ~ area ↓ 50%
• “haemodynamically significant”

• Diameter ↓ 70% ~ area ↓ 90%

• “clinically significant”

Flow

Velocity
reduction %
0 100
Simplified Bernoulli equation

Used to estimate pressure drop across stenosed valve

DP (peak pressure gradient in mmHg )= V2 (V measured in m.s-1 )

Velocity Profiles

Streamline - Parabolic
V mean = 0.5 V peak

Streamline - Plug
V mean = V peak

V mean = ???
Turbulence and Reynold’s Number
Whether or not flow is streamline (laminar) or turbulent depends on
factors such as the dimensions of the flow containing vessel, the
viscosity and density of the fluid, and its velocity.

The Reynolds number is a dimensionless quantity that expresses the

likelihood of turbulence.

For flow in a pipe of diameter D, experimental observations show that

laminar flow occurs when ReD < 2300 and turbulent flow occurs when
ReD > 2900.

Turbulence can occur as fluid emerges from a jet into a wider vessel.
 Notional spectrograms for various flow types
f f

PLUG
FLOW
t t
Peak
f f
Velocity
PARA-
BOLIC
FLOW
t t
f f
Mean
Velocity
BLUNT
FLOW t t
CONSTANT TIME-VARYING
Mean flow velocity

volume flow = V A
Colour flow Doppler
Effectively a multi-sampled PW from multiple sites
(100-400) superimposed on a 2D image→ low
FR!!!
Each area sampled minimum of 3 times to calculate
a Doppler frequency shift and estimate mean
velocity.

Frame rate determined by:

• Sector size ↓width/depth↑FR

• Packet size: The packet size is the number of

pulses transmitted per line. ↓packet size ↑FR

Same limitations as PW Doppler (i.e. Nyquist limit),

however as it is detecting mean velocity the
Nyquist limit is lower → aliases earlier
Colour Doppler
• Superimpose colour-coded velocity information on to conventional
grey-scale image
• Uses same transducer for imaging and Doppler shift information
• Produces fast mean frequency estimate in a group of pixels; can also
give an index of turbulence
• For each image pulse, require 3-15 pulses to obtain Doppler
information
Colour and TDI
Filters are used to discriminate between myocardium and tissue in colour imaging:

Blood is a low amplitude Myocardium is a high

scatterer (recall Rayleigh amplitude spectral
scattering) with relatively reflector with relatively
quick velocities. slow velocities.
Tissue Doppler (TDI)
Colour M mode
Physical and biological effects
of ultrasound

How ultrasound interacts with tissue

and the potential for biological
effects
 Safety

• Potential hazardous biological effects pul

– Heating and cavitation effects
• Measurement of beam intensity
– (Spatial Peak Temporal Average, [SPTA])
• Practical precautions:
– power levels,
– Peak intensities in different modes
– On-screen indices.
 Effects in tissue
Important to distinguish between
• physical effects
• biological effects
• clinical risk
• clinical outcomes
• relative importance of potential hazards in diagnosis and screening

Power and intensity

• Total power emitted by TDCR may be relatively low i.e. mW
• Local intensities, however may be relatively high i.e hundreds of mW
Tissue heating
Cooling
(unperfused tissues in vitro)
By conduction into
surrounding tissue
Tissue Intensity Time for 2 deg C
By local blood
mW cm-2 temp.rise perfusion
Liver 100 3.7 mins By thermoregulation

1000 22 secs
Bone 100 2.9 secs In-vivo
1000 0.3 secs Risk in thermally unregulated
tissues (e.g. foetus)
Eye lens 100 33 secs Risk in unperfused tissues (e.g.
lens of eye)
1000 3.3 secs
Intensity in a beam
ISPTA Spatial peak, temporal average
• Lowest for linear and convex array B-mode images.
• Higher in scan modes where beam is stationary
• Pulsed Doppler uses longer pulses and stationary beam, can
produce highest values
• Continuous wave and colour Doppler also higher than B
mode
• High at foci
• In sector scans can be high close to the transducer, where
all beams in the scan pass through during image formation.
Pulse pressure
• Pulse pressure – peak rarefaction pressure (P-)

• Cavitation – the catastrophic collapse of oscillating bubbles, or tissue

voids
• Cavitation can be destructive: ships’ propellers, aircraft fuel lines
• Causes DNA damage
• Possible effects at air interfaces in e.g. lung – impedance mismatch causes
large negative pressures
 On screen indicators
• Thermal index (TI)– proportion of possible 1 deg C temperature rise
under a range of scanning conditions. There are different TIs for
different tissues
• Mechanical index (MI) – frequency corrected proportion of
theoretical threshold of cavitation risk

Output Power
• Keep output power as low as practicable – use increased gain instead
wherever possible
 Goodbye!