RAD DNB

Syllabus · Ultrasound and Doppler physics

Ultrasound elastography

DNB Radiodiagnosis notes · a free sample topic by Dr. Anoop Phatnani, DNB Radiodiagnosis

Answer in a page

The core is always the same three blocks (principle, techniques, applications); what changes is which organ is asked in depth.

Question shapeWrite
"Principle, techniques and clinical applications" (3+3+4, the commonest)Definition, Young's modulus, the two diagrams; classification table of techniques; applications by organ with one line each
"Principle" alone (5 marks)Definition, stress and strain, the two ways of measuring stiffness, E = 3ρc², units, the diagram
"Applications in diffuse liver disease" or "role in liver disease" (3-4 marks)Technique of measurement, fibrosis staging, cut-offs with their source, portal hypertension, confounders
"Shear wave elastography" short notePrinciple, three types, units, applications, limitations
"Clinical applications of compression elastography in detail"Organ by organ for strain imaging: breast, thyroid, cervix, prostate, nodes, musculoskeletal, endoscopic
"Role in breast pathologies" (10 marks)Technique, Tsukuba score, strain ratio, E/B ratio, shear wave thresholds, use with BI-RADS, pitfalls
"Role of thyroid elastography" (5 marks)Scores, strain ratio, shear wave values, where it helps, limitations
"Sonoelastography in obstetrics and gynaecology" (5 marks)Cervix, placenta, fibroid versus adenomyosis, endometrium, ovary, safety
"Ultrasound versus MR elastography for liver fibrosis"Comparison table

Always state that cut-offs differ between vendors and techniques, and name the guideline you are quoting.

Principle

Definition. Elastography is imaging of the mechanical stiffness of tissue. It is the imaging equivalent of palpation. Disease changes stiffness: fibrosis, most cancers and inflammation make tissue harder. When ultrasound is used to do it, it is called ultrasound elastography or sonoelastography.

Why B-mode cannot do this. B-mode depends on the bulk modulus, which decides the speed of the ordinary (longitudinal) sound wave. It varies very little across soft tissues, which is why the scanner can assume 1540 m/s everywhere. The shear modulus, which decides how easily tissue deforms, varies many-fold between tissues and between normal and diseased tissue. Elastography images that property.

The physics

  • Stress (σ): force per unit area applied to the tissue. Unit: pascal (Pa = N/m²).
  • Strain (ε): the deformation produced, as change in length divided by original length. No unit.
  • Young's modulus (E) = stress / strain. A stiff tissue needs a large stress to produce a small strain, so E is high. Unit: kilopascal (kPa).
  • Shear modulus (G): resistance to a shearing (sideways) force. In soft tissue, which is almost incompressible, E = 3G.
  • A shear wave is a transverse wave: particles move at right angles to the direction of travel. It is slow (about 1-10 m/s in soft tissue, against 1540 m/s for the ultrasound beam), it is attenuated quickly, and it does not travel through simple (non-viscous) fluid.
  • Shear wave speed (c) depends on stiffness: G = ρc², so E = 3ρc², where ρ is tissue density (taken as 1000 kg/m³). With c in m/s, E in kPa is simply 3 × c². Example: 1.5 m/s is about 6.8 kPa.

Every elastography method has three steps

  1. Excite: apply a force to deform the tissue (hand pressure, physiological pulsation, a mechanical piston, or a focused ultrasound push pulse).
  2. Track: follow the tissue response with ordinary ultrasound pulses (speckle tracking by correlation of radiofrequency echoes before and after).
  3. Display: show the result as a colour map or greyscale map (elastogram) or as a number.

Two things can be measured, and this gives the two families

Strain imagingShear wave imaging
What is measuredHow much the tissue deforms along the beamHow fast a shear wave travels across the beam
StressUnknown, so E cannot be calculatedNot needed; speed alone gives E
ResultQualitative or semi-quantitative (colour, score, ratio)Quantitative (m/s or kPa)

Acoustic radiation force impulse (ARFI). A focused ultrasound pulse, much longer than an imaging pulse (a fraction of a millisecond), transfers momentum to tissue as it is absorbed and pushes the tissue at the focus by a few micrometres. The scanner can either measure that displacement on the beam axis (a form of strain imaging) or track the shear waves that spread sideways from the push (shear wave elastography). The same transducer pushes and tracks.

Assumptions behind the number: tissue is taken to be homogeneous, isotropic, incompressible and purely elastic. Real tissue is viscoelastic, so shear wave speed changes with the frequency of the wave. This is one reason why different machines give different values in the same patient.

Draw this

Four panels to reproduce: A, strain elastography, where a soft inclusion flattens under compression and a hard one does not; B, shear wave elastography, with the push pulse, shear waves spreading sideways and the tracking pulses; C, classification of techniques; D, the SRU 2020 rule of four for liver stiffness by ARFI methods in kPa and m/s.
Four panels to reproduce: A, strain elastography, where a soft inclusion flattens under compression and a hard one does not; B, shear wave elastography, with the push pulse, shear waves spreading sideways and the tracking pulses; C, classification of techniques; D, the SRU 2020 rule of four for liver stiffness by ARFI methods in kPa and m/s.

For a 3-mark "principle" part, panels A and B with their formulae are enough.

Techniques

TechniqueExcitationMeasuresOutputImage guidanceExample
Strain (compression) elastographyManual probe pressure, or cardiac, vascular or respiratory motionAxial strainColour or grey map, score, strain ratioYes, real timeBreast, thyroid, cervix
ARFI strain imagingUltrasound push pulseDisplacement on the beam axisQualitative mapYesSiemens Virtual Touch Imaging
Transient elastography (TE)Mechanical piston on the skin, 50 HzShear wave speed along one lineOne number, kPaNo (A-mode and M-mode only)FibroScan
Point shear wave elastography (pSWE)ARFI push at one focusMean shear wave speed in one small fixed boxOne number, m/s or kPaYes, B-modeSiemens VTQ, Philips ElastPQ
2D shear wave elastography (2D-SWE)ARFI pushes at many pointsShear wave speed pixel by pixelColour map plus numbers, m/s or kPaYes, real time or near real timeSuperSonic Imagine (Aixplorer), GE, Canon, Philips ElastQ

Transient elastography, pSWE and 2D-SWE are all shear wave methods. Many texts use "shear wave elastography" for the two ARFI-based methods only and keep "transient elastography" as a separate name; some use the terms interchangeably. Write the classification as above and you are covered either way.

Strain (compression) elastography

Also called static, quasi-static or real-time elastography.

How it is done

  • A high-frequency linear probe is held perpendicular to the skin. The lesion is placed in the centre of a colour box that also includes plenty of normal surrounding tissue (the lesion should fill no more than about a quarter to a half of the box).
  • The operator applies light, repeated compression and release, or simply holds the probe still and lets breathing and pulsation move the tissue. Most machines show a bar or a graph that confirms the compression is adequate.
  • The machine compares radiofrequency echoes before and after compression, calculates the displacement at each depth, and from its gradient the strain.
  • The elastogram is shown beside or over the B-mode image. On most systems blue is hard, green intermediate and red soft; some vendors use the opposite, so always read the colour bar. On a greyscale map, dark is hard.

How it is read

MethodWhat it is
Colour pattern scoreVisual score of the colours in and around the lesion: Tsukuba (breast), Rago or Asteria (thyroid)
Strain ratioStrain of reference tissue (fat in the breast, normal parenchyma in the thyroid) divided by strain of the lesion, both at about the same depth. A higher ratio means a stiffer lesion
E/B ratio (size ratio)Lesion length on the elastogram divided by its length on B-mode. Cancers look larger on the elastogram because of the stiff desmoplastic reaction around them
Fat-to-lesion ratio, strain histogram, hard-area ratioVendor-specific variants of the above

Strengths: available on most machines, cheap, real time, high spatial resolution, no extra acoustic output beyond B-mode.

Weaknesses

  • Relative only: it shows contrast between a lesion and its surroundings, so it is poor for diffuse disease where everything is equally stiff.
  • Operator dependent: too much pre-compression makes all tissue look stiff (stiffness rises non-linearly with pressure).
  • Stress falls with depth, so it works to about 3-4 cm only.
  • The lesion can slip sideways out of the plane during compression.
  • Ratios and scores are not comparable across vendors.
Strain elastography of a breast lesion (Hitachi), B-mode on the left and elastogram on the right. The lesion is oval, circumscribed and wider than tall, and on the elastogram it is mostly green with scattered blue, much like the surrounding tissue: a soft pattern. The colour bar runs from red (soft) to blue (hard), and the trace at the bottom shows the compression cycles.
Strain elastography of a breast lesion (Hitachi), B-mode on the left and elastogram on the right. The lesion is oval, circumscribed and wider than tall, and on the elastogram it is mostly green with scattered blue, much like the surrounding tissue: a soft pattern. The colour bar runs from red (soft) to blue (hard), and the trace at the bottom shows the compression cycles.

Shear wave elastography

Principle. A force is applied briefly and the shear waves it produces are tracked. Their speed is measured and reported in m/s, or converted to Young's modulus in kPa with E = 3ρc². Because the result does not depend on how hard the operator presses, it is quantitative and more reproducible than strain imaging.

Transient elastography (vibration-controlled transient elastography, FibroScan)

  • The first shear wave method, and the one with the most validation in liver disease. It is not an imaging ultrasound machine: there is no B-mode picture.
  • The probe carries a single-element transducer on a vibrating piston. The piston taps the skin at 50 Hz; the transducer tracks the resulting shear wave along its own axis.
  • M probe: 3.5 MHz, measures at 25-65 mm from the skin. XL probe: 2.5 MHz, 35-75 mm, for obese patients. An S probe exists for children.
  • It samples a cylinder of liver about 1 cm wide and 4 cm long, roughly 100 times the volume of a biopsy core.
  • The patient lies supine with the right arm raised; the probe is placed in a right intercostal space over the right lobe. Ten valid readings are taken and the median is reported in kPa (the machine reads up to 75 kPa).
  • Reliable when the interquartile range divided by the median (IQR/M) is 30% or less.
  • The same probe gives the controlled attenuation parameter (CAP) in dB/m (100-400), an estimate of steatosis.
  • It fails with ascites (the mechanical wave does not cross fluid), in obesity and with narrow intercostal spaces.

Point shear wave elastography

  • Done on a conventional scanner with a curved probe. The operator sees the B-mode image and places a small fixed box (about 1 cm) in liver free of vessels and ducts.
  • One ARFI push is delivered; shear wave speed is averaged over the box and shown as a single value.
  • Works through ascites, because the push is generated inside the liver.
  • Ten readings, median and IQR/M, as for transient elastography.

2D shear wave elastography

  • ARFI pushes are delivered at many points. In the SuperSonic method they are focused at successively deeper points faster than the shear wave travels, producing a cone-shaped ("Mach cone") wavefront; ultrafast plane-wave imaging at thousands of frames per second then tracks it.
  • A colour map of stiffness is laid over the B-mode image. One or more regions of interest are drawn in it, giving mean, maximum, minimum and standard deviation.
  • Many machines add a quality or propagation map (parallel, evenly spaced wavefront lines mean a good acquisition) and a stability index.
  • A 3D version exists for the breast.
2D shear wave elastography of the liver through a right intercostal window. Left: stiffness map (colour scale 0-70 kPa) with a circular region of interest placed below the capsule, away from vessels; the box is uniformly blue (soft). Right: the propagation map for the same box, with roughly parallel wavefront lines indicating a usable acquisition. Reading shown: 6.0 kPa.
2D shear wave elastography of the liver through a right intercostal window. Left: stiffness map (colour scale 0-70 kPa) with a circular region of interest placed below the capsule, away from vessels; the box is uniformly blue (soft). Right: the propagation map for the same box, with roughly parallel wavefront lines indicating a usable acquisition. Reading shown: 6.0 kPa.

Clinical significance in one line each: staging liver fibrosis without biopsy and predicting portal hypertension (the main use); adding specificity to breast and thyroid ultrasound; targeting prostate biopsy; assessing tendon, muscle, lymph node, kidney graft and spleen.

Limits: depth (ARFI methods are unreliable beyond about 6-7 cm), motion, obesity, vendor-specific values, and no signal in fluid.

Clinical applications

OrganMain useUsual technique
LiverStaging fibrosis, ruling compensated advanced chronic liver disease in or out, portal hypertensionTE, pSWE, 2D-SWE
SpleenPredicting clinically significant portal hypertension and varicesTE, pSWE, 2D-SWE
BreastBenign versus malignant; refining BI-RADS 3 and 4aStrain, 2D-SWE
ThyroidSelecting nodules for fine-needle aspiration; thyroiditisStrain, SWE
ProstateTargeting biopsy to stiff peripheral-zone foci (transrectal); about 35 kPa reported as a 2D-SWE cut-off. False positives: hyperplasia, calcification, prostatitisStrain, 2D-SWE
Lymph nodesMetastatic nodes are stiffer; choosing the node or the part of the node to sampleStrain, SWE
MusculoskeletalTendinopathy (a diseased tendon softens), muscle injury and spasticity, plantar fascia, soft tissue massesStrain, SWE
KidneyFibrosis in chronic kidney disease and in the transplant kidney (results inconsistent because the kidney is anisotropic)SWE
Pancreas and gutEndoscopic ultrasound elastography of pancreatic masses and nodes; fibrotic versus inflammatory stricture in Crohn diseaseStrain, SWE
Obstetrics and gynaecologyCervix, fibroid, adenomyosis, endometrium, placentaStrain, SWE
Testis, salivary glands, vesselsCharacterising focal lesions; plaque and thrombus ageStrain, SWE
InterventionChoosing the stiffest part of a lesion to biopsy; showing the ablation zone, which becomes stiffStrain, ARFI

Liver: diffuse liver disease

This is the best-validated use of elastography.

Why it works. Collagen deposited in fibrosis stiffens the liver. Stiffness rises with fibrosis stage, but not in a straight line: the steps are small between early stages and large between advanced fibrosis and cirrhosis. Elastography is therefore best at excluding or confirming advanced fibrosis and cirrhosis, and weaker at separating adjacent early stages.

Why it is wanted. Liver biopsy is invasive (pain, bleeding), samples about 1/50,000 of the liver, and suffers from sampling and observer variation. Fibrosis is graded on biopsy by METAVIR: F0 none, F1 portal fibrosis, F2 a few septa (significant fibrosis), F3 many septa (advanced fibrosis), F4 cirrhosis.

Technique for pSWE and 2D-SWE (SRU 2020 and EFSUMB 2017)

  • Fasting: at least 4 hours (SRU); EFSUMB accepts a minimum of 2 hours, with 10 minutes of rest first.
  • Supine or slight left lateral position, right arm raised above the head.
  • Right intercostal approach, right lobe (segments 7 and 8); the left lobe gives higher and more variable values.
  • Breath held in a neutral position: no deep inspiration and no Valsalva.
  • Probe perpendicular to the capsule. Box 1.5-2 cm below the capsule, about 4-5 cm from the skin, clear of vessels, ducts and rib shadow.
  • pSWE: 10 readings (at least 5 if IQR/M is within range). 2D-SWE: 5 readings when the machine has a quality map, 10 when it has none (SRU); EFSUMB accepts a minimum of 3. Report the median.
  • Reliability: IQR/M of 30% or less for kPa, 15% or less for m/s (SRU).
  • The report should state the machine and probe, the number of readings, median, IQR/M, and an interpretation that takes the cause of liver disease and the liver enzymes into account.
Result sheet from a liver shear wave study: ten readings with speed in m/s, stiffness in kPa and depth (about 3 cm). The median is 1.44 m/s or 6.1 kPa with an IQR of 0.6 kPa, so IQR/M is about 10%, well within the 30% limit.
Result sheet from a liver shear wave study: ten readings with speed in m/s, stiffness in kPa and depth (about 3 cm). The median is 1.44 m/s or 6.1 kPa with an IQR of 0.6 kPa, so IQR/M is about 10%, well within the 30% limit.

Cut-offs. Quote the source with every figure.

Whose figuresTechniqueThresholds
SRU consensus 2020, "rule of four" (viral hepatitis and fatty liver disease)pSWE and 2D-SWE, any vendor5 kPa (1.3 m/s) or less: normal. Under 9 kPa (1.7 m/s): compensated advanced chronic liver disease (cACLD) ruled out if there are no other clinical signs. 9-13 kPa (1.7-2.1 m/s): suggestive of cACLD, needs another test. Over 13 kPa (2.1 m/s): cACLD likely. Over 17 kPa (2.4 m/s): clinically significant portal hypertension likely
Baveno VII (2021), "rule of five"Transient elastographyUnder 10 kPa: cACLD excluded. 10-15 kPa: suggestive. Over 15 kPa: highly suggestive. 25 kPa or more: rules in clinically significant portal hypertension. 15 kPa or less with platelets 150 × 10⁹/L or more: rules it out. Under 20 kPa with platelets over 150: screening endoscopy for varices can be avoided
Castera 2005, chronic hepatitis CTransient elastographyF2 or more: 7.1 kPa. F3 or more: 9.5 kPa. F4: 12.5 kPa
Friedrich-Rust 2012 meta-analysispSWE (Siemens VTQ)F2 or more: 1.34 m/s. F3 or more: 1.55 m/s. F4: 1.80 m/s

A normal liver measures about 4-6 kPa on transient elastography. Thresholds differ with the cause of liver disease (they tend to be higher in alcohol-related and cholestatic disease), so a reading is always interpreted against the aetiology. cACLD is the term that now replaces "advanced fibrosis and compensated cirrhosis" in non-invasive work, because elastography cannot separate F3 from F4 reliably.

2D shear wave elastography of the liver on a different vendor's machine, with the speed map on the right (scale 0-4 m/s) and a second map of the same box on the left. The reading is 2.02 m/s, which falls in the SRU band of 1.7-2.1 m/s: suggestive of compensated advanced chronic liver disease and needing confirmation.
2D shear wave elastography of the liver on a different vendor's machine, with the speed map on the right (scale 0-4 m/s) and a second map of the same box on the left. The reading is 2.02 m/s, which falls in the SRU band of 1.7-2.1 m/s: suggestive of compensated advanced chronic liver disease and needing confirmation.

What it is used for

  1. Staging fibrosis in chronic hepatitis B and C, metabolic dysfunction-associated steatotic liver disease (MASLD, formerly NAFLD), alcohol-related liver disease, autoimmune and cholestatic disease, and so deciding who needs treatment and who needs biopsy.
  2. Ruling cirrhosis in or out, which decides who enters surveillance for hepatocellular carcinoma and varices.
  3. Portal hypertension: liver stiffness correlates with the hepatic venous pressure gradient up to about 10-12 mmHg. Spleen stiffness adds to it: on transient elastography, under 21 kPa argues against and over 50 kPa argues for clinically significant portal hypertension, and 40 kPa or less makes high-risk varices unlikely (Baveno VII, validated mainly in viral hepatitis).
  4. Prognosis: higher stiffness predicts decompensation, hepatocellular carcinoma and death.
  5. Monitoring: response to antiviral treatment, abstinence or weight loss; progression on serial studies. Use the same machine each time.
  6. Steatosis: CAP on FibroScan, and attenuation imaging on newer scanners, quantify fat at the same sitting.
  7. Special groups: congestive hepatopathy and the Fontan circulation, Budd-Chiari syndrome, biliary atresia, cystic fibrosis, and graft fibrosis after transplantation.
  8. Focal liver lesions: malignant lesions are on average stiffer than benign ones, but the overlap is large; elastography is not recommended for characterising them.

Role of strain elastography in the liver: small, because there is no normal reference tissue in diffuse disease. A liver fibrosis index derived from strain patterns exists on some machines but is not in routine use.

Confounders that raise liver stiffness without fibrosis (false positives)

  • Acute hepatitis and any flare with transaminases more than 5 times the upper limit of normal
  • Obstructive cholestasis
  • Hepatic congestion: right heart failure, outflow obstruction
  • Recent meal (stiffness stays raised for 2-3 hours) and recent strenuous exercise
  • Recent alcohol excess and acute alcoholic hepatitis
  • Infiltration: amyloid, lymphoma, extramedullary haemopoiesis
  • Deep inspiration or Valsalva; excess probe pressure on the left lobe
  • Whether steatosis by itself changes the reading is debated

Causes of failed or unreliable readings: obesity and a thick chest wall, narrow intercostal spaces, ascites (transient elastography), inability to hold the breath, a small shrunken right lobe, and an inexperienced operator.

Ultrasound versus MR elastography for liver fibrosis

The details of MR elastography are in its own note; only what the comparison needs is given here.

MR elastography in four lines. A passive pneumatic driver strapped over the right lobe delivers continuous 60 Hz vibration from an active driver outside the scan room. A phase-contrast sequence with motion-encoding gradients synchronised to the vibration images the shear waves. An inversion algorithm converts the wave images into a stiffness map (elastogram) with a confidence map. Regions of interest are drawn over reliable liver and the mean is reported in kPa.

Ultrasound elastographyMR elastography
Liver sampledSmall: a 1 × 4 cm cylinder (TE) or a box of about 1 cm (pSWE); a few cm (2D-SWE)Large: several slices through the whole liver, so least sampling error and best for patchy fibrosis
AccuracyVery good for cirrhosis, moderate for early fibrosisHighest of the non-invasive tests, particularly for early (F1-F2) fibrosis
Obesity and ascitesFail or become unreliable (2D-SWE and pSWE cope with ascites; TE does not)Not affected
Operator dependenceModerate; needs training and quality criteriaLow; highly reproducible between readers and scanners
FailureTE fails in about 3% and is unreliable in a further 15% or so, mostly from obesityA few percent at 1.5 T, mostly from iron overload (signal loss on gradient-echo sequences); more at 3 T
Cost and accessCheap, quick, portable, bedside, easily repeatedExpensive, limited availability, needs extra hardware and software
Time5-10 minutesA few breath-holds, usually added to a liver MRI
ContraindicationsNoneThose of MRI: some implants, claustrophobia
Added informationB-mode and Doppler of liver and spleen; CAP for fatMorphology, lesion detection and characterisation, fat fraction and iron in the same study
StandardisationValues differ between vendors and techniquesOne 2D gradient-echo technique at 60 Hz is used across vendors, so values are comparable
ConfoundersInflammation, congestion, cholestasis, foodThe same; these are biological, not technical
Number reportedYoung's modulus (or m/s)Shear stiffness (magnitude of the complex shear modulus)

The kPa values are not interchangeable. MR elastography reports shear stiffness, roughly a third of the Young's modulus reported by ultrasound, and it is measured at a different frequency. A normal liver is under about 2.5 kPa on MR elastography and cirrhosis is above about 5 kPa on the commonly used scale, against roughly 5 kPa and 13 kPa or more on ultrasound.

Bottom line: ultrasound elastography is the first-line screening and follow-up test; MR elastography is the problem-solver when ultrasound fails or is unreliable (obesity, ascites), when results disagree with the clinical picture, or when MRI is being done anyway.

Breast

Basis. Most cancers are harder than normal breast and harder than most benign lesions, because of the desmoplastic stromal reaction, which extends beyond the visible mass. Elastography is an adjunct to B-mode and BI-RADS, not a stand-alone test. Its main value is raising specificity and so reducing benign biopsies.

Technique. Patient positioned as for B-mode, linear probe perpendicular to the skin with the lightest possible contact, lesion in the centre of the box with fat and normal tissue around it, patient still or holding the breath. Excess pre-compression is the commonest error with either method.

Strain elastography: three ways to read it

  1. Tsukuba elasticity score (Itoh 2006), on a map where blue is hard
ScorePatternMeaning
1Whole lesion deforms like the surrounding tissue (all green)Benign
2Mosaic of green and blueProbably benign
3Centre blue, periphery greenProbably benign
4Whole lesion blueSuspicious
5Lesion and surrounding tissue blueMalignant

The cut-off lies between 3 and 4. In the original series this gave a sensitivity of about 87% and a specificity of about 90%.

  1. Strain ratio (fat-to-lesion ratio). One region of interest in the lesion, one in subcutaneous fat at a similar depth. Higher means stiffer. Published cut-offs range from about 2.3 to 4.5 depending on the machine and the study, so no single figure applies everywhere.

  2. E/B ratio. Lesion length on the elastogram divided by its length on B-mode. Less than 1: benign. 1 or more: malignant. It reflects the stiff halo of desmoplasia and tumour infiltration that B-mode does not show.

Strain elastography of a breast mass (Hitachi). Left: B-mode shows an irregular hypoechoic mass. Right: on the elastogram the mass is uniformly blue (hard), and the blue area is larger than the mass seen on B-mode, while the surrounding tissue is green and red (softer). This is the pattern of a high Tsukuba score and an E/B ratio above 1.
Strain elastography of a breast mass (Hitachi). Left: B-mode shows an irregular hypoechoic mass. Right: on the elastogram the mass is uniformly blue (hard), and the blue area is larger than the mass seen on B-mode, while the surrounding tissue is green and red (softer). This is the pattern of a high Tsukuba score and an E/B ratio above 1.
Strain ratio measurement in a breast mass. One region of interest (Z1) is placed in the lesion, which is blue (hard), and three reference regions (Z2-Z4) in the superficial tissue, which is red and green (soft). The machine lists the ratio of each reference region to the lesion at the bottom left. The colour bar on the right shows soft at the top and hard at the bottom.
Strain ratio measurement in a breast mass. One region of interest (Z1) is placed in the lesion, which is blue (hard), and three reference regions (Z2-Z4) in the superficial tissue, which is red and green (soft). The machine lists the ratio of each reference region to the lesion at the bottom left. The colour bar on the right shows soft at the top and hard at the bottom.

Cysts on strain imaging

  • Blue-green-red (BGR) sign: three horizontal colour layers inside a simple cyst on some machines.
  • Bull's-eye artefact on others: a black lesion with a bright centre and a bright spot behind it.
  • Both are artefacts of fluid moving under compression. They help to recognise a complicated cyst that looks solid on B-mode.

Shear wave elastography

  • Read from the colour map (homogeneity, shape, and stiffness of the lesion and of the tissue around it) and from the number. Maximum stiffness (Emax), measured over the stiffest part of the lesion or its rim, performs best.
  • BE1 multicentre study (Berg 2012, SuperSonic Imagine): a cut-off of 80 kPa (5.2 m/s). Adding shear wave findings to BI-RADS raised specificity from about 61% to about 78% without loss of sensitivity.
  • Stiff rim sign: a ring of high stiffness around the lesion, typical of cancer. Very hard cancers may show no signal or falsely low values in the centre because shear waves do not propagate there, so the rim must be measured.
  • Cysts show no signal (black), since shear waves do not travel in fluid.

Use with BI-RADS

  • The ACR BI-RADS atlas, 5th edition (2013), lists elasticity assessment among the associated features of the ultrasound lexicon, with three descriptors: soft, intermediate, hard. It warns that shape, margin and echo pattern outweigh elasticity.
  • Elastography changes management mainly in BI-RADS 3 and 4a:
    • A BI-RADS 4a lesion that is soft on elastography (Tsukuba 1-3, or Emax of 80 kPa or less) can be downgraded to 3 and followed instead of biopsied.
    • A BI-RADS 3 lesion that is hard (Tsukuba 4-5, or Emax of 160 kPa (7.3 m/s) or more in BE1) should be upgraded to biopsy.
  • BI-RADS 4b, 4c and 5 lesions are biopsied whatever the elastogram shows. A soft result never cancels suspicious morphology.

Other roles in the breast

  • Complicated cyst versus solid mass; fat lobule versus true lesion
  • Showing the true extent of a cancer, which is often larger on the elastogram
  • Choosing the stiffest part for biopsy
  • Axillary nodes: metastatic nodes are stiffer, especially in the cortex
  • Predicting and monitoring response to neoadjuvant chemotherapy (stiffness falls in responders)
  • Non-mass lesions and architectural distortion, as a pointer to where to biopsy

Pitfalls

False negative (soft cancers)False positive (hard benign lesions)
Mucinous, medullary and papillary carcinomaHyalinised or calcified fibroadenoma
Necrotic or cystic high-grade tumoursFat necrosis, surgical scar, radial scar
Ductal carcinoma in situ and very small invasive cancersMastitis, abscess wall, diabetic mastopathy
Lymphoma, some metastasesSclerosing adenosis, fibrosis

Technical limits: lesions deeper than about 4 cm, very large lesions that fill the box, lesions close to the chest wall or nipple, implants, and pre-compression.

Thyroid

Basis. Papillary carcinoma, the commonest thyroid cancer, is hard (fibrosis, psammoma bodies). Most benign nodules are soft. Elastography is added to greyscale ultrasound to decide which nodules need fine-needle aspiration.

Strain elastography

  • Transverse or longitudinal plane, light compression. Carotid pulsation can be used as the compressing force, but it also causes artefact, so a longitudinal plane away from the artery is preferred.
  • Rago score (5-point): 1, whole nodule elastic; 2, mostly elastic; 3, elastic at the periphery only; 4, whole nodule hard; 5, no elasticity in the nodule or in the area behind it. Scores 4 and 5 are suspicious.
  • Asteria score (4-point): 1, entirely elastic; 2, mostly elastic; 3, mostly hard; 4, entirely hard. Scores 3 and 4 are suspicious.
  • Strain ratio: normal thyroid (or strap muscle) strain divided by nodule strain. Reported cut-offs lie roughly between 2 and 4 and depend on the machine.

Shear wave elastography

  • Gives kPa or m/s without manual compression, so it is less operator dependent.
  • Normal thyroid is soft, roughly 10-20 kPa. Reported cut-offs for malignancy vary widely, from about 35 to 65 kPa or more for 2D-SWE and about 2.5-3 m/s for pSWE. No single cut-off is agreed.

Performance. Pooled sensitivity and specificity are each about 80% for both methods. The most useful property is a high negative predictive value: a nodule that is soft and has no suspicious greyscale feature is very unlikely to be malignant.

Role

  • Adjunct to greyscale features in selecting nodules for aspiration, and in choosing which nodule to sample in a multinodular gland
  • May help in nodules with indeterminate or non-diagnostic cytology, though results are mixed
  • Follow-up of nodules with benign cytology
  • Diffuse disease: stiffness is raised in Hashimoto, Graves, subacute (de Quervain) and Riedel thyroiditis
  • Cervical lymph nodes: metastatic nodes are stiffer

Position in guidelines. WFUMB (2017) and EFSUMB recommend it only as a complement to greyscale ultrasound, never a replacement. It is not a scoring item in ACR TI-RADS (2017); EU-TIRADS mentions it as an optional complementary tool; the earlier French TI-RADS included high stiffness as a suspicious sign.

Limitations

  • Follicular carcinoma is often soft and is missed; medullary and some other cancers can be soft too
  • Hard benign nodules: coarse or rim calcification, fibrosis, subacute thyroiditis
  • Cystic or predominantly cystic nodules are unreliable (fluid)
  • Nodules larger than about 3 cm, with no normal tissue left for reference; coalescent nodules
  • Isthmus (compressed against the trachea), deep or retrosternal nodules
  • Background thyroiditis makes the reference tissue stiff
  • Carotid pulsation, swallowing, breathing; pre-compression raises all values

Obstetrics and gynaecology

Most of these uses are still under evaluation; say so in the answer.

Obstetrics

UseWhat is found
Cervix: prediction of preterm birthThe cervix softens (collagen remodelling) before it shortens. A soft cervix, especially at the internal os, in the mid-trimester is associated with spontaneous preterm delivery, and adds to cervical length
Cervix: before induction of labourA soft cervix predicts successful induction and a shorter induction-to-delivery time; proposed as an objective addition to the Bishop score
Cervical insufficiency and cerclageAssessment of cervical consistency before and after cerclage
PlacentaStiffer in pre-eclampsia, fetal growth restriction and gestational diabetes in research series
Placenta accreta spectrum, caesarean scarInvestigational
Fetal organs (lung maturity, liver)Research only
Post-partumPelvic floor: levator ani and anal sphincter injury
  • Technique for the cervix: transvaginal probe in the anterior fornix, sagittal view as for cervical length, minimal pressure. Strain imaging is read as a colour pattern or as strain values at the internal and external os; the weakness is that there is no reference tissue and the applied force is unknown. Shear wave imaging gives an absolute value but is affected by depth and by probe pressure.
  • Safety: strain elastography uses the same output as B-mode. ARFI-based shear wave methods use longer, higher-energy pulses and raise the thermal index, so they should be used on the cervix and placenta with the ALARA principle, kept away from the fetus, and are not recommended for routine fetal imaging.

Gynaecology

UseWhat is found
Fibroid versus adenomyosisA fibroid is a well-defined, usually stiff lesion with a sharp border and a soft rim (pseudocapsule). Adenomyosis is ill-defined and heterogeneous, with no border. Both are generally stiffer than normal myometrium; which of the two is stiffer differs between studies, so the pattern matters more than the number
Fibroid treatmentMapping before, and showing the stiff treated zone after, high-intensity focused ultrasound or embolisation
EndometriumCarcinoma is stiffer than polyp or hyperplasia; may help in assessing myometrial invasion
Cervical cancerStiffer than normal cervix; helps to define tumour extent and parametrial spread and to monitor chemoradiation
Ovary and adnexaSolid parts of malignant masses are stiffer; overlap is large and it does not replace the standard ultrasound rules. Endometriomas and deep endometriotic nodules are stiff
Pelvic floorLevator ani stiffness in prolapse and incontinence
Breast in pregnancy and lactationAs for breast generally

Limitations in this field: no agreed cut-offs, different machines and methods across studies, deep pelvic organs beyond the reach of transabdominal elastography, bowel and bladder motion, and unresolved fetal safety for ARFI.

Clinical applications of compression elastography

Compression (strain) elastography shows relative stiffness, so its uses are in focal lesions surrounded by normal tissue in superficial organs or organs reached by an endocavitary probe. Write the technique and the three ways of reading it from the strain elastography section, then the following.

OrganApplicationHow it is read
Breast (the main use)Benign versus malignant; downgrading BI-RADS 4a and upgrading BI-RADS 3; complicated cyst versus solid; lesion extent; biopsy targeting; axillary nodesTsukuba score (cut-off between 3 and 4), strain ratio against fat, E/B ratio of 1 or more, BGR sign in cysts
ThyroidSelecting nodules for aspiration; high negative predictive valueRago score 4-5 or Asteria 3-4 suspicious; strain ratio against normal gland
Lymph nodesMetastatic versus reactive; which node to sampleProportion of hard area; strain ratio against muscle
Prostate (transrectal)Targeting biopsy of hard peripheral-zone foci; fewer coresReproducible hard focus on repeated compression
Cervix (transvaginal)Preterm birth risk, induction success, cervical cancer extentColour pattern, strain at internal os
UterusFibroid versus adenomyosis; endometrial lesionsBorder and pattern of the stiff area
MusculoskeletalAchilles, patellar and common extensor tendinopathy; plantar fasciitis; muscle tear; soft tissue massesSoftening of a normally hard tendon; colour grade
Endoscopic ultrasoundPancreatic mass (cancer versus chronic pancreatitis); mediastinal and abdominal nodes; rectal wallDominant colour, strain ratio, strain histogram
Testis and scrotumHard tumour versus soft benign lesion or haematomaColour score
Salivary glands, skin and subcutisCharacterising masses; sclerodermaColour score, ratio
LiverLimited: fibrosis index from strain pattern; showing the stiff ablation zone after radiofrequency ablationPattern analysis
VesselsPlaque composition; age of venous thrombus (older thrombus is harder)Strain within plaque or thrombus

End with its limits: relative, operator dependent, shallow, unsuitable for diffuse disease, and not comparable between vendors.

Advantages

  • Non-invasive, quick, no radiation, no contrast, bedside, repeatable
  • Added to a routine ultrasound examination on the same machine and probe
  • Shear wave methods give a number that can be followed over time
  • Reduces liver biopsies and benign breast and thyroid biopsies
  • Cheap compared with MR elastography

Limitations, artefacts and pitfalls

General

  • Values and cut-offs are specific to technique, vendor, probe and disease; follow-up should be on the same system
  • Depth: strain to 3-4 cm, ARFI to about 6-7 cm
  • Assumes simple elastic behaviour; tissue is viscoelastic, anisotropic (muscle, tendon, kidney) and stiffens under pressure
  • Operator training matters, most of all for strain imaging
  • Stiffness is not specific: fibrosis, tumour, inflammation, congestion and oedema all raise it

Artefacts

ArtefactCause
Pre-compression artefactProbe pressure stiffens superficial tissue and raises all values
BGR and bull's-eye artefactsFluid in a cyst on strain imaging
Signal voidNo shear wave in fluid, or in very hard or deep tissue; shown as black or uncoloured
"Bang" or near-field artefactHigh values just below the liver capsule from reverberation; hence the 1.5-2 cm rule
Stiff rim, soft centreShear waves fail to propagate through the centre of a hard cancer
Slip and out-of-plane motionLesion moves sideways under compression
ShadowingCalcification or ribs block the push and the tracking pulses
MotionBreathing, cardiac and carotid pulsation, swallowing
Boundary effectsShear waves reflect at capsules, vessels and thin structures and give falsely high or low values

Safety. ARFI pulses are longer than imaging pulses but remain within regulatory output limits; the main caution is the fetus.

Short note (5 marks)

  • Definition: imaging of tissue stiffness, the equivalent of palpation. Young's modulus E = stress / strain, in kPa.
  • Strain (compression) elastography: manual or physiological compression; strain is tracked and shown as a colour map. Relative only: score, strain ratio, E/B ratio.
  • Shear wave elastography: a piston (transient elastography, 50 Hz) or an ARFI push (point and 2D shear wave) produces shear waves; their speed (1-10 m/s) gives E = 3ρc². Quantitative, in m/s or kPa. Small diagram.
  • Liver: fibrosis staging and portal hypertension. Ten readings, median, IQR/M of 30% or less, after fasting. SRU 2020 rule of four for ARFI: 5, 9, 13, 17 kPa. Baveno VII for transient elastography: under 10 kPa excludes and over 15 kPa suggests advanced chronic liver disease.
  • Confounders: inflammation with high transaminases, cholestasis, congestion, a recent meal.
  • Breast: Tsukuba score 1-5 (cut-off between 3 and 4), E/B ratio of 1 or more, Emax above 80 kPa suspicious; refines BI-RADS 3 and 4a.
  • Thyroid and others: adjunct for nodule selection (follicular carcinoma is soft); cervix for preterm birth; prostate, nodes, tendon.
  • Limits: vendor-specific values, depth, operator dependence, obesity and ascites; MR elastography samples more liver and is used when ultrasound fails.