Thoracic ultrasound (TUS) is increasingly used for bedside evaluation of the pleural space and thorax. Given improved safety, NPSA and BTS guidance ‘strongly recommend’ TUS for pleural fluid procedure site selection. TUS also gives useful diagnostic information (see Table 72.1) which may alter management (e.g. a patient with a transudative effusion who has pleural nodularity, suggestive of malignancy).
Table 72.1 Diagnostic and therapeutic utility of TUS
Fluid quantification and characterization. Guided intervention. Obesity and rib crowding cause difficulties
Pleural thickening and nodularity
Detection and guided core biopsy. ‘Colour fluid sign’ may help differentiate between fluid and thickening
Assessment of function. Detection of thickening or nodularity (assessment limited by aerated lung)
Ruling out post-procedural pneumothorax. Unhelpful for assessing pneumothorax size. COPD, CF, and prior pleurodesis may mimic pneumothorax
Detection of atelectasis, consolidation, and peripheral lung lesions (abscess/tumour). Guided biopsy. Unable to assess structures deep to aerated lung
Detection of pericardial fluid, cardiomegaly
Rib fracture detection and FNA of metastases
Assessment and guided FNA
• Follow the Royal College of Radiology ‘level 1’ TUS syllabus
• Find a suitable mentor (with TUS level 2 or level 1 for ≥2y)
• Keep a log book, and maintain a record of video clips and still images
• Attend a practical and theoretical TUS course
• Level 1 practical training currently requires ≥1 session/week over ≥3 months (~5 scans/session)
• Maintain level 1 competency by: (1) doing ≥20 scans/y (≤3 months between scans), (2) maintaining contact with a named radiologist ‘mentor’, (3) auditing practice, (4) remaining current with literature and CPD.
Characteristics of the US wave
• US is a longitudinal wave in which particles move in the same direction as the wave (creating successive compressions and rarefactions)
• Frequencies used typically 2.5–12MHz (audible sound 20Hz–20kHz)
• Key formula:
Frequency (f) = constant (c)/wavelength (λ)
Where c = speed of US in soft tissue (~1,540m/s)
• With a typical frequency of 5MHz, TUS wavelength (which determines resolution) is ~0.3mm.
Changes to the US wave
• US wave can be transmitted, attenuated, or reflected
• Transmission occurs when particles in a tissue move together and have coherent vibration
• Attenuation (loss of US energy) occurs due to wave absorption, scatter, and refraction. Absorption occurs when particles do not move together and have chaotic vibration, generating heat. Higher frequencies (i.e. shorter wavelengths) are more likely to be absorbed than transmitted, giving poorer depth penetration but better resolution. Deeper structures lead to greater wave attenuation (travels further through tissue), and US machines compensate for this using time gain compensation (TGC), which can be fine-tuned using sliders on machine
• Reflection occurs at the interface between tissues with different impedence. High impedance at soft tissue-air interface and soft tissue-bone interface causes near-complete reflection (and explains inability to image aerated lung and the acoustic shadow cast behind ribs). Also, reason why coupling US gel required. Partial reflection is required to generate a return of signal to the US probe (and create an image).
Generation of the US wave
• US probes have piezoelectric crystals responsible for wave generation and detection. Timing and power of the returning wave generates the B/2D-mode image
• Pulse repetition frequency (PRF) determines interval between successive US pulses (must avoid collision of successive pulses). Slow PRF (e.g. simultaneous 2D/Doppler imaging) causes jerky images. Reducing size of scan field improves PRF.
• Numerous artefacts occur with TUS, including:
• Mirror artefact Smooth curved surfaces (e.g. diaphragm) reflect liver/spleen so that they seem within the thoracic cavity (appearing as consolidated lung)
• Horizontal reverberation artefact Tissue interfaces that have significant impedance mismatch (e.g. soft tissue-aerated lung interface) create successive reflections between the interface and the ultrasound probe itself, giving a series of echogenic parallel lines below the pleural stripe
• Comet tail artefact Another reverberation artefact seen at the pleural-aerated lung interface, which creates vertical ‘comet tails’, particularly at the lung bases
• Posterior acoustic shadowing Poor visualization of structures deep to interfaces with a high reflection coefficient (e.g. ribs, calcified gallstones) due to complete reflection/absorption
• Posterior acoustic enhancement Transmission through a medium which causes minimal attenuation (e.g. fluid-filled cyst) causes apparent enhancement of posterior structures.
Prior to the examination
• Examine any available radiology, particularly CT. Think—are there lesions on the CT which should be visible at TUS (including rib metastases, parenchymal pathology, or liver metastases)?
• Position the patient appropriately. For diagnostic TUS, sat up leaning forward, with arms resting on a table, gives excellent views laterally and posteriorly (where most pleural pathology lies). For pleural procedures, the lateral decubitus position prevents the patient from moving (recommended for real-time US-guided pleural intervention)
• Move the US machine to the patient, taking care not to run over the expensive (~£5,000) probe cables. Think about machine position, particularly if undertaking real-time intervention—the probe and the machine screen should be in a straight line. Clean the US machine and probe with an appropriate wipe.
Operating the US machine
• Confirm and enter patient details on the US machine
• Select 2D/B mode, using ‘abdominal’ preset if ‘thoracic’ not available (initial depth ~15cm; ensure that TCG sliders are arranged vertically)
• Use a 5MHz curvilinear probe for routine TUS (good compromise between depth of penetration and resolution). 10MHz linear probe better for vascular access and lymph node/rib metastasis FNA (high resolution but poor depth penetration)
• Hold the probe gently, like a pen. Three movements are important to get the most information from narrow intercostal spaces: rotation, angulation, and translation. TUS is a dynamic process; move the probe over both hemithoraces, while continuously optimizing the image, and make sure to first identify the costophrenic angle and liver/spleen to avoid mimics of a pleural effusion (e.g. ascites and loculated intra-abdominal collections).
Optimize machine controls while imaging
• Depth should be changed, depending on structure being imaged. Always ensure that you can initially see the full extent of any effusion and structures deep to the fluid
• Gain Avoid the temptation to set the gain too high
• Focal points Start with one focal point, positioned at the depth of maximal interest. Multiple focal points, whilst seemingly attractive, reduce the PRF and make the image jerky
• Frequency May need to reduce, particularly for larger patients, to increase depth penetration
• Colour Doppler Useful for assessing possible vascular structures and differentiating between pleural thickening and fluid (‘colour fluid sign’)
• TGC May need to increase gain at depth, particularly for larger patients. Conversely, a massive effusion (and accompanying posterior acoustic enhancement) may make deeper structures very bright, necessitating reducing gain at depth
• Freeze and store Always store at least one still image/video clip per patient. If performing an intervention, store a representative image from your intervention site
• Measure/calipers Useful for measuring depth of pleural effusion at site of intervention and the distance from skin to pleural effusion (Consider—will a standard 35mm 21G needle reach the fluid?)
• Poor image? Restart using the default settings, and ensure that the TGC sliders are vertical and that depth is appropriate. For larger patients, increase the US power to maximum, and consider reducing US frequency (to ~3MHz). Tissue harmonic imaging may improve tissue boundary differentiation. Some machines have an image ‘optimize’ button.
See Fig. 72.1.
• Bright echogenic pleural stripe caused by reflection at the soft tissue-air interface
• Pleural sliding/gliding—a shimmering at the pleural stripe as the visceral pleura slides over the parietal pleura
• Comet tail artefacts—fanning out vertically from the pleural stripe, particularly at the lung bases
• Horizontal reverberation artefacts—repeating periodic horizontal lines below pleural stripe
• Absent deep detail—high reflection coefficient at pleural stripe means that it is impossible to image aerated lung or structures deep to lung. The only apparent structures are artefacts.
• TUS may be a useful rule-out test for pneumothorax, but always get a CXR to confirm when pneumothorax suspected
• Absent pleural sliding/gliding
• Absent comet tail artefacts
• Horizontal reverberation artefacts—may be accentuated in pneumothorax
• Absent deep detail.
• See Fig. 72.2
• TUS has a higher sensitivity for pleural fluid detection than CXR
• Fluid is seen deep to parietal pleura as a relatively dark (hypoechoic) structure
• Anechoic effusions are black and featureless and may be transudative or exudative
• Echogenic effusions are exudative and appear speckled, due to protein/pus/blood/intrapleural air
• Septated effusions are also exudative, and septations can be caused by any pleural inflammation (e.g. pleural infection, malignancy)
• Size—measure maximal effusion depth. A variety of formulae have been proposed to estimate pleural fluid volume, particularly in an ITU setting. Practically, the following is suggested—small (only visible at one intercostal space), moderate (less than half the hemithorax), large (greater than half the hemithorax)
• Possible to image structures deep to fluid.
• Should be smooth, and it may be possible to discern five alternating hypo- and hyperechoic stripes (see Fig. 72.3; disrupted in pathology)
• Diaphragm poorly visualized without fluid but may be possible by imaging over the liver/spleen and angling probe upwards
• Diaphragm inversion occurs with large effusions and usually associated with significant dyspnoea
• Function may be assessed by watching movement with respiration and sniffing (and comparing both sides).
• ‘Compressive’ atelectasis is commonly seen with pleural fluid. Lung has a concave ‘hockey stick’ appearance with significant volume loss. Internal structure is visible due to lack of aeration
• Consolidated lung may look similar to liver (or spleen), and parenchymal structure is visible due to lack of aeration (see Fig. 72.4). Minimal volume loss and ill-defined boundaries. Hyperechoic (bright) branching structures and speckles represent air bronchograms. Branching hypoechoic structures are either pulmonary vessels (with demonstrable colour Doppler signal) or fluid bronchograms (without Doppler signal).
TUS features in malignancy
Sonographic features with high specificity (95–100%) for malignancy and overall sensitivity 79%:
• Parietal pleural thickening >1cm
• Nodular pleural thickening
• Visceral pleural thickening
• Diaphragmatic thickening >7mm
• Disruption of five diaphragmatic layers
• Diaphragmatic nodules.
No sonographic characteristics can rule out pleural infection, and fluid sampling is essential. Septated effusions may drain less well than non-septated effusions (although this should not discourage drain insertion, as many will still drain well). Densely echogenic fluid is likely to be pus or blood.
Pleural procedures (aspiration and drainage) can be guided sonographically, either by site marking or using real-time guided intervention. Site marking is easier and can be used for most effusions >2–4cm depth but must be performed immediately prior to intervention without patient repositioning. Real-time visualization of the needle in the pleural space requires sterile US gel and sheath and is technically more challenging but is required for smaller or loculated effusions. An in-plane oblique course is taken from the side of the probe, and the entire path of the needle is visible. A similar technique is used for pleural biopsy (using a Temno® cutting needle) and lymph node/rib metastasis FNA (using a linear probe and a 23 or 21G needle).
• Avoid risky sampling of small <1cm effusions (which may not even be apparent on CXR), unless there is a genuine diagnostic need
• Always identify the hemidiaphragm to ensure that the pleural space, rather than the upper abdomen, is being imaged
• Avoid posterior approaches for interventions whenever possible (even though this might be the site of maximal fluid depth), due to the relative exposure of the neurovascular bundle. Provided adequate fluid is present, always use the safe triangle.