Abstract
Pulmonary embolism (PE) is a pathological entity characterised by venous thromboembolism in the pulmonary arteries. It is a common entity in daily clinical practice and is the third most common cause of cardiovascular death. Correct diagnostic work-up is pivotal to ensure timely institution of appropriate therapy. This requires recognition of the characteristic imaging findings and awareness of the role and peculiarities of the different imaging techniques involved in the diagnostic and therapeutic process. This review aims to showcase the most commonly encountered imaging findings associated with pulmonary thromboembolism for each imaging modality, outline their advantages and disadvantages, and define the specific role of these techniques in the clinical setting. The latest innovations concerning the diagnostic work-up of PE will be also briefly discussed, yielding a perspective on the future developments and challenges that the radiologist will have to face in the coming years.
Shareable abstract
Acute and chronic pulmonary thromboembolic disease are disease entities with distinct imaging features that need to be recognised to ensure appropriate treatment https://bit.ly/3uqQ2j9
Educational aims
Learn the most common findings associated with acute and chronic pulmonary thromboembolism on multimodality imaging.
Understand the role of different imaging modalities in the diagnostic work-up of pulmonary embolism to guide therapeutic management.
Introduction
Pulmonary embolism (PE) occurs when venous emboli travel through the venous circulatory system to the right heart cavities, subsequently being caught within the pulmonary arteries [1]. Most PE originates from deep venous thrombosis (DVT) of the lower limbs, which may lead in ∼50% of cases to silent PE [2]. PE is a relatively common finding in daily clinical practice, with an estimate incidence in the general population of 60–120 cases per 100 000, an in-hospital mortality of 14% and a mortality at 90 days of ∼20% [3].
The predisposing factors leading to the formation of venous thrombi are defined by the Virchow's triad, which includes endothelial injury, local hypercoagulability and stasis [3]. These risk factors can be associated with a wide variety of conditions which can represent at the same time the underlying cause and the comorbidity of PE. These factors often interact with each other and can be inherited/nonmodifiable or acquired/modifiable. These include nonmodifiable factors such as thrombophilias and older age, as well as acquired conditions such as cancer, immobilisation (e.g. due to major trauma and surgery, long flights) or intake of pro-thrombotic drug or substances (e.g. contraceptive drugs, smoking) [3–5]. Moreover, the recent coronavirus disease 2019 pandemic has also highlighted how inflammatory state or infections could increase the thrombotic risk, with a reported incidence of ∼16.5% in affected patients [6].
Acute PE can clinically manifest with numerous signs and symptoms and hence is characterised by a broad spectrum of clinical presentation. Patients can be totally asymptomatic (i.e. PEs detected as an incidental finding in the diagnostic work-up of other diseases) or show common, although nonspecific signs and symptoms like dyspnoea, chest pain, haemoptysis and in more severe cases syncope and cardiac arrest [4, 7]. Moreover, clinical pictures suggestive for DVT (e.g. unilateral lower limb swelling and pain, tenderness of the soft tissues along the thrombosed venous axis) are also to be deemed as suspect for PEs, since the two pathologies are strictly intertwined [4, 7]. After acute episodes, up to 50% of patients may develop the so-called “post-PE syndrome”, which encompasses reduced exercise tolerance (also possibly affecting the psychological sphere) without identifiable pulmonary vascular disease, to the most severe long-term complication, i.e. chronic thromboembolic pulmonary hypertension (CTEPH) [8].
A correct interpretation of the clinical picture is essential for generating a diagnostic hypothesis concerning these clinical entities, so that an investigative pathway could be set up to confirm it and to rule out the other differential diagnosis. Therefore, the aim of this review is to depict a multimodality imaging approach to PE, showing the most characteristic imaging findings and describing the role each technique has in the diagnostic workflow.
Multimodality imaging approach to pulmonary thromboembolic disease
Chest radiography
Chest radiography is often the first-line imaging technique routinely performed on patients with acute chest pain. This imaging technique is often readily available in an emergency setting and is quite easy and fast to perform. However, as multiple studies already showed, the imaging findings associated with PE are often nonspecific and therefore chest radiography has primarily the purpose of excluding other differential diagnoses for chest pain (e.g. pneumonia or pneumothorax), rather than being considered the primary diagnostic tool [9, 10].
In many cases patients affected by PE have an abnormal chest radiography image, showing findings possibly indicative of the underlying disease. Among the most common findings, it is important to detect signs of increased pulmonary pressures secondary to obstruction of the pulmonary arteries, such as signs of pulmonary vascular redistribution, the Fleischner sign (dilation of central arteries) or the Palla sign (more specifically, dilation of descending right pulmonary artery), and the Westermark sign (focal area of hyperlucency of the lung parenchyma, secondary to oligaemia) [9–11]. Moreover, it is important to detect a possible pulmonary infarct (PI) secondary to PE, which can be detected as a pleural based, wedge-shaped lung opacity, also known as the Hampton's hump. Finally, these patients may show also some ancillary nonspecific findings, with cardiomegaly, pleural effusion and elevation of the hemidiaphragm among the most common (figure 1) [9, 10].
Computed tomography pulmonary angiography
Computed tomography pulmonary angiography (CTPA) is the diagnostic technique of choice in the diagnosis of acute PE, yielding high resolution images of the pulmonary arteries, allowing for the evaluation of structural and morphological abnormalities. This technique has multiple advantages, predominant being its ready availability in the emergency setting, but exposure to ionising radiation and the need for administration of iodinated contrast medium (CM) with its potential for nephrotoxicity should be taken into consideration [4]. CTPA has high sensitivity and specificity (both ∼90% according to multiple studies), a high positive predictive value for patients with intermediate to high pre-test probability of PE and high negative predictive value in patients with low to intermediate pre-test probability [12, 13]. CTPA also plays a pivotal role in the diagnostic work-up of CTEPH, with high sensitivity and specificity for disease detection (both >95% for expert radiologists, according to a recent meta-analysis) [14].
CTPA is conventionally performed by acquiring images of the whole thorax after intravenous CM administration in the pulmonary arterial phase, with the objective of optimal opacification of pulmonary arteries up to the subsegmental branches [15]. In a recent publication [16], shortening the scan range (upper edge of the aortic arch and the lower hemidiaphragm) could provide an estimated average radiation dose reduction by 48% without significant loss of sensitivity for PE detection. This is of particular importance in populations that are particularly radiosensitive such as breast irradiation in young females. Depending on the type of available equipment and the clinical indication, the CTPA protocol can be varied to suit the individual patient. For example, patients with acute chest pain, suspected of acute coronary syndrome, acute aortic syndrome and PE as possible underlying cause(s) of their clinical presentation could undergo computed tomography (CT) with the triple-rule-out protocol. This requires an ECG-gated acquisition with simultaneous optimised opacification of the pulmonary arteries, the coronary arteries and the thoracic aorta [17]. Although potentially useful to accurately evaluate a broad variety of pathological entities associated with chest pain in a single scan, some disadvantages of this protocol should be acknowledged, namely an increased exposure to radiation and an increase quantity of iodinated CM in comparison with dedicated CT acquisition protocols, without a well-established benefit concerning the diagnostic accuracy [17, 18]. Moreover, it is technically demanding and can result in suboptimal opacification of either the systemic and/or pulmonary vascular beds and involves time-consuming imaging interpretation. Hence it is not routinely incorporated into clinical practice in most institutions.
CTPA can also be performed with the dual-energy computed tomography (DECT) technique, only available using specific types of scanners. This technique is based on acquisition of images with two different spectra and using the principle of material decomposition, to both qualitatively and quantitatively evaluate the distribution of iodinated CM in the lung parenchyma. Dedicated post-processing software uses this data to create perfusion maps, that can provide additional information concerning the obstructive burden of PE and its effects on lung perfusion [19]. Although several authors highlighted the high sensitivity and specificity of DECT for detecting PE, the superiority of this technique in comparison with conventional single-source CT is still debated [20]. A possible advantage of DECT could be the detection of small perfusion defects secondary to subsegmental PE, which are difficult to detect on conventional CTPA, but there is also an ongoing debate on the clinical significance of these findings and how to treat these patients, with the main approaches being clinical surveillance and anticoagulation (figure 2) [4, 21]. For chronic thromboembolism, DECT can be useful for the diagnosis of proximal as well as distal disease. The sensitivity of DECT in CTEPH has recently been shown to be equivalent to ventilation/perfusion (Vʹ/Qʹ) single-photon emission computed tomography (SPECT) with the additional advantage of lower radiation dose and reduced imaging time [22].
Imaging findings of acute and chronic PE
The pathognomonic imaging finding for acute PE is a filling defect in a pulmonary artery after CM administration. Typically, in the setting of acute PE the embolus may totally or partially occlude the vessel or being caught in a peripheral arterial branch form an acute angle with the vessel wall. In the case of a partially occluding embolus surrounded by contrast, the appearance will be that of a central filling defect within the vessel, with the so-called “polo mint sign” (also known as “doughnut sign”) or “railway track sign”, respectively, if the image plane is orthogonal or parallel to the major axis of the vessel [23–25] The obstructed vessels may also appear slightly enlarged in comparison with the surrounding patent vessels. Moreover, in case of massive PE with thromboembolic material localised in the main pulmonary artery (MPA) at the level of the bifurcation, extending in the left and right pulmonary artery, the appearance is that of the so-called “saddle embolus”, often associated with dilation of the corresponding vessels (figure 3) [23–25]. Post-processing tools such as maximum intensity projection (MIP) reformats can be helpful in the detection of peripheral PE but should be used cautiously for diagnosis of a more centrally located PE [26].
Depending on the severity and the localisation of the emboli, acute PE can be associated with PI or signs of right heart strain. PI can be detected as a wedge-shaped subpleural consolidation (corresponding to the Hampton hump on chest radiography), ground-glass opacity, or ground glass opacity with a peripheral rim of consolidation (“reverse halo” sign) with internal reticulations or “bubbly” appearance. Subsequently, the PI decreases in volume, evolving towards a more nodular appearance (“melting sign”), with a possible residual fibrotic lung scar with various appearances (e.g. bands, small wedge-shaped or round nodular opacities) (figure 3) [23–25]. Using DECT it also possible to detect lung perfusion anomalies secondary to PE, usually detectable as peripheral, wedge-shaped areas of decreased perfusion in iodine maps (figure 2) [27]. Right ventricle (RV) dysfunction secondary to acute pressure overload is considered the primary cause of death in the setting of acute PE. The latter causes an increase in pulmonary vascular resistance through anatomical obstruction of the pulmonary arteries (which should lead to increased pulmonary arterial pressure if 30–50% of the cross-sectional area of the vessels is obstructed by emboli) and hypoxic neurohumoral mediated vasoconstriction. As a result, the RV dilates and adapts to the increased pressure up to a certain extent, after which RV failure ensues. The hallmarks of RV dysfunction are RV dilation in comparison with the left ventricle (right ventricular to left ventricular ratio >1), flattening or leftward bowing of the interventricular septum and reflux of CM in the vena cava inferior and the hepatic veins (figure 3) [28].
As already mentioned, acute PE can evolve towards CTEPH, a severe and possibly fatal complication implying the development of pulmonary hypertension and ultimately also RV dysfunction. Chronic emboli may appear as filling defects causing the total occlusion of pulmonary arteries or can partially obstruct the lumen of the vessel, appearing as small linear filling defects (“webs”) or vessel wall thickenings due to the residual thrombotic material lying eccentrically [23, 25]. Moreover, the vessels involved by chronic PE can be seen as narrowed in comparison with the healthy surrounding vasculature, sometimes ending abruptly and being visually undetectable in the peripheral regions of the parenchyma; these findings are also often coupled to a mosaic lung attenuation, with superimposition of areas of decreased perfusion (oligaemia) amidst normally perfused lungs. Rarely emboli can also evolve calcifying inside the pulmonary arterial branches, which will appear slightly dilated (figure 4) [23, 25].
The hallmark of pulmonary hypertension is the enlargement of the MPA: more than 29 mm in men and 27 mm in women, and a pulmonary artery to ascending aorta ratio >0.9, as determined by a sub-study of the Framingham Heart study and defined in the latest European Respiratory Society (ERS)/European Society of Cardiology (ESC) guidelines on pulmonary hypertension, although these values are still debated concerning their sensitivity in detecting the disease [5, 29, 30]. As ancillary findings, it is also possible to detect compensatory hypertrophy of bronchial or intercostal arteries [5]. The signs of RV dysfunction are those already described concerning acute PE, with the possible additional presence of ventricular hypertrophy (i.e. wall thickness at the level of the right ventricular outflow tract ≥6 mm) as a hallmark of chronic adaptation to increased pulmonary vascular resistance (figure 4) [30].
Prognostic value of CTPA and therapeutic implications
Beside the diagnostic work-up of patients affected by PE, CTPA also has a pivotal role in the prognostic evaluation of patients affected by PE, thus also playing a role in the therapeutic management of these patients. For example, the right ventricular to left ventricular ratio has been acknowledged as the imaging sign with strongest predictive value regarding the development of adverse clinical outcomes in patients affected by acute PE [23, 25]. Moreover, several authors have developed scoring systems aimed at evaluating the severity of disease, aiming to perform risk stratification, and possibly guiding the therapy plan. For example, it is worth citing the clot burden quantification systems elaborated by Qanadli et al. [31] and Mastora et al. [32], which proved useful to perform outcome prediction. However, despite the potential merits of such scores, they are not commonly used in daily clinical practice, since this kind of quantitative evaluation of disease is time consuming and complex and their role in the prognostic evaluation of PE is still debated [23, 25].
A technically good quality CTPA can not only provide precise anatomical evaluation of chronic thromboembolism disease but also can be used as a roadmap to help assess operative suitability. However, CTPA cannot yet be considered as a standalone technique for the pre-operative evaluation of patients with CTEPH, and a thorough evaluation with other techniques is advised (i.e. Vʹ/Qʹ lung scintigraphy and angiography) [4, 30].
Vʹ/Qʹ scintigraphy
Lung scintigraphy is a well-established imaging technique in the diagnostic work-up of PE. It can be performed with different tracers and techniques depending on multiple factors, such as the clinical indication, the availability of the necessary equipment and user preference. The technique is based on the administration of radioactive tracers (e.g. intravenous technetium-99m (99mTc)-macroaggregated albumin for perfusion scans and Xenon-133 (133Xe) for ventilation scans) in order to detect anomalies concerning perfusion and ventilation of the lung parenchyma in the context of PE [4, 23]. Traditionally a planar perfusion and ventilation is performed to detect a mismatched perfusion defect indicative of thromboembolic disease. The introduction of Vʹ/Qʹ SPECT has improved the diagnostic accuracy of the technique and with the addition of a low-dose CT (Vʹ/Qʹ SPECT-CT), there is also improvement in the specificity [23, 33]. The pathophysiological background on which the interpretation of Vʹ/Qʹ scans is based implies a reduced blood flow in the region of the lung parenchyma involved with PE with a persisting normal ventilation, thus yielding an image of Vʹ/Qʹ mismatch (high Vʹ/Qʹ). Unfortunately, the underlying physiology of the lung is not so straightforward. Indeed, these anomalies may also be caused by different conditions and not all cases of PE cause a Vʹ/Qʹ mismatch (e.g. in the case of combined PE and PI, with decreased ventilation and perfusion); therefore, making the diagnosis a matter of probability of disease (although this approach could be considered obsolete with the new SPECT and SPECT-CT scanners) [23, 33, 34].
Regarding the interpretation of planar Vʹ/Qʹ scans, previously the suggested approach was to deem a scan as positive for PE if ≥2 segmental perfusion defects were detected in a patient with an appropriate pre-test probability while a normal scan excluded PE; in between these two diagnostic categories, a grey area is defined, including findings inconclusive or nondiagnostic for ruling out PE [34]. However, with Vʹ/Qʹ SPECT, which is the advised technique for diagnosing PE according to the European Association of Nuclear Medicine, a scan should be evaluated with a dichotomic approach (positive or negative for PE). Specifically, a case is defined as positive if at least a segmental or two subsegmental defects are detected, but the specific criteria to define the size of these mismatch areas are still debated [33, 34]. Mismatch areas are usually detected as peripheral, wedge-shaped areas of reduced lung perfusion with a pleural basis, similarly to perfusion defects described on the DECT-derived iodine maps (figure 5) [23, 25].
Among the main advantages of Vʹ/Qʹ SPECT and SPECT-CT it is important to cite the low radiation exposure (<2 mSv effective dose), low-cost of the technique and high values of sensitivity and specificity (both >90% for SPECT and both >95% for SPECT-CT according to a recent meta-analysis) [35]. However, scintigraphy is often not readily available in an emergency setting, the diagnostic criteria are still debated (as already stated above) and it does not provide alternative differential diagnoses [4, 36]. According to the latest ERS/ESC guidelines, Vʹ/Qʹ scintigraphy could be safely used to rule out acute PE if the perfusion scan is negative and it could also be a valuable alternative to CTPA in selected cases, for example, patients with a history of anaphylaxis induced by iodinated CM administration [4, 34]. In the setting of CTEPH, lung scintigraphy is the diagnostic test of choice for ruling out the disease; it can also provide additional information regarding the severity of disease, though in this case its role is still matter of discussion [30].
Catheter pulmonary angiogram
For decades, catheter pulmonary angiogram (CPA) was considered the gold standard in diagnosing PE, although it has now been replaced by CTPA as the first-line diagnostic technique for acute PE. CTPA has a similar level of diagnostic accuracy with a significantly reduced radiation exposure and the advantage of being a noninvasive technique [4, 37]. Hence CPA is now very rarely used for acute PE diagnosis.
The usual approach for performing CPA involves introducing a catheter through the right common femoral vein or the right jugular vein and then reaching the MPA through the right atrium and ventricle, also registering blood pressure in the right atrium and the MPA. This is a delicate phase of the procedure, because a one of the most common complications is the onset of transient arrhythmias (namely, right bundle branch block), which could be potentially lethal in predisposed patients (i.e. those with pre-existing left bundle branch block), with a possibility of total heart block and ensuing cardiac arrest. After reaching the MPA, the operator can perform selective catheterisation of the pulmonary arteries and the patency of the vessels is evaluated by injecting iodinated CM and obtaining images with digital subtraction angiography (DSA) in anteroposterior and oblique planes [23, 37].
The characteristic findings of acute PE in CPA-derived images are filling defects of the pulmonary arteries (partial obstruction or occlusion), similar to what has been described for CTPA, with possibly ancillary signs like reduced pulmonary flow and peripheral angiogram (depending on the localisation of the thromboembolic material) and delayed venous return. In the setting of chronic PE, the hallmarks of the disease, detected with CPA, are the presence of web or bands inside the vessel lumen, irregularities of the vessel's contours with possible stenotic tracts, pouch defects and total occlusion of a vessel with abrupt disappearance of the peripheral branches; signs of CTEPH can be also detected, such as dilated and tortuous appearance of the patent pulmonary arteries (figure 6) [23, 38].
CPA is considered not only an important diagnostic technique, but it also plays an essential role in the treatment of PE. In the acute setting, percutaneous revascularisation might be indicated in patients with haemodynamic deterioration on anticoagulation treatment, with signs of RV dysfunction detected on CTPA images. This technique involves the selective catheterisation of the involved vessel and subsequently the use of specific catheters with different approaches aimed at restoring the patency. Among the most used techniques, it is worth mentioning the use of catheters allowing the local infusion of a thrombolytic drug (commonly tissue plasminogen activator) to induce the degradation of the embolus, or the use of catheters for fragmentation or aspiration of the thromboembolic material [4, 37]. In the setting of CTEPH, CPA with balloon angioplasty could be performed in patients unsuitable for pulmonary endarterectomy (PEA) or with persistent/recurrent pulmonary hypertension after PEA. Balloon pulmonary angioplasty (BPA) has proven useful in improving the haemodynamics, the RV function and therefore also increasing the tolerance to exercise [30, 38]. BPA is involves performing selective catheterisation of the affected vessel, identifying the target stenotic lesion, and subsequently using an inflatable balloon to dilate the area, restoring the blood flow [30, 38].
Beside the two-dimensional (2D) projections provided by conventional CPA, three-dimensional (3D) techniques have been also applied during the interventional procedure, aimed at providing additional information, especially concerning the anatomy of the pulmonary arteries. These techniques are 3D rotational angiography (performed with the direct injection of contrast in the pulmonary arteries and using flat-panel technology) and cone-beam CT. The latter provides 3D images characterised by high spatial resolution although low contrast resolution, and is therefore, susceptible to beam hardening artefacts and noise. Merging these images with those obtained through real-time fluoroscopy allows a 3D roadmap to be obtained, possibly being of guidance in the most complex interventional procedures [39].
Magnetic resonance angiography
Magnetic resonance angiography (MRA) is a noninvasive diagnostic technique used to evaluate the patency of the pulmonary arteries and lung perfusion following the administration of gadolinium-based CM. A MRA protocol usually includes 3D sequences with different approaches (e.g. free breathing or breath-hold), launched before and after the administration of contrast imaging, and 2D T1- and T2-weighted and cine steady-state free precession sequences [40, 41]. A combination of dynamic contrast-enhanced perfusion followed by time-resolved contrast-enhanced angiography is useful for assessment of the pulmonary vasculature. The direct findings of acute PE are filling defects of the pulmonary arteries (from partial obstruction up to total occlusion), the vessel cut-off sign (amputated appearance of a vessel on MIP images), the double bronchus sign (on axial plane, the presence of two low-signal adjacent structures, i.e. the bronchus and the thrombus), the central dot sign (a central area of hyperintensity within the thrombus), the ghost vessel sign (enhancement of the vessel wall in the delayed post-contrast phase, surrounding an obstructive thrombus), and the bright clot sign (hyperintense appearance of the thrombus in T1-weighted pre-contrast images, due to the presence of methaemoglobin). The ancillary findings of PE are signs of venous stasis, areas of atelectasis, pulmonary infarctions, contrast-enhancement of the visceral pleura, and wedge-shaped perfusion defects [41]. Similar to what was described for CTPA, MRA also allows for detection of the hallmarks of CTEPH with similar imaging criteria (i.e. the presence of dilated MPA and RV, filling defects with the shape of webs or bands in the pulmonary arteries, perfusion anomalies of the lung parenchyma, etc.). However, magnetic resonance could further be used to perform a more in-depth evaluation of the RV function, performing a cardiac magnetic resonance (figure 7) [30, 40].
Among the many advantages of MRA it is important to mention the possibility of performing a scan without exposing the patient to radiation, the high specificity and sensitivity (both >90%), and the superior contrast resolution compared with CTPA. However, magnetic resonance also has disadvantages that should be acknowledged, like the limited availability (especially in an acute clinical setting), the complexity and the long duration of the scan, and therefore the need for an optimal setting to avoid the risk of nondiagnostic examinations, and contraindications related to the technique (i.e. claustrophobia and the presence of non-compatible implanted devices) [42]. Due to all the reasons discussed, the role of MRA in the diagnostic work-up of PE is still debated in the guidelines, although it represents more and more not only an important diagnostic technique but also a powerful tool for prognostic evaluation of these patients [4, 30].
PE mimics and the role of positron emission tomography-CT
The hallmark of PE, as already stated, is the presence of filling defects in the pulmonary arteries due to the presence of emboli, as detected by different imaging techniques. However, this finding can also be secondary to multiple non-thromboembolic pathologies, that are less common than PE, although still important in the differential diagnosis. Among the most important mimics, it is important to take into consideration the possible presence of primary tumours of the pulmonary arteries, namely primary pulmonary artery sarcomas (PAS), or metastatic emboli [43]. PAS are rare malignant tumours originating from the intimal layer of the pulmonary arteries, that are usually detected in adult patients with a slight predominance in women. These tumours can manifest with a clinical picture and imaging findings that may overlap with the characteristics of PE, obliterating the lumen of the pulmonary arteries in a similar way. However, some additional findings could be helpful in differentiating between the two entities, such as the presence of metastases, nodular filling defects, contrast uptake and/or heterogeneous attenuation of the intraluminal mass and the so-called “wall eclipse sign” in CTPA images [43]. Fluorodeoxyglucose (FDG) positron emission tomography (PET)-CT can have a role in differentiating PAS from PE, since PAS usually show a higher FDG uptake in comparison with emboli and may be associated to metastatic disease (figure 8) [39, 43]. However, caution must be exercised as local increased influx of inflammatory cells in acute PE can cause a focal increase in uptake of FDG at the embolic site resulting in increased maximum standardised uptake value (SUVmax) [44]. Beside neoplasms, there are also inflammatory conditions involving the pulmonary arteries (e.g. Takayasu arteritis and Behçet disease), possibly causing diffuse or focal narrowing of the vessels due to wall thickening and thus mimicking PE. FDG PET-CT could be useful in diagnosing large vessel vasculitides, with increased FDG uptake involving the wall of the thoracic vessels seen in the acute phase of disease [39, 43].
Future perspectives
The diagnostic workflow of PE involves the use of an increasing number of resources, especially because of a constant increase in the demand for scans aimed to rule out the disease. In particular, the number of requested CTPA scans has considerably increased in the past decade, posing a challenge for cardiothoracic radiologists. The major issues that need to be tackled involve multiple aspects, such as training in radiology and therefore the availability of well-trained personnel for the interpretation of the images, the need to properly interpret a high number of scans in a limited time (especially concerning the emergency setting) and the ensuing clinical decision process [45]. These issues are the rationale behind the implementation of artificial intelligence (AI) in the diagnostic workflow of PE. Among the most important applications in the daily clinical practice, AI software could aid the radiologist in the detection of PE findings (computer-aided diagnosis systems) and could even rearrange the list of examinations to be analysed to prioritise positive scans. As an example, a study was conducted on more than 10 000 chest CT scans performed in a European cancer centre to evaluate the diagnostic efficacy of an AI software for PE detection. The software showed a high level of sensitivity and specificity (both >90%) and reduced the rate of missed incidental PE from 44.8% to 2.6% when used as aiding tool; moreover, the AI-based prioritisation of scans in the list allowed the detection and notification time to be reduced from a median time of 7714 min for a routine workflow to 87 min, therefore potentially impacting the therapeutic management of these patients [46]. However, a recent publication showed that using an AI triage system to detect PE on CTPA examinations improved wait times but did not affect radiologist accuracy, miss rate or report turnaround times [47]. Therefore, despite some promising results, the introduction of AI-based software in the daily routine is still far from being a well-established practice in most hospitals due to multiple lingering issues, related not only to software development, cost and ease of workplace integration, but also to the generalisability of the algorithms and the training of the radiologists concerning AI. Therefore, further studies will be needed to evaluate the impact of AI in a real clinical scenario and to ultimately define its role in the diagnostic workflow.
Aside from AI, the technical innovations regarding the diagnosis of PE also concern the acquisition of images. In particular, the recent introduction of the first clinical photon counting CT (PCCT) allowed images to be obtained with a higher spatial resolution and preserved radiation dose efficiency compared with conventional CT systems with energy-integrating detectors. Moreover, PCCT could also allow performance of scans with a reduced CM administration as well as a reduced effective dose, while maintaining a good to optimal image quality [48].
However, despite these promising results, the use of PCCT is still limited. Indeed, the recent introduction of this new technology also implies a limited diffusion of such systems, with it currently available mostly in university hospitals. Therefore, in the next few years, a progressive expansion of PCCT scanners is expected into peripheral centres, possibly becoming a new standard in the diagnosis of PE.
Conclusion
PE is a common pathological entity that radiologists must deal with in daily clinical practice and knowledge of its imaging characteristics is essential for a correct diagnosis and subsequent therapeutic management of these patients. Furthermore, each imaging technique available to perform the diagnosis has its own role in the clinical setting, with different peculiarities and relevance in the diagnostic process. Therefore, it is of the utmost importance that the radiologist as well as respiratory physicians can correctly use all available imaging techniques, ultimately leading to the optimal treatment for PE, while at the same time being aware of the latest advances in the field.
Key points
Acute and chronic pulmonary thromboembolic disease are disease entities with distinct imaging features that need to be recognised to ensure appropriate treatment.
The reference standard for the noninvasive diagnosis of acute PE is multidetector CTPA, but not for chronic thromboembolism as a negative CTPA cannot confidently exclude distal disease.
Right to left ventricular ratio on CT is a strong predictor of adverse clinical outcomes in acute PE.
A normal perfusion scan carries high sensitivity for ruling out thromboembolic disease and is currently the recommended test to exclude chronic thromboembolic pulmonary disease as a cause of pulmonary hypertension.
Self-evaluation questions
Hampton's hump on chest radiograph is due to:
Pleural effusion
Pericardial effusion
Dilated pulmonary artery
Peripheral infarction
Right ventricular dilatation
Which of the following is not a feature of chronic thromboembolic disease on CTPA?
Eccentric clot
Calcified clot
Doughnut sign
Intravascular web
Stenotic artery
Which of the following is the typical abnormality of PE on Vʹ/Qʹ scan?
Normal perfusion, abnormal ventilation
Normal perfusion, normal ventilation
Abnormal perfusion, abnormal ventilation
Abnormal perfusion, normal ventilation
None of the above
Which of the following is not a finding of acute PE on magnetic resonance imaging?
Vascular webs
Vessel cut-off sign
Double bronchus sign
Ghost vessel sign
Bright clot sign
PET-CT can show FDG avidity in which of the following conditions?
a) Pulmonary artery sarcoma
b) Takayasu arteritis
c) Behçet disease
d) Sarcoidosis
e) All the above
Suggested answers
d. Hampton's hump is a well-defined peripheral, pleural-based opacity representing haemorrhage and necrotic lung tissue in a region of pulmonary infarction caused by acute PE.
c. Acute PE is characterised by an intraluminal filling defect surrounded by contrast medium. This produces the so-called “Doughnut” sign when viewed in short axis. The other choices are findings of chronic thromboembolic disease.
d. Mismatched perfusion defect is a characteristic feature of PE. Perfusion scan is sensitive for thromboembolic disease, but not specific as similar appearances can be seen in other conditions such as pulmonary artery malignancy and pulmonary veno-occlusive disease.
a. Intravascular obstruction of pulmonary arteries by unresolved chronic clots can result in the formation of thin fibrotic webs on vascular imaging.
e. FDG PET-CT is a sensitive marker of inflammation and malignancy and is useful to differentiate conditions that can mimic CTEPH.
Footnotes
Conflict of interest: The authors have nothing to disclose.
- Received September 22, 2023.
- Accepted November 14, 2023.
- Copyright ©ERS 2024
Breathe articles are open access and distributed under the terms of the Creative Commons Attribution Non-Commercial Licence 4.0. For commercial reproduction rights and permissions contact permissions{at}ersnet.org