This chapter struggles against the ocean of knowledge expected for CXR interpretation in Section 2.1.20, "Radiology in Intensive Care", from the CICMSecond Part General Exam Syllabus (First Edition). "Chest Xrays seem like such a fixed and finite topic of discussion" is the kind of statement one might soothe oneself with right up until the point that one tries to develop a chapter like this. In actual fact, there was more material here than could ever fit into a little grey summary box.
For pre-exam reading, one cannot go pastRaoof et al (2012)as a big broad overview, as well as the excellent hardware-themed Xray interpretation papers byGodoy et al (Part 1andPart 2); but the best by far is the 2024Glossary of Terms for Thoracic Imaging from the Fleishner society, which is not just a glossary but a beautifully illustrated image-rich tour of chest imaging terminology. If one had little time to spend on this, one should spend it among those references. On the other hand, if one were merely thinking about the CICM Second Part Exam as a vague threat in the far distance, one should instead take time to read through Reed'sChest Radiology: Patterns and Differential Diagnoses(at least Part 2, Pulmonary opacities). The author, in this autumn of his career, regrets that he only found the time for this so late, as it would have helped the younger him appear 30% more clever.
Indications for chest Xrays in the ICU
Why do we even do these? What if one had to justify this routine practice in less than 2 minutes? One may struggle, as many such Xrays are done "routinely" with minimal justification, and the literature on the subject is firmly directed towards minimising or abandoning the practice, rather than supporting it. Without revisiting the chapter on theutility of the routine daily chest Xray, one may summarise the justifications as "we might miss something because the patient is hard to examine, sedated and unable to report symptoms" or "some kind of tube might be in the wrong spot". It is much easier to justify the use of the chest Xray in ICU patients when it is performed to explain some detected change in the clinical state of the patient, or to ensure the position of some kind of device. In fact one might say that those are the only indications.
Relevant anatomy and interpretation of the ICU chest Xray: a structured examination
"Structure" is the word CICM examiners occasionally use to attack candidates who approach the interpretation of data in a way which should make the wider society anxious. This manifests as either a complete failure to make note of important negative findings or a fixation on the most obvious pathology that distracts from subtle but important features. The best way to guard against this is to have a structured approach, i.e. to always do the Xray thing the same way, and ideally following some kind of logical pattern. In the case of the CICM radiology viva, structure becomes essential to score the maximum marks in the shortest time, as the time is usually short indeed. For the imaging studies included in the 10-minute viva session, the trainee will usually have about 2 minutes for a chest Xray, and there will usually be at least 2-3 of these, which means the interpretation needs to be compressed into an information-dense point-form report.
Which structure? There are many. The reader is invited to google it for themselves and adopt the most effective strategy that they find logical. The author himself has used the DRABCDEFG mnemonic device during his own tortuous origin:
- D: "do it again": adequacy of the film
- R: random plastic: the hardware
- A: airway structures
- B:bones
- C:cardiac silhouette
- D: diaphragms
- E:extrathoracic tissues
- F:lung Fields
- G:gastric bubble
The attentive reader will note that this structure, if followed to the letter, will result an incredibly long interpretation time, and is probably unsuitable for the CICM exams. An abbreviated version could look like this:
- Adequacy: comment if the film is wildly rotated
- Bits: mention the hardware
- Crap: scan the boring outsides of the film, otherwise you will forget the obvious stable line or subcutaneous emphysema
- Distractions: spend a moment on the most obvious stand-out pathology, as this is where the marks probably are
- Essentials: look at the lung parenchyma and compare the lungs to each other
This is vaguely similar to what is recommended by the highly veneratedFelson's Principles of Chest Roentgenology (2020), where the mnemonic device is "Are There Many Lung Lesions":
- Abdomen
- Thorax
- Mediastinum and airway
- Lung
- the other Lung
The authors of Felson'spose that you should do it this way specifically to prevent the interpreter from imedietaly stampeding directly at the most obvious abnormality:
"Startreading every radiograph—chest or otherwise—by scanning the areas of leastinterest first, working toward the more important areas. You are less likely tomiss secondary but important findings this way. For the chest x-ray, start in theupper abdomen, then look at the thoracic cage (soft tissues and bones), then themediastinal structures, and finally the lung. Look at each lung individually, then
compare left lung and right lung."
The fact that Felson's seems to be the foundational text recommended by authorities on the subject, as well as the fact that Benny Felson had singlehandedly formed radiology from raw clay,lends authenticity to these suggestions.
Adequacy of the chest Xray
Is it a good Xray? Is it good enough to reveal pathology? The usual methods of assessing the technical adequacy of a film are described below, but the CICM exam candidates are reminded that most of the time a grossly inadequate film will not make its way into the exam because even the examiners will agree that this would be unfair. In other words, if the film is of such poor quality that one would not make decisions on the basis of it, then one should not expect the trainees to interpret it either.
The usual mnemonic used to describe this is PIRMA (Mowery & Singh, 2020):
- Penetration
- Inspiration
- Rotation
- Magnification
- Angulation
Penetration is generally described as a film where the thoracic spine is"just visible"or "faintly visible" behindthe heart. Others include more detail, eg "Exposure should be adequate if you are able to see approximately T4 vertebra and spinal process", and overexposure is where "you are able to see all vertebral bodies with obvious intervertebral spaces". Another source, a paper that assessed the adequacy of films in an ED (Chand et al, 2013), had used "slightly visible lower intervertebral disc below T9" as the criterion. Given thatan official ACR-SPR-STR document for mobile radiographymerely mentions that "on an optimally exposed chest radiograph, the lung parenchyma is displayed at a mid-gray level",it is clear that the adequacy of the penetration is in the eye of the beholder.
Rotation is generally assessed using the clavicle heads: they should be equidistant from the spinous processes.
Rotation in other directions is also possible (eg. when the imperfectly conscious patient in the ICU bed slumps forward when they are sat up for their morning Xray), which is referred to as "angulation". Again the clavicals can be used for this; they should be "S" shaped and their medial ends should be projected over the posterior third or fourth ribs.
Some things in this quality assessment are impossible unreachable ideals. For example, inspiration is often beyond the command of the intensive care radiographer, as the patient may be either uncontrollably tachypnoeic or mechanically ventilated and therefore (theoretically) having carefully controlled small tidal volumes. ICU Xrays will therefore often have inadequate inspiration, and we will often forgive them. Similarly the magnification of an AP film is hard to control for because it is performed using a mobile device with very limited control over the precise distance or angle between the radiation source and the detector plate. The resulting unpredictable distance between the heart and the source makes it hard to estimate the degree to which the cardiac contour is magnified by the effects of AP projection.
The hardware on the ICU chest Xray
No ICU radiological investigation is more cluttered with medical waste than the chest Xray. Often it feels as if the nurses and radiographers pile additional ECG leads and ventilator circuit components on the chest before taking the film to ensure that every detail is obscured by coiled wiring. The author is especially fond of seeing the internal organs of telemetry monitors and personal electronic devices left casually on the patient's chest (or in their actual hands as they play with them while being irradiated). Underneath the extrathoracic hardware, the most important structures to identify are:
- Endotracheal tube
- CVC, vas cath or PA catheter
- Nasogastric tube
- Chest drains
- Other/sundry(eg. ECMO cannulas, fancy catheter-directed thrombolysis lines, pacemakers, VADs, and so forth)
Do we care about this? On a fundamental level most people would agree that looking at the ICU patient's Xray and neglecting to notice that the ETT has migrated into the right main bronchus is something of a fail. Similarly the incorrectly positioned NG could produce aspiration, the misplaced central line could produce thrombosis, the partly dislodged ICC will produce subcutaneous emphysema, and so on. In other words, hardware positioning is important because poor hardware positioning has serious implications. Presumably on the basis of this, some marks in the radiology viva seem to be allocated to the equipment, though this must surely be viewed as less important than the pathology.
What can you say about these devices in the timeframe of 2 minutes where most of the minutes should be spent reporting on the obvious pneumonia? Not a lot. Practice is essential, as there is nothing worse than being stressed and having no words. One should develop a way of putting a set of normal findings into formal phrases, such as "endotracheal tube tip is appropriately positioned at the level of T4, right IJ CVC tip is in the lower SVC near the cavoatrial junction, the NGT side hole is well below the diaphragm". For each device, there is more detail in specific sections, as linked below.
Endotracheal tube position
The ETT should be 5cm from the carina, with the head in the neutral position (a"neutral" head position is where the mandible is somewhere around the level of C5-C6).Alternatively, T3-T4 level, going by fixed bony landmarks. This corresponds to a carina position at around T5-6.
One may attempt to look for specific recommendations for the position of double lumen tubes, but (as is logical) the only radiological finding worth noting is whether or not they are in the correct bronchus.
CVC position
Central venous devices should be in central venous structures, but not in cardiac structures, nor in any of the smaller venous structures in the chest. Specifically, the tip should be:
- At a level corresponding to the carina on a chest radiograph (which corresponds tothe cephalic limit of the pericardial reflection)
- Ideally above the pericardial reflection
"Innominate" in some ways is the opposite of nomenclature, but the author felt it rolled off the tongue better than "brachocephalic", and kept it.
Nasogastric tube position
Logically, this item should be in the stomach, i.e the tip (and all the side holes of a sump tube) should be inside the gastro-oesophageal junction. Disappointingly, the stomach and oesophagus are both relatively radiolucent structures and their position can only be estimated.Cohen et al (2011)offer measurements derived from 200 barium swallows, claiming that tube tips lower than the disk between 11th and 12th thoracic vertebra and more than 16 mm from the left side of the spine lie in the stomach and not the lower oesophagus. To help, manufacturers usually include a radiopaque line in the tube, which is interrupted at the level of the sidehole, as shown:
There are of course an endless array of possible intra and extrathoracic devices that one might find on the chest Xray of an ICU patient and it would be a fool's errand to attempt a catalogue of them all. Instead the reader is left with a link to"Uncommon Iatrogenic Devices Seen on Chest Radiographs" by Raj et al (2021). The presence of ECMO cannulae, PA catheters, IABP balloons and pacing wires should be noted.
The chest Xray findings
Not all ICU patients have a full shed of hardware in their chest, the CICM trainee may occasionally be faced with an Xray which is unadorned by any lines and tubes, forced to scavenge marks from the barren fields of their own radiological expertise. Fortunately, there are a finite number of ways for pathologies to manifest on a plain radiograph. One might even say that the pathology either increases, or decreases, the permeability of the tissues to radiation, and so all of Xray radiology can be reduced to recognising different patterns of opacities and lucencies. In fact there's not many different abnormal lucencies (they are mostly limited to gas in unexpected places), which means we are mostly interested interpretation of opacities.
Airspace opacities
Yes, the pedantic reader might point out that bronchi are also spaces filled with air, but when the Fleishner society defined "aisprace" they decided to only include the spaces that participate in gas exchange. Thus, airspace opacities are areas of increased Xray attenuation, i.e. spaces where some occupying substance is absorbing or deflecting radiation. These present as patches of lighter shade on the background of darker aerated lung. Historically, these would have also been referred to as "infiltrate", and people (including well educated senior people) will still occasionally be seen referring to them as such, but in fact this is obsolete terminology with vaguely sinister connotations of incursion or espionage.Patterson & Sponaugle (2005)killed this Cold War expression through the dry comedy of their article, which begins with the gentle admonition that "language is a tool" and proceeds to accuse radiologists of offering an uncertain "hedge" in lieu of an image interpretation. This was based on the finding thatfully 50% of surveyed physicians gave at least six possible things that could be meant by "inflitrate" when they read it from a radiology report. "Opacity", the F society argues, is better because it clearly represents what is described (it's a thing that is opaque to radiation) without sounding like any specific pathological process.
On the other hand "opacity" is still a very general descriptor. It refers to literally anything with greater average electron density than normal lung. The list of possibilities is impossibly broad, and to relate it here would be an affront to the reader's attention span. It will suffice to say that the nature of the opacity can be divined from:
- The edges (diffuse or discrete)
- The behaviour of surrounding lung (pulled towards, pushed away)
- The density (isodense with solid organs, or partially translucent)
- Details of internal structure (presence or absence of vascular markings, air bronchograms, nodules, lines, etc)
- Distribution (confluent or diffuse, bounded by anatomical borders such as fissures, distributed widely across the lungs or confined to characteristic regions)
- History over serial images (transient or persistent)
As these descriptors are many, it would be impossible to draw a table that lists differentials for every combination. Or rather, it would not be helpful to anybody, least of all the CICM trainee who is trying to develop a professional-sounding chest Xray interpretation technique. The better approach would be to learn some common patterns that present as airspace opacification and that relate to discrete groups of diagnostic possibilities, preparing to
As for most of the radiology syllabus in the CICM second part exam, this preparation mostly involves learning the language.Those opacities: are they diffuse or multifocal? Are they well-defined, or are they ill-defined? And if they are well-defined, would you describe them as alveolar and segmental or lobar, or are they more nodular, or linear? Reticular?Reticulonodular?The reader may wish to retreat to the safety of a classification system, but none exists. What follows is an attempt to create one, using several sources (notably Radiopedia, Reed (2018), and the Fleischner society glossary).
- Normal lung markings,which are shadows of vessels, mostly the branches of the pulmonary artery (Vehmas et al, 1993)
- Discrete localised opacities, few in number:
- Effusion
- Atelectasis
- Consolidation
- Tumour
- Widespread opacities, delocalised ormultiple
- Multifocal:i.e. the opacities are inhomogeneous and may coexist with patches of healthy or minimally diseased lung
- Well-defined: i.e. masses or abscesses, with distinct boundaries
- "Ill-defined":the airspac opacity with indistinct borders, blending into normal lung
- Confluent:patches of alveolar opacification that cluster together and become more dense centrally (a sub-trope of the "ill defined" genre)
- Diffuse,which is a variation on the theme of multifocal confluent ill-defined opacification, except so poorly defined that individual foci cannot be identified
- Nodular, which can be fine or coarse, and when they are coarse their edges can be seen and described as:
- Rounded
- Lobulated
- Spiculated
- Linear or"Reticular"opacities, which again come in fine and coarse flavours:
- Fine reticular
- Coarse reticular
- Reticulonodular, when you can't make up your mind whether you are looking at nodules or prominent interstitial markings
- "Ground glass" changeswhich are too diffuse and indistinct to be linear or nodular, and which are not dense enough to merit the term "consolidation"
- Multifocal:i.e. the opacities are inhomogeneous and may coexist with patches of healthy or minimally diseased lung
Pleural effusionis not strictly speaking "airspace", but is listed here because it is often on the list of the answers to the question "why is this base so opacified".
Atelectasis
Atelectasis is ubiquitous in ICU chest radiography and it is surprising that there is not more of it in the theoretical exam content. The radiological features of a large atelectatic region (to discriminate it from an area of consolidation or effusion) are:
- Displacement of thoracic structurestowardthe affected lung (if much of that lung has collapsed)
- Compensatory overinflation of unaffected lung
- Crowding of the ribs, where the intercostal spaces are visibly narrower on the collapsed side:
How is this not pneumonia, one might ask? How can you discriminate consolidation from atelectasis? Well: first of all, atelectasis usually does not contain air bronchograms. In fact quite the opposite: the bronchi entering an atelectatic area are often seen to stop suddenly because they are blocked (and this is how the lung became atelectatic in the first place):
Atelectasis needs not be whole-lung, to retract structures: it can retract the borders of the diaphragm (sometimes referred to as "tenting"), it can gather together some bronchovascular structures which would otherwise remain unbunched, or it can pull the edges of a visible intrlobular fissure.
Small areas of atelectasis can exist which appears as thin linear densities parallel to the diaphragm. They are otherwise known asFleischner's lines, or"plate-like" atelectasis:
Pleural effusion
Pleural effusion would now be the most logical topic to follow a discussion of atelectasis, as atelectasis and effusion can often coexist in ICU patients, and there is often debate as to which any specific missing hemidiaphragm might be attributed to. As an intrinsically lazy person, the author is often reluctant to immediately attend the bedside with an ultrasound probe to quickly discriminate between the two, preferring to scrutinise the radiograph and engage in conjecture."Effusion" here means any fluid. It could literally be anything.Pleural effusion is usually isodense with the densest structures, such as the heart and liver, as the electron density of the fluid is usually similar to soft tissue (whereas atelectasis contains some aerated regions and is therefore diffusely hypodense compared to muscle and fat):
Where there might be confusion, it is important to remember that effusion is something that occupies space, whereas atelectasis is thelossof space. As such, an effusion is expected to push structures out of the way:
The most classical feature is that an effusion tends to form a meniscus in the erect Xray:
How much fluid is there? Best not to guess, if you plan to impress the examiners with your attachment to precise reporting. In general it is said that about 500ml is required to completely flood the hemidiaphragm, 200ml of fluid is required to make a normal erect PA film look abnormal, and about 50ml is enough to blunt the costophrenic angles on full inspiration (Blackmore et al, 1996). A large enough effusion starts to creep up the wall of the chest cavity, separating the lung edge from the pleura:
Of course menisci form only with gravity, in an erect film. In a supine film, and presumably also in zero gravity, a modest pleural effusion will simply opacify the hemithorax by forming an even layer in the background of the lung:
A truly monstrous effusion on a supine film may even appear to encircle the lung:
A basal position of an effusion is not mandatory, particularly if the lung is adherent to the chest wall, or if the effusion is formed from something thick and gelatinous. For example, the apical "cap" haemothorax that develops following a complicated central line insertion is often localised to the upper lobe and stays in that position, either because it is subpleural or because it is no longer liquid (i.e. clot):
Pleural effusion is also not confined to the pleural margins, and can make excursions deeper into the chest as it extends along fissure boundaries:
Effusion can mimic other forms of Xray opacity when its location becomes atypical, such as scenarios where the fluid is trapped by adhesions along the boundary between the parietal and visceral pleura. In this fashion the pleural fluid can hang suspended from the chest wall or a fissure like a goon sack. The effect can resemble consolidation, as the effusion would be dense enough to obscure lung vascular markings and would have discrete edges that might make one think that it is trying to obey lobar boundaries:
Which is an awkward segue into:
Consolidation
The Fleishner society defines consolidation as "an exudate or other product of disease that replaces alveolar air, rendering the lung solid".The most important radiological features of this are:
- It obscures vascular markings, because it surrounds them with isodense material (whereas interstitial opacities and atelectasis usually accentuate vascular marking)
- It obscures the walls of bronchi by surrounding them
- It does not obscure the airfilled cavities of bronchi, which produces "air bronchograms", a term coined by BenFelson because Felix Fleischner first described it as "the visible bronchial tree", an equally descriptivename which didnot roll off the tongue.
- It often obeys lobar boundaries, sticking to one side of the boundary. This is not mandatory (consolidation can be patchy and widespread) but it does appear often in the sort of pathological processes that cause consolidation, such as pneumonia.
Anything that replaces or removes alveolar air can be described as consolidation, which means theoretically atelectasis could pass for consolidation. Except:
- Vascular markings are often intact in atelectasis, merely getting bunched closer together
- Volume is lost (whereconsolidation tends to fill the space, pushing surrounding structures aside, atelectasis obliterates it). This can sometimes help distinguish situations where the atelectasis appears to be "lobar", i.e. where it develops a sharply outlined boundary which one might associate with lobar or segmental pneumonia. Sure, there is a boundary, but it tends to bowtowardsthe collapsed lung, i.e. the area of atelectasis appears to have concave borders, whereas pneumonia should bulge out. This overdistension of nearby aerated lung segments is occasionally referred to as "compensatory emphysema"
Is this really "compensation", or is that in fact a huge bleb of an emphysematous bullus? suggest that a comparison of an expiratory and inspiratory films can help discriminate between these; a lung expanded in compensation will usually grow and shrink with the respiratory cycle, whereas a large abnormal airspace will usually not. - Air bronchograms are sometimes absent in atelectasis (specifically wherebronchial obstruction is the origin of the atelectasis), though not necessarily so. Normal spacing of bronchi is preserved in pneumonia, whereas in atelectatic lung with preserved bronchial patency, the aerated bronchi tend to bunch together, as in this classical image from the original Fleishner article:
Of course if the reader were to argue that the consolidation of pneumonia and the collapse of atelectasis may coexist, making these supposedly distinctive signs unreliable, they would be entirely correct. However this would do nothing to relieve them of the need to learn these features. It merely reminds us that a lot of radiology is an informed guess, because we are scrying the behaviour of three-dimensional dynamic structures from two-dimensional static images. Many will have encountered the terms in formal reports as a combination hedge ("collapse/conslidation"), as far back as the classic paper on localisation of lesions byHodson (1956). Which, as we reach the end of localised parenchymal disease, brings us to the question:
Which lobe is consolidated or collapsed?
Yes, the pedant will cry for a lateral view to compare, but this is ICU and we have no time for this sort of preachy moralising. The AP film can gives clues as to which lobe is collapsed by the pattern of structures that are accentuated or obscured. The ancient paper by Johnson & Bauer (1961), from an era before CT and ultrasound, is a classic of the literature and contains the best summary of these features which is presented as an annotated pictorial atlas below. Unfortunately the original images scanned so poorly that Elsevier should be ashamed of their attempts to charge you for this manuscript, but fortunately a Google search and Radiopedia were able to produce enough replacement images to satisfy our requirements.
Right upper lobe collapse or consolidation
- The right heart border is distinct
- The right upper mediastinal border is obscure
- The horizontal fissure is demonstrated, whereas it would otherwise not be seen.
This image fromradiologymasterclass.co.uk is an excellent example, even demonstrating an air bronchogram which the untrained eye might mistake for a preserved mediastinal contour (except it branches, which the mediastinal contour should not do):
RUL anterior segment: the opacity will obscure the outlineof the ascendingaorta and will create a sharp edge of thehorizontal interlobar fissure. This thing could stretch all the way across, reaching the chest wall.
Occassionally one may be faked out by an enlarged thymus, which looks somewhat similar:
But whenever one catches oneself thinking like this, one must remember that a huge thymus is almost entirely the province of the infant (as above), and an adult intensivist should never see such a thing.
RUL posterior segmentalso demonstratesthe horisonal fissure, and can also stretch across the entire horizontal hemithorax, but unlike the anterior segment, a distinct mediastinal contour should be well preserved. Consolidation in this segment tends to hug the chest wall:
RUL apical segment:the opacity will hide the right border of the trachea, but the hilumwill remain visible. A part of this segment is always behind the dense solid heads of the clavicle and the anterior parts of the first rib, which makes it somewhat challenging to identify conslidation here. Moreover this is not a segment known for retaining large wads of sputum or collapsing in response to pressure (being apical, and therefore mostly inflated at all times), which made it hard to find an example image. Most of the time these areas of the lung are affected by tuberculosis or cancer, leaving findings that are not exactly classical of other atelectasis or pneumonia, such cavitation and fibrosis.This one is of a RUL (mostly apical segment) collapse due to a bronchial tumour, from the excellentClassic Imaging Signsby Gao & McKinney (2021).
Right middle lobe collapse or consolidationis characterised by the disappearance of the right heart border, and the visibility of the hemidiaphragm, which usually preserves its edge.
Specifically, that is theRML medial segment:the right heart border is obscure and the right hemidiaphragm remainsdistinct. The horizontal fissure forms the upper limit for consolidation occurring in this segment:
RML lateral segmentlays laterally (they wouldn't have called it "lateral" otherwise) and also tends to involvehorizontal fissure (in the sense that it becomes visible as the boundary of this segment if the segment collapses or becomes filled with pus). In contrast, the right heart border and theright hemidiaphragm remain distinct, which is how you tell it apart from consolidation in the medial segment or the apical segments of the right lower lobe:
Right lower lobe collapse or consolidation usually leaves asharp silhouette of the right cardiac border , but obscures thehemidiaphragm and costophrenic angle:
RLL anterior and lateral basal segments are hard to distinguish from one another. As you can see from their anatomical positions, they superimpose over one another. Moreover the fissure boundary of these structures is oblique to the anteriorly facing observer, and does not present a crisp border. If the rest of the RLL is spared, you should still be able to see the hemidiaphram medially.
RLL medial basal segmentcollapse or consolidation leaves the medial third of the diaphragm obscured, but has no other distinguishing features. It tends to produce opacity in the same position as the right middle lobe medial segment, but the cardiac border should be preserved. Also unlike the right middle lobe, the horizontal fissure does not form a crisp upper boundary for this structure, and its opacity appears to be diffusely smeared along the hilum:
RLL posterior basal segmenton its own usually cannot be seen on an AP view, as it is obscured by the diaphragmatic border and the right heart, as you can see from this CT section.
Apical and subapical segmentsof the RLL usually produce opacities with distinct upper or lower boundaries and occupy the middle medial position. Consolidation here looks like RUL pneumonia, and it can even have a hard lower boundary (because the segmental boundary is horizontal), except unlike RUL pneumonia this one leaves a hard crisp edge for the right heart and mediastinum.
Left upper lobe collapse or consolidation tends to obliterate the edges of the left superior mediastinum and the left edge of the trachea. This lobe occupies a large amount of real estate on the AP film, extending rather low for an "upper" lobe. Observe, collapse of the whole LUL created by a left upper lobe bronchial mass, from Radiopedia:
The left upper lobe has four segments:
Left upper lobe apicoposterior segmenthugs the upper borders of the pleura and consolidation in this space takes out the radiolucent window created by the first rib and clavicle:
Left upper lobe anterior segmenthugs the mediastinal border and consolidation here presents as a localised process that spares the periphery of the lung and the uppermost apex:
The lingular segments of the LUL (superior and inferior lingular segments) exist along the left heart border and are difficult to define radiologically from an AP film, in the sense that both segments tend to occupy the same space and give the same sign, which is a loss of the middle left heart border, abutting the hilum:
The left lower lobe is the most frequently collapsed lobe in the ICU patient, owing to the oppressive tyranny of heavy mediastinal structures. This manifests as the loss of the retrocardiac hemidiaphragm with relative preservation of the left heart border:
The left lower lobe has lateral, anteromedial, superior, and posterior segments, but they usually go down together, which makes it somewhat difficult to find images of segmental collapse. Occasional unique opportunities present themselves (usually in the form of cancer), such as this posterior segmental collapse from learnradiology.com:
"Widespread alveolar opacities"
They aren't always politely lobar, are they; and in any case only obsessive people would insist for consolidation to be loyal to anatomical boundaries. Pulmonary disease is often a chaos of inhomogeneous tissue involvement, and in fact, come to think of it, that is exactly the kind of disease process that would land you in the ICU and make you sick enough to generate a truly horrific exam-worthy chest Xray. Ergo, CICM trainees should prepare to report Xrays with widespread abnormalities. And it would help to be able to report the distribution pattern:
- Subpleural (peripheral predominance)or perihilar (central predominance)
- Apical predominance vs basal predominance
- Anteroposterior or apicobasal gradient
What the hell are those?Martínez (2023)presents an absolutely delightful (though extremely busy) diagram to help guide differential diagnosis:
And of the common causes of widespread opacities, the commonest is pulmonary oedema, which has some characteristic features everyone should be familiar with:
Pulmonary oedema
Grainger (1958)orGrainger (1977)are great resources for this, but the best is surelyBarile (2020). A lot of the images that follow were extracted from this excellent paper. In summary:
- First you get upper lobe diversion
- Then you get interstitial oedema
- Then you get fluid in the fissures, Kerley lines and peribronchial cuffing
- Lastly you get alveolar oedema and "batwings"
First, the upper lobe vessels become prominent. The venous structures of the upper lobes should be subtle and thin, as these drain into the left atrium and are under low pressure under normal circumstances (being well above the atrium, constantly emptied under the effects of gravity). These act as something of a manometer, expanding and becoming prominent when the left atrial pressure increases. Another term occasionally used for this is "cephalisation".
Next, small pulmonary vessels lose their definitionon the radiograph; blurring occurs. This can be mistaken for something like ground glass. This is interstitial oedema, and is usually the first finding. It happens when the capillary pressure is modestly increased. Here, the excellent images from Barile demonstrate these changes in the same patient, before and after developing acute heart failure.
The vessel shadows, distinct and sharp in the "before" image, become both more prominent and less clearly edged, with fuzzy borders. Whereas previously they looked like branching forked vessels and could be easily identified as vascular markings, their oedematous appearance is now less clearly branching out from a centre, and more like a network of intertwining vaguely radial lines (reticulations, if you will). Barille associates this with a higher transmural pressure than what would be required to merely dilate the upper lobe vessels.
Peribronchial cuffingoccurs: the blood vessels which follow the bronchi become engorged, and thus causes a "cuffed" appearance. These cuffs are usually the most prominent around the hila. Some authors also refer to an increased artery-to-bronchus ratio, referring to the increased size of the vessels seen edge-on when compared to bronchi (the bronchi should usually be larger).
Kerley Blinesmay appear - these are linear shadowscreatedby collections of septal fluid; they tend to be most prominent near the peripheries of the lungs, because the fissures are seen side-on. Kerley described these in 1933butthe "B" letter appeared in the 1950 revision of his book, where he describe them as "short, sharp lines seen only at the bases, usually less than an inch long and running transversely out to touch the pleural margin".The "B" means nothing, they were arbitrarily the second of three linear opacities he defined.
Thickening of fissuresalso occurs for the same physiological reason; think of them as really big Kerley lines.
Kerley"A lines", "lines several inches long, rather ragged and radiating from the hilum. They do not bifurcate and they do not follow the normal branching pattern of bronchi and vessels". Barille explains these as fluid in the large interlobular lymphatics.Again,this excellent imageis offered of a patient with extremely poor cardiac function before and after developing fluid overload, but in fact these can also occur in the setting of lymphangitis carcinomatosis.
Kerley "Clines" were "fine interlacing lines giving the network appearance. It is the fine interlacing lines which have given rise to the term reticulation". Other authors describe them as "mesh-like", "polygonal" or "spider web-like". Some authors believe these are subpleural lymphatics, others believe they are superimposed B lines, yet others suggest they represent Kerley’s B lines en face."Poorly defined diffuse reticular pattern" is probably how the Fleischner society would want these to be described, as we are trying to de-eponymise the field.
"Bat Wing" oedemais alveolar oedema in a non-gravity-dependent distribution, and is usually seen in hyper-acute heart failure, eg. if there is acute mitral incompetence due to papillary muscle rupture. The oedema is alveolar, i.e. it resembles consolidation, and there may even be air bronchograms. These two images from thecommonveinare classical and require no annotation:
This classic sign was recognised very early; the first paper describing it wasHodson (1950), who remarked that "One of its characteristic and peculiarfeatures is that it does not reach to the lung periphery but stops short, leaving a clear zone betweenit and the chest wall, which is said to correspond with more or less normal lung tissue in thepost-mortem specimen". The mechanisms behind this subpleural sparing is surprisingly unknown;Gluecker et al (1999)list about four different hypotheses, including weird perihilar mucopolysaccharide accumulations and "the contractile property of alveolar septa".
Interstitial disease
It would not be in anybody's interest for the author to defraud his readership by presenting himself as an expert in chest radiology, and that is exactly the kind of person one would need to be to authoritatively teach anybody about the different radiological presentations of interstitial lung disease. For the CICM trainee, fortunately, a detailed understanding is not expected. Recognition of common patterns and the ability to identify abnormalities and describe them is the most we can expect.
In general the interstitium of the lung gives rise to linear opacities, recalling Kerleys lines. The Fleischner society defines these as
"a collection of innumerable small linear opacities that, by summation, produce an appearance resembling a net (synonym: reticulation)"
Innumerable. Without number. This implies that you should see a large number of these structures, rather than one or two, as those may actually represent normal lung markings. There is a way to characterise these further as "fine", "medium", "honeycombing" or "coarse" but these are usually used in reference to CT findings. They refer to different sized cavities, the shadows cast by walls of which are the linear threads that form the reticular "network". Because HRCT is the most helpful imaging modality for these structures the majority of modern articles deal with CT interpretation, and to find some examples and comments on chest Xray findings, one needs to delve deep into the pre-CT ages. As the result, some of what follows had to come fromKerr (1984)andJohnson Jr et al (1970). The psittacosis is fromBonello et al (2020)and the amiodarone toxicity is fromWolkove & Baltzan (2009):
"Honeycombing", which is variably thrown into the "medium reticular" or "coarse reticular" group by various authors, requires "loss of architecture and well-defined adjacent cystic structures". This is very difficult to define on a chest Xray and most normal people would wait for a CT, but the CICM trainee is usually posed with a two-minute time pressure and a chest Xray without much history. If the term is going to be used in this context, thick walls and large clear spaces should be prominent, as in this cystic fibrosis Xray from radiologykey:
Fine reticular lines are a chaos of collisions, and where they intersect (or where the image resolution begins to fail) their confluent outlines can produce the appearance of little dense blots. This is the "reticulonodular" pattern:
"the reticulonodular pattern is usually the result of the summation of points of intersection of innumerable lines, creating the effect on chest radiographs of superimposed micronodules"
So, this gives the effect of micronodules; it is often not actual micronodules. Or at least so the Fleischner society holds. From the perspective of the reader, that opinion should be regarded as sound as a papal dictum; but it is clear from reading the literature that this society is in the minority among radiology professionals. For example, the entry fromradiopedia.comseems to suggest that this pattern really exists as a separate entity apart from both reticular and nodular patterns, and moreover that it has some distinct pathological correlations ("very few diseases are confirmed to show this pattern pathologically",namelysilicosis,pulmonary sarcoidosis andlymphangitis carcinomatosis). It appears that this classification is resorted to when the author is unable to completely commit to either the totally reticular nor the wholly nodular description. Below, one histiocytosis is fromthecommonveinand the other is fromChai et al (2013).
Anyway: whether or not you believe in the uniqueness of the reticulonodular pattern, there is definitely such a thing as genuine true nodules, and these are a different definition again:
"the presence of innumerable small rounded opacities that are discrete and range in diameter from 2 to 10 mm"
To characterise them even further,
- "A micronodule is less than 3 mm in diameter"
- "A ground-glass nodule (synonym, nonsolid nodule) manifests as hazy increased attenuation in the lung that does not obliterate the bronchial and vascular margins."
- "A solid nodule has homogenous soft-tissue attenuation"
- "A part-solid nodule (synonym, semisolid nodule) consists of both groundglass and solid soft-tissue attenuation components"
It is obvious that nobody could ever expect the CICM exam candidate to be measuring the diameters of innumerable small round nodules, nor does anyone ever really trust the measurements created using an AP film with an unknown source-to-image distance. However, even an entirely disinterested reader will recognise the value of knowing that the smallest solitary lesion that can be resolved on a CXR is generally held to be 3mm on "real" analog films (Newell & Garneau, 1951), and that the resolution of digital Xrays is usually about 2000 x 2500 pixels, with each 35 × 43 cm cassette therefore having about nine pixels for every millimetre. Fine nodules are therefore the ones which are at the limits of the resolution of the image you have been provided with, and coarse ones are anything larger.
"Ground glass"
The term refers to the appearance of glass that has been incompletely opacified by grading with an abrasive, which leaves it sufficiently transparent to transmit diffuse light but sufficiently opaque to obscure all but the crudest, most general shape of anything behind. One would typically have this sort of glass in their bathroom window if they wanted to frustrate a perverted neighbour without fully closing the door on the possibility of being voyerised. The Fleischner society describes this pattern as
"an area of increased attenuation that does not completely obscure the underlying bronchial and vascular structures"
To produce such ahazy increase in opacity while preserving bronchial and vascular markings, the disease process would have to partially opacify structures which are below the level of the resolution of the image, whether alveolar or interstitial. The key feature is that whatever the process is, it does not completely displace alveolar air, and so is not as dense as "proper" consolidation. Any disease that does this will inevitably produce all kinds of other radiological findings and the ground glass often coexists with consolidation, nodules, reticular opacities and cavitating lesions. To find a "pure" example is impossible and the reader is enjoined to look for this white whale themselves.
Above, the PJP pneumonia from emcrithas enough ground glass between all the reticulonodular changes to be a satisfactory example. As with many other such findings in chest radiology, the AP or PA film is no match for the high resolution of the modern CT scan, and nobody could expect the CICM exam candidate to make a diagnosis on the basis of just the radiograph. Especially considering how broad the differentials might be; which probably goes for most of the interstitial nodular and alveolar patterns described above. Which brings us to:
"What might account for these findings?"
The worst nightmare of the stressed-out CICM exam candidate is the radiology follow-up question, where the examiner sitting next to them will enquire as to the possible clinical or diagnostic meaning of whatever imaging thing they are looking at. This might occasionally even be interpreted as a sign that they have somehow failed to spontaneously produce a diagnosis during their report. That is usually not the case; the Xray probably has a low-scoring addendum question attached to it and the trainee may be able to score some fractional marks by being able to list some differentials. One must consider these questions in the radiology viva as squandered opportunities to test image interpretation, as lists of differentials are much better interrogated in written answers and hot cases, and in any case represent a lower tier of assessment (mostly involving recall and only some minimum of analysis). Still, it would be nice to carry around a dozen differentials for easily recognised patterns.