The Core Curriculum: Cardiopulmonary Imaging, 1st Edition (2004)

Chapter 3. Approach to the Chest Radiograph

Especially for neophyte interpreters of radiographs, it is essential to approach the images in the proper way. For $29.95 we will send you the ancient secret approach to the chest radiograph discovered by Leonardo da Vinci. It not only improves your interpretation of chest radiographs, it helps you to win friends and influence people.

In reality, only the first sentence of the last paragraph is true. Any approach to the chest radiograph is likely to be successful if it results in a reproducible systematic scheme for evaluating radiographs. We detail one such approach that may help you to keep things organized. You are welcome to use this or to adapt it or to create your own system.

The key thing to avoid, especially as you begin to look at radiographs, is the “gestalt” approach. By gestalt we mean looking at the radiograph and hoping to receive an overall impression in a mysterious way (perhaps via radiowaves, ESP, or ESPN). We confess that as we have become more experienced at interpreting chest radiographs, we do make an initial rapid scan of a study and often come up with a quick gestalt of normal or abnormal. Even so, we supplement this initial impression with a more systematic review of the image for a more complete diagnosis. This initial gestalt is not always correct, but with experience it has been reassuring how often relatively small or subtle abnormalities on a radiograph will, in essence, announce their presence, calling attention to themselves even at the first cursory review of the image. This is presumably what experience does for a radiologist. Many years of experience do not decrease the necessity for a formal review scheme that reduces errors of carelessness. Still, it is probably nice to realize that image interpretation will not always be a purely mechanical exercise, that there will be an element of art as the image itself “talks to you.” We now present a recommended approach.

Approach to the Radiograph

The approach is based on the observation that many inexperienced film readers think of a chest radiograph as a lung radiograph. This leads to interesting errors of perception, such as missing major abnormalities completely when the lungs are normal and missing key extrapulmonary abnormalities that completely elucidate the diagnosis when the lungs are abnormal.

For that reason, we recommend saving the lungs for last. One way to make this easier to accomplish is to start on the periphery of the radiograph and work one’s way in. Thus, we recommend beginning the assessment of the chest radiograph by examining the bones and soft tissues of the chest wall and also by assessing the neck and the upper abdomen. Next come the pleural spaces bilaterally. We then skip across the lungs to examine the mediastinum, followed by the heart and great vessels. With all that out of the way, the lungs are finally assessed. A more detailed examination of this approach follows.

A Brief Consideration of Position and Projection

Before we get into the detailed approach to the radiograph, a brief discourse on position and projection is advisable. Because radiographs are generally two-dimensional representations of three-dimensional anatomy, we recommend two right-angle projections of a structure for best depiction of that structure. In the chest, this means that we try to obtain a frontal chest radiograph, either posteroanterior (PA) (x-ray beam enters the patient from behind and emerges through the front, where the x-ray cassette or digital plate awaits) or anteroposterior (AP) (beam from in front traverses the anterior thorax and emerges through the back) and a lateral radiograph. We prefer a PA radiograph when the patient’s condition permits, in part to limit the portion of the radiograph obscured by the soft tissues of the heart; anterior structures like the heart are more magnified on an AP view than on a PA view. We similarly prefer a left lateral radiograph (beam enters via the patient’s right side and emerges from the left side), which also limits the magnification of the heart. A standard chest radiograph is obtained with a distance of 72 inches from the x-ray source to the cassette or digital plate and with the patient in the upright position, taking and holding a deep inspiration.

This idealized version of reality is not always reproducible in real life. Some patients are unable to follow instructions (language barrier, decreased mentation) or unable to take a deep breath (dyspnea, pleuritic chest pain). Acutely ill patients may be unable to stand; semiupright radiographs can sometimes be obtained with the patient in bed, but supine radiographs are another potential outcome. Patients who cannot come to our department for radiographs can be examined with portable equipment but almost always in the AP projection and seldom with a lateral view. The distance from x-ray source to image may have to be less than 72 inches, resulting in magnification of the entire chest. These technical modifications have consequences with regard to visualization and diagnosis of thoracic abnormalities.

Outside the Thorax: Soft Tissues

Careful assessment of the extrathoracic soft tissues occasionally plays a critical role in the evaluation of the chest. One of the most important things to which we try to pay attention is symmetry of the soft tissues of the neck and the breasts (Fig. 3.1). This is because asymmetry may indicate prior surgery, and this may in turn indicate prior cancer.

Assess symmetry of breasts.

Assess symmetry of neck soft tissues.

Assess for soft tissue calcifications, gas, or metal.

You might believe that interest in prior cancer is based on generating a differential diagnosis for abnormalities that are subsequently detected. You would be only partly right. The differentiation of normal from abnormal is, in our opinion, the single most important job of the chest radiologist and also the single hardest job of the chest radiologist. Although malpractice attorneys paint a picture of missed abnormalities as “obvious,” such abnormalities do not come prospectively labeled (except in textbooks such as this). In many instances a question arises as to whether a finding is an overlap of normal structures, a normal variant, a benign abnormality, or a more important finding.

Pretest probability is an important factor in deciding whether a finding is real or artifact and important or unimportant. Experience teaches that pre-employment chest radiographs on patients in their twenties are virtually certainly normal. Chest radiographs on “twenty-somethings” with chest pain are almost as frequently normal. In such a setting, a questionable nodule is virtually never a real nodule, and we almost never bring up the need for follow-up radiographs or oblique radiographs or limited thin-section computed tomography or positron emission tomography.

Figure 3.1 Prior right mastectomy. Left breast outline is visible (arrows), and there is also asymmetric lucency over the right lung base laterally.

But what if the patient has been treated for cancer? Although cancer is not commonly found in such young patients, we do occasionally see breast cancer in that age group, which could result in mastectomy. We do occasionally see thyroid cancer in that age group, which could result in neck dissection. So the observation of asymmetric neck or breast soft tissues could change the pretest probability significantly and could make us read the radiograph and interpret the questionable abnormalities very differently.

Of course, the patient is certainly aware that he or she has had previous breast or thyroid cancer. Unfortunately, for studies like chest radiographs we as interpreting radiologists almost never see the patients. The referring physician is also certainly aware of this diagnosis. You may be tempted to believe it is ridiculous that a radiologist’s knowledge of such diagnoses would have to rely on his or her ability to make observations on the radiograph. All we can say is, “Just wait until you start looking at radiographic requisitions.” The best we can often hope for is clinical history that is not out-and-out incorrect or frankly misleading. Many requisitions are completed by overworked clerks who know very little medicine. If they are told to order chest radiographs, they write a history like “rule out infiltrate.” Even an accurate history that details symptoms and current clinical concerns may not include relevant past information thought not to be of current clinical concern (such as remote breast cancer).

The bottom line is that observations such as previous mastectomy are important, clinically relevant, influential in deciding whether a study is normal or abnormal, and critical to differential diagnosis of abnormalities. It is likely that many health systems will go to a computerized order entry system in the not-so-distant future. One day, it may be easy to retrieve all relevant clinical information for every ordered radiograph. Until that day, the systematic approach to the radiograph requires an assessment of the symmetry of neck and breast soft tissues.

Outside the Thorax: Abdomen

A complete consideration of abdominal diseases is outside the scope of this text. Nevertheless, some observations are relevant. Because of the centering of an upright chest radiograph and its technique, it is often far better than an upright abdominal radiograph at depicting free intraperitoneal gas (Fig. 3.2), usually under the right hemidiaphragm. Although larger amounts of gas are required, supine chest radiographs can also be evaluated for intraperitoneal gas (Fig. 3.3).

Assess lucent structures: bowel gas pattern, free intraperitoneal gas, bowel wall gas, and other extraintestinal gas (retroperitoneal, biliary, etc.).

Figure 3.2 Upright free intraperitoneal gas. The central tendon of the diaphragm is visible (arrows), and there is also visualization of inner and outer walls of left upper quadrant bowel (Rigler’s sign) (arrowheads).

Splenomegaly can be an important finding for explaining intrathoracic abnormalities. It may raise the possibility of leukemia or lymphoma, helping to explain mediastinal and hilar lymph node enlargement or thoracic opportunistic infection. In a trauma patient it may indicate splenic hematoma or laceration, tying in with left pleural effusion or pneumothorax. It may go along with hepatomegaly and/or ascites in a patient with liver disease, elucidating right pleural effusion. Splenomegaly is generally diagnosed via medial displacement of gas in the stomach and caudal displacement of gas in the splenic flexure along with mass-like soft tissue in the left upper quadrant of the abdomen (Fig. 3.4).

Assess soft tissue structures: organomegaly and soft tissue masses.

Figure 3.3 Supine free intraperitoneal gas. There is a large central lucency in the upper abdomen (arrows).

Figure 3.4 Splenomegaly. A. Posteroanterior chest x-ray: homogeneous opacity in the left upper quadrant of the abdomen (S). B. High resolution chest computed tomography: diffuse parenchymal abnormality, largely ground glass opacity and interlobular septal thickening. Young patient + splenomegaly + diffuse lung disease = storage disease; this is Niemann-Pick disease.

Although most bowel gas will be outside the field of view of a chest radiograph, inspection of the bowel contained in a chest radiograph may yield evidence of bowel obstruction, infarction, or other abnormalities. At a minimum, this can lead the alert chest radiologist to recommend standard abdominal radiographs for better evaluation. Similarly, patients with lower thoracic or shoulder pain may actually be experiencing referred pain from upper abdominal disease. Such abnormalities may result in mass effect on upper abdominal bowel.

Outside the Thorax: Bones

As with abdominal findings, skeletal findings could constitute a separate textbook entirely. Nevertheless, a few observations are in order. One is that it is very difficult to detect skeletal abnormality without careful examination of individual bones; look at the trees, forget the forest. Especially when there is chest pain, try to compare the ribs from side to side, one by one. Trauma is another setting where painstaking inspection of visualized bones can prevent important mistakes.

It is interesting to note that the absence of a structure usually seen may be much harder to detect than the presence of abnormality of that structure (Fig. 3.5). Similarly, inexperienced interpreters may have difficulty distinguishing between variants of normal and abnormalities. Side-to-side comparison is particularly helpful in these settings; unlike abnormalities, normal variants are often bilateral and are reasonably symmetric.

Assess opaque structures: skeletal abnormalities, calculi, and metallic foreign bodies.

Compare for symmetry of bilateral structures.

If there are two views, assess both (sternum and spine better seen on lateral view).

With abnormalities at the periphery of the chest, it may be difficult to distinguish between those originating in the pleura and those arising in the chest wall (extrapleural). Shape of the lesion may be helpful (Chapter 17), but adjacent skeletal abnormality is absolute proof that the lesion is in the extrapleural space.

Pleura

The pleural spaces are potential spaces at the periphery of the lungs. For the purposes of this chapter the key pleural abnormalities we want to detect are pneumothorax and pleural effusion. Both illustrate the importance of patient position on the radiographic appearance of abnormality. Air and fluid in the pleura migrate in opposite directions; fluid moves to the dependent aspect of the space, whereas air moves to the nondependent aspect. With a patient upright, we look for pneumothorax at the thoracic apex and for pleural effusion in the caudal hemithorax (Fig. 3.6). However, many patients being examined for these findings are hospitalized critically ill patients. They may be unable to assume the upright position, in which case fluid will spread out along the posterior aspect of the chest (Fig. 3.7) and air will rise to the anterior aspect of the chest. Leading to more confusion, such patients may be quickly propped upright for a radiograph without allowing air or fluid the few minutes it may take to redistribute as expected.

Figure 3.5 Asymmetric bones. A. Posteroanterior chest x-ray: what is missing? B. Earlier posteroanterior chest x-ray in same patient: expansile neoplasm of distal right clavicle (E) is Ewing sarcoma and right clavicle is surgically absent in A.

Evaluate all the way to the edge of the radiograph.

Fortunately, a solution is available. Even very ill patients are often able to assume the lateral decubitus position. For suspected pleural effusion, we put the side of presumed abnormality down to get fluid to layer along the inner aspect of the ribs (Fig. 3.7). For suspected pneumothorax we put the side of presumed abnormality up to get air to rise under the inner aspect of the ribs. This also works for free intraperitoneal gas. In that setting we do a left side down decubitus radiograph so that intraperitoneal gas rises over the hepatic margin.

Concentrate on thoracic apex and base.

The key observation in the diagnosis of pneumothorax is visualization of the lung edge. When the potential pleural space contains no gas or fluid, the lung edge is so closely applied to the inner aspect of the chest wall that it cannot be visualized. Gas in the pleural space displaces this edge and provides good contrast for its visualization (the gas is black, the edge is white). Especially in supine patients (but even sometimes in upright patients) we need to look for gas in the caudal aspect of the chest, the least dependent portion of the pleural space in the supine position (Fig. 3.8).

Supplement liberally with decubitus radiography.

 

Pleural effusion classically produces a “meniscus” appearance of soft tissue opacity (fluid), displacing the lung away from the lateral and posterior costophrenic sulci (Fig. 3.6A). Because the posterior costophrenic sulcus is the most dependent portion of the upright pleural space, blunting of the normally sharp posterior costophrenic angle on the lateral radiograph is the best indicator of a small effusion. On the frontal radiograph the lateral costophrenic sulcus is often evaluated, but subpulmonic effusion can leave that angle unblunted (Fig. 3.9). In such cases we need to look for three signs of effusion: uniform opacity of the ipsilateral hemidiaphragm, instead of the usual gradation of opacity from very white in the upper abdomen to very dark in the lower lung; loss of visualization of the posterior lung vessels through the hemidiaphragm on the frontal radiograph; and displacement of the apparent dome of the hemidiaphragm, usually found one-third of the way from the midline to the lateral chest wall, more laterally.

Figure 3.6 Upright right pleural effusion. A. Posteroanterior chest x-ray: typical appearance of effusion (E), with meniscus appearance in right lateral costophrenic angle (arrows)B. Posteroanterior chest x-ray after thoracentesis: there is now an air-fluid level (arrows), indicating hydropneumothorax. The right lung edge is visible in the more cephalad pleural space (arrowheads).

Figure 3.7 Supine right pleural effusion. A. Anteroposterior chest x-ray: hazy increased opacity throughout the right hemithorax, despite which right lung vessels are easily visualized. B. Right lateral decubitus radiography: effusion (E) layers between the lung edge(arrows) and the inner margins of the ribs (arrowheads).

Figure 3.8 Supine right pneumothorax. A. Anteroposterior chest x-ray: lucency (L) at the base of the right hemithorax and overlying the right upper quadrant of the abdomen. B. Anteroposterior chest x-ray after chest tube placement: lucency has resolved.

Figure 3.9 Subpulmonic right pleural effusion. A. Anteroposterior chest x-ray: right hemidiaphragm is uniformly opaque, with lateral displacement of apparent dome (arrow)B. Anteroposterior chest x-ray after effusion resolved: normal gradation of diaphragmatic opacity and more medial diaphragmatic dome.

Mediastinum

Skipping past the lungs, we evaluate the midline space between the two lungs (the mediastinum). One important indicator of mediastinal abnormality is widening of the mediastinum. It would be very helpful to identify as abnormal any mediastinum wider than a given measurement. Unfortunately, it is very difficult to come up with such a measurement. Mediastinal width depends on projection (wider on AP views than PA views), x-ray source to image distance (wider with shorter distance), depth of inspiration (wider with poorer inspiration), patient rotation (the aorta is more prominent when the patient is rotated in the left anterior oblique projection), and patient age (atherosclerotic unwinding of the aorta, common with increasing patient age, accentuates mediastinal width), among other factors. Although it is sometimes obvious that a mediastinum is too wide (Fig. 3.10), that is not always the case. As in most cases of trying to distinguish normal from abnormal, experience is very helpful (another way of saying that you can expect to be poor at making such distinctions when you first begin to interpret chest radiographs).

Look carefully for the lung edge.

Mediastinal contours may also be helpful. Focal contour bulges can be important clues to the presence of mediastinal abnormality. Such contour abnormalities may be surprisingly subtle (Fig. 3.11), and bitter experience teaches that mediastinal masses of astounding size are sometimes virtually invisible.

Assess normal contours.

It can be useful to evaluate normal mediastinal structures for displacement, which may allow diagnosis of noncontour deforming mediastinal abnormalities. The trachea is usually in the midline in the upper thorax (Fig. 3.12) and usually deviates slightly to the right at the level of the aortic arch. Rightward deviation tends to increase with atherosclerotic aortic disease. Tracheal deviation can be simulated by patient rotation; this is best evaluated by comparing the position of the medial clavicles (anterior structures) against that of the spinous processes of the vertebrae (posterior structures). The clavicles should be equidistant from the spinous processes. Head turning can also be misleading and is evaluated by following the spinous processes up into the neck and by assessing chin position. Similarly, evaluating a contrast-filled esophagus for deviation may allow diagnosis of an aberrant right subclavian artery coursing between the esophagus and the spine; other periesophageal abnormalities can be similarly demonstrated, but oral contrast is virtually always required.

Evaluate for deviation of normal structures.

Figure 3.10 Unequivocal mediastinal widening (arrows). This should call to mind lymphoma or metastatic disease but actually represents amyloidosis.

Figure 3.11 Subtle contour abnormalities. A. Posteroanterior chest x-ray: small bump (arrow) over aortopulmonary window. B. Lateral chest x-ray: surprisingly large anterior mediastinal mass (T) is thymoma. C. Posteroanterior chest x-ray in a different patient: soft tissue mass at apex of left hemithorax (N). D. Apical lordotic radiography demonstrates mass (N), a neurofibroma, to better advantage.

Figure 3.12 Mass deviating trachea (arrows). Typical appearance of a goiter.

Cardiopericardial Silhouette

Analysis of the mediastinum readily leads to evaluation of the heart. Note the intentional use of the term “cardiopericardial silhouette”; this reminds us that enlargement of this structure may be seen with a normal heart but with significant pericardial abnormality. In some instances a large pericardial effusion has a distinctive appearance (described as a “water bottle” shape) (Fig. 3.13), but the differentiation of cardiomegaly from pericardial effusion by shape is seldom easy. Pericardial effusion deserves special consideration when the cardiopericardial silhouette enlarges rapidly (Fig. 3.14) or when catheters or pacemakers are displaced from the apparent edge of the silhouette (Fig. 3.15).

Figure 3.13 Pericardial effusion with classic “water bottle” configuration.

Assess for calcifications, gas, or metal.

Cardiomegaly is also a diagnostic challenge. As a rough rule of thumb we often measure the maximum transverse cardiac diameter and compare it with the maximum transverse thoracic diameter; a ratio of greater than 0.5 is used to indicate cardiomegaly. Inexperienced observers tend to rely on this parameter; those of us with more experience have been disappointed too often. The cardiothoracic ratio is influenced by rotation, position, and projection; the heart will be magnified on an AP view of the chest, and the cardiac apex may project further laterally on a relatively lordotic view of the chest. The transverse cardiac diameter also appears larger on an expiratory view or in diastole. It is preferable to use both views (when a frontal and lateral are obtained) to get a handle on cardiac size. Even then, a research project in the 1980s convinced us that experienced chest radiologists looking at radiographs in patients who had concurrent echocardiography were not reliably able to distinguish normal from abnormal heart size. If there is a question about cardiac size, presence or absence of pericardial effusion, or even ejection fraction, echocardiography amplifies our diagnostic capabilities enormously.

Assess size and shape (two views are helpful for both).

It is also important to view the cardiac margins for evidence of specific chamber enlargement. The normal right heart border is formed by the right atrium and the superior vena cava, with a variable contribution from the aging ascending aorta. The normal left heart border is formed by the aortic knob, the main and left pulmonary artery, and the left ventricle. Even when the overall transverse cardiac diameter is normal, the so-called fourth mogul (an additional left heart contour) and the double density sign below the right hilum may indicate left atrial enlargement (Fig. 3.16). Other signs of left atrial enlargement include splaying of the carina, increased subcarinal opacity, and posterior displacement of the mid-portion of the heart border on the lateral radiograph. Left ventricular enlargement results in more caudal posterior displacement of the heart border on the lateral view. This may be particularly evident when comparing the position of the posterior heart border with that of the inferior vena cava. Although left atrial enlargement has useful signs, we remain unimpressed by the ability to predict specific chamber enlargement (based on the 1980s research project discussed previously).

Assess normal contours.

Figure 3.14 Pericardial effusion. A. Anteroposterior chest x-ray 12-20: typical postsurgical findings after mitral valve repair. B.Anteroposterior chest x-ray 12-30: larger cardiopericardial silhouette; patient in distress. Echocardiography confirmed pericardial effusion, clinically causing pericardial tamponade.

Figure 3.15 Pericardial effusion diagnosed by catheter position. A. Anteroposterior chest x-ray in patient without pericardial effusion: enlarged cardiopericardial silhouette reflects known cardiomyopathy, but pulmonary artery catheter (arrows) is not deviated from right cardiac margin. B. Anteroposterior chest x-ray in a different patient with large pericardial effusion. Note displacement of pulmonary artery catheter (arrows).

Figure 3.16 Cardiomegaly with left atrial enlargement. Diagnostic signs include fourth mogul (arrow), increased subcarinal opacity (O), and double density sign (arrowheads).

Intracardiac calcifications may indicate cardiac abnormality even in the absence of cardiopericardial silhouette enlargement (Chapter 20). When frontal and lateral views are available, differentiating between calcified aortic and mitral valves is not generally challenging. However, on isolated frontal views (especially in postsurgical patients) this may be surprisingly difficult (Fig. 3.17).

Assess for calcifications.

Figure 3.17 Confusing location of prosthetic valves on anteroposterior chest x-ray. Patient with previous placement of prosthetic mitral and aortic valves (V), which are completely superimposed.

Great Vessels

As noted above, aortic size and configuration change with age. Gross enlargement (aneurysm) may be evident but is generally better evaluated by thoracic computed tomography or magnetic resonance imaging. Nevertheless, it is important to evaluate the aorta for certain particular signs of aortic dissection, especially in patients with chest pain or pain radiating to the back. Dissection may result in an irregular aortic margin (Fig. 3.18). It occasionally results in displacement of intimal calcification away from the aortic edge, although this sign is difficult to evaluate with chest radiographs because apparent displacement may only reflect our two-dimensional representation of a structure moving through space both from front to back and from right to left. Many patients with aortic dissections have a normal chest x-ray appearance of the aorta, reminding us that with appropriate symptomatology, any chest x-ray is suspicious for dissection.

Assess aortic contour, size, and calcification.

Evaluation of pulmonary vascular abnormalities is covered at greater length in Chapter 21. In this setting we consider the diagnosis of congestive heart failure in very simplistic terms. Although pulmonary vascular redistribution (relative enlargement of upper lobe pulmonary veins, which are usually smaller than lower lobe veins because of gravity) can be an important sign of pulmonary venous hypertension in ambulatory patients, those are not the patients we are usually asked to assess for congestive heart failure. Pulmonary vascular redistribution is not helpful in patients who spend most of their time supine or semisupine, where the gravitational effect makes posterior veins bigger than anterior veins (on a frontal radiograph, who knows which are which?). We are therefore far more concerned with signs of interstitial pulmonary edema, such as pulmonary vascular indistinctness, fissural thickening, and septal lines (Fig. 3.19); pleural effusion can be an important accompanying finding. It is a mistake to think of fissural thickening as being a result of pleural effusion. Instead, it is generally a manifestation of subpleural interstitial edema; when subpleural lung on either side of a fissure develops edema, we do not resolve this as a perifissural abnormality but as fissural thickening.

Assess distinctness of pulmonary vessels and other signs of interstitial pulmonary arteries.

Assess size and symmetry of central pulmonary arteries.

Assess size and symmetry of peripheral pulmonary arteries.

Figure 3.18 Aortic dissection. A. Posteroanterior chest x-ray: descending aorta is enlarged and irregular in contour (arrows)B.Posteroanterior chest x-ray in a different patient: descending aortic contour irregularity is more subtle.

Figure 3.19 Interstitial pulmonary edema. A. Posteroanterior chest x-ray with patient in congestive heart failure: peribronchial cuffing (C), fissural thickening (arrows), and vascular indistinctness. B. Posteroanterior chest x-ray after heart failure resolved: resolution of findings, with return of bronchial wall thickness to normal (arrows).

Lungs

After you have waited so patiently for so long, here is disappointing news: You have to read most of the rest of the book to learn how to diagnose lung disease at chest radiography (you probably thought that one chapter would do the trick, didn’t you?). For the beginning observer, it is useful to think of lung diseases as focal or diffuse and to look at focal diseases as nodular or non-nodular and diffuse diseases as alveolar or interstitial. Nodular or non-nodular is relatively easy: Does it look like a mass? Does it look like it has well-defined margins? If you had tweezers or tongs, could you easily lift it out of the chest? Would you even want to?

Compare side to side.

Assess for focal or diffuse lucency or opacity.

Inspect closely but also from afar.

Supplement with additional views (oblique, lordotic) as needed.

As for alveolar and interstitial disease, their characteristics are described in Chapter 14. Instead of describing specific lung abnormalities here, we conclude this section with a description of selected signs in chest radiography:

Silhouette Sign

Loss of visualization of a normal anatomic landmark (such as the left hemidiaphragm) indicates an abnormality in the immediately adjacent lung (in this instance, the left lower lobe) that would otherwise outline that landmark. Similarly, loss of visualization of the right heart border indicates disease in the medial segment of the right middle lobe (Fig. 3.20). Not all silhouette signs are equally reliable; disease that overlaps both the right heart border and the right hemidiaphragm and silhouettes neither is in the right lower lobe. The longer contact between right lung and hemidiaphragm than between left lung and hemidiaphragm, because of the heart, means that right lower lobe disease will not always silhouette the right hemidiaphragm.

Hilum Overlay Sign

In differentiating an enlarged main, left, or right pulmonary artery from an overlying mediastinal mass, the demonstration of pulmonary vessels converging medial to the apparent lateral margin of the “mass” indicates that it is not the artery (Fig. 3.21). The converse, the hilum convergence sign, suggests that when pulmonary vessels converge to the edge of a “mass” (but not medial to it), the mass is pulmonary artery.

Figure 3.20 Silhouette sign. A. Posteroanterior chest x-ray: right heart border is not visible (arrows)B. Lateral chest x-ray: middle lobe air-space disease (A) is responsible.

Figure 3.21 Hilum overlay sign. A. Posteroanterior chest x-ray: mass (M) overlies main and left pulmonary artery, but arteries converge medial to edge of mass (arrows)B. Lateral chest x-ray: mass is teratoma (T) anterior to right (R) and left (L) pulmonary arteries.

Air Bronchogram

When alveoli are filled with blood, pus, water, or cells, air-filled bronchi may stick out as branching black structures against a white backdrop (Fig. 3.22); otherwise, the walls of bronchi are generally too thin to allow demonstration of the bronchi. Air bronchograms thus indicate that a given abnormality is alveolar; in addition, they prove that an abnormality on the chest radiograph is parenchymal (e.g., not pleural or extrapleural).

Figure 3.22 Air bronchograms. Branching lucencies (arrows) surrounded by right upper lobe airspace disease that represents pneumonia.

Superior Triangle Sign, Inferior Triangle Sign, and Luftsichel

These are signs of lobar atelectasis or collapse. When the right lower lobe collapses, the superior vena cava may rotate laterally, producing a triangular configuration (superior triangle sign). In contradistinction, upper lobe collapse may produce a triangular configuration at the diaphragmatic dome (inferior triangle sign). With left upper lobe collapse, the hyperinflated left lower lobe adjacent to the aortic arch results in a crescent-like configuration known by its German name (luftsichel) (Fig. 3.23).

Golden S Sign

A collapsed lobe occurring in the absence of a central mass often has a concave border. When there is a central mass associated with atelectasis there is often a resultant focal convexity that results in an “S” shape (or more often, a backward S) (Fig. 3.24).

Comet-Tail Sign

Rounded atelectasis is a mass-like lesion of collapsed lung, typically occurring after a pleural effusion or other pleural abnormality. Lung seemingly becomes adherent to the adjacent pleural abnormality, so that when pleural disease recedes, the lung may twist into a rounded configuration. The diagnosis is aided considerably by demonstration of vessels and bronchi that supply the atelectatic portion of lung “swirling” into the lung with a comet-tail configuration (Fig. 3.25).

Cervicothoracic and Thoracoabdominal Signs

Because there is no anterior lung above the level of the clavicles, a lesion that extends above that level and remains sharply outlined must be located posteriorly in the chest. This does not apply only to lesions in the lung, because this sign can be seen with posterior mediastinal and pleural lesions as well. A lesion that simply disappears at the level of the clavicle is generally anterior. This is the cervicothoracic sign. Similarly, the posterior costophrenic sulcus of lung extends far more caudally than anterior basilar lung. Therefore, a lesion that extends below the dome of the diaphragm and remains sharply outlined must be in the posterior chest, whereas a lesion that abruptly terminates at the dome of the diaphragm must be anterior (thoracoabdominal sign).

Figure 3.23 Luftsichel in left upper lobe collapse. A. Posteroanterior chest x-ray: luftsichel (arrows) outlines aortic knob. B. Computed tomography: collapsed left upper lobe (C) with sliver of expanded left lower lobe (arrows) insinuating adjacent to posterior aortic arch(A).

Figure 3.24 Golden S sign. Central convexity (M) results in backward S shape of elevated minor fissure (arrows).

Figure 3.25 Comet-tail sign. Vessels and bronchi (arrows) swirl into area of round atelectasis (A).

Approach to the Radiograph in the Picture Archiving and communication system Era

In general, interpreting chest images on monitors instead of chest radiographs on alternators necessitates no change in the approach to the study. A reproducible systematic approach remains of the utmost importance. However, experience with viewing studies on the monitor leads to the following recommendation: Always take the time to view digital images in a magnified mode. This tends to slow the interpreter down (at least a little), but there are definitely lesions that cannot be seen without magnifying the image. Do not use optical magnification (magnifying glass in many systems), because it results in visual distortion.

An added advantage of viewing images this way is that it automatically gets the viewer out of gestalt mode, requiring immersion in the local details of the image. It also forces one to scrutinize the entire image carefully. We do not only view the image in a magnified form. We liken this to the practice (when viewing radiographs on alternators) of leaning forward and looking closely at the films, using a high intensity light as necessary, but also leaning back in an effort to see the forest for the trees before signing off on a case. The magnified image at the monitor is the correlate of leaning forward, but a final survey of the restored original image is like leaning back for the overview.

One other helpful tool for viewing images at the monitor is an image-sharpening algorithm (bone reconstruction or large edge or the like). This is very helpful for questionable pneumothorax and for location of a difficult-to-visualize catheter tip.