Why Is Cardiac Radiology So Difficult?
World Small Animal Veterinary Association World Congress Proceedings, 2004
Christopher R. Lamb, MA, VetMB, DACVR, DECVDI, MRCVS, ILTM
Department of Veterinary Clinical Sciences, The Royal Veterinary College, University of London

Not all aspects of cardiac radiology are difficult

Radiography is widely used for diagnosis of congestive heart failure because it enables non-invasive assessment of the pulmonary veins, and may be used to distinguish pulmonary oedema from other conditions that can cause similar clinical signs, such as bronchopneumonia.

In each pulmonary lobe, the lobar arteries and veins are normally equal in diameter and slightly smaller than their accompanying bronchus in an inspiratory radiograph. In a lateral radiograph, pulmonary veins are ventral to their corresponding lobar artery; in dorsoventral or ventrodorsal radiographs, pulmonary veins are medial to the corresponding lobar artery. Pulmonary congestion may be recognised radiographically when the pulmonary veins appear larger than either the corresponding lobar artery or the bronchus. When examining animals with suspected cardiac disease, the dorsoventral radiograph may be preferred to the ventrodorsal because it usually provides a clearer view of the caudal lobar vessels. Alternatively, the left lateral recumbent radiograph usually provides a good view of the right cranial lobar vessels. The right cranial lobar vessels are normally thinner than the thinnest part of the right fourth rib. In some animals with pulmonary congestion, pulmonary vessels appear to be more numerous than normal, which probably reflects enlargement of vessels that are normally too small to be clearly visualised.

Pulmonary oedema develops in stages. Initially, oedema fluid leaks into the loose tissue around pulmonary vessels and bronchi, and its radiographic appearance may mimic bronchial wall thickening. There is a tendency for oedema to collect first at the hilum, although this may be difficult to recognise radiographically because the hilar region may already have an increased opacity as a result of superimposition of enlarged vessels and the left atrium. Oedema fluid then accumulates in the alveolar septa, which become thicker, producing a hazy, diffuse interstitial pattern. Finally, fluid leaks through the epithelium of the alveolar ducts and floods the alveoli. If sufficient alveoli are flooded, the lung appears consolidated, sometimes with air bronchograms, and hence is classified radiographically as an alveolar pattern. In dogs, pulmonary oedema is usually most marked radiographically in the caudal lobes, but may affect the entire lung in severely affected individuals. Pulmonary oedema tends to obscure the heart and pulmonary vessels, making their evaluation more difficult.

Note that if an animal with cardiac failure becomes hypovolaemic, for example because of concurrent disease, the reduction in circulating blood volume may mask signs of cardiac enlargement and pulmonary congestion and its radiographs may appear normal. In such a case, rehydration may result in rapid development of pulmonary congestion and oedema.

Radiographic signs that may be observed in dogs with right-sided cardiac failure include enlarged caudal vena cava, hepatomegaly, pleural and/or peritoneal fluid. Radiography is more sensitive than physical examination for detection of pleural or peritoneal fluid. Pleural fluid is commonly observed in cats with congestive cardiac failure, but is unusual in dogs.

Radiology of animals that are not in overt cardiac failure

Radiography is often used to examine animals with non-specific clinical signs that might be associated with cardiac disease or some other condition. These non-specific signs include cough, exercise intolerance and "collapse." Interpretation of thoracic radiographs in these cases should be based on an open-minded assessment of all structures--it is important not to put too much emphasis on the cardiac silhouette alone.

Basic cardiac pathophysiology

Cardiac diseases can impose different loads upon the heart depending on their pathophysiology, and the effect on the myocardium is variable depending on the load. Diseases that impose a volume load, such as mitral insufficiency, result in eccentric hypertrophy or dilatation of cardiac chambers with a corresponding increase in the external dimensions of the heart. On the other hand, diseases that impose a pressure load, such as aortic stenosis, tend to result in concentric hypertrophy, i.e., thickening of the myocardium that encroaches upon the ventricular lumen with little or no change in the external dimensions. Hence, the type of cardiac disease that is present will determine what radiographic signs will develop.

Structural changes affecting the heart may occur gradually, sometimes over a period of years, and the rate of development of a cardiac lesion also influences the radiographic signs. For example, the most marked left atrial enlargement occurs in dogs with chronic mitral valve disease in which the left atrial wall and surrounding pericardium gradually stretch in response to a chronic moderate increase in left atrial pressure and left ventricular end-diastolic pressure. Animals with acute mitral insufficiency (e.g., as a result of ruptured chorda tendinea') can have a sudden marked increase in left atrial pressure and develop severe pulmonary oedema before significant left atrial enlargement has occurred.

Assessment of cardiac size

In animals that are not in cardiac failure, such as a puppy with a murmur identified at vaccination, radiographic examination of the heart is focused on the evaluation of cardiac size and shape. Assessment of cardiac size is usually done by comparing mentally the appearance of the cardiac silhouette in a patient with examples of normal retained in memory; however, veterinarians (including specialists) making such a subjective assessment often experience difficulty deciding if the cardiac silhouette is enlarged or misshapen. For example, there is a tendency to falsely assume there is cardiomegaly when examining puppies, brachycephalic breeds or obese dogs because these dogs usually have a relatively broad, rounded cardiac silhouette. When evaluating the heart, it may be better to compare the patient's radiographs with those of a normal dog of the same breed; however, the search for suitable comparison radiographs can be time-consuming and inconvenient.

Objective methods for evaluating the cardiac silhouette involve measurement of various cardiac dimensions and cardiothoracic ratios; however, these methods are undermined by the marked interbreed and individual variations in thoracic conformation in dogs and variations in the appearance of the heart resulting from inconsistent positioning for radiography, phase of the respiratory or cardiac cycle, and any concurrent other thoracic diseases. Similarly, a rule of thumb, such as "a normal cardiac silhouette in the dog...usually ranges from 2.5 to 3.5 times the width of intercostal spaces" is ineffective because it is too crude to be sensitive, and makes no allowance for these variations.

The vertebral heart scale (VHS) is a method for cardiac measurement that compares the dimensions of the cardiac silhouette with the length of thoracic vertebral bodies. Based on analysis of 100 dogs of various breeds, the generic normal range is 8.7-10.7. VHS measurements tend to increase in dogs with cardiac disease.

Measuring the cardiac silhouette might be expected to aid radiographic diagnosis of canine cardiac disease; however, this does not appear to be the case. For example, in a recent study observers' ability to correctly identify dogs with cardiac disease did not improve when using the VHS method compared to subjective radiographic interpretation alone. When observers changed their initial impression on the basis of a VHS measurement, it was just as likely to result in an incorrect diagnosis as a correct diagnosis.

Measuring the cardiac silhouette does not aid diagnosis of cardiac disease because there is considerable overlap in results from dogs with cardiac disease and normal dogs. This overlap occurs partly because dogs with concentric hypertrophy and those examined in the early stages of their disease may not have any significant cardiac enlargement and partly because certain breeds have relatively large-appearing hearts. Normal Boxer dogs have significantly higher mean VHS than normal dogs of other breeds and Labrador retrievers have significantly higher mean VHS than other breeds except the Boxer and the King Charles cavalier spaniel (table 1). There is also evidence that females have smaller mean VHS than males. Clearly, interbreed differences, and possibly gender, should be taken into account when interpreting the significance of a cardiac measurement.

Table 1. Results of VHS measurements on lateral thoracic radiographs of six canine breeds


Normal ranges

VHS cut-off

at cut-off


10.3-12.6 a,b,c,d,e



Labrador retriever

9.7-11.7 a,f,g,h,I



German shepherd dog

8.7-11.2 b,f,j




9.0-10.8 c,g,k



King Charles cavalier spaniel

9.9-11.7 d,h,j,k,l



Yorkshire terrier

9.0-10.5 e,i,l



The normal ranges encompass the 5th-95th percentiles.
Ranges with the same superscript are significantly different (p<0.03).

Even when using breed-specific normal VHS ranges, there is still significant overlap between normal dogs and dogs with cardiac disease. At the optimal VHS value for separation of cardiac from non-cardiac diseased dogs of each breed the accuracy is relatively low, in the range 58-83% (table 1). VHS is an inaccurate method for diagnosis of cardiac disease in Boxer dogs because of their high incidence of aortic stenosis, which tends to result in concentric hypertrophy of the left ventricle with no visible increase in the external cardiac dimension until the condition is advanced. VHS is more accurate for cardiac diagnosis in small breeds of dog that are affected frequently by mitral insufficiency, which is more likely to be recognised radiographically because it leads to eccentric hypertrophy or cardiac dilatation, both of which increase the external cardiac dimensions.

Assessment of cardiac shape

Just as described for assessment of cardiac size, veterinarians usually reach their conclusions about the shape of the cardiac silhouette on the basis of a subjective assessment. There is limited potential for use of measurements when assessing cardiac shape, although attempts have been made to distinguish left-and right-sided chamber enlargement using measurements.

Each of the cardiac chambers and great vessels contribute to the cardiac silhouette, hence enlargement of one or more of these structures may change the shape of the cardiac silhouette, sometimes being visible as a localised bulge. For example, left atrial dilatation frequently results in a bulge in the cardiac silhouette that is visible on both lateral and dorsoventral radiographs. However, in dogs with right or left ventricular enlargement there is only fair agreement between the degree of chamber enlargement as assessed subjectively by radiography and measurements made by echocardiography. This lack of agreement reflects inaccuracy in radiographic interpretation that occurs because of various factors, including individual and interbreed variations in cardiac conformation, variations in positioning for radiography, phase of the respiratory and cardiac cycles, the lack of change in external cardiac dimensions as a result of concentric thickening of the myocardium, and the tendency of the pericardium to smooth over any bulge on the surface of the heart. As a result, it is unlikely that radiographic attempts to identify enlargement of these cardiac chambers will be reliable.

Despite these limitations, many textbooks and articles on the subject of canine congenital cardiac anomalies describe their radiographic features with little emphasis on the difficulties of this assessment. Retrospective studies have described abnormal cardiac shape as a sign of enlarged cardiac chambers or great vessels in the majority of dogs with various congenital anomalies, which suggests that it should be possible to diagnose many congenital cardiac anomalies by survey radiography. An exception to this appears to be aortic stenosis, in which the majority of affected dogs have no abnormalities on survey radiographs.

In a recent study two experienced observers examined the radiographs of 57 dogs with common congenital cardiac anomalies without access to any clinical information in order to avoid biasing their interpretations. Under these conditions, the observers reached the correct diagnosis in less than 40% cases. This poor result reflects the difficulty observers had identifying shape changes that can occur in radiographs of dogs with enlarged cardiac chambers. Radiographic signs of specific cardiac chamber enlargement (or pulmonary vascular abnormalities) were recognised by both observers in only 20% instances in which they might be expected. Abnormal cardiac shape was recognised more frequently in dogs with anomalies that volume-loaded the heart than in dogs with anomalies that induced a pressure load on a cardiac chamber, which again emphasises the influence of pathophysiology on the radiographic appearance of the heart.


 Survey radiography is a useful method for diagnosis of congestive cardiac failure

 Measuring the cardiac silhouette does not usually aid diagnosis of cardiac disease in animals that are not in failure

 Survey radiography is not an accurate method for identification of congenital cardiac anomalies because the expected signs of chamber enlargement are not present in most cases


1.  Lamb CR, Tyler M, Boswood A, Skelly BJ, Cain M: Assessment of the value of the vertebral heart scale in the radiographic diagnosis of cardiac disease in dogs. Vet Record 146:687-690, 2000.

2.  Lamb CR, Wikeley H, Boswood A, Pfeiffer DU: Use of breed-specific ranges for vertebral heart scale in dogs as an aid to radiographic diagnosis of cardiac disease. Vet Record 148:707-711, 2001.

3.  Lamb CR, Boswood A, Volkman A, Connolly D: Assessment of survey radiography as a method for diagnosis of congenital cardiac diseases in dogs. J Small Anim Pract 42: 541-545, 2001.

Speaker Information
(click the speaker's name to view other papers and abstracts submitted by this speaker)

Christopher R. Lamb, MA, VetMB, DACVR, DECVDI, MRCVS, ILTM
Department of Veterinary Clinical Sciences
The Royal Veterinary College
University of London

MAIN : Diagnostic Imaging : Cardiac Radiology
Powered By VIN