Thoracic ultrasonography in adults and foals (Proceedings)

Thoracic ultrasonography yields information about the lung and pleural cavity heretofore unavailable to the veterinarian. Almost the entire thorax can be evaluated ultrasonographically, including the cranial mediastinal region. The side or sides of the thorax affected, as well as the precise location of lesions, can be determined in most horses because the involved lung segment is usually pleural based (exceptions are lesions located in the axial portion of the lung with no peripheral lung involvement, or a hernia in the axial portion of the diaphragm with no gastrointestinal viscera against the thoracic wall or displacing the lung dorsally). The character of pleural fluid can be determined ultrasonographically, as can the type and severity of underlying pulmonary parenchymal disease in horses with pleural effusion.

Normal structures

There is a large difference between the acoustic impedance of air and soft tissue resulting in air being a near perfect reflector of ultrasound. Therefore, the normal visceral pleural edge of the lung appears as a straight hyperechoic line with characteristic equidistant reverberation air artifacts indicating normal aeration of the pulmonary periphery. Watching the lung as the horse breathes; the visceral pleural edge of the lung is imaged gliding ventrally across the diaphragm with respiration. In most normal horses there is no pleural fluid visualized. However, small accumulations (up to 3.5 cm) of anechoic pleural fluid have been detected in clinically normal horses.

Pleural abnormalities

Pleural effusion

The most common pleural abnormality (pleural effusion) appears as an anechoic to hypoechoic space between the lung, thoracic wall, diaphragm, and heart. This fluid causes compression of normal healthy lung parenchyma (compression atelectasis), retraction of the lung towards the pulmonary hilus, and a ventral lung tip that floats in the surrounding fluid. The amount of pleural fluid in the thorax can be roughly estimated by the fluid level and the amount of pulmonary parenchymal consolidation or abscesses present. With pleural effusions, another normal structure, the pericardial-diaphragmatic ligament, a normal pleural reflection of the parietal pleura over the diaphragm and heart is imaged as a thick membrane floating in pleural fluid. This membrane runs from the thoracic side of the diaphragm over the heart and appears as a 3-6 mm thick, undulating sheet of homogeneous tissue. Similarly, scanning in the cranial mediastinum in horses with pleural effusion usually reveals sonolucent fluid with dorsal displacement of the lung. A thick echogenic band of tissue is imageable dividing the mediastinum into right and left sides.

Pleural fluid character

The sonographic pattern of pleural effusions includes anechoic, complex non-septated, and complex septated fluid. Composite fluids are complex and more echogenic than normal, containing fibrin, cellular debris, a higher cell count and total protein concentration, and gas. Anechoic sonolucent fluid represents a transudate or modified transudate with a relatively low cell count and total protein concentration. Increased echogenicity of the fluid indicates an increased cell count or total protein concentration. Blood within the pleural cavity (hemothorax), or within any body cavity often has a hypoechoic to echogenic swirling pattern, may be septated, and can usually be differentiated from a more purulent exudate by the latter's more homogeneous echogenic appearance, and the tendency for layering to occur with pyothorax. Clotting may be imaged in pleural fluid as soft, echogenic masses.

Hemangiosarcoma should always be considered in the differential diagnosis of hemothorax in adult horses, as this is one of the more common thoracic neoplasms in horses. Thoracic trauma should be considered in neonatal foals with hemothorax and the ribs carefully evaluated for fractures. An ultrasonographic evaluation of the thorax is indicated in all neonatal foals with known or suspected thoracic trauma to look for pericardial, cardiac, pulmonary or diaphragmatic injury and hemothorax or pneumothorax, as any of the preceding conditions can be life threatening. The cells and cellular debris in pyothorax are more echogenic, heavier, and in the most ventral location while the less cellular fluid or gas cap is detected dorsally. Free gas within the fluid (polymicrobullous fluid) is imaged as small, very bright hyperechoic echoes within pleural fluid with more free gas echoes imaged dorsally in the pleural fluid. The free gas echoes often adhere to the fibrinous pleural surfaces and may be detected here initially without being mixed into the pleural fluid. Free gas echoes may also be compartmentalized in only one portion of the thorax when initially imaged, but usually spread rapidly to all portions of the thorax. Free gas echoes are usually caused by an anaerobic infection within the pleural cavity.

Fibrin has a filmy to filamentous or frond-like appearance and is usually hypoechoic. Fibrin is deposited in layers or in web-like filamentous strands on the parietal and visceral pleural surfaces. Loculations between the parietal and visceral pleural surfaces of the lung, diaphragm, pericardium, and inner thoracic wall limit pleural fluid drainage. As these fibrin strands become more organized and fibrous they become more rigid and echogenic, often distorting the structures to which they are attached during one phase of respiration and possibly restricting pulmonary mechanics. This fibrin may eventually organize in the cranial mediastinum and wall this area off from the rest of the thorax, resulting in a cranial mediastinal abscess.

Pneumothorax

A gas-fluid interface is detected in horses with hydropneumothorax (pleural effusion and pneumothorax). This pneumothorax is usually caused by a bronchial-pleural fistula that occurs in horses with severe pulmonary parenchymal necrosis. The gas fluid interface can be imaged moving simultaneously in a dorsal to ventral direction with respiration, the "curtain sign", reproducing the movements of the diaphragm. The curtain sign is best visualized with pleural effusion, parenchymal consolidation or atelectasis. The lung echo changes position relative to the pleural fluid, while the dorsal free gas echo moves with pleural fluid movement and respiration. The lung is imaged floating in the pleural fluid retracted towards the pulmonary hilus. A pneumothorax without pleural effusion is more difficult to detect ultrasonographically because gas free in the pleural cavity and air within the lung have the characteristic hyperechoic reflection and reverberation artifacts with periodicity. The discrete small hypoechoic irregularities with comet tail artifacts seen in the visceral pleural surface of horses are absent at the site of the pneumothorax. To detect dorsal pneumothorax in horses without pleural effusion, the scan should begin at the most dorsal aspect of the thorax and continue ventrally, looking for a break in the characteristic reverberation air artifact. A soft tissue density echo may be detected at the site of pulmonary atelectasis between the dorsal free gas echo and the ventral air echo from aerated lung.

Non-effusive pleuritis

A dry pleuritis is more difficult to detect ultrasonographically because there is no fluid separating parietal and visceral pleural surfaces. Careful examination of the interface between the parietal and visceral pleural surfaces should be performed during inspiration and expiration, evaluating movement of the visceral pleural lung surface relative to the parietal pleural surface of the thoracic wall and diaphragm. If movement of the lung across the parietal pleural surfaces is rough or erratic, a dry pleuritis is probably present. Absence of any movement between these surfaces during respiration is also an indication of a dry pleuritis or adhesions between parietal and visceral pleural surfaces, but is occasionally seen in normal horses taking very shallow breaths.

Pulmonary abnormalities

Compression atelectasis

Compression atelectasis occurs whenever the lung parenchyma is collapsed by fluid, air, or viscera occupying space normally containing lung. The compressed lung is collapsed and smaller airways are no longer aerated, leaving this portion of lung hypoechoic (echogenicity of soft tissue). The atelectic lung is retracted towards the hilus. Linear air echoes may be imaged in larger airways and appear crowded together as they converge towards the root of the lung. Normal lung is also lighter than fluid and floats on top of and within pleural fluid.

Consolidation

The earliest sign of consolidation may be dimpling or an irregularity of the visceral pleural surface of the lung, a nonspecific change caused by nonuniform aeration of the lung periphery. Comet tail artifacts radiate from these nonaerated areas. In horses with pneumonia, sonolucent areas representing pulmonary parenchymal consolidation appear, surrounded by normally aerated portions of lung. These areas of pulmonary parenchymal consolidation usually have an irregular margin with hyperechoic artifacts deep to the lesion. The ultrasonographic diagnosis of pulmonary parenchymal consolidation is based upon the detection of sonolucent pulmonary parenchyma and visualization of one or more of the lung's anatomical features: sonographic air bronchograms, sonographic fluid bronchograms, pulmonary vessels, or scattered echogenic foci due to residual air in consolidated lung parenchyma. Sonographic air bronchograms are imaged as distinctive hyperechoic linear air echoes in sonolucent lung. These hyperechoic linear echoes correspond to the traditional air bronchograms detected on thoracic radiographs. These strongly echogenic branching lines of the air-filled bronchi converge towards the root of the lung, becoming larger as they merge. Sonographic fluid bronchograms are non-pulsatile, anechoic tubular structures that also converge towards the root of the lung, becoming larger as they converge. In contrast to air bronchograms, fluid bronchograms are only detectable sonographically, not radiographically. Although the diameter of sonographic fluid bronchograms normally decreases towards the superficial fluid alveologram, an increase in its diameter towards the periphery suggests pulmonary consolidation with fluid bronchiectasis. Sonographic fluid bronchograms can be differentiated from pulmonary vessels that are pulsatile, tubular structures that also enlarge as they converge towards the root of the lung. A large area of consolidated lung is usually wedge-shaped, poorly defined, and sonolucent. Large areas of consolidation often appear heterogeneous sonographically. The anechoic areas represent very fluid-filled or necrotic areas, hypoechoic areas are the typical consolidated areas containing fluid bronchograms, and hyperechoic areas represent air bronchograms. Hepatization of lung parenchyma occurs with severe consolidation, resulting in an ultrasonographic appearance similar to liver. Multiple small hyperechoic gas echoes in a severely consolidated or hepatized lung are suggestive of an anaerobic pneumonia. A rounded or bulging area of consolidation suggests severe consolidation, often progressing to pulmonary necrosis or abscess formation. A gelatinous-appearing lung occurs with parenchymal necrosis; the affected lung is usually sonolucent and bulging, although collapse of this area may follow. These necrotic areas then both cavitate and form an abscess or rupture into the pleural space creating a bronchial-pleural fistula. Pulmonary infarcts should be suspected when a clearly demarcated hypoechoic to echoic area of lung is imaged. The infracted area often appears more echoic than the adjacent consolidated lung and has a segmental appearance. Color flow and power Doppler ultrasound can be used to evaluate pulmonary blood flow in suspected areas of infarction.

Bronchial pleural fistula/abscess

A bronchial-pleural fistula is diagnosed ultrasonographically when the visceral pleural edge of the lung is no longer present, a cavitation is imaged involving the visceral edge of the lung and hyperechoic air echoes, and sonolucent fluid echoes can be imaged in real time moving from the gelatinous area of pulmonary necrosis into the pleural space. This results in a pneumothorax, as a bronchus communicates with the pleural space. The pneumothorax may occur with or without a concomitant pleural effusion. Horses with bronchial-pleural fistulas, if they survive, usually develop a large bronchial-pleural abscess surrounding the site of the bronchial-pleural fistula.

Pulmonary abscess

Abscesses are identified ultrasonographically in the lung by their cavitated appearance and the absence of any normal pulmonary structures (vessels or bronchi) detected within. An anechoic area lacking air or fluid bronchograms with acoustic enhancement of the wall or lung deep to the sonolucent area is the initial sonographic appearance of an abscess. Abscesses may be encapsulated with an echogenic fibrous capsule, but are more frequently imaged without any ultrasonographic evidence of encapsulation. Most abscesses are more sonolucent than the surrounding pulmonary parenchyma, but may appear more echogenic if thick purulent or caseous exudate is present. Hyperechoic free gas echoes may be imaged mixed in with the exudate, again suggesting the presence of anaerobic organisms. The material within the abscess tends to be layered with the heaviest, most echogenic debris in the most ventral portion of the abscess, followed by more sonolucent fluid in the center, with the hyperechoic gas echoes in the most dorsal portion of the abscess. The detection of a dorsal gas cap within the abscess is indicative of a bronchial communication and probable anaerobic infection. In foals with multiple Rhodococcus equi abscesses, many do involve the pulmonary periphery and therefore are detectable ultrasonographically.

Pulmonary neoplasia/granulomatous disease/pulmonary fibrosis

The detection of small multifocal sonolucent to echogenic masses distributed randomly throughout the lung is consistent with granulomatous disease or metastatic neoplasia and rarely, with primary pulmonary neoplasia or pulmonary fibrosis. These soft tissue masses are usually small and diffusely scattered throughout the lung field. The majority of neoplastic pulmonary masses are homogeneous and hypoechoic, compared to the surrounding normal lung, but may be isoechoic, or have heterogeneous echogenicity. Neoplastic masses can usually be differentiated from parenchymal consolidation by the absence of bronchial and normal vascular structures within the masses. Cystic necrotic areas or areas of dystrophic calcification casting acoustic shadowing may be imaged within neoplastic masses.

Cranial mediastinal abnormalities

Pleural effusion/cranial mediastinal abscess

If a large amount of fluid is present in the cranial mediastinum, the heart will be pushed caudally one or two ICS. With chronic complex effusions, the fluid in the cranial mediastinum may wall off and become encapsulated as an abscess, occasionally causing signs of cranial vena caval obstruction.

Cranial mediastinal neoplasia

Soft tissue masses may be imaged in the cranial mediastinum and are most common in horses with thoracic lymphosarcoma although may be detected in horses with mesothelioma or hemangiosarcoma. Lymphosarcoma masses in the cranial mediastinum are usually associated with large pleural effusions, making these large soft tissue masses easier to image. These masses usually occupy the entire cranial mediastinum, obliterating the normal thick membranous division imaged in horses with pleural effusion. The mass usually displaces the right apical lung lobe dorsally and the heart caudally.

Diaphragmatic hernias

Diaphragmatic hernias are likely to result in viscera occupying a portion of the caudal mediastinum or caudal thorax. A diaphragmatic hernia can be diagnosed ultrasonographically when viscera is imaged in the thoracic cavity immediately adjacent to the lung, or floating within pleural fluid without the diaphragm separating the thoracic and abdominal viscera.

Patient management and prognosis

The thoracic ultrasound examination can be used to help form a more accurate prognosis for survival and select appropriate treatment at the horse's initial presentation, as well as monitoring response to therapy. Survival of horses with pleuropneumonia is more likely if pleural fluid, fibrin, loculations, free gas echoes, or parenchymal necrosis are not detected on the initial ultrasonographic examination. If free gas echoes are detected in pleural fluid, a guarded to grave prognosis should be given and broad spectrum antimicrobial therapy, including coverage effective against anaerobic microorganisms (metronidazole), initiated immediately before results of culture and sensitivity testing are available. The cost effectiveness of treatment must be considered because horses with anaerobic pleuropneumonia or parenchymal necrosis are likely to require a longer period of antimicrobial treatment and are unlikely to return to their prior performance level, if they survive. The number of treatment days was longer for horses with pleuropneumonia when pleural fluid, fibrin, loculations, free gas echoes, pulmonary parenchymal necrosis, or abscesses were detected ultrasonographically at initial examination. Initial treatment duration can be estimated from the initial ultrasonographic findings and their severity.