Thoracic radiography: Pulmonary interstitial patterns (Proceedings)

The thorax is traditionally examined by a compartment approach—6 basic compartments or "spaces" plus the thoracic wall should be considered during the radiographic examination. The compartments include the mediastinum, the pleural space and four pulmonary divisions-bronchial, vascular, interstitial and alveolar. The heart is part of the mediastinal compartment, but often discussed separately. The advantages of the compartment approach over haphazard assessment of the thoracic radiograph are twofold: 1) a systematic evaluation is performed minimizing incomplete assessment and bias. Pulmonary lesions are categorized according to anatomic location. Next, gamuts or lists of diseases that might be considered for a particular anatomic location can be made and rank-ordered in the context of the signalment, history and clinical signs. Although the above paradigm is widely used in human and veterinary medicine, an alternate method of evaluation the pulmonary compartments has been proposed (Nykamp, et al).

The concept of an "interstitial pattern' is abstract because the interstitium contains numerous sub-gross structures, including capillaries, lymphatics and alveolar septae and walls. In addition, structures in other compartments, i.e. e., vessels and bronchioles, traverse the interstitial space. Pathologic interstitial patterns are due to increased opacity in the interstitium, resulting from abnormal cellular or fluid accumulation. Underinflation (expiration) can also cause interstitial pattern in a normal patient because the components of the interstitium are less widely separated by aerated lung. Classification includes structured and un-structured interstitial patterns. Un-structured patterns are typified by "veiling" and "smudging" of vascular structures while structured patterns result in formation of nodules and linear markings. Peribronchial accumulation of cells and fluid are truly in the interstitium and are included in the structured classification and referred to as cuffs or ring shadows. Interstitial patterns do not compromise the alveolar space, thus air is still present within the affected lung, unlike that seen with alveolar lesions.

Technical considerations

Three radiographic projections are recommended for evaluating the thorax. Either ventrodorsal or a dorsoventral projection and two lateral projections are made. Both left and right lateral radiographic projections are necessary because the dependent lung is compressed by overlying tissues, which reduces aeration and masks soft tissue changes such as nodules and alveolar opacities in the dependent lung. General anesthesia exaggerates the above gravity dependent effects and should be avoided when making thoracic radiographs.

Non-uniform shape causes variation in the thickness of the thorax. Therefore, an optimal radiographic exposure cannot be selected for the entire thorax-some areas will be relatively under- or overexposed on the radiographic image. This problem can be minimized with careful selection of film screen combinations and using a technique chart that produces a wide range of opacities, e.g., a long scale of contrast. This is obtained by using a high peak kilovoltage (kVp) low milliampere-second (mAs). Digital radiography virtually eliminates this problem.

The respiratory cycle affects the appearance of a thoracic radiograph. Motion is a common problem, thus short exposure times are recommended (typically less than 1/60 sec). An attempt should always be made to expose thoracic radiographs during inspiration. At peak inspiration, the airways and vasculature are widely separated by end air spaces, producing a relative "black" lung field. The compliant pulmonary vessels, especially the veins and capillary beds, are less distended compared to peak expiration because of increased intrathoracic pressure. The opposite is true on a radiograph made at peak expiration where airways and vessels are not widely separated by air spaces. The pulmonary vessels also contain more blood and are larger. These factors, along with less air being in the lungs, leads to a more opaque lung field and a 'lighter" radiograph with prominent bronchial and vascular structures. A radiograph made during expiration can give misleading information.

Interstitial diseases

• Interpretation-what are we seeing?

     o Unstructured

• Difficult to assess, often over interpreted.

• Look for smudging of vascular structures and bronchiolar walls and "haziness of the lung field"

• Unstructured patterns are due to subgross infiltrates of cells or fluid that "break up" or scatter the incident x-ray beam. Scattering produces loss of clarity of normal structures and haziness. Therefore, the actual pathologic change is not seen.

• The ventrodorsal or dorsoventral is the most useful view for making this assessment. Using the lateral view is error prone because both lungs are superimposed and may be giving a false impression of increased interstitial opacity.

• Examples

• Early edema

• Interstitial pneumonia (pneumonitis), often viral

• Diffuse pulmonary metastasis

• Hemorrhage

• Pulmonary fibrosis due to aging

     o Structured

• Structured interstitial patterns consist of nodules masses, linear or ring shadows cast by cellular and tissue accumulations that can be resolved by the x-ray beam (as opposed to scattering with unstructured patterns.).

• Most radiographic systems will only resolve structures 3 mm and larger. Focal opacities that appear small are usually due to quantum mottle or "noise" and should not be considered true lesions.

• Nodules

• Pulmonary nodules may be single, multiple or coalescing. Pulmonary nodules are usually of fluid opacity, containing cells and or fluid. However, they may occasionally communicate with airways becoming air filled cavities as with paragonimiasis or necrotic tumor nodules. These are termed cavitary lesions because their center is not solid. Fluid filled nodules are also cavitary but cannot be distinguished from tissue nodules with conventional radiography. Small 3-4 mm soft tissue nodules are termed miliary nodules and are often seen with disseminated mycosis and some neoplasia. However, dystrophic mineralization and or pulmonary osteoma formation may cause mineralized miliary opacities that are benign and should not be confused with malignancy. Radiographic differentiation of the various causes of pulmonary nodules is nearly impossible and should be discouraged.

• Examples

     o Neoplasia, primary and secondary

     o Mycosis

     o Abscess

     o Cyst

     o granuloma

• Masses

• Nodular type lesions over 5 cm are frequently referred to as masses. Most masses are thought to arise from the interstitium.

• Ring shadows

• Peribronchial infiltrates-cells or fluid accumulations produce ring shadows when viewed end on. While these shadows can be confused with bronchial wall lesions they are actually located in the interstitium although they often reflect bronchial disease.

• Linear markings

• The interstitium contains potential spaces including perivascular and peribronchial areas which may fill with exudates, cells or edema fluid and may be large enough to be seen radiographically. When this occurs, there is generally overlapping and summation producing a reticular pattern that is most often seen with mycotic pneumonia and discrete cell neoplasia.

• Pitfalls to interpretation

     o Pseudo interstitial pattern due to incomplete pulmonary expansion has already been mentioned.

     o End on vessels can give a false appearance of pulmonary nodules, this can be avoided by using two 90 degree orthogonal projections and following vessels in the area to determine the nodule in question is actually a branch of a vessel.

     o Fluid filled bronchi and tortuous pulmonary arteries can be confused with nodular disease.

• The value of the compartmental approach and the value of differentiation of interstitial form alveolar disease have been debated. Detractors say overlap of lesions from one compartment to the other is common with many diseases and thus the gamut method does not work. In addition, when alveolar disease is present, the interstitial compartment is obscured by the alveolar lesion. While the above arguments have merit, some value of this classification remains. Gamuts for the various compartments do not totally overlap. Classifying lesions to compartments will refine differential diagnosis. Identifying a true intestinal versus an alveolar lesion affects case management—airway sampling methods are less likely to provide diagnostic information with an interstitial lesion versus an alveolar lesion.

Interpretation of thoracic radiographs is challenging because of the various opacities in the thoracic cavity, particularly those associated with the lungs. Dynamics of the respiratory cycle and inherent limitations of the two dimensional radiographic image are also problematic. Nevertheless, careful radiographic technique and a systematic approach to interpretation can improve the diagnostic accuracy of the thoracic radiograph.

References

Suter, PF: Thoracic radiograph: a text atlas of thoracic diseases of the dog and cat. Westwil, Switzerland, 1984, Peter F. Suter.

Berry CR, Graham JP, Thrall DE: Interpretation paradigms for the small animal thorax. In Thrall DE, editor: Textbook of Veterinary Diagnostic Radiology 5th edition, Philadelphia, 2007, Saunders, pp 462-485.

Lamb CR, The canine and feline lung. In Thrall DE, editor: Textbook of Veterinary Diagnostic Radiology 5th edition, Philadelphia, 2007, Saunders, pp 591-608.

Pechman RD. Effect of dependency versus non dependency on lung lesion visualization. Vet Radiol and Ultrasound, 28:185-190, 1987.

Felson B. A new look at pattern recognition and diffuse pulmonary disease. Am J Roentgenol, 133:183-189, 1979.

Nykamp SE, Scirvani PV, Dykes NL: Radiographic signs of pulmonary disease: an alternative approach. Compend for Contin. Ed. Pract Vet, 24:25-35, 2002.