The role of imaging in the assessment of emphysema has assumed increased importance since the advent of surgery as one of the potential therapeutic options. The chest radiograph had tended to be dismissed because it did not correlate well with pulmonary function tests, but pulmonary function tests may be normal in mild emphysema.¹ The chest radiograph may be normal too, but more often in chronic bronchitis than in true emphysema. Signs of hyperinflation include a flattened diaphragm, particularly one depressed to the level of the seventh rib anteriorly or below, while on the lateral film there may be an increased anteroposterior diameter of the chest and increased retrosternal and retrocardiac lucency. Other signs of emphysema include peripheral pruning of vessels, although this is lost with the onset of cor pulmonale when the vessels appear to become more numerous and larger.^2 3

Ventilation and perfusion isotope scanning have been used for some years to assess patients with limited cardiorespiratory reserve undergoing lung resection for carcinoma, when the percentage contribution of different areas of the lung can be calculated to predict postoperative lung function. As perfusion scanning alone correlates with pulmonary function tests, the ventilation scan may not be required.⁴ More recent studies^5 6 have shown that single photon emission computed tomographic (SPECT) scanning, possibly with surface rendered images, can provide further detail in patients being assessed for lung volume reduction surgery (LVRS).

Computed tomographic (CT) scanning has advanced the radiographic investigation of emphysema by demonstrating areas of low attenuation along with reduction in the vessels. Different thresholds have been used to define the level below which emphysema is said to be present, ranging from -900 HU to -960 HU. CT scans also give information on the type of emphysema. Paraseptal emphysema is represented by a series of thin walled cysts that affect the peripheral 1-2 cm of the lung. This type of emphysema does not usually affect respiratory function, except that these thin walled cysts may enlarge to become significant bullae. Spontaneous pneumothorax may occur. In comparison, centrilobular (CLE) and panlobular emphysema (PLE) do affect lung function once a significant proportion of the lung is affected. CLE is the more common type and tends to occur in the upper third of the lungs, while PLE occurs more commonly in the lower third of the lung. Alpha-1-antitrypsin deficiency is one of the causes of PLE. In the early stages of CLE involvement of the central portion of the lobule is seen which progresses to involve the entire lobule as the disease becomes worse. PLE affects the whole lobule and mild cases may be difficult to distinguish from normal lung, but as it progresses the low attenuation areas and the reduction in vessels become more obvious. Recent studies have correlated the distribution of emphysema with their effect on lung function and in this issue of Thorax Nakano et al⁷ have contributed further to this knowledge.

In assessing patients for LVRS several papers have correlated the findings of preoperative radiology with outcomes.^8-11 Hyperinflation must be present and this is best shown by the plain film. Factors favouring a good outcome include upper lobe emphysema and marked heterogeneity in the pattern of emphysema, although Gierada et al⁹ emphasise that quantification of heterogeneity is difficult. Whilst uniformly severe emphysema represents low heterogeneity, and large bullae with normal lungs elsewhere represent high heterogeneity, in between these extremes visual assessment is subjective and more quantitative work of the type performed by Nakano et al is required. Part of the reason why upper lobe emphysema is more favourable in LVRS is that it has been shown, by a comparison of HRCT scanning with pulmonary function tests, that lower zone emphysema affects lung function more than upper zone emphysema.¹² Nakano et al found that the inner half of the lung is more often affected by emphysema than the outer half, and that abnormality of the inner segment may have a greater effect on lung function. This could usefully be studied in relation to outcomes in LVRS. Whilst reviewing HRCT scanning, negative factors affecting outcomes such as pleural disease and bronchiectasis should be noted. The HRCT scan should be supplemented by either consecutive thick section or spiral CT scans to search for lung nodules. In 148 patients Rozenshtein et al¹³ found pulmonary nodules in 11% of their patients, and just under half of these (5%) were found to be stage 1 lung cancers. These nodules can be resected at the time of LVRS.

Radiographic assessment is only part of the work up of patients with emphysema for LVRS. Nevertheless, the role of imaging is emphasised by these recent papers correlating outcomes of surgery with radiographic findings. By further subdividing the lung into inner and outer segments the work by Nakano et al may in time further refine the role of imaging in patient selection for LVRS.