The impact of computed tomography in veterinary medicine has increased during the last few years. Especially in small animals and horses CT is useful because of its routine feasibility and diagnostic reliability [23, 24]. In contrast, CT has not been established as a routine imaging method in livestock medicine so far due to high investigation costs, time-consuming and personnel-intensive procedures and limited availability of equipment. In addition, CT examination of large animals is limited by the gantry diameter of most CT devices (600 mm) as well as size and capacity of the examination table with a maximum load of 160 kg . Therefore, as described by Barbee and Allen (1990), CT examinations in horses are only carried out on the head, neck and limbs . Computed tomography in pigs as animal models for human medicine has been used in several studies [26–28].
The aim of this study was to investigate the characteristic morphological alterations of lung tissue after experimental App challenge using computed tomography in pigs. Furthermore, the diagnostic potential and limitations of CT in pigs were compared with those of digital radiography. On the basis of the collected CT findings in living animals, a quantitative scoring system for determining the lung status of diseased pigs was derived. This scoring system is of high practical importance for follow-up studies during the course of experimental challenges with App but can not be transferred to other respiratory pathogens. The CT protocol used, based on 7 mm slices and a reconstruction interval of 5 mm was found appropriate for the detection of macroscopic lung lesions attributed to App. The detection of microscopic lung changes such as interstitial pneumonia (which may be a finding in viral pneumonias) is expected to require slices as narrow as 1 mm [29, 30]. These conditions could not be achieved with the type of CT equipment and the short examination times used in our study. The examination time limit has not been determined in this study. It depends on the length of time the pigs remain anaesthetised and the speed of data capture by the CT equipment.
The clinical signs after experimental App challenge of 24 pigs ranged from nil to severe symptoms, with two affected animals requiring euthanasia due to severe dyspnoea on days 1 and 3, respectively. These results were reflected in the gross pathological findings which ranged from normal lungs to severe pleuropneumonia.
CT examination of animals in the challenged group showed ground-glass opacities and consolidations on days 7 and 21. Ground-glass opacities are caused by partial filling of the alveolar spaces, by thickening of the interstitium, partial collapse of alveoli, or an increased capillary blood volume [31, 32]. As shown in several other studies, App infections are accompanied by swelling of capillary endothelia and thickening of alveolar septa in the first hours after challenge, followed by fibrinous exudation, migration of inflammatory cells and haemorrhage [16, 17]. These pathological changes were found by CT examination in the form of hazy areas with an increase in lung density as well as more prominent bronchial and vascular outlines.
Consolidations are caused by an accumulation of transudate or exudate, which replace the alveolar air or by a collapse (atelectasis) in lobules. Therefore, they can be distinguished from ground-glass opacities because of the obscured underlying vessels and airway walls . In the present study CT-diagnosed consolidations appeared as focal, homogeneous compaction of the lung with obliteration of the vascular and bronchial walls. These changes are due to focal necrosis, edematous and widened interstitia and serofibrinous exudates. According to the cross-sectional CT images, typical lesions caused by App were characteristically distributed throughout the lungs with a higher prevalence in the diaphragmatic lobes [5, 8, 10]. A quantitative measurement of lung tissue density for characterising macroscopic and microscopic lung tissues alterations using CT is possible, however, with some limitations [33, 34]. The measurement of tissue density was used to characterise consolidations typical for App infection such as abscess-like nodules encapsulated in connective tissue. Galanski (1998) described the density value of abscesses ranging between 0 and 80 HU while pure pus has a CT value of 30 HU . Others have found connective tissue formation and necrotic processes after experimental piglet infection with Bordetella bronchiseptica and Pasteurella multocida were of high densities (0 to 150 HU) in CT scans . The average density of consolidations occurring in this study was also in the positive range of the Hounsfield scale compared to the physiological density of lung tissue which ranges from −550 HU to −950 HU . Therefore, density measurement enabled a tentative diagnosis of the presence of accumulated pus, abscess-like nodules and necrotic areas.
Density measurements in areas with non-homogeneous, partially compacted and poorly defined ground-glass opacities cannot be performed . Within these areas varying densities at the observation time occurred due to spontaneous breathing with varying inspiration depth. This variability resulted in high standard deviations of the measured values. Others have indicated that reproducible density measurements over time are only feasible at controlled inspiration depth .
Both the CT and the x-ray examinations strongly correlated with each other and with the clinical findings on days 7 and 21. However, for assessing the extent of lung alterations, a clinical scoring scheme alone is considered insufficient , due to the fact that severe tissue damage can be compensated by remaining parenchyma without clinical manifestation [1, 21]. This observation is supported by our finding that clinical scores showed only a weak correlation with gross pathological findings at the end of the experiment.
Frequencies of cases within the severity classifications by CT, x-ray and necropsy were not consistent in all instances. As shown in this and other studies, pleurisy and pleural effusions could not be sufficiently differentiated from other gross pathological changes by CT and radiographic examination [21, 39]. In some cases, false positive radiographic findings may occur especially in spontaneously breathing animals if the recordings are not carried out at the point of maximum inspiration [40–42]. False negative results may be due to the curvature of the diaphragm, hiding lung changes, including enlarged lymph nodes within the caudal lobes . On the other hand, CT examination without overlapping organs, combined with cross-sectional images and a precise knowledge of anatomical structures enables the recognition of small lung tissue abnormalities. CT score graduation and its assignment to disease severity were chosen to ensure the best concordance with the disease classes defined by the lung lesion score. Therefore, the disease severity score graduation developed in the present study should be considered a preliminary guide for the quantification of App induced lung lesions and should be further validated in a larger cohort study. CT of the thorax could also be evaluated in caesarian-derived, colostrum-deprived or in gnotobiotic piglets, because these are widely used for challenge experiments in human medicine.
Our current approach compared the CT methodology with other methods that have been established in conventional pigs. Based on our findings, it is possible to detect gross pathological lung tissue changes in living pigs, which can be applied to efficacy studies of drugs and vaccines, or pathogenesis studies at certain points of time. In the context of ethical considerations, the number of animals can be reduced, since animals do not have to be sacrificed and necropsied at different time points to monitor the course of infection.