The results of this study are promising and LUS could be a reliable method for the diagnosis of PTB. We have seen pathological findings in 79/82 (96.3%) and only in 3 patients we didn’t detect any alteration: one of these patients had only a very small area of central ground glass opacity (to be referred to alveolar hemorrhage of non-tuberculous origin), another one had only large cavitation with surrounding thickening that did not reach the pleura in the posterior segment of the Superior Upper Lobe behind the scapula and the last one had only a small thickening at the center of the lung parenchyma. Even with CXR we weren’t able to highlight findings in 4 patients: however, in all those four patients we were able to highlight pathological findings at LUS so, by adding LUS and CXR, we were able to detect pathological findings in all 82 patients analyzed. Although the two populations are similar in clinical, epidemiological and anamnestic features and CT signs, there is a greater disease severity and a lesions extension (both in CXR, LUS and CT) in PTB patients than in non-PTB ones. From this data, we can assume that the detection of multiple and diffuse micronodules at LUS (in the presence of clinical-epidemiological features), allows us to presume with good confidence the diagnosis of PTB.
Regarding lesions distribution, we have seen that TB patients had a higher distribution of lesions in the Posterior-Superior and Anterior-Superior zones; moreover, statistical correlation was also seen in the Antero-Inferior zone in relation to the failure of detecting lesions in no-PTB patients. In addition to the fact that the total number of lesions in Anterior-Inferior zone is poor, as it is not characteristic of TB lesions, this data may appear partly distorted by the fact that this region appears more difficult to explore in women due to adipose and dense glandular tissue which obscure partially the visualization of the pleural surface. Comparing LUS to CT by zones, was evident that we obtain a lower sensitivity of LUS in Posterior-Superior zone due to the presence of the scapula which limits the visualization of the pleural interface with a sensitivity of 68% (compared to the overall sensitivity of 80%). Two other zones with lower sensitivity than average turned out to be the Anterior-Inferior due to the presence of the breast in women, and the Anterior-Superior maybe due to the presence of the clavicle and to a not always easy access in the supraclavicular fossa.
The overall LUS sensitivity of 80% was found to be lower than the values described in other US studies of pulmonary infectious diseases [10, 17, 18]. This happens in our opinion for two reasons: first of all because in our study we searched systematically for all lung lesions and we made an exact comparison by lung zones between the three methodical, instead of reporting the number of patients with any lung lesion despite to lesion number or the zone they were located. Another reason is certainly that TB lesions are often localized, in a higher percentage than other infectious diseases, in the Posterior-Superior regions where many lesions do not appear visible at LUS due to the presence of the scapula. Sensitivity was found to be slightly different depending on the lesions examined (Table 5). The highest sensitivity was obtained for the ground-glass opacity/white lung even if this data was not fully reliable due to the mismatch of these findings in many regions and different patients. We have found regions of white lung at the margins of pulmonary consolidations, corresponding to small areas of perilesional ground-glass not always reported on CT, or as regions of pulmonary consolidations that reach the pleura only as partial alveolar filling (therefore with a image of the peripheral white lung).
Pleural effusions deserve a separate discussion. As described extensively in the literature, also our study confirms that LUS is more sensitive than CT in identifying pleural diseases (33 pleural effusion at LUS with 5 of them complicated versus 22 pleural effusion at CT with one of them complicated) [19, 20]. In 2 of our patients with mild and non-specific lung lesions at CT, it was LUS that suggest a high suspicion of PTB due to the high number of corpuscles and fibrinopurulent stratifications in the pleural space (Fig. 3).
The most interesting data comes from the fact that combining CXR and LUS we obtain a sensitivity that reaches 90% compared to CT. CXR and LUS can detect different findings, in such a complementary way: LUS can better identify micronodules, interlobular septal thickening, pleural thickening and effusions. These findings are visible at the pleural interface while in a methodical such as CXR, more panoramic and less sensitive for smaller lesions, they appear more difficult to identify. On the other hand, CXR is more sensitive for consolidations, nodules and cavitations (21–24). The greater panoramic view of CXR has made it possible to detect even those exclusively central densities that do not reach the pleura, while for cavitations the CXR has shown a much lower sensitivity than the CT (58%), however higher than the sensitivity demonstrated by the LUS (33%).
Our study showed had some limitations. As widely known, ultrasound examination of the chest does not identify the deepest lesions, or it can determine them only through the observation of superficial manifestations not always present. Moreover, LUS is an operator-dependent technique, and the observed findings may change based on the experience of the operator. Although, to the best of our knowledge, there is no comparative data in the literature related to the diagnosis of PTB between CT and LUS, even if it was performed on a limited number of patients. Another important limitation is the difficulty of making a confident radiological diagnosis in patients with a similar clinical-epidemiological feature. The two populations in the study were very homogeneous and the differential diagnosis was difficult also in CT. Another limit of the study was the impossibility of obtaining CXR lateral view to fully compare plain radiograph with CT and LUS without forcing us to merge CT and ultrasound findings to allow comparability.
Nevertheless, LUS examination has some important advantages: it is easier to learn, more available, safer and cheaper compared to the classic radiological techniques. Considering its low costs, LUS may be an important diagnostic tool, especially in developing countries where the incidence of PTB is higher and the delay in diagnosis represents a critical point in disease control. Furthermore, LUS does not require patient transport nor the presence of expensive CT machines in dedicated spaces, nor dedicated technicians.