Radial endobronchial ultrasound with a guide sheath for diagnosis of peripheral cavitary lung lesions: a retrospective study

Although EBUS-GS has improved the diagnostic yield for peripheral lung lesions, reports on the use of EBUS-GS specifically for cavitary lesions are unavailable. To the best of our knowledge, this is the first report on the efficacy and safety of EBUS-GS for diagnosis of PCLLs. In addition, the utility of the practical sampling method of EBUS-GS for malignant and benign PCLLs was carefully evaluated in this study. Two important clinical observations were made. First, EBUS-GS had a high diagnostic yield (80 %) for PCLLs with a low risk of significant complications (2.0 %). Second, the EBUS-GS sampling method combined with cytologic, histopathologic, and microbiologic samples was effective for diagnosis of both benign and malignant PCLLs.

A cavity is the result of such pathologic processes as necrosis, cystic dilation of pulmonary structures, and displacement of lung tissue by cystic structures [16]. In addition, malignant lesions can become hollowed out due to treatment-related necrosis, internal cyst formation, or internal desquamation of tumour cells [17]. Active infectious and malignant diseases require a prompt and accurate diagnosis to minimize adverse outcomes and apply appropriate management. Although thin-section chest CT findings, such as cavity wall thickness, are helpful for diagnosis of cavitary lesions, sufficient pathologic or microbiologic samples are often necessary to ensure correct diagnosis [16]. TTNA is a useful nonsurgical method for examination of peripheral lung lesions, but its risk of complication is generally higher than TBB, and thus, it may have concerns about safety for cavitary lesions because of internal necrosis and inflammatory reactions [18, 19]. Actually, Zhuang et al. reported that although TTNA had an accuracy of 96.1 % for diagnosis of 102 cavitary lung lesions, complications included pneumothorax (8.8 %), alveolar haemorrhage (13.7 %), and hemoptysis (1 %) [20].

Currently, EBUS-GS is becoming a standard procedure for diagnosis of peripheral lung lesions based on its useful and safe profile. Therefore, confirmation of its diagnostic utility for PCLLs is also demanded in clinical practice. In the present study, we first reported a high diagnostic yield (80 %) for EBUS-GS for PCLLs. This result is comparable to the diagnostic yield for R-EBUS for peripheral lung lesions reported in recent studies. Therefore, EBUS-GS can be considered as a standard procedure for diagnosis of peripheral lung lesions and PCLLs [9, 21‐23]. Moreover, we demonstrated that the diagnostic yield was better when the R-EBUS probe was within the lesion, which has been similarly shown in previous studies [24‐26]. However, in the present study, no other important factors, such as bronchus sign, lesion size, or cavity wall thickness, were found to be associated with increased diagnostic yield, despite the small sample size. Especially in larger lesions, the inflammatory reaction around the cavity wall might interfere with obtaining diagnostic samples even if bronchus sign is positive. A previous study on the use of TTNA for diagnosis of cavitary lung lesions reported that more nondiagnostic samples were obtained from lesions with thin walls [20]. Repeated biopsies at precise locations might contribute to accurate diagnosis even for PCLLs with thin walls, as shown in the representative case.

The complication rate in the present study was 2.0 % (pulmonary infection), which is lower than that reported in previous studies on the use of TTNA for diagnosis of cavitary lung lesions [3, 20]. There was no case of pneumothorax or significant haemorrhage, which are major complications of TBB for peripheral lung lesions. Precise localisation by R-EBUS and the tamponade effect of the guide sheath might reduce the risk of complications, even for PCLLs. Although TBB using R-EBUS has been reported as a safe procedure for diagnosis of peripheral lung lesions, a previous study showed that among 965 peripheral lung lesions examined by using EBUS-GS, pulmonary infection occurred in 5 patients (0.5 %), 2 of which had PCLLs [8]. Therefore, although EBUS-GS was found to be safe for diagnosis of PCLLs in the present study, we should be mindful of the risk of pulmonary infection.

This study also demonstrated that the EBUS-GS sampling method combined with cytologic, histopathologic, and microbiologic samples was effective for diagnosis of both benign and malignant PCLLs. In the present study, diagnostic tissue samples were obtained by using TBB for 74.1 % of malignant lesions and 69.6 % of benign lesions. Recently, histopathologic diagnosis of lung cancer is becoming important for definitive treatment including chemotherapy and tyrosine kinase inhibitors [27]. Moreover, because cavitary lung cancer has a worse prognosis than non-cavitary lung cancer, successful histopathologic diagnosis using a safe technique is demanded [28]. This should also be applied to benign cavitary lesions. Chan et al. recently reported that necrotizing granulomatous inflammation was demonstrated by using EBUS-GS in 10 of 22 patients with pulmonary TB, and that histopathologic diagnosis of TB might contribute to prompt treatment [9]. Certainly, histopathologic diagnosis is necessary for appropriate treatment not only for malignant lesions but also for benign lesions. Even for PCLLs, EBUS-GS has the advantages of locating lesions accurately, controlling bleeding, and obtaining adequate tissue samples.

In the present study, microbiologic samples demonstrated a diagnostic yield of 47.8 % for benign diseases; the diagnostic yield was particularly high for mycobacterium infection, which is the main cause of a cavitary lung lesion (100 % in TB, 66.7 % in NTM). Microbiology results were negative for infectious cases with lung abscess or Aspergillosis, though anaerobic or Aspergillus culture was not performed routinely. Because the culture sensitivities of these diseases using bronchoscopy samples are not satisfactory [29, 30], routine bronchial wash or bronchoalveolar lavage after removing the guide sheath may not be necessary. Instead, obtaining adequate tissue samples can contribute to definitive treatment. Although the guide sheath has to be introduced into the lesion for adequate tissue sampling, the R-EBUS probe was located within the lesion in only 26 patients (52 %) in the present study, despite a positive bronchus sign in the majority of cases (82 %). Although this may be due to a lack of skills in the trainees, additional TBNA may be important for improved diagnosis of PCLLs, especially when they are not located by using R-EBUS [13, 31‐33].

The present study has several limitations. First, this was a retrospective, nonrandomised, single-institution study with a small sample size. Second, the procedures were not performed by the same operator. Thus, it remains unclear whether the results can be generalized to other institutions.