This randomised trial investigating endoscopically nonvisible SPL up to 6 cm diameter by SB showed no diagnostic benefit of switching to UB for the diagnostic bronchoscopic sampling when compared to the procedure performed with the SB alone, when used under fluoroscopic guidance. There is only limited evidence from randomised controlled trials comparing the diagnostic yield of UB with that of SB. We found that UB had no diagnostic advantage compared to SB but was more cumbersome for operators. Moreover, the procedure time was nearly five minutes longer with UB, although the difference was not statistically significant. The clinical relevance of four minutes time difference can maybe neglected, since the change from the standard to the ultrathin bronchoscope is omitted in daily routine but was part of procedure time here. Although the classical definition of SPLs considers lesions ≤ 3 cm, we have chosen to investigate lesions ≤ 6 cm, since tuberculomas of nearly double that size have been reported [
10].
Data concerning the diagnostic yield of fluoroscopic guided UB compared to SB in the literature is scarce, especially for the evaluation of SPL. Considering the existing evidence in this field, the diagnostic yield of fluoroscopic guided UB in our study (55.0 %) was within the published range from 30.0 to 69.0 % reported by others [
11‐
13], although these studies used different sampling methods and were in different populations. In a small prospective, single-arm study published by Rooney et al. on 17 patients, the diagnostic yield of UB was only 30.0 %, perhaps because only a small biopsy brush was used for the ultrathin bronchoscope [
11]. The combined use of histological and cytological biopsy techniques has led to an improved diagnostic yield obtained with UB of 60.0 % [
12]. Using a slightly thicker bronchoscope (outer diameter 3.5 mm) in a prospective study on 102 patients, Oki et al. reported a diagnostic yield of 69.0 % for the thin bronchoscopy irrespective of lesion size. This result was obtained without using a brush; only forceps biopsy and bronchial washing were performed [
13]. We used forceps, brushing, and bronchial washings under fluoroscopic guidance as sampling methods. We did not investigate the diagnostic utility of the combined use of UB and SB in this context as sampling was only performed with the allocated bronchoscope (UB or SB). The diagnostic yield in our study may have been influenced by the prevalence of TB in our cohort (27.5 %), which is clearly higher than in the three other studies ranging from 5.9 to 12.5 %.
Technological advances in recent years have led to the availability of newer modalities helping to navigate to SPLs [
14]. The weighted diagnostic yield of ENB, VB, and R-EBUS for the work-up of SPLs without the use of UB is in the range of 67–74 % which is comparable with conventional methods mentioned above [
5,
15]. However, in the context of the new navigational bronchoscopic techniques the utility of UB needs to be reinvestigated. The usefulness of UB may play an important role when it is used with other new modalities such as CT-fluoroscopy or virtual bronchoscopic navigation. The combination of UB and CT-fluoroscopy is reported to have a similar diagnostic yield of 78 % in one study [
16]. Data regarding VB combined with UB are conflicting. In the randomised study by Asano et al, there was no significant difference of the diagnostic yield between the VB-assisted group (67 %) and the non-VB-assisted group (60 %) [
17]. However, under the combined use of EBUS, fluoroscopy, and virtual bronchoscopic navigation guidance, a recent randomised study published by Oki et al. could show that the diagnostic yield for malignant lesions was significantly higher with a 3.0-mm ultrathin bronchoscope (74 %) compared to a 4.0-mm thin bronchoscope (59 %) [
18]. The authors concluded that the combination of UB with navigational technology and endobronchial ultrasound (EBUS) seems to combine the best two modalities for evaluating peripheral pulmonary lesions [
18]. Interestingly, the diagnostic yield of UB was markedly higher in the study by Oki et al. (74 %) compared to our study (55 %). In contrast, the yield of SB was higher in our study (80 %) compared to 59 % by Oki et al., although a slightly thinner bronchoscope was used in their study [
18]. On one hand, this underlines the conclusion that advanced navigational technologies improve the diagnostic yield of UB. On the other hand, these differences raise the issue of the operator’s experience and a possible influence by the higher TB prevalence in our cohort. An additional aspect to consider is that VB- or ENB-guided biopsy techniques increase average costs, which is not feasible in many parts of the world [
19]. Compared to video-assisted thoracoscopic surgery (VATS) or (18)F-fluoro-deoxyglucose positron emission tomography (FDG-PET), ENB is more cost-effective to diagnose lung cancer in moderate- to high-risk patients [
20]. Another issue, which should be considered, is training. The recommended training requirement for R-EBUS is at least 50 supervised procedures [
21]. Issues of costs and training need to be considered before acquisition and introduction of such modalities. Some of these diagnostic modalities may not be feasible in less affluent settings. In addition, the benefit of these highly sophisticated diagnostic tools in an area with high incidence of TB remains questionable.
The limitations of this study are its small sample size, since it was originally designed as pilot trial, and that not all bronchoscopic results were confirmed by surgical sampling. Interestingly, none of the individuals in whom observation was considered the appropriate strategy by the clinical board had a malignant lesion. Furthermore, a new ultrathin bronchoscope with a larger working channel has been introduced which most probably leads to higher yields [
13]. All of the UB procedures were done with a fiberscope, which might have influenced diagnostic yield and could have caused a bias. We did not evaluate the distance of SPL from hilum or presence of bronchus sign as possible predictors of the diagnostic yield. Lastly, the bronchoscopies were performed by five different operators, which may have introduce a certain bias. However, three of the five operators had a documented track record from a previous study, and at least one of them was present in the bronchoscopy suite for all study related procedures [
22].