Radial probe endobronchial ultrasound‐guided transbronchial lung biopsy for the diagnosis of cavitary peripheral pulmonary lesions

Abstract Background Cavitary peripheral pulmonary lesions (PPLs) are often diagnosed via transthoracic needle biopsy. However, today, radial probe endobronchial ultrasound (RP‐EBUS) is widely used to diagnose PPLs. The efficacy and safety of RP‐EBUS‐guided transbronchial lung biopsy (RP‐EBUS‐TBLB) used to diagnose cavitary PPLs remain poorly known. We investigated the utility of RP‐EBUS‐TBLB using a guide sheath (GS) without fluoroscopy to diagnose PPLs. Methods Of 743 RP‐EBUS procedures conducted to diagnose PPLs performed at our institution from January 2019 to October 2020, we analyzed 77 cavitary PPLs. TBLB was performed using RP‐EBUS with a GS without fluoroscopy. The diagnostic accuracy and complications were assessed. All lung lesions with a definitive diagnosis were included in analyses. Results The overall diagnostic accuracy was 85.7% (66/77). Of malignant lesions (n = 34), 29 (85.3%) were diagnosed successfully. Of benign lesions (n = 43), 37 (86.0%) were diagnosed successfully. In multivariate analyses, a thicker cavity wall (≥10 mm, odds ratio [OR] 14.22, 95% confidence interval [CI] 2.58–78.35, p = 0.002) and EBUS imaging with the probe within the lesion (OR 12.02, 95% CI 1.91–75.53, p = 0.008) independently affected diagnostic success. The likelihood of success increased with increasing thickness of the cavity wall (p < 0.001, test for trend). The specimens obtained for molecular confirmation of malignancy were satisfactory. There were four cases of infection (5.2%) and three cases of pneumothorax (3.9%). Conclusions RP‐EBUS‐TBLB of cavitary PPLs affords high diagnostic accuracy with acceptable complication rates.


INTRODUCTION
A pulmonary cavity is a gas-filled space within a zone of pulmonary consolidation, a mass, or a nodule. A cavity may reflect necrosis, cystic dilation of a lung region in patients with benign diseases and necrosis, internal cyst formation, or internal desquamation of tumor cells in patients with malignant diseases. 1 The differential diagnosis is broad and includes various infections, autoimmune conditions, and primary and metastatic malignancies. 2 Radiological findings of benign and malignant cavitary peripheral pulmonary lesions (PPLs) often overlap; tissue analyses are required when diagnosing a PPL.
Transthoracic needle biopsy (TTNB) has commonly been used to diagnose PPLs. The diagnostic accuracy is close to 90%, 3 but a meta-analysis found that the overall complication rates after core biopsy and fine needle aspiration (FNA) are 38.8% and 24.0%, respectively. 4 The same is true after diagnosis of cavitary PPLs. TTNB is associated with a diagnostic rate of over 80%, but the complication rate is elevated by more than 20%. 5,6 Recently, radial probe endobronchial ultrasound (RP-EBUS) has become widely used to diagnose PPLs. The pooled sensitivity of RP-EBUSguided transbronchial lung biopsy (RP-EBUS-TBLB) is 0.72 and the safety profile is excellent. 7 Diagnosis of the cavitary PPLs is different from that of the noncavitary PPLs, which means that the cavitary wall should be targeted for biopsy. However, its diagnostic efficacy and the safety of RP-EBUS for diagnosing cavitary PPLs remain poorly known. In this study, we analyzed the clinical utility of RP-EBUS in terms of diagnosing cavitary PPLs.

Study design and subjects
Of 743 RP-EBUS procedures performed to evaluate PPLs at Yeungnam University Hospital from January 2019 to October 2020, we performed a retrospective observational study on 81 consecutive patients who underwent RP-EBUS-TBLB to explore cavitary PPLs. Of the biopsied lesions, 77 were definitively diagnosed according to biopsy results (n = 66), clinical course (n = 1), sputum microbiology (n = 5), liver biopsy (n = 1), EBUS-transbronchial needle aspiration (n = 3), and surgical resection (n = 1). Four cases with indefinite diagnoses (neither benign nor malignant) were excluded from the diagnostic accuracy analyses. A study flowchart is presented in Figure 1.

CT preparation and biopsy
All patients underwent thin-section chest computer tomography (CT) (0.75-mm slice thickness at intervals of 0.75 mm; Somatom Definition AS 64-slice CT system, Siemens Healthcare) prior to biopsy. Three pulmonologists (Kyung Soo Hong, Jong Geol Jang, and June Hong Ahn) reviewed the images before the procedure and defined the pathways to be used when accessing cavitary PPLs. The bronchus sign of CT was defined as a bronchus leading to the target lesion. The distance from the lesion to the pleura was the shortest distance evident in the axial CT scan. 8 The cavitary wall thickness was the greatest thickness in the region of the biopsy site apparent in the axial image. All bronchoscopy procedures were performed by one of the three pulmonologists (above). A 4-mm diameter bronchoscope (BF P260F, Olympus) was used to attain the target lesions. After the lesions were revealed by RP-EBUS, biopsy specimens were collected using a brush and biopsy forceps. All RP-EBUS-TBLB procedures were performed using a guide sheath (GS) but without X-ray fluoroscopy.
Representative data on RP-EBUS-TBLB patients with cavitary PPLs are shown in Figure 2.

Diagnostic classification
A final diagnosis of malignancy was made based on the definite histological or cytological evidence of malignancy. Benign lung lesion was diagnosed according to the following criteria: identification of definite benign features (at least one of the histology, cytology, or microbiology tests was diagnostic), regression of the lesion with medical treatment, and a stable size for at least 12 months. The lesions initially diagnosed as malignant by RP-EBUS and finally diagnosed with malignancy were considered true positives. The lesions initially diagnosed as benign by RP-EBUS and finally diagnosed as benign were considered true negatives. Initial benign results that were finally diagnosed with malignancy were considered false negatives. To explore factors affecting diagnostic accuracy, we divided the study population into two groups: a diagnostic success group (true-positive and true-negative results) and a diagnostic failure group (false-positive and false-negative results). We calculated the diagnostic accuracy by dividing the number of diagnostic successes by the total number of cases.

Statistical analyses
Continuous variables were compared using the Student t test or the Mann-Whitney U test and are expressed as means AE standard deviations. Categorical variables were compared using the chi-square or Fisher's exact test and are described as frequencies (percentages). Univariate and multivariate logistic regression analyses were performed to identify factors affecting the diagnostic accuracy. The factors evaluated in multivariate analyses were those significant at the p < 0.1 level in univariate analyses. We used the linearby-linear association test to analyze the diagnostic accuracy by cavity wall thickness. We employed the Kruskal-Wallis test to compare the numbers of specimens by cavity wall thickness. In all analyses, a two-tailed p value <0.05 was considered statistically significant. All statistical procedures were performed using SPSS software (version 24.0, IBM).

Ethics statement
This study was conducted in accordance with the tenets of the Declaration of Helsinki, and the protocol was reviewed and approved by the Institutional Review Board of Yeungnam University Hospital (no. 2020-09-025). The requirement for informed patient consent was waived because of the retrospective study design.

Baseline characteristics
Of the 743 RP-EBUS procedures performed for PPL diagnosis in our institution from January 2019 to October 2020, 77 cavitary PPLs were included in this study. The baseline characteristics of the 77 patients are listed in Table 1. The mean age was 68.0 AE 11.9 years and 59 were men (76.6%). Chronic obstructive pulmonary disease (COPD, n = 23, 29.9%) was the most common underlying lung disease. The mean predicted FEV1% was 80.9 AE 17.8. Most lung lesions were in the right lower lobe (n = 23, 29.9%) and the mean distance from the pleura to the lung lesion was 8.9 AE 10.2 mm. The mean cavity wall thickness was 12.3 AE 5.4 mm. The bronchus sign was positive on CT in 68 patients (88.3%) and EBUS images within the lesion were apparent in 67 (87.0%) patients. The mean procedural time was 17.7 AE 7.1 min.

Factors affecting diagnostic success
We sought factors affecting PPL diagnostic success (  (Figure 3(a)). Diagnosis failed in all patients with cavity wall thicknesses <5 mm. Six patients with cavity wall thicknesses <5 mm were finally diagnosed as lung cancer (two patients), pulmonary tuberculosis (one patient), nontuberculous mycobacteria (two patients), and fungal ball (one patient). The diagnostic success rate was 86.7% in patients with cavity wall thicknesses ≥5 mm and <10 mm. In patients with cavity wall thicknesses ≥15 mm, the diagnostic success rate was 95.2%. However, the number of acquired specimens did not significantly differ by cavity wall thickness (Figure 3(b)).

Complications
We encountered four cases of infection (5.2%) and three of pneumothorax (3.9%) in the 77 patients who underwent RP-EBUS-TBLB. Among the four cases of infection, three recovered with antibiotics and one developed empyema requiring chest tube insertion. All three patients with pneumothorax recovered following oxygen therapy.

DISCUSSION
We found that that RP-EBUS-TBLB was very accurate and safe when used to diagnose cavity PPLs. The diagnostic accuracy was 85.7% and the complication rate was 9.1% (infection 5.2% and pneumothorax 3.9%). Analyses of brushing cytology and microbiology samples, in addition to histology samples, increased the diagnostic rates of both malignant and benign lesions. A thicker cavity wall (≥10 mm) and an EBUS image within the lesion independently enhanced success. The probability of success increased with increasing thickness of the cavity wall, and all obtained specimens allowed molecular evaluation. To the best of our knowledge, this is the largest study to date that has analyzed the utility of RP-EBUS-TBLB in terms of diagnosing cavitary PPLs. Furthermore, this is the first study to reveal the positive relationship between cavity wall thickness and diagnostic rate in the diagnosis of cavitary PPLs. A pulmonary cavity is a gas-filled space within a zone of pulmonary consolidation, a mass, or a nodule. Cavities in patients with benign lung disease may be caused by necrosis (suppurative, caseous, or ischemic), cystic dilation of a lung structure, or displacement of lung tissue by a cystic structure. Cavities in patients with malignant lung disease may be caused by treatment-related necrosis, internal cyst formation, or internal desquamation of tumor cells with subsequent liquefaction. 1 Radiologically, a notch and an irregular internal wall were more common in malignant cavitary lesions; a linear margin, satellite nodules surrounding the cavity, bronchial wall thickening, consolidation, and ground-glass attenuation were more frequent in benign cavitary lesions. 9 However, biopsy is often required to confirm diagnosis because the chest CT findings of benign and malignant cavitary PPLs overlap.
Prior to the introduction of RP-EBUS, TTNB was commonly used to diagnose cavitary PPLs. In a previous study, CT-guided fine needle aspiration biopsy of cavitary pulmonary lesions was associated with a high diagnostic accuracy (96.1%). 6 However, the complications included pneumothorax (8.8%), alveolar hemorrhage (13.7%), and hemoptysis (1%). Although fine needle aspiration biopsy (rather than core needle biopsy using cutting needles) was employed, the complication rates were relatively high. In another study, the overall accuracy of TTNB in terms of diagnosing cavitary lesions was 81%. 5 The overall complication rate was 28%, with pneumothorax being the most common problem (24.5%); 9.4% of patients required chest tube insertion, 1.9% developed mild hemoptysis, and 1.9% had a small hemothorax. In summary, TTNB used to diagnose cavitary PPLs affords a relatively high diagnostic rate (>80%), but the procedure-related complication rate is also high (≥20%).
Only one prior study has analyzed the utility of RP-EBUS in terms of cavitary PPL diagnosis (50 lesions). 10 In that study, the overall diagnostic accuracy was 80%, 77.8% for malignant and 82.6% for benign lesions. An EBUS probe within the lesion was the only factor that enhanced diagnostic accuracy. Of the 50 procedures, only one pulmonary infection occurred. No other major complication, such as pneumothorax or significant hemorrhage, was reported.
In this study, we analyzed 77 cavitary PPLs. In agreement with the previous study, 10 we found that localization of the probe within the lesion independently enhanced diagnostic success. However, our most interesting result is that a thicker cavity wall (≥10 mm), not the size of the cavity per se, increased the diagnostic success rate, which increased significantly as the cavity wall thickness increased. In Zhuang et al., the diagnostic accuracy was not associated with cavity wall thickness, but more nondiagnostic samples occured in wall thicknesses <5 mm than ≥5 mm lesions in fine needle aspiration biopsies of cavitary PPLs. 6 Kiranantawat et al. reported that wall thickness at the biopsy site is a significantly independent factor in terms of diagnostic success. 5 Although the number of acquired specimens did not differ significantly by cavity wall thickness in the present study, we recommend targeting the thickest wall when guiding RP-EBUS for TBLB; this optimizes tissue samples. In our experience, cavitary walls of thickness <5 mm tend to yield unsatisfactory samples. Although the reason for this is unclear, not using fluoroscopy may have affected these results. Further study is needed to determine the relationship between cavitary wall thickness and tissue sample quality.
As in the previous study, 10 we found that analyses of brushing cytology and microbiology samples (in addition to histology samples) increased the diagnostic rate. The final histology diagnosis rates for malignant and benign lesions were 79.4% and 55.8%, respectively. However, after adding brushing cytology and microbiology data, these figures increased to 85.3% and 86.0%. In particular, microbiology tests were very valuable in terms of diagnosing benign lung disease. After biopsy and GS removal, efforts must be made to obtain enough bronchial washing samples for microbiology tests with the bronchoscope wedging as close to the lesion as possible.
Only a few studies have explored the feasibility of molecular studies of lung cancer confirmed via RP-EBUS. One study reported that specimens obtained via RP-EBUS-TBLB using a 1.8-mm diameter biopsy forceps allowed analyses of EGFR mutational status, and ALK and PD-L1 IHC, in more than 90% of adenocarcinoma cases. 11 One French study revealed that multigene molecular analyses could be performed on almost 80% of specimens sampled via RP-EBUS-TBLB using a 1.5-mm diameter microbiopsy forceps. 12 We used such forceps, and the statuses of EGFR mutations, ALK, ROS-1, and PD-L1 IHC were evaluable in more than 80% of samples, comparable to a previous study. 12 Our RP-EBUS-TBLB complication rates (3.9% for pneumothorax and 5.2% for infection) are higher than those of previous RP-EBUS-TBLB studies. One study on 965 PPLs reported an overall complication rate of 1.3%: 0.8% pneumothorax and 0.5% pulmonary infection. 13 A meta-analysis revealed that pneumothorax developed in 1.0% of patients, of whom 0.4% required chest tube insertion. 14 Several complications were possible in the present study, for two reasons. First, in terms of pneumothorax, many patients had underlying lung disease (in particular, 29.9% had COPD). A previous study found that pneumothorax developed in 3.3% of such patients and that pulmonary emphysema was an independent risk factor for pneumothorax. 15 Second, in terms of infection, a recent study showed that infections developed in 4.47% of patients, and cavitation in the lesion was significantly associated with infectious complications after RP-EBUS-TBLB (OR 3.63, 95% CI 1.51-8.71). 16 Cavities are sometimes caused by necrosis; biopsy itself may exacerbate infection in a necrotic area caused by pathogens transported (via the bronchoscope) from the oral cavity and upper airway. However, our complication rates were much lower than those of studies that biopsied cavitary PPLs using TTNB. 5,6 Our work had several limitations. First, this was a single-center retrospective study on a small number of cavitary PPL patients; the results cannot be generalized. Second, we performed biopsy without fluoroscopic guidance to diagnose PPLs and this may have affected the diagnostic accuracy. The use of fluoroscopy help to check whether the forceps are within the lesion and assist TBLB in the lower lobe lesions, increasing the diagnostic accuracy. Third, all procedures were performed by three pulmonologists, who may have differed (to some extent) in competence. However, all had at least 5 years of experience in performing many respiratory procedures and all used standardized protocols. Thus, among-examiner differences are unlikely to be significant.

CONCLUSIONS
RP-EBUS-TBLB of cavitary PPLs affords a high diagnostic accuracy with an acceptable complication rate. A thicker cavity wall (≥10 mm) and EBUS imaging when the probe lay within the lesion independently enhanced the diagnostic success rates. The success rate increased with increasing thickness of the cavity wall. Most specimens were sufficiently large to allow molecular malignancy confirmation.

DECLARATIONS
This study was conducted in accordance with the tenets of the Declaration of Helsinki, and the study protocol was reviewed and approved by the Institutional Review Board of Yeungnam University Hospital (YUH IRB no. 2020-09-025). The need for informed patient consent was waived because the study was retrospective in nature.
CONSENT FOR PUBLICATION Not applicable.