This retrospective study was conducted using the bronchoscopy databases of Pusan National University Hospital (a university-affiliated tertiary referral hospital in Busan, Republic of Korea) and Ulsan University Hospital (a university-affiliated secondary referral hospital in Ulsan, Republic of Korea) for the period December 2015 to November 2018. During the study period, 993 patients with PLL (797 and 196 patients at Pusan National University and Ulsan University Hospitals, respectively) received bronchoscopy using radial probe EBUS. Among them, 23 patients were excluded because lung cancer was previously confirmed and the radial probe EBUS bronchoscopy was performed as a re-biopsy for the identification of T790M mutation of epidermal growth factor receptor . This left 970 patients for selection for the study. Some of the clinical data on the patients enrolled between 2015 and 2018 was included in two articles published in 2018 and 2019 [8,19]. This study was approved by the institutional review boards of Pusan National University Hospital (IRB no. 1906-033-080) and Ulsan University Hospital (IRB no. 2020-07-011). As this study was performed retrospectively, the requirement for informed patient consent was waived.
Analysis of the CT scans
PLL was defined as an intrapulmonary lesion existing beyond the segmental bronchus that is generally invisible on conventional bronchoscopy . The mean diameter of a PLL was defined as the average of its maximum and vertical diameter. A positive bronchus sign was defined as the presence of a bronchus leading directly to the PLL . The margin of the PLL was classified as follows [22,23]: 1) smooth when the margin was well-demarcated with round or oval-shape curves; 2) lobulated when the margin was distinguished by some smooth and relatively large convexities; 3) spiculated when the margin was irregular and had multiple radiating strands; and 4) pneumonic consolidation when the margin could not be distinguished because of surrounding consolidation. PLLs were classified as solid, part-solid, or ground-glass opacity according to a visual assessment method based on CT attenuation . The distance from the pleura was measured as the closest distance between the PLL and the visceral pleura. The Hounsfield units (HU) of the PPL on pre- and post-enhancement CT phases were measured using mediastinal window images to minimize volume averaging .
If the lesion showed cavitation, the maximal thickness of the cavity wall was measured, and it was classified as concentric cavitation if the wall thickness was uniform, or eccentric cavitation otherwise . The following accompanying CT findings were also analyzed: 1) presence of a satellite centrilobular nodule; 2) bronchiectasis; 3) anthracofibrosis of the airway; 4) pulmonary emphysema; 5) fibrocalcific tuberculosis; 6) interstitial lung diseases; 7) atelectasis; and 8) pleural effusion (Additional file 1).
Bronchoscopic procedure and specimens
All patients underwent bronchoscopy with a 4.0-mm-sized flexible bronchoscope (BF-P260F; Olympus, Tokyo, Japan) after conscious sedation with midazolam and fentanyl. For local anesthesia, 2% lidocaine was injected into the tracheobronchial tree via the working channel of the bronchoscope. After airway inspection, the bronchoscope was advanced as far as possible into the bronchus of the target lesion under CT image guidance. Thereafter, a 20-MHz radial probe EBUS (UM-S20-17S; Olympus, Tokyo, Japan) covered by a guide sheath (K-201; Olympus, Tokyo, Japan) was advanced through the working channel until resistance was met. Then, under X-ray fluoroscopic guidance, the radial probe EBUS was pulled back slightly to acquire the ultrasound images . Radial probe EBUS images were classified as: 1) within; 2) adjacent; or 3) invisible, as described previously . When the location of the target lesion was identified, the radial probe was retrieved with the guide sheath being kept in place. Brushing cytology and a forceps biopsy were then performed through the guide sheath under fluoroscopic guidance. After obtaining the cytology and biopsy samples, the guide sheath was removed and bronchial washing of the target lesion was performed through the working channel with 5 ml of sterile saline.
The bronchial washing fluids were used to perform an acid-fast bacillus smear with culture and a real-time polymerase chain reaction (PCR) for mycobacterium. Fluorescence microscopy with auramine-rhodamine staining was used for the acid-fast bacillus smear. Both solid (3% Ogawa medium) and liquid medium (BACTEC MGIT 960 system; Becton Dickinson Microbiology Systems, Sparks, MD) were used for the mycobacterium culture. A real-time PCR for mycobacterium (AdvanSure TB/NTM real-time PCR kit; LG Life Science, Seoul, Republic of Korea) was also performed on the bronchial washing fluid.
If a final diagnosis could not be determined from the bronchoscopic samples, an additional percutaneous needle biopsy or surgical biopsy was performed. When the pathological findings of the percutaneous needle biopsy or surgical biopsy specimens were suspicious for tuberculosis, such as the presence of chronic granulomatous inflammation, PCR (MTB-PCR kit; Biosewoom, Seoul, Republic of Korea) was additionally performed on the tissue specimen, at the discretion of the pathologist.
Diagnosis of pulmonary tuberculosis
Tuberculosis was diagnosed as follows: 1) Mycobacterium tuberculosis was cultured from the bronchial washing fluid; and 2) PCRs for Mycobacterium tuberculosis using either the bronchial washing fluid or tissue specimen were positive and clinicoradiological improvements were achieved after standard anti-tuberculosis treatment .
Statistical analyses were performed using R version 3.6.6 (R-Project, GNU GPL). All results are presented as number and percentage for categorical variables and median with interquartile range [IQR] for continuous variables. Data comparisons were made using χ2 or Fisher’s exact tests for categorical variables, and independent t-tests or Wilcoxon rank-sum tests for continuous variables. Multivariate logistic regression analysis was performed using factors with a P-value < 0.1 in the univariate analysis, to identify independent factors related to unexpected diagnosis of pulmonary tuberculosis. P-values < 0.05 were considered statistically significant. Receiver operating characteristics curves were plotted to calculate the area under the curve, Youden index, sensitivity, and specificity.