Comparison of computed tomography‐guided percutaneous needle biopsy and endobronchial biopsy in the diagnosis of multifocal pulmonary lesions

Abstract Background The retrospective study aimed to compare computed tomography (CT)‐guided percutaneous needle biopsy (PNB) and endobronchial biopsy (EB) in the diagnosis of multifocal pulmonary lesions with endobronchial involvement. Methods Between November 2014 and June 2017, consecutive patients who had underwent both CT‐guided PNB and EB via bronchoscopy for diagnosis of pulmonary lesions were evaluated retrospectively. Tissue samples were submitted for pathological examination, acid‐fast bacilli, TB RT‐PCR, and mycobacterial culture. Sensitivities of the two methods alone or in combination were calculated and compared using Fisher's exact test. Results Sixty‐seven patients (46 men and 21 women) were enrolled and could be diagnosed (32 malignant, 18 TB, and 17 benign). A final diagnosis of either malignant or TB diseases was made in 34 (68.0%) patients for CT‐guided PNBs, 19 (38.0%) patients for EBs, and 42 (84.0%) patients for the combination of both methods. Further statistical analysis showed significant difference in sensitivity between CT‐guided PNBs, or the combination of both methods, and EBs (all P < 0.05), and no difference between CT‐guided PNBs and the combination (P > 0.05). However, the combination of both methods appears to have the highest sensitivity in the detection of malignancies or TB diseases. Conclusion Compared with EB, CT‐guided PNB has a high diagnostic yield for the detection of TB and malignancy in patients with multifocal pulmonary lesions with endobronchial involvement. When the two biopsies are combined, it appears to provide an incremental diagnostic value for the pulmonary lesions.


| INTRODUC TI ON
Computed tomography (CT)-guided percutaneous needle biopsy (PNB) is a minimally invasive procedure that is used to evaluate pulmonary lesions. CT-guided PNB can establish a malignant diagnosis, establish a benign diagnosis, or obtain material for culture. 1 Although there are several known complications including pneumothorax and hemorrhage, careful attention during biopsy planning, technique performance, and postprocedural care can help to prevent or minimize most potential complications. 2 CT-guided PNB has been considered an accurate approach for diagnosis. Geraghty PR et al reported the overall accuracy of CT-guided PNB was 93.5%. 3 However, central lesions adjacent to the bronchi remain one of the main relative contraindications to PNB. 1 As an important tool in the diagnosis of bronchopulmonary diseases, bronchoscopy is preferred for central endobronchial or peribronchial lesions. Several studies supported that peripheral pulmonary lesions with endobronchial involvement are easily evaluated via bronchoscopy. 4,5 Moreover, endobronchial biopsy (EB) is a useful technique to diagnose infection when less invasive methods to diagnose the underlying etiology are not suitable; this may enhance the diagnostic utility of bronchoscopy in high tuberculosis (TB) burden countries, such as China and India. 6 Imaging can aid in the selection of a technique to obtain tissue, including in the choice between bronchoscopic biopsy and PNB; however, in a proportion of patients who require tissue diagnosis, the optimal technique remains unclear for patients with multifocal pulmonary lesions with endobronchial involvement. In this study, we compared the diagnostic performance of these two methods to guide clinicians in the selection of the optimal technique.

| Subjects
At our institution, between November 2014 and June 2017, all consecutive patients who had undergone both CT-guided PNBs and EBs via bronchoscopy for diagnosis of pulmonary lesions within 7 days were evaluated retrospectively. Demographic data were recorded, including age and sex. All data were collected using a medical record number unique to each member. Thus, all records used were de-identified and anonymized.
F I G U R E 1 Computed tomographyguided percutaneous needle biopsy of a lung lesion (A), and bronchoscopic finding revealing an endobronchial abnormality (B). A 65-year-old patient underwent computed tomography-guided percutaneous needle biopsy (C), and endobronchial biopsy (D) was confirmed squamous cell carcinoma. A TB patient underwent computed tomographyguided percutaneous needle biopsy (E), and endobronchial biopsy (F) was diagnosed with supportive histological evidence, such as epithelioid cells A and multinucleated giant cells B (hematoxylin and eosin stain, 200 × magnification)

| CT-guided PNB
The CT-guided PNBs were performed using a semiautomated

| Endobronchial biopsy
Flexible bronchoscopy was performed using a fiberoptic bronchoscope (BF-260, Olympus). EB (at least four specimens) was performed using biopsy forceps in the areas that appeared abnormal, if any, or from secondary carinal areas. Moreover, bronchial brushing was performed before EB. Biopsies were stained for mycobacteria.
In addition to the stains, the specimens were also sent for TB RT-PCR and mycobacterial culture.
The biopsy results were categorized as malignant, TB, and benign ( Figure 1).

| Statistical analysis
Continuous variables are expressed as the mean ± standard deviation. Categorical variables are presented as simple proportions.
Sensitivities of the two methods alone or in combination were calculated separately. Fisher's exact test was used to evaluate the differences in sensitivity between pairs of categorical data. A P < 0.05 was considered significant. Analyses were performed using SPSS 16.0 (SPSS). In terms of detecting malignancy, the sensitivity of CT-guided

| RE SULTS
PNBs was superior to that of EBs (P < 0.05), and all patients with positive brushing cytology had positive needle biopsies. There was no difference in the sensitivity rates between EBs and bronchial brushings (P > 0.05). Compared with EBs, the sensitivity of CT-guided PNBs was markedly higher (P < 0.01). Although the sensitivity of the combination of both methods (CT-guided PNBs and EBs) appears to be higher than that of each one alone, there was no significant difference in sensitivity between the combination of both methods and CT-guided PNBs (P = 0.082 > 0.05).
In terms of detecting TB, there was no difference in sensitivity rates between CT-guided PNBs and EBs (P > 0.05); the combination was not superior to that of each one alone (P > 0.05).

| D ISCUSS I ON
Because of its advantages, which include being minimally invasive, easy and simple to handle, and low rate of complications, CT-guided PNB has been widely used in the diagnosis of pulmonary lesions, especially for lung cancer, with high accuracy and safety. 8 In the present study, we investigated the performance of CT-guided PNB in the diagnosis of multifocal pulmonary lesions with endobronchial involvement, while comparing CT-guided PNB with EB via bronchoscopy. In high TB burden countries, we found that the sensitivity of CT-guided PNBs for a final diagnosis of either malignant or TB diseases was superior to that of EBs (68.0% vs 38.0%). If the two techniques were combined, the combination will greatly improve the diagnostic sensitivity for malignancies or TB diseases and lead to a more satisfactory detection result (42/50, 84.0%).
In contrast to bronchoscopic procedures, needle biopsy avoids contamination of the specimen. Additionally, with image guidance, the needle can be specifically directed toward the target lesion.
However, when target lesions were adherent to the bronchi, it was considered a contraindication to needle biopsy. During bronchoscopy, a variety of sampling methods has been used. Usually, when visible airway abnormalities are present, endobronchial biopsy was preferred for a better assessment. Obtaining a tissue biopsy for pathological examination is key to ensuring a prompt and appropriate diagnosis and effective treatment for the pulmonary lesions. In a recent study that evaluated the diagnostic accuracy of CT-guided PNB for pulmonary lesions, the diagnostic accuracy, sensitivity, and specificity of CT-guided PNB were reported to be 92.9%, 95.3%, and 95.7%, respectively. 9 In our study, CT-guided PNB also showed good performance in the diagnosis of pulmonary lesions, and the detec- and was not included in the analysis that estimated the association between tumor size and the diagnostic results. Fifth, it should be noted that some, more central lung lesions, are more amenable to EB. Therefore, patients with multifocal pulmonary lesions with peripheral and endobronchial lesions may benefit from the combination of the two methods.

| CON CLUS IONS
In conclusion, pulmonary lesions can be of various etiologies.
However, in high TB burden countries, TB and malignancy were more common and more severe. The present study indicates that

CO N FLI C T O F I NTE R E S T
The authors declare that they have no conflict of interest.