Detection of tumor-derived DNA dispersed in the airway improves the diagnostic accuracy of bronchoscopy for lung cancer

Abstract

The diagnostic accuracy of bronchoscopy for detecting lung cancer, especially peripheral lung cancer with lesions outside the endoscopically visible range, remains unsatisfactory. The aim of this study was to perform next-generation sequencing on bronchoscopic specimens to determine whether this improves the accuracy of bronchoscopy for diagnosing lung cancer and to identify factors influencing sensitivity. The bronchoscopic sensitivity for diagnosing lung cancer was initially evaluated in 191 patients who underwent lobectomy after bronchoscopy at our hospital. Sputum, bronchial wash fluid, and resected lung cancer specimens were subsequently collected from 18 patients with peripheral small cell lung cancer for genomic analysis. DNA was extracted from formalin-fixed, paraffin-embedded surgical tissue specimens and the supernatant and cell fractions of sputum and bronchial wash fluid. Deep sequencing was performed using a lung cancer panel covering all exons of 53 lung cancer-related genes. The bronchoscopic sensitivity for diagnosing lung cancer at our hospital was 60.7%. Multivariate analysis revealed that this was influenced by tumor size and location, but not histological type or lymph node metastasis. The sensitivity was the highest for biopsy followed by curettage and bronchial wash specimens. DNA mutations homologous to those identified in the primary lesions were detected in the bronchial wash fluid of 10 patients (55.6%), while only 2 patients (11.1%) were diagnosed with lung cancer based on conventional cytological examinations. In conclusion, the addition of genomic analysis to routine pathological examinations improves the diagnostic accuracy of bronchoscopy.