A comparison of cytology and fluorescence in situ hybridization for the detection of lung cancer in bronchoscopic specimens

Study objectives: To determine the relative sensitivity and specificity of cytology and fluorescence in situ hybridization (FISH) for the detection of lung cancer in bronchoscopically obtained specimens. Design: Cytology and FISH were performed on brushing and washing specimens obtained from patients undergoing bronchoscopy for suspected lung cancer. FISH utilized the LAVysion probe set (Abbott Molecular; Des Plaines, IL), which contains locus-specific probes to 5p15, 7p12 (EGFR), 8q24 (C-MYC), and a centromeric probe to chromosome 6. Setting: Single-center, academic, tertiary medical center. Participants: One hundred thirty-seven patients referred for bronchoscopy for suspicion of lung cancer. Interventions: Cytology and FISH were performed on bronchoscopic brushings and washings. Measurements and results: One hundred thirty-seven patients undergoing bronchoscopy had pathology, FISH, and cytology results. FISH and cytology were performed on 123 washing and 78 brushing specimens. Sensitivities of FISH and cytology were 71% and 51% (p = 0.007), respectively, for brushing specimens, and 49% and 44% (p = 0.541) for washing specimens. When FISH and cytology results were combined, sensitivities were 75% and 61%, respectively, for brushing and washing specimens, which was significantly better (p < 0.001) than cytology alone. Specificities of FISH and cytology for patients with negative findings at the time of initial bronchoscopy were 83% and 100% (p = 0.125), respectively, for brushing specimens, and 95% and 100% (p = 0.500) for washing specimens. Conclusions: These findings show that FISH is significantly more sensitive than conventional cytology for detecting lung cancer in bronchial brushing specimens; when combined with cytology, FISH can improve the diagnostic sensitivity of detecting malignancy in bronchial brushing and washing specimens.