Background: In diagnosing lung cancer, evidence-based guidelines recommend choosing the procedure that is safe, least invasive, and provides information about stage of the cancer together with the diagnosis. This study was done to assess the quality of practice patterns for diagnosing lung cancer.
Methods: Retrospective review of patients diagnosed with lung cancer from January 2011 - December 2013.
Results: Intra-parenchymal mass & mediastinal abnormality (discrete lymph node and mediastinal infiltration) were the common intra-thoracic radiological abnormality seen in 80 (64.5%) and 71 (57.2%) of patients respectively. TTNA was the most commonly performed procedure. No radiological difference in size & location of the mass was found in TTNA or bronchoscopy group. The yield of TTNA was higher than bronchoscopy (95% vs 68%, p=0.001) and the cost per patient was lower (S$581 vs S$1122, p=0.001). However TTNA correlated with missed opportunity of nodal staging in (52.5%) of patients, greater complication rate (48%), and delayed diagnosis by 14 (1-337) days. In bronchoscopy, the delay was shorter, and complication rate was lower. However 72.7% of patients missed opportunity of nodal staging, and the number of procedures needed per patient for diagnosis was higher (1.34 vs 1.05, p=0.02).
Conclusion: Bronchoscopy was timely and safer technique than TTNA, but fraught with need for repeat procedures, and higher cost. TTNA on the other hand was more diagnostic and inexpensive than bronchoscopy but had high rate of complication, and delayed diagnosis. Both procedures had high rate of discordance with the guidelines & missed information regarding stage. Reserving TTNA for small peripheral lesions without mediastinal abnormality or bronchus sign, greater adoption of convex probe EBUS-TBNA, and availability of daily TTNA were the factors identified to improve quality.