Intro: Flexible bronchoscopy is routinely used to assess lung cancer and collect specimens, with bronchoalveolar lavage(BAL) estimated to have a diagnostic yield of 43-47% (Detterbeck, FC et al. Chest 2013). However, there is no clear guidance on choosing bronchoscopy as an initial diagnostic test. Therefore, we aimed to identify the factors influencing the pre-pandemic performance of bronchoscopies with BAL when investigating suspicious pulmonary lesions.

Methods: We reviewed the charts of 111 patients who underwent bronchoscopy with BAL at our center between January and June 2018 for suspected lung cancer. Efficacy of BAL was assessed for neoplastic diagnosis and subtyping.

Results: Mean age of patients was 70(49-90) years, with 59 males. Mean size of lesion was 31mm and 72(64.9%) had a final diagnosis of lung cancer. Of 111 BALs, 20(18%) showed suspicious or cancer cells. Higher diagnostic yield was found with lesions larger than 2cm(27.8% vs 0%, p<0.001), central tumors(41.7% vs 5.8%, p<0.001), bronchus sign(41.7% vs 15.2%, p=0.024) and lymphangitic carcinomatosis(66.7% vs 15.2%, p=0.001). Higher T(p=0.001) with TNM staging system, abnormal endoscopic appearance(p<0.001) and intense positron emission tomography uptake(p=0.010) were other factors linked with better performance. A high malignancy risk score per Mayo Clinic Model was associated with a diagnostic BAL(22.7% vs 3.7% for low-intermediate risk, p=0.084), but not routinely calculated by clinicians in our center.

Conclusion: Despite its frequent use in evaluating lung cancer, BAL had a low diagnostic yield of 18% at our center before the COVID-19 era. Several factors affect its efficacy and could guide a more selective approach for future use.