BACKGROUND: Given that transbronchial lung cryobiopsy (TBLC) is recommended as a surrogate for surgical lung biopsy (SLB) in the diagnosis of interstitial lung disease (ILD), few studies have evaluated the optimal number of biopsy sample or the impact of radiologic pattern on the diagnostic yield and postoperative complications. This study aimed to investigate how biopsy sample number and radiologic patterns affect diagnostic yield and complications. METHODS: We conducted a multi-center retrospective study including 334 consecutive ILD patients who underwent TBLC. The impact of number of biopsy sample and radiologic pattern on diagnosis yields and postoperative complications was assessed. Logistic regression analyses were used to explore the risk factors for complications. RESULTS: The histopathologic and multidisciplinary discussion (MDD) diagnostic yields were 70.06% (234/334) and 86.23% (288/334). Moderate-severe bleeding and pneumothorax occurred in 39 (11.68%) and 29 (8.68%) cases, respectively. The MDD diagnostic yield was higher in patients with non-fibrotic pattern on high-resolution computed tomography (HRCT) compared to those with a fibrotic pattern (88.11% . 75.00%, P=0.02). However, the diagnostic yields or postoperative complications were comparable when obtaining three or more than three biopsy samples. Multiple lobe biopsy and number of biopsy samples were associated with bleeding [odds ratio (OR) =3.675, 95% confidence interval (CI): 1.414-9.553, P=0.008; OR =0.649, 95% CI: 0.470-0.895, P=0.009], while a fibrotic pattern increased the risk of pneumothorax (OR =3.479, 95% CI: 1.210-10.005, P=0.02). CONCLUSIONS: This study showed that obtaining three biopsy samples is appropriate to achieve an adequate diagnostic yield, while procedure methods are associated with complications. Further well-designed studies are required to standardize TBLC procedures.