To verify the clinical effectiveness and safety of the airway tree navigation system constructed by artificial intelligence (AI) in the navigation diagnosis of peripheral pulmonary nodules (PPLs).

Early diagnosis and treatment of lung cancer is of great significance, in which navigated tracheoscopic biopsy is an important tool for confirming the diagnosis of early lung cancer. Conventional navigation software realizes airway reconstruction and guides biopsy by recognizing differences in HU values on computed tomography scans. It is difficult for conventional navigation software to recognize the reconstruction due to the special characteristics of small airways that are susceptible to interference and collapse. Therefore, an AI deep learning approach can realize accurate construction of small airways and guide accurate biopsy. This study intends to validate the clinical effectiveness and safety of the AI-constructed airway tree navigation system in the navigational diagnosis of peripheral pulmonary nodules (PPLs).

Inclusion Criteria:

1. Aged 18 years or above.
2. Patients with one or more peripheral lung nodules suspected to be lung cancer or poorly absorbing lesions on conventional anti-infective therapy.
3. Patients with nodule diameters ≤30 mm (diameters mentioned in the text are the average of the maximum and minimum diameters).
4. The nodules were pure ground glass nodules, partially solid nodules, or solid nodules.
5. The nodule is surrounded by lung parenchyma and is not visible in the bronchial lumen above the segment.

Exclusion Criteria:

1. Preoperative judgment that it is difficult for the patient to benefit from bronchoscopic biopsy (e.g., high risk of bleeding due to perivascular encasement of the lesion, difficulty in reaching the airway adjacent to the lesion due to previous lung surgery, etc.).
2. Those with incomplete clinical data.
3. Those with missing visits after biopsy.