Learning Curve of Shape-Sensing Robotic-Assisted Bronchoscopy (ssRAB) for Peripheral Pulmonary Lesions in a Thoracic...

Background: Robotic-assisted bronchoscopy enables precise navigation to peripheral pulmonary lesions and expands minimally invasive diagnostic options in thoracic surgery. At our institution, the ION™ Endoluminal System (Intuitive Surgical, Sunnyvale, CA, USA) was introduced to improve diagnostic accuracy in challenging peripheral targets. It is widely recognized that a defined number of procedures is required to achieve procedural proficiency and optimal clinical outcomes when adopting a novel platform. Therefore, this retrospective single-center study aimed to evaluate the learning curve associated with the implementation of this technology in a thoracic surgery center. Methods: In this retrospective study, all consecutive patients who underwent robotic-assisted bronchoscopies performed using the ION™ Endoluminal System (Intuitive Surgical, Sunnyvale, CA, USA) for the diagnosis of peripheral pulmonary lesions between August 2024 and March 2026 were analyzed. A total of 128 lesions in 89 patients were initially identified. Cases involving marker placement without diagnostic biopsy, as well as procedures not performed by the primary operator, were excluded. After applying exclusion criteria, 109 procedures in 76 patients were included. The mean patient age was 65.4 ± 9.1 years, and 44 patients were female (57.9%). To assess the learning curve, procedures were chronologically divided into three groups: early (cases 1-36), intermediate (37-73), and late (74-109). Outcome measures included procedure time, number of biopsies per lesion, tumor size, and diagnostic yield. Group comparisons were performed using non-parametric and chi-square tests. Procedural learning was assessed by cumulative sum (CUSUM) analysis of procedure time. Results: The overall diagnostic yield was 85.3% (93/109). The diagnostic yield increased over time from 73.0% in the early phase to 83.3% in the intermediate phase and 94.6% in the late phase. The overall comparison was statistically insignificant (χ²p = 0.117); however, there was a significant linear trend across phases, indicating progressive improvement with exposure to the application of this technology. Procedure time decreased significantly from a median of 49.0 min in the early phase to 31.0 min in the intermediate phase and 30.0 min in the late phase (p < 0.001). At the same time, the number of biopsies per lesion increased significantly (p < 0.001). Tumor size did not differ significantly between groups (p = 0.170). Conclusions: Robotic-assisted bronchoscopy demonstrates a clear learning curve, characterized by increasing diagnostic yield and significantly reduced procedure time during the implementation phase. The technique can be effectively integrated into the thoracic surgical diagnostic workflow and represents a valuable addition to minimally invasive diagnostics for peripheral pulmonary lesions.