Abstract
BACKGROUND: Precise segmentectomy has become the first choice of surgical treatment for pulmonary nodules and early lung cancer, and the key and difficult point of the surgery lies in the precise location and resection of the lesion. DeepInsight is an auxiliary software for precise lung surgery jointly developed by our center and Neusoft Company, which can determine the precise anatomy of the lung and locate the location of lung lesions before operation. This study is to verify the authenticity and reliability of DeepInsight lung bronchial angiography assisted surgery. METHODS: In this study, 1,020 patients with pulmonary nodules <2.0 cm in diameter were included in the Department of Thoracic Surgery Jiangsu Provincial People's Hospital from August 1, 2016 to December 31, 2019. Computed tomographic angiography (CTA) was performed on all the included patients before surgery. The DeepInsight software was used to perform preoperative bronchial angiography on the operative side of the lung to identify the affected pulmonary segments, pulmonary arteries and pulmonary veins. Two thoracic surgeons independently assessed the visibility of the affected pulmonary vessels using the 5-point method, and the χ² test assessed the consistency between observers. In addition, virtual imaging and real anatomy of pulmonary vessels on the operative side were performed during the operation, and the involved pulmonary vessels were finally determined by 2 chief physicians of thoracic surgery. RESULTS: There were no statistically significant differences between the number and spatial anatomy of the vessels involved in the pulmonary virtual imaging using DeepInsight software before operation and the number of vessels involved during operation in 1,020 patients. And the consistency among observers is quite satisfactory. CONCLUSIONS: The DeepInsight software virtual imaging of pulmonary bronchial vessels can accurately reconstruct the actual pulmonary vessels and assist the completion of pulmonary segmental resection.