The learning curve of uniportal video-assisted thoracoscopic lobectomy assisted with the three-dimensional CT reconstruction of the pulmonary vessels and bronchus

张淼,武文斌,杨敦鹏,张辉,胡正群,刘冬,李敏,王其斌

DOI: https://doi.org/10.3760/cma.j.cn101202-20190211-00050

2020-01-01

Abstract:Objective:To explore the learning curve of lobectomy by uniportal video-assisted thoracoscopic surgery (U-VATS) assisted with three-dimensional computed tomography (3D-CT) reconstruction of the pulmonary vessels and bronchus for resectable lung cancer.Methods:A retrospective cohort study was conducted. The data of 59 lung cancer patients undergoing intercostal U-VATS lobectomy by the same surgeon team in Xuzhou Central Hospital between January 2017 and January 2019 were retrospectively reviewed. This study consisted of 32 male and 27 female patients with the age of 34-81 (62.8±9.5) years. The patients were divided into subgroups of A (15 cases), B (15 cases), C (15 cases) and D (14 cases) according to the order of surgery. Preoperative 3D-CT bronchography and angiography (3D-CTBA) were performed by OsiriX software using thin-slice CT data for each patient, with the aim to observe the numbers and drainage of pulmonary arteries, veins, and bronchus branches, followed by lobectomy simulation. Then, U-VATS lobectomy and systemic lymphadenectomy of hilar and mediastinum were performed according to the previous surgical plan. The operative time, intraoperative blood loss, stations and numbers of dissected lymph nodes, rate of transfer to multi-port VATS or thoracotomy, chest tube duration and total drainage volume, complication rate, visual analog scale (VAS) of pain on the 14th postoperative day, and hospital stay were compared, respectively. Cumulative sum (CUSUM) analysis of the operation time and blood loss was utilized to verify the number of operation cases needed for the stabilization of the learning curve.Results:The 3D-CTBA indicated 12 cases (20.3%, 12/59) had variation in pulmonary arteries, veins or bronchus. No thoracotomy, massive hemorrhage, or mortality was observed in this cohort. One case in group B reported an injury of bronchial artery. These groups indicated similar intraoperative blood loss, stations and numbers of harvested lymph nodes, conversion rate of surgical approaches, complications, chest drainage duration, and postoperative hospital stay (all P values>0.05, respectively). The operation time of the group A, B, C, D was (130.7±17.7) min, (103.7±11.1) min, (87.7±5.9) min, (88.9±6.3) min, respectively. Meanwhile, the postoperative hospital stays of these subgroups was (4.5±3.0) d, (3.8±2.2) d, (2.5±1.0) d, (2.5±0.8) d, respectively. The differences of these parameters between the four groups were statistically significant ( F=45.807、3.530, all P values<0.05). The operation time of group C and D was significantly shorter than that of group A and B ( P<0.05), whereas there was no significant difference between group C and D( P>0.05). All the patients were followed up for 6-30 (average, 22) months. During the follow-up, there were no recurrence or metastasis of the tumor during the radiological exams. The learning curve was drawn with the operation time and intraoperative blood loss as indicators, and the fitting formula of the scatter plot were as follows: =127.7+ 0.8 X-0.1 X2+ 1.4e -3X3, R2=0.705; =84.6 - 2.5log X, R2=0.019. Accordingly, the CUSUM fitting formulae were as follows: =37.2+ 2.3 X-0.2 X2+ 2.8e -3X3, R2=0.701; =19.3-2.8 X+ 9.5e -2X2-9.1e -4X3, R2=0.090. Conclusions:The learning curve of U-VATS lobectomy assisted with 3D-CTBA is about 30 cases. Preoperative resection simulation using 3D-CTBA might be helpful to reduce the risk of vascular injury, shorten the operation time and the learning curve of the doctors.