Original Article
Video-assisted thoracoscopic lobectomy: which is the learning curve of an experienced consultant?
Abstract
Background: This study evaluates the number of video-assisted thoracic surgery-lobectomies (VATS-lobectomies) required for an experienced consultant thoracic surgeon to obtain competence and to perform standard quality surgery.
Methods: We have analysed the initial VATS-experience (January 2012 to September 2014) of a confirmed senior consultant who has performed 145 consecutive anatomic resections by thoracoscopy. After excluding bilobectomies, segmentectomies, and lobectomies for infectious disease, we have focused into 119 consecutive lobectomies, classified into 4 chronologic groups of 30 each. We have considered: demographics; pathology; postoperative outcomes; conversion rate; morbidity. We compared the 4 groups in a Bayesian inference model (very strong probability of a difference if Pr>95% or <5%; strong probability if 95%>Pr>80% or 5%Results: There was a very strong probability of difference of group 1 (first 30 lobectomies) compared to the 3 other groups: less incomplete fissures (Pr1<2=0.019, Pr1<3=0.037, Pr1<4=0.046), more node samplings (Pr1>2=0.977, Pr1>3=0.96, Pr1>4=0.997) and, conversely, less radical dissections (Pr1<2=0.022, Pr1<3=0.039, Pr1<4=0.003), less harvested nodes (Pr1<2≤0.001, Pr1<3≤0.001, Pr1<4≤0.001), less pleural adhesions (Pr1<2=0.077, Pr1<3=0.044). Instead, there was a very strong probability of difference of group 4 compared to the first three groups (first 90 lobectomies): lower conversion rate (Pr1>4=0.992, Pr3>4=0.996, Pr2>4=0.995), lower duration of the operation (Pr1>4=0.946, Pr2>4=0.901, Pr3>4=0.932), less air leak (Pr1>4=0.936, Pr2>4=0.97) and shorter chest tube drainage (Pr1>4=0.94, Pr2>4=0.94, Pr3>4=0.937), as well as shorter hospital stay (Pr2>4=0.94, Pr3>4=0.937).
Conclusions: The learning curve was bimodal. After the initial 30 lobectomies, oncologic quality of the procedure improved and stabilized. The surgeon became less selective and accepted to proceed with more complex cases (incomplete fissures, pleural adhesions). Efficiency was obtained after 90 lobectomies (shorter operative time and lower conversion rate).
Methods: We have analysed the initial VATS-experience (January 2012 to September 2014) of a confirmed senior consultant who has performed 145 consecutive anatomic resections by thoracoscopy. After excluding bilobectomies, segmentectomies, and lobectomies for infectious disease, we have focused into 119 consecutive lobectomies, classified into 4 chronologic groups of 30 each. We have considered: demographics; pathology; postoperative outcomes; conversion rate; morbidity. We compared the 4 groups in a Bayesian inference model (very strong probability of a difference if Pr>95% or <5%; strong probability if 95%>Pr>80% or 5%Results: There was a very strong probability of difference of group 1 (first 30 lobectomies) compared to the 3 other groups: less incomplete fissures (Pr1<2=0.019, Pr1<3=0.037, Pr1<4=0.046), more node samplings (Pr1>2=0.977, Pr1>3=0.96, Pr1>4=0.997) and, conversely, less radical dissections (Pr1<2=0.022, Pr1<3=0.039, Pr1<4=0.003), less harvested nodes (Pr1<2≤0.001, Pr1<3≤0.001, Pr1<4≤0.001), less pleural adhesions (Pr1<2=0.077, Pr1<3=0.044). Instead, there was a very strong probability of difference of group 4 compared to the first three groups (first 90 lobectomies): lower conversion rate (Pr1>4=0.992, Pr3>4=0.996, Pr2>4=0.995), lower duration of the operation (Pr1>4=0.946, Pr2>4=0.901, Pr3>4=0.932), less air leak (Pr1>4=0.936, Pr2>4=0.97) and shorter chest tube drainage (Pr1>4=0.94, Pr2>4=0.94, Pr3>4=0.937), as well as shorter hospital stay (Pr2>4=0.94, Pr3>4=0.937).
Conclusions: The learning curve was bimodal. After the initial 30 lobectomies, oncologic quality of the procedure improved and stabilized. The surgeon became less selective and accepted to proceed with more complex cases (incomplete fissures, pleural adhesions). Efficiency was obtained after 90 lobectomies (shorter operative time and lower conversion rate).
