Learning curve of totally thoracoscopic pulmonary segmentectomy

Abstract

Totally thoracoscopic pulmonary segmentectomy (TTPS) is a feasible and safe technique that requires advanced thoracoscopic skills and knowledge of pulmonary anatomy. However, data describing the learning curve of TTPS have yet to be obtained. In this study, 128 patients who underwent TTPS between September 2010 and December 2013 were retrospectively analyzed to evaluate the learning curve and were divided chronologically into three phases, namely, ascending phase (A), plateau phase (B), and descending phase (C), through cumulative summation (CUSUM) for operative time (OT). Phases A, B, and C comprised 39, 33, and 56 cases, respectively. OT and blood loss decreased significantly from phases A to C (P <0.01), and the frequency of intraoperative bronchoscopy for target bronchus identification decreased gradually (A, 8/39; B, 4/33; C, 3/56; P = 0.06). No significant differences were observed in demographic factors, conversion, complications, hospital stay, and retrieved lymph nodes among the three phases. Surgical outcomes and techniques improved with experience and volume. CUSUM_OT indicated that the learning curve of TTPS should be more than 72 cases.

Keywords

thoracoscopic / segmentectomy / learning curve / CUSUM

Introduction

Since pulmonary segmentectomy was first applied to treat bronchiectasis in 1939 [¹], it has been considered a versatile resection procedure that can be diagnostic and therapeutic in the setting of indeterminate pulmonary nodules, metastatic cancer, and stage IA non-small cell lung cancer (NSCLC) [²]. Compared with lobectomy, segmentectomy is associated with equivalent oncologic outcomes in stage IA NSCLC [³–⁵], and it provides various advantages, such as preservation of postoperative pulmonary function and increased tolerance for the resection of possible secondary cancers [⁶–⁸].

With the development of video-assisted thoracoscopic surgery (VATS), many surgeons have actively performed thoracoscopic pulmonary segmentectomies. The hospital stay of patients undergoing a thoracoscopic approach is shorter and their morbidity is lower than those of patients subjected to an open technique [⁹,¹⁰].

Total thoracoscopic pulmonary segmentectomy (TTPS) is feasible and safe, but it requires more demanding surgical skills than lobectomy [¹¹–¹⁶]. A steep learning curve must be overcome before a surgeon achieves proficiency in TTPS because it is a complex procedure. A learning curve displays a temporal relationship between a surgeon’s mastery of a specifically assigned task and the number of cases performed [¹⁷].

Data describing the learning curve of TTPS have yet to be obtained. The authors consecutively performed 128 cases covering different types and difficulty levels. In this study, cumulative summation (CUSUM) was applied to analyze the learning curve of TTPS and present the relationship between surgical outcomes and technical progress.

Study patients

Waiving of individual patient consent was approved by the Nanjing Medical University Institutional Review Board because of the retrospective design of this study. A total of 128 consecutive TTPS were performed by a team, who has performed 300 VATS lobectomies, at the First Affiliated Hospital of Nanjing Medical University from September 2010 to December 2013. The following inclusion criteria for TTPS were considered:

(1) Small peripheral nodules (diameter≤2 cm) that were (i) adenocarcinoma in situ and (ii) nodules with≥50% ground glass opacity on CT;

(2) Poor pulmonary reserve or another major comorbidity that contraindicated lobectomy;

(3) Deep indeterminate pulmonary nodules and solitary metastases that were unable to be removed by wedge resection.

Surgical methods

TTPS was performed under general anesthesia with single-lung ventilation. The first port (15 mm) was placed on the seventh or eighth intercostal space (ICS) on the mid-axillary line to insert a 10 mm 30° thoracoscope (Karl Storz, Germany) without CO₂ insufflation. The second port (20–40 mm) was located on the fourth or fifth ICS between the anterior axillary line and the mid-clavicle line. The third port (20 mm) was positioned on the seventh or eighth ICS between the sub-scapular line and the posterior axillary line. Target pulmonary veins, arteries, and bronchus were dissected individually. Bronchoscopy was conducted when the target bronchus was difficult to identify. The intersegmental veins were preserved, and the intersegmental plane was determined on the basis of the inflated–deflated line and then separated anatomically along the intersegmental veins. The inflated–deflated interface up to the outer one-third of the pulmonary parenchyma was divided using an endo-stapler. The sufficient margins of malignant lesions (margin width≥2 cm or≥tumor diameter) were ensured. N1 and N2 lymph nodes were sampled intraoperatively for frozen section analysis (Table S1). Lobectomy was carried out if the nodes were diseased or if the margin was insufficient. All of the primary malignant cases received hilar and mediastinal lymphadenectomy. Operative time (OT) was defined as the time from the first incising skin to the last closing port.

CUSUM analysis

CUSUM is the continuous accumulation of differences between the mean of all data points and an individual data point [¹⁷,¹⁸]. All of the 128 patients were sorted chronologically. The CUSUM of the first patient was the difference between the mean OT of all of the patients and the OT of the first patient. The CUSUM of the second patient was the CUSUM of the first patient added to the difference between the mean OT of all of the patients and the OT of the second patient. The CUSUM of the remaining patients were calculated in the same manner until the CUSUM of the last patient was determined. No deaths were recorded in this study. Therefore, a risk-adjusted CUSUM was unnecessary.

The row data of OT are scattered in Fig. 1. The plotted CUSUM_OT was parabolic, which demonstrated a typical learning curve (Fig. 2). The curve of best fit is a second-order polynomial equation: y = −0.3227x² + 38.667x + 96.173, where y represents CUSUM_OT and x denotes the consecutive case number. The curve yielded a high R value of 0.93. Three phases constituted the learning curve. The first peak point (the 39th case) confirmed phase A (the first 39 cases) that formed an ascending slope and represented the initial phase of the learning curve. The last peak point (the 72th case) was verified by phase B (the following 33 cases) that corresponded to the plateau of the learning curve. A descending slope was described by phase C (the remaining 56 cases) that indicated the mastery of TTPS.

Clinical data

The characteristics of the patients and nodules are summarized in Table 1. Segmentectomies were performed for 79 women and 49 men, and their mean age was 59 years (age ranging from 31 years to 82 years). The types of TTPS are summarized in Table 2 and Fig. S1.

No significant differences were found in the demographic factors and surgical procedure among the phases (Tables 1 and 2). The OT of the three phases significantly decreased (A, 217.2±58.9 min; B, 187.0±39.9 min; C, 158.7±35.9 min; P<0.01; Table 3). The OT was significantly shorter in phase C than in phase A (P<0.01) but was not significantly different between phases A and B (Table 4).

Blood loss (BL) significantly declined among the phases (A, 51.2±27.8 mL; B, 37.4±24.9 mL; C, 24.6±33.6 mL; P<0.01; Table 3). The BL in phase C was significantly lower than that in phase A (P<0.01; Table 4). No significant differences were observed in the BL between phases A and B and between phases B and C. The frequency of intraoperative bronchoscopy to identify the target bronchus decreased gradually from phase A to phase C (A, 8/39; B, 4/33; C, 3/56; P = 0.066; Table 3). Bronchoscopies were performed during the following segmentectomies: segmentectomy on the right upper lobe, further segmentectomy on the left upper division, and bilateral basilar segmentectomies. Bronchoscopy was unnecessary during the following segmentectomies: superior, lingular, basilar, and left upper division segmentectomies.

The length of hospital stay was comparable among the phases, and the number of retrieved lymph nodes was similar among the phases. Seven cases (2, 2, and 3 in phases A, B, and C, respectively) were converted to a different operative thoracoscopic procedure (Table 3): left S⁸ to S⁸⁻¹⁰ and left S³ to S¹⁺²⁺³ in phase A; right S⁹⁺¹⁰ to S⁷⁻¹⁰ and left S^1+2c to S¹⁺² in phase B; and left S¹⁺²⁺³ to an upper lobectomy, left S⁶ to low lobectomy, and right S⁹⁺¹⁰ to S⁶⁺⁹⁺¹⁰ in phase C. The changes in the strategy were due to oncological factors, such as inadequate margins, lymph node involvement, or technical reasons, such as difficulty in determining vessels and bronchus or difficulty in conducting parenchymal division. No conversion to open procedures was registered.

No severe complications or deaths occurred. The complication rates among the three phases were comparable: A, 15.4% (6/39); B, 15.2% (5/33); and C, 12.5% (7/56). The total complication occurrence was 14.1% (18/128): 6 cases of prolonged air leakage (>7 days), 3 cases of atrial fibrillation, 4 cases of pulmonary infection, and 5 cases of hemoptysis (>3 days).

Debate

The introduction of new techniques into medical practice involves additional training for health professionals [¹⁹]. Clinical outcomes have improved as surgeons increase the volume of thoracoscopic procedures. Zhao et al. [²⁰] found that the learning curve for a VATS lobectomy for lung cancer requires 30–60 consecutive patients. Meyer et al. [²¹] discovered that the learning curve of a robotic lobectomy includes 18±3 cases. Guo et al. [²²] suggested that at least 30 cases are necessary to reach general competence in thoracoscopic esophagectomy.

CUSUM was initially designed as an industrial quality control test in the early 20th century and was first described by Page in 1954. Since the 1970s, this technique has been applied in the medical field to analyze the learning curve. CUSUM can also reveal changes over time. As such, surgeons have adopted it as a form of quality control to monitor their performance and consequently validate a particular skill [²³].

In this study, CUSUM_OT demonstrated a typical learning curve, such as a parabola, which consisted of phases A, B, and C that comprised 39, 33, and 56 cases, respectively. Phase A was the initial phase of the learning curve. As the number of cases increased, the OT and BL decreased. Although outcomes, such as OT and BL, in phase B improved, phases A and B did not significantly differ. Thus, phase B was still in the accumulation phase. Compared with those in phase A, the OT and BL in phase C decreased significantly (P<0.01). Phase C showed a steadily descending slope and achieved the appropriate outcomes among the three phases. After performing between 39 and 72 cases, surgeons reached sufficient proficiency to enhance the postoperative variables.

OT is a direct measure of the complexity of a case and the basic mastery of a surgical task and surgeon comfort. Previous studies on learning curves showed that OT and BL usually decrease when professionals attempt to master a new skill [¹⁷,¹⁸,²⁰–²²]. Targeted bronchial identification and parenchymal dissection represent the most time-consuming and difficult steps of TTPS. Technical maturity in identifying a bronchus and dissecting an intersegmental plane decrease the OT and BL in the three phases.

In the early learning stages, surgeons who have been proficient in lobectomies are familiar with various segmentectomies, including those of the superior, lingular, basilar, and left upper division, whose anatomical structures are similar to those of the lobe. The left upper division and lingular segment constitute the left upper lobe that resembles the right upper lobe connected to the middle lobe with an incomplete fissure. The superior artery and the bronchus are the first branches of the lower lobe artery and the bronchus and are distally located from the basilar artery and the bronchus. The posterior pulmonary lobe with an aberrant fissure on the lower lobe frequently corresponds to the superior segment [²⁴].

Surgeons should possess additional knowledge about segmental anatomy if they plan to perform other segmentectomies, such as segmentectomy on the right upper lobe and further segmentectomy on the left upper division and basilar segment. Bronchoscopy was also applied to these segmentectomies during our learning process. In addition to studying professional anatomy books, practicing intraoperative bronchoscopy helped elucidate segmental bronchial distribution. We found that bronchoscopy intraoperatively played a decisive role when the target bronchus was difficult to identify. With the accumulation of experience, we became familiar with the segmental bronchial anatomy and gradually reduced our dependence on bronchoscopy. However, the frequency of intraoperative bronchoscopy did not significantly differ among the phases (P = 0.066) likely because of the small sample size.

Surgeons can subjectively control the degree of difficulty based on personal capacity. With the appropriate selection of patients and surgical procedures, along with increased surgical experience, excellent outcomes can be achieved. On the basis of our findings and experiences, we recommend a stepwise approach. (1) A surgeon who proficiently performs a VATS lobectomy can start with simple TTPS, such as superior and lingular segmentectomy. (2) After mastering the separation of the single intersegmental plane and understanding the segmental anatomy, the surgeon should conduct complex TTPS to explore the separation of multiple intersegment planes. (3) Finally, the smallest possible size of a target segment can be removed anatomically to maximize the preservation of healthy pulmonary parenchyma.

Several limitations of this study should be addressed. First, this research was inherent to a retrospective and single center study. Second, the procedures were performed by one team composed of three surgeons with mastery skills for minimally invasive thoracic procedures. Before conducting segmentectomy, they completed 300 VATS lobectomies within 2 years. The average OT is 134 min (ranging from 92 min to 258 min) and the average BL is 35 mL (ranging from 10 mL to 380 mL).

Summary

This study suggests that surgical outcomes and techniques improve with experience and number of cases. The appropriate selection of patients and surgical procedures is beneficial to initial TTPS. CUSUM_OT indicates that the learning curve of TTPS should be more than 72 cases. Intraoperative bronchoscopy plays an important role in technical progress.

References

[1] Churchill ED, Belsey R. Segmental pneumonectomy in bronchiectasis: the lingual segment of the left upper lobe. Ann Surg 1939; 109(4): 481–499

[2] Schuchert MJ, Abbas G, Awais O, Pennathur A, Nason KS, Wilson DO, Siegfried JM, Luketich JD, Landreneau RJ. Anatomic segmentectomy for the solitary pulmonary nodule and early-stage lung cancer. Ann Thorac Surg 2012; 93(6): 1780–1787, discussion 1786–1787

[3] Landreneau RJ, Normolle DP, Christie NA, Awais O, Wizorek JJ, Abbas G, Pennathur A, Shende M, Weksler B, Luketich JD, Schuchert MJ. Recurrence and survival outcomes after anatomic segmentectomy versus lobectomy for clinical stage I non-small-cell lung cancer: a propensity-matched analysis. J Clin Oncol 2014; 32(23): 2449–2455

[4] Okada M, Nishio W, Sakamoto T, Uchino K, Yuki T, Nakagawa A, Tsubota N. Effect of tumor size on prognosis in patients with non-small cell lung cancer: the role of segmentectomy as a type of lesser resection. J Thorac Cardiovasc Surg 2005; 129(1): 87–93

[5] Yamashita S, Tokuishi K, Anami K, Moroga T, Miyawaki M, Chujo M, Yamamoto S, Kawahara K. Thoracoscopic segmentectomy for T1 classification of non-small cell lung cancer: a single center experience. Eur J Cardiothorac Surg 2012; 42(1): 83–88

[6] Harada H, Okada M, Sakamoto T, Matsuoka H, Tsubota N. Functional advantage after radical segmentectomy versus lobectomy for lung cancer. Ann Thorac Surg 2005; 80(6): 2041–2045

[7] Yoshimoto K, Nomori H, Mori T, Ohba Y, Shibata H, Tashiro K, Shiraishi S, Kobayashi T. A segmentectomy of the right upper lobe has an advantage over a right upper lobectomy regarding the preservation of the functional volume of the right middle lobe: analysis by perfusion single-photon emission computed tomography/computed tomography. Surg Today 2010; 40(7): 614–619

[8] Wisnivesky JP, Henschke CI, Swanson S, Yankelevitz DF, Zulueta J, Marcus S, Halm EA. Limited resection for the treatment of patients with stage IA lung cancer. Ann Surg 2010; 251(3): 550–554

[9] Atkins BZ, Harpole DH Jr, Mangum JH, Toloza EM, D’Amico TA, Burfeind WR Jr. Pulmonary segmentectomy by thoracotomy or thoracoscopy: reduced hospital length of stay with a minimally-invasive approach. Ann Thorac Surg 2007; 84(4): 1107–1113

[10] Schuchert MJ, Pettiford BL, Pennathur A, Abbas G, Awais O, Close J, Kilic A, Jack R, Landreneau JR, Landreneau JP, Wilson DO, Luketich JD, Landreneau RJ. Anatomic segmentectomy for stage I non-small-cell lung cancer: comparison of video-assisted thoracic surgery versus open approach. J Thorac Cardiovasc Surg 2009; 138(6): 1318–25.e1

[11] Watanabe A, Ohori S, Nakashima S, Mawatari T, Inoue N, Kurimoto Y, Higami T. Feasibility of video-assisted thoracoscopic surgery segmentectomy for selected peripheral lung carcinomas. Eur J Cardiothorac Surg 2009; 35(5): 775–780, discussion 780

[12] Oizumi H, Kanauchi N, Kato H, Endoh M, Takeda S, Suzuki J, Fukaya K, Sadahiro M. Total thoracoscopic pulmonary segmentectomy. Eur J Cardiothorac Surg 2009; 36(2): 374–377, discussion 377

[13] Leshnower BG, Miller DL, Fernandez FG, Pickens A, Force SD. Video-assisted thoracoscopic surgery segmentectomy: a safe and effective procedure. Ann Thorac Surg 2010; 89(5): 1571–1576

[14] Gossot D, Ramos R, Brian E, Raynaud C, Girard P, Strauss C. A totally thoracoscopic approach for pulmonary anatomic segmentectomies. Interact Cardiovasc Thorac Surg 2011; 12(4): 529–533

[15] Yang CF, D’Amico TA. Thoracoscopic segmentectomy for lung cancer. Ann Thorac Surg 2012; 94(2): 668–681

[16] Oizumi H, Kanauchi N, Kato H, Endoh M, Suzuki J, Fukaya K, Sadahiro M. Anatomic thoracoscopic pulmonary segmentectomy under 3-dimensional multidetector computed tomography simulation: a report of 52 consecutive cases. J Thorac Cardiovasc Surg 2011; 141(3): 678–682

[17] Bokhari MB, Patel CB, Ramos-Valadez DI, Ragupathi M, Haas EM. Learning curve for robotic-assisted laparoscopic colorectal surgery. Surg Endosc 2011; 25(3): 855–860

[18] Liao HJ, Dong C, Kong FJ, Zhang ZP, Huang P, Chang S. The CUSUM analysis of the learning curve for endoscopic thyroidectomy by the breast approach. Surg Innov 2014; 21(2): 221–228

[19] McKenna RJ Jr. Complications and learning curves for video-assisted thoracic surgery lobectomy. Thorac Surg Clin 2008; 18(3): 275–280

[20] Zhao H, Bu L, Yang F, Li J, Li Y, Wang J. Video-assisted thoracoscopic surgery lobectomy for lung cancer: the learning curve. World J Surg 2010; 34(10): 2368–2372

[21] Meyer M, Gharagozloo F, Tempesta B, Margolis M, Strother E, Christenson D. The learning curve of robotic lobectomy. Int J Med Robot 2012; 8(4): 448–452

[22] Guo W, Zou YB, Ma Z, Niu HJ, Jiang YG, Zhao YP, Gong TQ, Wang RW. One surgeon’s learning curve for video-assisted thoracoscopic esophagectomy for esophageal cancer with the patient in lateral position: how many cases are needed to reach competence? Surg Endosc 2013; 27(4): 1346–1352

[23] Williams SM, Parry BR, Schlup MM. Quality control: an application of the cusum. BMJ 1992; 304(6838): 1359–1361

[24] Mawatari T, Murakami G, Koshino T, Morishita K, Abe T. Posterior pulmonary lobe: segmental and vascular anatomy in human specimens. Clin Anat 2000; 13(4): 257–262

RIGHTS & PERMISSIONS

Higher Education Press and Springer-Verlag GmbH Germany, part of Springer Nature