Abstract
Objective:
The objective of this study was to evaluate the impact of the video-assisted thoracoscopic (VATS) approach on the outcomes of patients who underwent pneumonectomy.
Methods:
The effect of the surgical approach on perioperative complications and survival in patients who underwent pneumonectomy for non-metastatic non-small cell lung cancer (NSCLC) across 3 institutions (2000-2016) was assessed using multivariable logistic regression, Cox proportional hazard analysis and propensity-score matching. Completion pneumonectomies were excluded from this study and an “intent-to-treat” analysis was performed.
Results:
During the study period, 359 patients met inclusion criteria and underwent pneumonectomy for non-metastatic NSCLC: 124 (35%) VATS and 235 (65%) thoracotomy. Perioperative mortality (VATS 7% [n=9] vs open 8% [n=19]; p=0.75) and morbidity (VATS 28% [n=35] vs open 28% [n=65], p=0.91) were similar between the groups, even after multivariable adjustment. Using a VATS approach, there was a significantly greater number of lymph nodes (p < 0.001) and N2 stations dissected (p < 0.001). Further, there was a significantly shorter chest tube duration for the VATS group (VATS 1.2±1.4 days vs open 1.5±1.4 days, p = 0.01). VATS showed similar 5-year survival when compared with thoracotomy in unadjusted analysis (47% [95% CI: 36-56] vs 33% [95% CI: 27-40]; p = 0.19), even after multivariable adjustment (HR = 0.76 [95% CI: 0.50-0.1.18]; p=0.23). In a propensity-score matched analysis that balanced patient characteristics, there were no significant differences found in overall survival between the two groups (p=0.69).
Conclusion:
Although the role of VATS pneumonectomy will likely become clearer as more surgeons report results, this multicenter study suggests that the VATS approach for pneumonectomy can be performed safely, with at least equivalent oncologic outcomes compared to thoracotomy.
Keywords: lung cancer, MIS surgery, thoracic surgery
Introduction
The video-assisted thoracoscopic (VATS) approach to lobectomy has been demonstrated to be a safe and appropriate treatment modality for early stage non-small cell lung cancer (NSCLC).1, 2 There are few studies, however, regarding the safety and feasibility of the VATS approach to pneumonectomy. To date, there has been only one single-center study that has compared the outcomes of VATS vs open pneumonectomy for NSCLC; in that study, no significant differences in perioperative complications, 30-day mortality and overall survival were found.3
The purpose of this study was to build upon the current data regarding the safety and feasibility of VATS pneumonectomy, by assessing the short- and long-term outcomes among patients undergoing VATS versus open pneumonectomy for NSCLC in a multicenter setting. Our objective was to test the hypothesis that the VATS approach in appropriately selected patients is not associated with significant differences in perioperative and oncologic outcomes when compared to pneumonectomy by thoracotomy.
Methods
A retrospective analysis was performed of patients with non-metastatic NSCLC who had a pneumonectomy across three institutions (Duke University Medical Center, Roswell Park Cancer Institute and Cedars Sinai Medical Center) from 2000 to 2016. The study received Institutional Review Board approval from all three institutions with individual patient consent being waived; all records were from Duke University Medical Center, Roswell Park Cancer Institute and Cedars Sinai Medical Center. Patients who underwent a completion pneumonectomy or who had clinical M1 disease were excluded from the analysis. Baseline variables and outcome variables included demographics, pulmonary function (forced expiratory volume in 1 second [FEV1] and diffusing capacity of the lung for carbon monoxide [DLCO]), comorbidities, histology, preoperative clinical stage, pathological stage, chest tube duration, length of hospitalization, estimated blood lost during surgery, length of surgery, acute pain syndrome, number of lymph nodes dissected, number of stations dissected, number of N2 stations dissected, atrial fibrillation, empyema, pneumonia, pericardial effusion, myocardial infarction, and overall and recurrence-free survival. VATS pneumonectomy was defined as anatomic pneumonectomy using videoscopic guidance, an access incision (4-7 cm), and 1 to 3 other port incisions, without using a rib retractor.
Overall survival was determined from the date of pneumonectomy to death from any cause, with patients censored at the time of last known follow-up. Recurrence-free survival was determined from the date of pneumonectomy to death or recurrence as evaluated by biopsy or imaging seen during follow-up, with patients censored at the time of last follow-up.
Post-discharge follow-up data were collected through direct contact with patients and physicians, clinic notes, and a national death registry. Preoperative staging was determined prior to resection based on computed tomographic (CT) scan, positron emission tomographic (PET) scan, brain imaging with CT or magnetic resonance imaging, bronchoscopy, mediastinoscopy, thoracoscopic staging, and needle biopsy. The final pathology report was used to determine postoperative pathological stage.
Statistical analysis
Baseline characteristics and outcomes were compared between the VATS and open groups using the Student’s unpaired t-test or the Wilcoxon rank-sum test when applicable for continuous variables and the Pearson’s χ2-test or Fisher’s exact test when applicable for categorical variables. Patients who underwent conversions from VATS to open pneumonectomy were assessed using an intent-to-treat analysis.
Overall and recurrence-free survival was evaluated using the Kaplan–Meier method and the log-rank test. A Cox proportional hazards model was used to compare overall survival between open and VATS groups, adjusting for variables chosen based on clinical relevance, which included age, sex, FEV1, DLCO, diabetes, congestive heart failure, chronic obstructive pulmonary disease (COPD), coronary artery disease, clinical T status, clinical N status, tumor size, neoadjuvant chemotherapy, neoadjuvant radiation, history of smoking, and operative year. Complete case analysis was used for this adjusted model. The proportional hazards assumption was tested for all Cox models using smooth scaled Schoenfeld residual plots—no violations of assumptions were found—with linearity confirmed for all continuous predictors included in Cox regression analysis using Martingale residuals. A logistic regression model was developed to identify predictors of conversions in the VATS pneumonectomy group, adjusting for variables chosen based on clinical relevance, which included age, sex, FEV1, DLCO, diabetes, chronic obstructive pulmonary disease (COPD), coronary artery disease, clinical T status, clinical N status, tumor size, neoadjuvant chemotherapy, neoadjuvant radiation, and operative year.
A propensity-matched analysis of open versus VATS pneumonectomy patients was performed to further evaluate potential differences in perioperative outcomes and survival between groups. Propensity scores were developed and defined as the treatment probability with the VATS approach versus open approach conditional on measured covariates. Variables included in the propensity score model were age, sex, race, FEV1, DLCO, tumor size, pathologic T status, pathologic N status, histology, diabetes, coronary artery disease, history of smoking, neoadjuvant chemotherapy, neoadjuvant radiation, hypertension, and operative year. A 1:1 nearest neighbor matching algorithm with a caliper distance of 0.01 and no replacement was then used to match patients. Following propensity-score matching, patient demographics and outcomes were assessed using standardized differences, as well as Pearson’s χ2-test or Fisher’s exact test when applicable for categorical variables and Student’s unpaired t-test or the Wilcoxon rank-sum test when applicable for continuous variables. Additionally, the Kaplan–Meier method and the log-rank test were used to evaluate overall and recurrence-free survival.
Statistical significance for all tests was set as a P-value of less than 0.05. All statistical analyses were performed using Stata version 13.0 (StataCorp LP, College Station, TX, USA).
Results
Patient and treatment characteristics
Of the 359 patients who met study inclusion criteria, 124 (65%) underwent pneumonectomy via thoracotomy, and the remaining 235 (35%) patients underwent thoracoscopic pneumonectomy. Table 1 reports the baseline characteristics of the patients in each group. When compared with patients in the open group, patients in the VATS group were more likely to be female and had higher FEV1 (% predicted) (Table 1). Operative year also differed significantly between the groups; there was a higher percentage of patients who underwent VATS in more recent years (Figure S1). Eighty-nine percent (n = 318/359) of the cases were done by surgeons who have performed both thoracoscopic and open pneumonectomies. There were no other significant differences in baseline characteristics between the two groups.
Table 1.
Patient Characteristics
| Characteristic | Open approach (n = 235) | VATS approach (n = 124) | P-value |
|---|---|---|---|
| Female sex (n, %) | 77 (33) | 57 (46) | 0.01 |
| Age (mean ± SD) (years) | 60.2 ± 11.6 | 62.3 ± 11.3 | 0.07 |
| Ethnicity (n, %) | 0.31 | ||
| White | 186 (79) | 89 (72) | |
| Black | 24 (10) | 6 (5) | |
| Hispanic | 0 (0) | 0 (0) | |
| Native American | 3 (1) | 0 (0) | |
| Asian | 4 (2) | 4 (3) | |
| Middle Eastern | 3 (1) | 2 (2) | |
| Unknown | 15 (6) | 23 (19) | |
| FEV1 (mean ± SD) (% predicted)1 | 71.6 ± 19.8 | 77.4 ± 19.3 | 0.03 |
| DLCO (mean ± SD) (% predicted)2 | 71.2 ± 20.3 | 75.2 ± 18.7 | 0.06 |
| History of diabetes (n, %) | 22 (9) | 8 (6) | 0.34 |
| Hypertension (n, %) | 81 (34) | 54 (44) | 0.12 |
| COPD (n, %) | 47 (20) | 33 (27) | 0.09 |
| Congestive heart failure (n, %) | 2 (1) | 2 (2) | 0.61 |
| Coronary artery disease (n, %) | 31 (13) | 22 (18) | 0.25 |
| Stroke (n, %) | 4 (2) | 0 (0) | 0.30 |
| Pulmonary embolism (n, %) | 2 (1) | 0 (0) | 0.55 |
| History of smoking (n, %) | 200 (85) | 106 (85) | 0.82 |
| Smoking pack years (mean ± SD)3 | 39.3 ± 27.8 | 38.4 ± 28.2 | 0.47 |
| Neoadjuvant chemotherapy (n, %) | 73 (31) | 37 (30) | 0.81 |
| Neoadjuvant radiation (n, %) | 44 (19) | 16 (13) | 0.16 |
| Surgical Laterality (n, %) | 0.41 | ||
| Left | 135 (57) | 66 (53) | |
| Right | 88 (37) | 52 (42) | |
| Unknown | 12 (5) | 6 (5) | |
| Operative Year (n, %) | < 0.001 | ||
| 2000-2005 | 107 (46) | 22 (18) | |
| 2006-2010 | 78 (33) | 45 (36) | |
| 2011-2016 | 50 (21) | 57 (46) |
Data only available for 318 of 359 patients (open: n = 200, VATS: n = 118)
Data only available for 302 of 359 patients (open: n = 190, VATS: n = 112)
Data only available for 347 of 359 patients (open: n = 225, VATS: n = 122)
Table 2 presents tumor characteristics for the two groups. Patients in the open group had lower clinical N status and higher pathological T status and when compared with patients in the VATS group. There was a significant difference in histology between the two groups with adenocarcinoma being the most common histology in the VATS group and squamous cell carcinoma being the most common histology in the open group. There were no significant differences in pathological N status or pathological stage between the two groups (Table 2).
Table 2.
Tumor Characteristics
| Characteristic | Open approach (n = 235) | VATS approach (n = 124) | P-value |
|---|---|---|---|
| Clinical T status (n, %) | 0.57 | ||
| 1a | 9 (4) | 7 (6) | |
| 1b | 10 (4) | 8 (6) | |
| 2a | 70 (30) | 39 (31) | |
| 2b | 33 (14) | 13 (10) | |
| 3 | 58 (25) | 25 (20) | |
| 4 | 21 (9) | 7 (6) | |
| Unknown | 34 (14) | 25 (20) | |
| Clinical N status (n, %) | 0.001 | ||
| 0 | 110 (47) | 51 (41) | |
| 1 | 37 (16) | 30 (24) | |
| 2 | 49 (21) | 14 (11) | |
| 3 | 0 (0) | 4 (3) | |
| Unknown | 39 (17) | 25 (20) | |
| Clinical stage (n, %) | 0.72 | ||
| IA | 7 (3) | 6 (5) | |
| IB | 44 (19) | 22 (18) | |
| IIA | 36 (15) | 15 (12) | |
| IIB | 39 (17) | 26 (21) | |
| IIIA | 62 (26) | 28 (23) | |
| IIIB | 10 (4) | 6 (5) | |
| Unknown | 37 (16) | 21 (17) | |
| Pathological T status (n, %) | 0.01 | ||
| 0 | 0 (0) | 4 (3) | |
| 1a | 9 (4) | 8 (6) | |
| 1b | 4 (2) | 5 (4) | |
| 2a | 84 (36) | 55 (44) | |
| 2b | 25 (11) | 8 (6) | |
| 3 | 80 (34) | 29 (23) | |
| 4 | 22 (9) | 10 (8) | |
| Unknown | 11 (5) | 5 (4) | |
| Pathological N status (n, %) | 0.12 | ||
| 0 | 71 (30) | 41 (33) | |
| 1 | 102 (43) | 64 (52) | |
| 2 | 52 (22) | 17 (14) | |
| Unknown | 10 (4) | 3 (2) | |
| Pathologic stage (n, %) | 0.05 | ||
| 0 | 0 (0) | 4 (3) | |
| IA | 4 (2) | 5 (4) | |
| IB | 20 (9) | 16 (13) | |
| IIA | 49 (21) | 28 (23) | |
| IIB | 49 (21) | 25 (20) | |
| IIIA | 94 (40) | 37 (30) | |
| IIIB | 8 (3) | 4 (3) | |
| IV | 2 (1) | 0 (0) | |
| Unknown | 9 (4) | 5 (4) | |
| Tumor size (mean ± SD) (cm) | 5.6 ± 3.7 | 4.8 ± 2.3 | 0.07 |
| Histology (n, %) | 0.07 | ||
| Adenocarcinoma | 60 (26) | 47 (38) | |
| Squamous | 113 (48) | 51 (41) | |
| Adenosquamous | 1 (<1) | 1 (1) | |
| Large cell | 9 (4) | 5 (4) | |
| Carcinoid & Neuroendocrine | 5 (2) | 7 (6) | |
| Poorly Differentiated | 7 (3) | 1 (1) | |
| Other | 22 (9) | 10 (8) | |
| Unknown | 18 (8) | 3 (2) |
Perioperative outcomes
Table 3 reports perioperative outcomes for patients in the two groups. Patients who underwent thoracoscopic pneumonectomy had shorter chest tube duration (P = 0.01), longer operative time (P < 0.001), a greater number of lymph nodes dissected (P < 0.001), and a greater number of N2 stations dissected (P < 0.001) compared with patients who underwent pneumonectomy by thoracotomy. A VATS approach was associated with a higher incidence of pneumonia (p = 0.05) and higher incidence of pericardial effusion (p = 0.02). There were no significant differences in 30-day or 90-day all-cause mortality, length of hospitalization, and overall perioperative complications between the two groups.
Table 3.
Perioperative Outcomes
| Characteristic | Open approach (n = 235) | VATS approach (n = 124) | P-value |
|---|---|---|---|
| 30-day mortality (n, %) | 19 (8) | 9 (7) | 0.75 |
| 90-day mortality (n, %) | 32 (14) | 19 (15) | 0.66 |
| Length of hospitalization, median (IQR)4, days | 5 (4 – 7) | 5 (4 – 7) | 0.39 |
| Chest tube duration, median (mean ± SD)5, days | 1.5 ± 1.4 | 1.2 ± 1.4 | 0.01 |
| Estimated blood loss (mean ± SD) (mL)6 | 404 ± 321 | 467 ± 508 | 0.70 |
| Length of surgery (mean ± SD) (min)7 | 234 ± 109 | 321 ± 149 | < 0.001 |
| Acute pain syndrome (n, %)8 | 15 (7) | 7 (7) | 0.77 |
| Lymph nodes dissected (mean ± SD)9 | 13.1 ± 9.4 | 22.0 ± 15.3 | < 0.001 |
| Lymph node stations dissected (mean ± SD)10 | 4.2 ± 2.3 | 4.5 ± 3.7 | 0.96 |
| N2 stations dissected (mean ± SD)11 | 3.7 ± 1.8 | 4.7 ± 2.1 | < 0.001 |
| Overall morbidity (n, %) | 65 (28) | 35 (28) | 0.91 |
| Atrial fibrillation | 48 (20) | 20 (16) | 0.39 |
| Bronchopleural fistula | 12 (5) | 5 (4) | 0.65 |
| Empyema | 8 (3) | 2 (2) | 0.35 |
| Pneumonia | 16 (7) | 16 (13) | 0.05 |
| Pericardial effusion | 2 (1) | 6 (5) | 0.02 |
| Myocardial infarction | 3 (1) | 2 (2) | 1.00 |
Data only available for 315 of 359 patients (open: n = 209, VATS: n = 106)
Data only available for 269 of 359 patients (open: n = 150, VATS: n = 87)
Data only available for 184 of 359 patients (open: n = 73, VATS: n = 111)
Data only available for 214 of 359 patients (open: n = 103, VATS: n = 111)
Data only available for 309 of 359 patients (open: n = 202, VATS: n = 107)
Data only available for 289 of 359 patients (open: n = 189, VATS: n = 100)
Data only available for 312 of 359 patients (open: n = 207, VATS: n = 105)
Data only available for 293 of 359 patients (open: n = 206, VATS: n = 87)
Conversions
Twenty-three (19%) of the VATS cases were converted to open during surgery. The baseline characteristics and tumor characteristics for the patients who underwent a conversion are detailed in Supplemental Table S1. The conversion rate across all three centers decreased over time (Figure S1). The 30-day mortality for patients who underwent conversions was 17% (n = 4); of note, there were no intraoperative deaths and none of these 4 patients died from massive intraoperative hemorrhage. The laterality, tumor size, and T and N status of these four patients were as follows: left, T2bN1, 6.1 cm; right, T1bN0, 2.7 cm; right, T2aN2, 5.0 cm; right, T2aN1, 4.5 cm. The cause of death was post-operative pulmonary embolism, postoperative right sided heart failure, pneumonia and postoperative cardiac arrest. There were no significant differences in 30-day mortality between the conversions group and the open group (p=0.14). In multivariable logistic regression modeling, there were no variables found to be significantly associated with increased likelihood of converting from VATS to open (Supplemental Table S2)
Survival analysis
The median follow-up for the thoracotomy group was 21.0 (interquartile range [IQR]: 8.8–63.9) months and the median follow-up for the VATS group was 22.1 (IQR 8.9–54.6) months. There were no significant differences in median overall survival between the VATS and open groups (45.7 months [95% CI, 29.0–65.5] vs 24.6 months [95% CI,19.1–30.9]). The VATS approach did not significantly differ in long-term overall survival when compared with the open approach (5-year survival: 47% [95% CI, 36–56%] vs 33% [95% CI, 27–40%]; log-rank, p = 0.19) (Fig. 1a). In multivariable analysis, the VATS approach was not associated with improved survival when compared with the open approach (hazard ratio [HR], 0.76; 95% CI: 0.50–1.18; p = 0.23).
Figure 1:
Overall and recurrence-free survival for patients undergoing VATS vs open pneumonectomy for non-small-cell lung cancer. (A) Overall survival. This figure depicts the Kaplan–Meier overall survival estimates of the open (solid line) versus VATS group (dashed line). (B) Recurrence-free survival. This figure depicts the Kaplan–Meier recurrence-free survival estimates of the open (solid line) versus VATS group (dashed line). Tick marks represent censored subjects.
There were no differences in recurrence rates between the VATS and the open groups (26% vs 29%, p = 0.57). With regard to recurrence-free survival, there were no significant differences between the VATS and open groups in median survival (24.1 months [95% CI, 15.4–40.9] vs 16.5 months [95% CI, 13.8–22.3]) or 5-year survival (36% [95% CI, 26–46%] vs 31% [95% CI, 25–37%]; log-rank, p=0.52) (Fig. 2b). In multivariable analysis, a VATS approach was found to have no significant difference in recurrence-free survival when compared to an open approach (HR, 0.83; 95% CI, 0.55–1.25; P = 0.38).
Figure 2:
Overall and recurrence-free survival for patients undergoing VATS vs open pneumonectomy for non-small-cell lung cancer: propensity score-matched analysis. (A) Overall survival. This figure depicts the Kaplan–Meier overall survival estimates of the open (solid line) versus VATS group (dashed line) for patients included in the propensity score-matched analysis. (B) Recurrence-free survival. This figure depicts the Kaplan–Meier recurrence-free survival estimates of the open (solid line) versus VATS group (dashed line) for patients included in the propensity score-matched analysis. Tick marks represent censored subjects.
Propensity Analysis
Propensity-score matching was used to create two groups of 50 patients each who had an open or VATS pneumonectomy. Comparison of baseline patient characteristics after propensity score-matching between patients who had VATS versus open pneumonectomy demonstrates no statistically significant differences in baseline characteristics between the two groups (Table 5). After propensity matching, there were no significant differences in 30-day and 90-day mortality, length of hospitalization, estimated blood loss during surgery, length of surgery, and perioperative complications between the two groups (Table 6). Chest tube duration was significantly shorter when using the VATS approach compared to using the open approach (median: 1 day [IQR, 1–2] vs 2 days [IQR, 1–2]; P = 0.04). In addition, the VATS approach was associated with a greater number of lymph nodes dissected (P = 0.01) and a greater number of N2 stations dissected (P = 0.001) when compared with pneumonectomy by thoracotomy.
Table 5.
Patient and tumor characteristics after propensity score-matching
| Characteristic | Open approach (n = 50) | VATS approach (n = 50) | P-value |
|---|---|---|---|
| Female sex (n, %) | 20 (40) | 21 (42) | 0.84 |
| Age (mean ± SD) (years) | 61.3 ± 11.0 | 63.5 ± 10.5 | 0.32 |
| Ethnicity (n, %) | 1.00 | ||
| White | 45 (90) | 46 (92) | |
| Black | 5 (10) | 4 (8) | |
| FEV1 (mean ± SD) (% predicted) | 73.9 ± 16.7 | 76.5 ± 19.3 | 0.46 |
| DLCO (mean ± SD) (% predicted) | 71.8 ± 18.1 | 71.5 ± 20.1 | 0.93 |
| History of diabetes (n, %) | 0 (0) | 1 (2) | 1.00 |
| Hypertension (n, %) | 20 (40) | 25 (50) | 0.32 |
| COPD (n, %) | 13 (26) | 16 (32) | 0.51 |
| Clinical T status (n, %) | 0.87 | ||
| 1a | 1 (2) | 4 (8) | |
| 1b | 2 (4) | 2 (4) | |
| 2a | 23 (46) | 21 (42) | |
| 2b | 5 (10) | 5 (10) | |
| 3 | 14 (28) | 14 (28) | |
| 4 | 5 (10) | 4 (8) | |
| Clinical N status (n, %) | 0.13 | ||
| 0 | 31 (62) | 23 (46) | |
| 1 | 10 (20) | 19 (38) | |
| 2 | 9 (18) | 7 (14) | |
| 3 | 0 (0) | 1 (2) | |
| Clinical stage (n, %) | 0.96 | ||
| IA | 2 (4) | 2 (4) | |
| IB | 12 (24) | 10 (20) | |
| IIA | 10 (20) | 8 (16) | |
| IIB | 12 (24) | 15 (30) | |
| IIIA | 11 (22) | 13 (26) | |
| IIIB | 3 (6) | 2 (4) | |
| Congestive heart failure (n, %) | 0 (0) | 0 (0) | 1.00 |
| Coronary artery disease (n, %) | 7 (14) | 12 (24) | 0.20 |
| Pulmonary embolism (n, %) | 0 (0) | 0 (0) | 1.00 |
| History of smoking (n, %) | 46 (92) | 46 (92) | 1.00 |
| Smoking pack years (mean ± SD) | 41.8 ± 32.6 | 42.4 ± 24.3 | 0.73 |
| Neoadjuvant chemotherapy (n, %) | 14 (28) | 17 (34) | 0.52 |
| Neoadjuvant radiation (n, %) | 6 (12) | 6 (12) | 1.00 |
| Surgical Laterality (n, %) | 1.00 | ||
| Left | 18 (36) | 18 (36) | |
| Right | 32 (64) | 32 (64) | |
| Operative Year (n, %) | 0.70 | ||
| 1999-2003 | 10 (20) | 8 (16) | |
| 2003-2009 | 21 (42) | 19 (38) | |
| 2010-2016 | 19 (38) | 23 (46) | |
| Pathological T status (n, %) | 0.94 | ||
| 0 | 0 (0) | 0 (0) | |
| 1a | 3 (6) | 5 (10) | |
| 1b | 1 (2) | 2 (4) | |
| 2a | 25 (50) | 26 (52) | |
| 2b | 4 (8) | 4 (8) | |
| 3 | 14 (28) | 11 (22) | |
| 4 | 3 (6) | 2 (4) | |
| Pathological N status (n, %) | 0.92 | ||
| 0 | 15 (30) | 14 (28) | |
| 1 | 27 (54) | 29 (58) | |
| 2 | 8 (16) | 7 (14) | |
| Pathologic stage (n, %) | 0.91 | ||
| 0 | 0 (0) | 0 (0) | |
| IA | 2 (4) | 2 (4) | |
| IB | 6 (12) | 10 (20) | |
| IIA | 15 (30) | 14 (28) | |
| IIB | 11 (22) | 8 (16) | |
| IIIA | 15 (30) | 15 (30) | |
| IIIB | 1 (2) | 1 (2) | |
| IV | 0 (0) | 0 (0) | |
| Tumor size (mean ± SD) (cm) | 4.93 ± 2.64 | 4.77 ± 2.41 | 0.91 |
| Histology (n, %) | 0.83 | ||
| Adenocarcinoma | 18 (36) | 20 (40) | |
| Squamous | 22 (44) | 22 (44) | |
| Large cell | 4 (8) | 2 (4) | |
| Carcinoid & Neuroendocrine | 2 (4) | 4 (8) | |
| Poorly Differentiated | 2 (4) | 1 (2) | |
| Other | 2 (4) | 1 (2) |
Table 6.
Perioperative outcomes after propensity matching
| Characteristic | Open approach (n = 50) | VATS approach (n = 50) | P-value |
|---|---|---|---|
| 30-day mortality (n, %) | 3 (6) | 2 (4) | 1.00 |
| 90-day mortality (n, %) | 4 (8) | 7 (14) | 0.53 |
| Length of hospitalization, median (IQR), days | 5 (4-6) | 5 (4-6) | 0.10 |
| Chest tube duration, median (IQR)12, days | 2 (1-2) | 1 (1-2) | 0.04 |
| Estimated blood loss (mean ± SD) (mL)13 | 378 ± 221 | 395 ± 408 | 0.48 |
| Length of surgery (mean ± SD) (min)14 | 239 ± 92 | 272 ± 93 | 0.14 |
| Acute pain syndrome (n, %)15 | 5 (10) | 4 (8) | 0.75 |
| Lymph nodes dissected (mean ± SD)16 | 12.9 ± 9.8 | 20 ± 15.5 | 0.01 |
| Lymph node stations dissected (mean ± SD) | 4.8 ± 2.3 | 4.9 ± 4.1 | 0.56 |
| N2 stations dissected (mean ± SD) | 3.7 ± 1.6 | 4.9 ± 1.9 | 0.001 |
| Overall morbidity (n, %) | 12 (24) | 14 (28) | 0.65 |
| Atrial fibrillation | 10 (20) | 11 (22) | 0.81 |
| Bronchopleural fistula | 1 (2) | 1 (2) | 1.00 |
| Empyema | 0 (0) | 0 (0) | 1.00 |
| Pneumonia | 3 (6) | 4 (8) | 1.00 |
| Pericardial effusion | 1 (2) | 3 (6) | 0.62 |
| Myocardial infarction | 0 (0) | 0 (0) | 1.00 |
Data only available for 79 of 100 patients (open: n = 41, VATS: n = 38)
Data only available for 68 of 100 patients (open: n = 19, VATS: n = 49)
Data only available for 77 of 100 patients (open: n = 28, VATS: n = 49)
Data only available for 99 of 100 patients (open: n = 50, VATS: n = 99)
Data only available for 90 of 100 patients (open: n = 44, VATS: n = 46)
After propensity score-matching, the median follow-up for the VATS group was 40.4 (IQR: 16.9, 61.2) months and the median follow-up for the thoracotomy group was 26.9 (IQR: 12.0, 66.7) months. There were no significant differences in median survival between the VATS (63.1 months [95% CI, 42.0–81.3]) and open (38.4 months [95% CI, 21.7–100.2]) approaches. In addition, there were no significant differences in long-term survival between the VATS and open approach (5-year survival: 60% [95% CI, 43–74%] vs 44% [95% CI, 29–58%]; log-rank, p = 0.69 (Fig. 2a). With regard to recurrence-free survival, there were no significant differences between the VATS and open groups in median survival (39.7 months [95% CI, 17.4–65.5] vs 27.0 months [95% CI, 14.9–70.8]) or 5-year survival (34% [95% CI, 19–50%] vs. 39% [95% CI, 25–53%]; log-rank, p=0.65) (Fig. 2b).
Discussion
This study compared outcomes of patients who underwent VATS or open pneumonectomy for NSCLC across three institutions. The VATS approach was associated with longer operative time, a greater number of lymph nodes dissected, and a greater number of N2 stations dissected compared with pneumonectomy by thoracotomy. There were no significant differences between the groups with regard to perioperative complications and 30-day mortality. In unadjusted analysis, the VATS approach was associated with improved overall survival but after multivariable adjustment, there were no significant differences in overall and recurrence-free survival between VATS and open pneumonectomy. In addition, in propensity score-matched analysis, there were no significant differences in overall and recurrence-free survival between the two approaches in unadjusted analysis, whereas the VATS approach was associated with improved survival via multivariable analysis. These results suggest that the use of the VATS approach for pneumonectomy is feasible in appropriately selected patients.
This study is the first multicenter study reporting the outcomes of thoracoscopic pneumonectomy. The first reports demonstrating technical feasibility of thoracoscopic pneumonectomy were published in the early 1990’s, but since then, studies have been limited largely to case reports.4–14 There have been only two other studies that have evaluated the outcomes of VATS vs open pneumonectomy in patients with NSCLC,3, 15 with findings similar to those reported in the present study. Battoo and colleagues found that in their single-center experience of 107 patients, of which 67 (63%) patients underwent a VATS approach, there were no significant differences in perioperative complications and overall survival between VATS and open pneumonectomy.3 Their conversion rate reported was 16% which is similar to the conversation rate of 19% in this study. Liu and colleagues similarly found no significant differences in perioperative complication rates between VATS and open pneumonectomy.15
In the present study, the VATS approach was associated with a greater number of lymph nodes dissected and lymph node stations sampled. We speculate that the reason why fewer lymph nodes were dissected in the open group may have been because more patients in the open group underwent induction radiation. In the open group, use of radiation may have obliterated planes and/or lymph nodes, which may have led to fewer lymph nodes dissected. Of note, the median number of N2 lymph node stations dissected in both groups was over 3, suggesting that both approaches allowed for appropriate hilar and mediastinal dissection.
In the present study, 19% of the VATS group underwent a conversion to an open pneumonectomy. With regard to appropriate patient selection for VATS pneumonectomy, we performed a logistic regression to identify predictors of conversions, and found that none of the included variables were associated with increased likelihood of converting from a VATS to open approach. Of note, in the present study, the majority of the VATS and open procedures (89%) were performed by the same surgeons, albeit during their learning curve. In general, the technical difficulty of a thoracoscopic approach alone does not determine the candidacy of the patient, as the comfort level of an individual surgeon changes over his or her learning curve. Skills learned in the VATS approach for central tumors that require lobectomy are transferrable to VATS pneumonectomy and ongoing improvements in technology, specifically optics and tools that promote dissection, make more patients eligible for VATS pneumonectomy than before.
The patients who are not good candidates for VATS pneumonectomy are those in whom it is difficult to determine, by VATS, the technical feasibility of a sleeve / pulmonary arterioplasty. In addition, very large tumors that are central requiring thoracotomy for removal is also a relative contraindication. However, even in such cases, with experience, a VATS approach can be feasible. Exposure to the hilum can be enhanced by the use of a camera and can allow for central clamping and dissection, including stapling of the left main pulmonary artery at its origin, for a central left pneumonectomy. At the very least, an initial thoracoscopic approach may allow additional views of the hilum that are less obstructed by the tumor, thereby enhancing dissection.
In the present study, all cases were first assessed to see whether a sleeve lobectomy would be feasible. Our approach is to prioritize parenchymal sparing procedures such as the sleeve lobectomy over a pneumonectomy in patients with central lung cancers (e.g. we would readily perform an open sleeve lobectomy over a VATS pneumonectomy). Assessment for sleeve lobectomy by VATS was the same as by open approaches and involved the use of imaging analysis and intraoperative dissection of interlobar structures. With regard to the 19% of cases that underwent a conversion, these were all cases where it was felt that a pneumonectomy was necessary. One important limitation of this study is that we were not able to precisely capture the decision making process of sleeve vs pneumonectomy (e.g. data on the cases where a conversion led to a sleeve lobectomy are unavailable).
Although several studies have shown that a VATS approach to lobectomy is associated with decreased pain16 and a decreased length of hospital stay,17 in the present study, no significant differences in perioperative mortality and length of hospital stay were found between the VATS and open pneumonectomy groups. This may be because compared to a lobectomy, the adverse physiological impact of a pneumonectomy may affect certain outcomes (e.g. perioperative mortality, length of stay) more significantly than the impact of the surgical approach. In this study, other important outcomes such as chronic pain scores, extent of narcotic use, length of time to return to work, inflammatory and immunologic responses, time to administration of adjuvant chemotherapy, chemotherapy tolerance, and non-cancer specific mortality were not measured.
Of note, in prior single-center reports of VATS vs open pneumonectomy, a significantly greater percentage of patients in the VATS group had reduced acute pain (as measured by a visual analogue pain scale)15 or were pain free at 1 year when compared to the open group (53% vs 19%, p=0.03)3. In the present study, we did not measure chronic pain and although we did include acute pain as a variable, this outcome was determined from physician clinic notes and not by specific visual analogue scales and other instruments. We hypothesize that evaluation of other outcomes not included in this study may show differences between VATS and open approaches to pneumonectomy and that assessment of these outcomes will be important in future studies. Furthermore, since this report describes the early experience of VATS pneumonectomy, it may be reasonable to expect that outcomes will improve with experience and enhancements that reliably occur with greater adoption.
The strength of this study is its multi-institutional design, with a large cohort of over 400 patients that allowed for evaluation of short- and long-term outcomes of VATS vs pneumonectomy using multivariable and propensity score-matched analysis. The study does have several limitations. First, it is a retrospective study, and although multivariable adjustment and propensity-matching were employed to decrease the potential effects of selection bias, unobserved confounding could still be present. Second, the power of the propensity analysis was limited because there were only 50 patients in each arm of the propensity-matched analysis; thus, the study may have been underpowered to detect small but meaningful differences. Third, the findings from this study are from three institutions that perform a high volume of VATS pneumonectomies and may be limited in its generalizability. Fourth, there has been an evolution in practice over the past 16 years which may have affected our outcomes, although we did try to account for differences in treatment “eras” by including the operative year in multivariable and propensity score-matched analysis. Fifth, the learning curve of VATS pneumonectomy was not formally assessed. Sixth, we did not have data on the exact location of the tumor (e.g. peripheral vs central) and the percentage of intrapericardial pneumonectomy in the VATS and open groups. Qualitatively, an intrapericardial approach did not deter the conduct of a VATS pneumonectomy. From early in our experience of both lobectomies and pneumonectomies, opening the pericardium thoracoscopically to achieve fresh tissue planes to avoid tedious vascular VATS dissection in an amalgamated hilum was a useful maneuver for all central, post induction cases. Thus, we do not expect that the open approach to pneumonectomy would have been associated with a much higher rate of intrapericardial pneumonectomies when compared to a VATS pneumonectomy. Seventh, data on clinical stage was not available for some patients and cancer specific survival was not available for any of the patients. Eighth, although we do have lymph node data that was presented in this study, the data is limited by the fact that these data can contain inaccuracies, given that protocols from pathologists across institutions were not uniform and details regarding how pathologists counted individual nodes are unavailable. Finally, the follow-up time for patients was relatively short and longer follow-up time would strengthen findings.
Conclusion
In this multi-institutional study, VATS pneumonectomy for NSCLC was found to be feasible and effective. When compared to an open approach, the VATS approach to pneumonectomy was found to be associated with a similar perioperative complication rate and similar overall survival and does not appear to compromise oncologic outcomes. Based on these results, we conclude that surgeons with experience in using VATS for locally advanced lung cancer may consider the thoracoscopic approach for patients with NSCLC who require pneumonectomy.
Supplementary Material
Video Legend: Left video-assisted thoracoscopic pneumonectomy.
Figure S1: This bar graph demonstrates the conversion rate over four time periods contained in the study: 1992-1999, 2000-2005, 2006-2010, and 2011-2016. It should be noted that there were no conversions in the range 1992-1999, and there were only four VATS pneumonectomies performed during this time.
Central Picture.
Overall survival for patients undergoing VATS vs open pneumonectomy for non-small-cell lung cancer.
Table 4.
Multivariable Cox Proportional Hazards Model
| Characteristic | Hazard Ratio | 95% CI | P-value |
|---|---|---|---|
| VATS | 0.76 | [0.50, 1.18] | [0.23] |
| Age | 1.03 | [1.01, 1.05] | [0.002] |
| Sex | |||
| Male | Ref | Ref | Ref |
| Female | 0.96 | [0.66, 1.39] | [0.81] |
| FEV1 (% predicted) | 1.00 | [0.99, 1.01] | [0.46] |
| DLCO (% predicted) | 1.00 | [0.99, 1.01] | [0.55] |
| Diabetes | 2.27 | [1.27, 4.06] | [0.01] |
| Congestive heart failure | 1.17 | [0.27, 5.03] | [0.83] |
| Coronary artery disease | 1.03 | [0.64, 1.64] | [0.90] |
| COPD | 1.37 | [0.90, 2.09] | [0.14] |
| Clinical T status | |||
| 1a | Ref | Ref | Ref |
| 1b | 2.72 | [0.97, 7.62] | 0.06 |
| 2a | 1.70 | [0.72, 3.99] | 0.23 |
| 2b | 2.16 | [0.81, 5.80] | 0.13 |
| 3 | 1.21 | [0.47, 3.12] | 0.69 |
| 4 | 1.91 | [0.71, 5.14] | 0.20 |
| Clinical N status | |||
| 0 | Ref | Ref | Ref |
| 1 | 1.22 | [0.79, 1.88] | 0.37 |
| 2 | 1.41 | [0.90, 2.22] | 0.13 |
| Tumor Size | 1.04 | [0.97, 1.12] | 0.29 |
| Neoadjuvant Chemotherapy | 1.99 | [1.21, 3.25] | 0.01 |
| Neoadjuvant Radiation | 0.58 | [0.32, 1.02] | 0.06 |
| History of smoking | 0.85 | [0.45, 1.62] | 0.63 |
| Operative Year | 0.96 | [0.92, 1.00] | 0.07 |
Perspective statement.
The safety and feasibility of the VATS approach to pneumonectomy is not well characterized. In this multi-institutional study, VATS pneumonectomy for non-small-cell lung cancer was found to be associated with an increased number of lymph nodes and N2 stations dissected, as well as similar short- and long-term survival, when compared to an open approach.
Acknowledgments
Sources of Support/Conflicts of Interest: The authors have no conflicts of interest to declare. TAD is a consultant for Scanlan (<$10,000). TD is a consultant for Medtronic (<$5,000).
Glossary of Abbreviations:
- VATS
video-assisted thoracoscopic surgery
- NSCLC
non-small cell lung cancer
Footnotes
Meeting Presentation: To be presented at the 98th Annual Meeting of the American Association for Thoracic Surgery, April 28-March 1, 2018, San Diego, CA
Reprints will not be available from the authors
Central Message
In this multicenter study, the VATS approach for pneumonectomy was found to be feasible and effective, with at least equivalent oncologic outcomes compared to an open approach.
References
- 1.Ettinger DS, Wood DE, Aisner DL, et al. Non-Small Cell Lung Cancer, Version 5.2017, NCCN linical Practice Guidelines in Oncology. Journal of the National Comprehensive Cancer Network : JNCCN. 2017;15:504–535. [DOI] [PubMed] [Google Scholar]
- 2.Gaudet MA, D’Amico TA. Thoracoscopic Lobectomy for Non-small Cell Lung Cancer. Surg Oncol Clin N Am. 2016;25:503–513. [DOI] [PubMed] [Google Scholar]
- 3.Battoo A, Jahan A, Yang Z, et al. Thoracoscopic pneumonectomy: an 11-year experience. Chest. 2014;146:1300–1309. [DOI] [PubMed] [Google Scholar]
- 4.Cadiere GB, Himpens J, Torres R, et al. Multimedia article. Entirely thoracoscopic pneumonectomy using the prone position: a new technique. Surgical endoscopy. 2005;19:1282–1283. [DOI] [PubMed] [Google Scholar]
- 5.Conlan AA, Sandor A. Total thoracoscopic pneumonectomy: indications and technical considerations. The Journal of thoracic and cardiovascular surgery. 2003;126:2083–2085. [DOI] [PubMed] [Google Scholar]
- 6.Craig SR, Walker WS. Initial experience of video assisted thoracoscopic pneumonectomy. Thorax. 1995;50:392–395. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Gonzalez-Rivas D, Delgado M, Fieira E, Mendez L, Fernandez R, de la Torre M. Uniportal video-assisted thoracoscopic pneumonectomy. Journal of thoracic disease. 2013;5 Suppl 3:S246–252. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Nakanishi R, Hirai A, Yamashita T, Oka S. Video-assisted thoracoscopic completion pneumonectomy for a second primary cancer: a case report. The Journal of thoracic and cardiovascular surgery. 2008;135:945–946. [DOI] [PubMed] [Google Scholar]
- 9.Nwogu CE, Glinianski M, Demmy TL. Minimally invasive pneumonectomy. The Annals of thoracic surgery. 2006;82:e3–4. [DOI] [PubMed] [Google Scholar]
- 10.McKenna RJ Jr. Lobectomy by video-assisted thoracic surgery with mediastinal node sampling for lung cancer. The Journal of thoracic and cardiovascular surgery. 1994;107:879–881; discussion 881-872. [PubMed] [Google Scholar]
- 11.Piwkowski C, Gabryel P, Kasprzyk M, Dyszkiewicz W. Video-assisted thoracic surgery pneumonectomy: the first case report in Poland. Wideochir Inne Tech Maloinwazyjne. 2012;7:197–201. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Podbielski FJ, Marquez GD, Nelson DG, Diettrich NA, Connolly MM. Thoracoscopic assisted pneumonectomy. JSLS. 1997;1:75–77. [PMC free article] [PubMed] [Google Scholar]
- 13.Roviaro G, Varoli F, Rebuffat C, et al. Major pulmonary resections: pneumonectomies and lobectomies. The Annals of thoracic surgery. 1993;56:779–783. [DOI] [PubMed] [Google Scholar]
- 14.Walker WS, Carnochan FM, Mattar S. Video-assisted thoracoscopic pneumonectomy. Br J Surg. 1994;81:81–82. [DOI] [PubMed] [Google Scholar]
- 15.Liu Y, Gao Y, Zhang H, et al. Video-assisted versus conventional thoracotomy pneumonectomy: a comparison of perioperative outcomes and short-term measures of convalescence. Journal of thoracic disease. 2016;8:3537–3542. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Bendixen M, Jorgensen OD, Kronborg C, Andersen C, Licht PB. Postoperative pain and quality of life after lobectomy via video-assisted thoracoscopic surgery or anterolateral thoracotomy for early stage lung cancer: a randomised controlled trial. The Lancet. Oncology 2016;17:836–844. [DOI] [PubMed] [Google Scholar]
- 17.Yan TD, Black D, Bannon PG, McCaughan BC. Systematic review and meta-analysis of randomized and nonrandomized trials on safety and efficacy of video-assisted thoracic surgery lobectomy for early-stage non-small-cell lung cancer. J Clin Oncol. 2009;27:2553–2562. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Video Legend: Left video-assisted thoracoscopic pneumonectomy.
Figure S1: This bar graph demonstrates the conversion rate over four time periods contained in the study: 1992-1999, 2000-2005, 2006-2010, and 2011-2016. It should be noted that there were no conversions in the range 1992-1999, and there were only four VATS pneumonectomies performed during this time.