Clinical outcomes and changes of lung function after Segmentectomy Versus Lobectomy for Lung Cancer Cases

Abstract

Objective

We aimed to compare the clinical outcomes and changes of pulmonary function tests (PFTs) after segmentectomy or lobectomy for non-small-cell lung cancer patients.

Patients and methods

The retrospective study included 212 patients who underwent segmentectomy (group S) and 2336 patients who underwent lobectomy (group L) between 1997 and 2012. Follow-up and medical records data was collected. We used all the longitudinal PFT data within 24 months after operation, and performed linear mixed modeling. We analyzed the 5-year overall survival (OS) and disease free survival (DFS) in stage IA patients. We used a propensity score case matching procedure to minimize the bias due to imbalanced group comparisons.

Results

one death (0.4%) in group S and seven (0.3%) in group L occurred in the perioperative period. The hospital stays of the two groups were similar (Median: 5.0 vs. 5.0 days; range: 2-99 vs. 2-58). Mean OS time and DFS time of T1a after segmentectomy or lobectomy seemed to be similar (4.2 years vs. 4.5 years, P=0.06; and 4.1 years vs. 4.4 years, P=0.07), respectively. Compared with segmentectomy, lobectomy yielded marginally significantly better OS (4.4 years vs. 3.9 years; P=0.05) and DFS (4.1 years vs. 3.6 years; P=0.05) in T1b cases. We did not found significantly different impact of PFT after segementectomy or lobectomy.

Conclusion

Both of the surgical types are safe. We advocate lobectomy in stage IA cases, especially in T1b cases. A retrospective study with a large sample size and more detailed information should be conducted for PFT evaluation with further stratification into lobe and side.

Introduction

Lobectomy is traditionally considered as the standard surgical procedure for primary non–small-cell lung cancer [NSCLC] (1) until segmental resection was reported.(2) Since the introduction of segmentectomy, controversy remains regarding the optimal surgical approach for early stage NSCLC.(3) Advocates for lobectomy demonstrated reduced risk of local recurrence and better prognosis in comparison to segmentectomy;(1, 4) for instance, recurrence rates were appreciably higher in the cases who underwent sub-lobar resection as compared to lobectomy (17.2% vs. 6.4%).(4) Supporters for segmentectomy believe the two operations have the similar curative effects,(3, 5-7) but segmentectomy offers better pulmonary functional preservation.(8, 9)

Our recent retrospective study on a cohort of 113 NSCLC patients (Stage IA to IIIB) , who underwent segmentectomy for primary lung cancer between 1999 and 2004, reported no perioperative mortality, significant comorbidities in 62 patients (55%) and tumor recurrence in 39 patients (35%).(10) Herein, we sequentially compare the clinical outcomes and evaluate changes of pulmonary function tests (PFTs) after segmentectomy or lobectomy on the cohort of 2548 cases who were enrolled from 1997 to 2012.

Because surgical approaches, i.e., thoracotomy or video-assisted thoracic surgery (VATS) also can potentially lead to significant discrepancy of complications or PFTs,(11, 12) in this study we stratified the cases into thoracotomy and VATS for the analysis, respectively.

Materials and Methods

Patients and Data Collection

The study protocol was reviewed and approved by the Mayo Clinic Institutional Review Board. A detailed study protocol was reported previously.(13) Briefly, between 1997 and 2012 at Mayo Clinic (Rochester, Minnesota, U.S.A.), all patients underwent CT imaging before the operations. PFTs were performed in the majority of patients as well as standard investigations for preoperative lung cancer staging, such as positron emission tomography/CT fusion scans.(14) Medical records data included age, sex, smoking status, operative procedure, mortality, and complications as well as length of hospital stay, histopathology, and preoperative and postoperative PFTs. Patients were staged postoperatively according to the 7^th edition of the TNM staging system of the American Joint Committee on Cancer (AJCC).

We stratified the cases into open thoracotomy, e.g., muscle-sparing thoracotomy (posterolateral or serratus anterior incision), or VATS for analysis, respectively, according to our published protocols.(10, 15) All the operations were performed at Mayo Clinic. Importantly, simple wedge resections without anatomical dissection of the bronchovascular structures were not incorporated as segmentectomy. Mediastinal lymphadenectomy was a standard procedure after either lobectomy or segmentectomy. Perioperative mortality included all deaths occurring within 30 days of the operative procedure and deaths that occurred later during the initial perioperative hospitalization.

PFTs included (1) % predicted FEV1 (Forced Expiratory Volume in one second), (2) % predicted FVC (Forced Vital Capacity) and (3) % predicted DLCO (Diffusion Capacity of the Lung for Carbon Monoxide). Specifically, we focused on PFTs at two time points: (1) Pre-operation: PFTs were performed within one month before the operation; (2) Post-operation: PFT data were conducted within 24 months after the operation. Thereafter, we used all the longitudinal data (repeated measures) we have, and performedsss linear mixed modeling, including the variables, i.e., surgery types and time from treatment to PFT test.

Follow-up and recurrence data was collected through detailed medical records data abstraction and self-administered questionnaires including present health and treatment update starting within six months postdiagnosis and then annually, thereafter. For deceased patients, follow-up questionnaires were sent to the next-of-kin to collect the information regarding the patients' death. When conducting survival analysis, we included the cases with stage IA, and used 5-year overall survival (OS) and 5-year disease free survival (DFS) data, i.e., the survival freedom from recurrence or progression. The patients whose follow-up time was less than 30 days or who died within 30 days after surgery were excluded from OS and DFS analyses.

Data Analysis

Descriptive statistics for categorical variables are reported as frequency and percentage, and continuous variables are reported as mean (standard deviation). Positive rates were compared using Chi-squared test or Fisher's exact tests. Between groups, analyses of variance were performed by the Student's t-test according to normal distribution or Wilcoxon's signed-rank test. OS or DFS was estimated using the Kaplan-Meier survival method. We matched the cases between the imbalanced groups by propensity scores (PS) using the Greedy algorithm.(16) Prior to matching, we calculated the PS incorporating the variables that can potentially impact clinical outcomes or PFTs, including age, gender, smoking status, Body Mass Index (BMI), lung side and lobe, cell type, tumor grade, and surgical approach. Thereafter, we used 3:1 matching between groups. All data entry and analyses were performed with SAS 9.3 software (SAS institute Inc., Cary, NC, U.S.A.).

Results

Demographic, pathological and clinical characteristics

The study included 212 patients who underwent segmentectomy (group S) and 2336 patients who underwent lobectomy (group L). The demographic and pathological characteristics of the patients are shown in Supplementary Table 1. Figure 1 indicates year-proportion of each approach and surgery type, showing VATS operations increased remarkably since 2005.

Figure 1. Proportions of surgery types of 2509 cases, 1997-2012, Mayo Clinic (Minnesota).

One death (0.4%) in group S and seven (0.3%) in group L occurred in the perioperative period. In addition, the hospital stay of groups S and L were similar (median: 5.0 vs.5.0; range: 2-99 vs.2-58, P=0.31).

Perioperative complications

Table 1 shows the complications after the operations via VATS (N=301) or thoracotomy (N=2247). The Clinical features of these cases were presented in supplementary table 2. We matched two groups with the cases of the two approaches by PS incorporating the variables, i.e., age, gender, smoking status, BMI, tumor stage, and other comorbidities which can potentially impact perioperative complications (17, 18). After PS matching, there was no significant difference of above mentioned variables between the VATS and thoracotomy groups. Before matching, 73.8% cases in the VATS group had no complications; however, , 64.0% cases in thoracotomy group had no complications (P<0.01). The trend was similar after matching (73.8% vs. 62.2%, P<0.01). In addition, Table 1 indicates incidence of atrial fibrillation and pulmonary complications in the thoracotomy group were remarkably higher than in the VATS group, before or after matching.

Table 1. Perioperative Complications of VATS vs. Thoracotomy.

	Total Cases			Matched Cases by Propensity Scores
	VATS (N=301)	Thoracotomy (N=2247)	P value	VATS (N=301)	Thoracotomy (N=903)	P value
No complications	222 (73.8%)	1439 (64.0%)	<0.01	222 (73.8%)	562 (62.2%)	<0.01
Atrial fibrillation 1	28 (9.3%)	325 (14.5%)	0.01	28 (9.3%)	142 (15.7%)	<0.01
Air leak >7 day 2	33 (11.0%)	218 (9.7%)	0.49	33 (11.0%)	96 (10.6%)	0.87
Pulmonary complications 3	8 (2.7%)	165 (7.3%)	<0.01	8 (2.7%)	64 (7.1%)	<0.01
Anaesthetic complications 4	13 (4.3%)	64 (2.8%)	0.16	13 (4.3%)	25 (2.8%)	0.18
Chylothorax or hemothorax	5 (1.7%)	35 (1.6%)	0.89	5 (1.7%)	8 (0.9%)	0.26
Acute renal failure	2 (0.7%)	18 (0.9%)	0.80	2 (0.7%)	6 (0.7%)	1.00
Gastrointestinal system 5	2 (0.7%)	22 (1.0%)	0.59	2 (0.7%)	11 (1.2%)	0.42
Urinary retention	3 (1.0%)	16 (0.7%)	0.59	3 (1.0%)	6 (0.7%)	0.56
Acute myocardial infarction	0 (0.0%)	12 (0.5%)	0.20	0 (0.0%)	5 (0.6%)	0.19
Wound 6	0 (0.0%)	10 (0.4%)	0.25	0 (0.0%)	6 (0.7%)	0.16
Embolism 7	2 (0.7%)	9 (0.4%)	0.51	2 (0.7%)	3 (0.3%)	0.43
Empyema	0 (0.0%)	5 (0.2%)	0.41	0 (0.0%)	0 (0.0%)	--

Note:

¹Development of postoperative atrial fibrillation.

²Prolonged air leak lasting >7 days.

³Pneumonia and its consequences, i.e., ARDS or respiratory failure.

⁴Postoperative confusion, vocal cord paralysis, or alcohol withdrawal symptoms.

⁵Gastroparesis or ileus.

⁶Wound dehiscence or infection.

⁷Deep venous thrombosis or pulmonary embolism.

Similarly, we evaluated the complications of groups S and L by thoracotomy approach and VATS, respectively. We matched the cases of groups S and L by PS incorporating age, gender, smoking status, BMI, tumor stage, and other comorbidity (thoracotomy approach : supplementary table 3; VATS: supplementary table 4). After PS matching, there was no significant difference of the clinical variables between groups S and L. With respect to the thoracotomy approach, there was no significant difference of incidence of complications between groups S and L before matching (Table 2). However, segmentectomy seemed to render less complications than lobectomy after matching (P=0.02, Table 2). With regard to the VATS approach, Table 3 indicates that the incidence of air leaks or pulmonary complications were significantly higher in the group S than in the group L before or after matching.

Table 2. Postoperative Complications of Segmentectomy vs. Lobectomy (Thoracotomy Approach).

	Total Cases			Matched Cases by Propensity Scores
	Lobectomy (N=2070)	Segmentectomy (N=177)	P value	Lobectomy (N=531)	Segmentectomy (N=177)	P value
No complications	1318 (63.7%)	120 (67.8%)	0.27	310 (58.4%)	120 (67.8%)	0.02
Atrial fibrillation 1	301 (14.5%)	24 (13.6%)	0.72	98 (18.5%)	24 (13.6%)	0.14
Air leak >7 day 2	204 (9.9%)	13 (7.3%)	0.28	56 (10.5%)	13 (7.3%)	0.21
Pulmonary complications 3	149 (7.2%)	16 (9.0%)	0.37	52 (9.8%)	16 (9.0%)	0.77
Anaesthetic complications 4	59 (2.9%)	5 (2.8%)	0.98	19 (3.6%)	5 (2.8%)	0.63
Chylothorax or hemothorax	34 (1.6%)	1 (0.6%)	0.27	4 (0.8%)	1 (0.6%)	0.79
Acute renal failure	18 (0.9%)	0 (0.0%)	0.21	3 (0.6%)	0 (0.0%)	0.31
Gastrointestinal system 5	20 (1.0%)	2 (1.1%)	0.83	6 (1.1%)	2 (1.1%)	1.00
Urinary retention	15 (0.7%)	1 (0.6%)	0.81	5 (0.9%)	1 (0.6%)	0.64
Acute myocardial infarction	12 (0.6%)	0 (0.0%)	0.31	2 (0.4%)	0 (0.0%)	0.41
Wound 6	8 (0.4%)	2 (1.1%)	0.15	1 (0.2%)	2 (1.1%)	0.10
Embolism 7	7 (0.3%)	2 (1.1%)	0.11	1 (0.2%)	2 (1.1%)	0.10
Empyema	5 (0.2%)	0 (0.0%)	0.51	1 (0.2%)	0 (0.0%)	0.56

Note:

¹Development of postoperative atrial fibrillation.

²Prolonged air leak lasting >7 days.

³Pneumonia and its consequences, i.e., ARDS or respiratory failure.

⁴Postoperative confusion, vocal cord paralysis, or alcohol withdrawal symptoms.

⁵Gastroparesis or ileus.

⁶Wound dehiscence or infection.

⁷Deep venous thrombosis or pulmonary embolism.

Table 3. Postoperative Complications of Segmentectomy vs. Lobectomy (VATS Approach).

	Total Cases			Matched Cases by Propensity Scores
	Lobectomy (N=266)	Segmentectomy (N=35)	P value	Lobectomy (N=105)	Segmentectomy (N=35)	P value
No complications	197 (74.1%)	24 (68.6%)	0.49	76 (72.4%)	24 (68.6%)	0.67
Atrial fibrillation 1	26 (9.8%)	2 (5.7%)	0.44	11 (10.5%)	2 (5.7%)	0.40
Air leak >7 day 2	25 (9.4%)	8 (22.9%)	0.02	8 (7.6%)	8 (22.9%)	0.01
Pulmonary complications 3	5 (1.9%)	3 (8.6%)	0.02	0 (0.0%)	3 (8.6%)	<0.01
Anaesthetic complications 4	11 (4.1%)	2 (5.7%)	0.66	7 (6.7%)	2 (5.7%)	0.84
Chylothorax or hemothorax 5	4 (1.6%)	0 (0.0%)	0.47	0 (0.0%)	0 (0.0%)	--
Acute renal failure	1 (0.4%)	1 (2.9%)	0.09	0 (0.0%)	1 (2.9%)	0.08
Gastrointestinal system 6	2 (0.8%)	0 (0.0%)	0.61	2 (1.9%)	0 (0.0%)	0.41
Urinary retention	3 (1.1%)	0 (0.0%)	0.53	0 (0.0%)	0 (0.0%)	--
Embolism 7	2 (0.8%)	0 (0.0%)	0.61	0 (0.0%)	0 (0.0%)	--

Note:

¹Development of postoperative atrial fibrillation.

²Prolonged air leak lasting >7 days.

³Pneumonia and its consequences, i.e., ARDS or respiratory failure.

⁴Postoperative confusion, vocal cord paralysis, or alcohol withdrawal symptoms.

⁵Gastroparesis or ileus.

⁶Wound dehiscence or infection.

⁷Deep venous thrombosis or pulmonary embolism.

Postoperative changes of PFT

We analyzed the cases only after thoracotomy. We did not analyze the data after VATS, due to the small sample size. Therefore, we had 264 cases who had both preoperative PFTs conducted one month before operation, and postoperative PFT data conducted within 24 months after operation. Among them, there were 17 and 247 cases in groups S and L, respectively. Supplementary table 5 suggested that there was no significant difference of age, gender, smoking status, BMI, tumor stage, lobe and side, and COPD between groups S and L after we matched these variables. Additionally, there was no significant difference of preoperative PFT data after PS matching (supplementary table 5).

Linear mixed modeling indicated that there was no significant trend within each group or difference of the trend of postoperative PFT data, i.e., FEV1, FVC and DLCO following segementectomy or lobectomy (Figure 2). As a consequence, we did not found significantly different impact of PFT after segementectomy or lobectomy.

Figure 2.

There was no significant difference of the trend of postoperative PFT data, i.e., FEV1, FVC and DLCO following segementectomy or lobectomy. A: FEV1; B: FVC; and C: DLCO

Survival analysis

To evaluate the possible impact of tumor size on local recurrence and survival (19, 20), we stratified the cases in stage IA into two subsets, T1a and T1b, according to tumor size.

We matched the PS of groups S and L incorporating the variables, i.e., age, gender, smoking status , BMI, tumor differentiation, side and lobe, surgical approach and other comorbidities, which can potentially impact survival ,(21-23) (Supplementary table 1). Thereafter, there was no significant difference regarding the abovementioned variables between groups S and L after PS matching (Supplementary table 1). Figures 3 A and 3B indicate that mean overall survival time and disease free survival time of T1a after segmentectomy or lobectomy seemed to be similar (4.2 years vs. 4.5 years, P=0.06; and 4.1 years vs. 4.4 years, P=0.07), respectively. However, compared with segmentectomy, lobectomy yielded marginally significantly better OS (4.4 years vs. 3.9 years; P=0.05) and DFS (4.1 years vs. 3.6 years; P=0.05) in T1b cases (Figures 3C and 3D).

Figure 3. Overall survival and disease free survival of T1a and T1b cases after segmentectomy or lobectomy.

Note: A and B: There was no significant difference on overall survival and disease free survival of cases of T1a after segmentectomy or lobectomy, respectively; C and D: There was marginally significant difference on overall survival and disease free survival of cases of T1b after segmentectomy or lobectomy, respectively.

Discussion

The surgical strategies of early NSCLC should include two contradictory aspects: maximum resection to reduce the risk of recurrence and minimum invasion to preserve pulmonary function. Thus far, segmentectomy and lobectomy are the two major operations for early NSCLC; however, the optimal surgical approach is still inconclusive. A controversy began regarding the best surgical treatment option for early NSCLC since the publication of a prospective randomized study that compared sublobar resections (segmentectomy and wedge resections) with lobectomy for stage IA NSCLC.(4) The study indicated that recurrence rates were appreciably higher in the cases who underwent sublobar resection as compared to lobectomy (17.2% vs 6.4%). However, there was no statistically significant survival difference between the cases after the sublobar resection or lobectomy. A population-based study in stage I cases demonstrated lobectomy conferred superior unadjusted overall (p<0.0001) and cancer-specific 5-year survival compared with segmentectomy (p=0.0053).(1) Nevertheless, other studies have obtained the different results.(3, 5, 6) In a Japanese study, Yamashita et al (5) indicated the similar recurrent and metastatic rate in stage I cases between segmentectomy and lobectomy groups (7.9% vs. 5.6%, p>0.05 and 5.3% vs. 5.6%, p>0.05). Additionally, in an American study, Shapiro et al (3) demonstrated the similar recurrence rate in stage I cases between the aforementioned groups (17.2% vs 20.4%, p=0.71). In a Chinese study, Zhong et al (6) unveiled that no significant difference was observed in 5-year overall or disease-free survival in stage I cases after segmentectomy or lobectomy. Okada et al (24) and Warren et al (25) have provided the compromises that segmentectomy could be safe and considered for stage I patients with tumors < 3cm and <2 cm, respectively.

With regard to the impact of the two operations on postoperative function, the results and conclusions are very complicated. Harada et al(8) have found that the extent of removed lung parenchyma directly affected postoperative functional loss, and segmentectomy offered significantly better functional preservation compared with lobectomy. In another study, Yoshimoto et al (9) demonstrated that postoperative FEV1 of increased after lobectomy or segmentectomy on left upper lobe, whereas it decreased after right upper lobectomy. However, lobectomy or segmentectomy on bilateral lower lobe had no impact on postoperative FEV1. (9) Kashiwabara et al (26) concluded that segmentectomy should be considered in patients with cT1N0 NSCLC with a normal (>80%) predicted postoperative FEV1. In patients with a predicted normal value of FEV under 70%, segmentectomy offers no functional advantages over lobectomy. Collectively, the impact of PFT after the operation depends on emphysema status, operation side and lobe, as well as surgical types.

Herein, we retrospectively studied clinical outcomes and postoperative changes of PFTs based on a cohort of 2548 cases. We acknowledge the limitation of the study: it is a retrospective comparative analysis rather than a prospective randomized comparison. To minimize the selection bias, we used a propensity score case matching procedure and then compared the matched groups. Propensity score case is regarded as the best method to address the selection bias issue in the context of a retrospective non randomized comparison. (16) Generally, many-to-one (M:1) matching on the propensity score ranges from 4:1 to 1:1. (27) Increasing the “M” tends to increase the bias in the estimated treatment effect; conversely, increasing the “M” decreases the sampling variability, (27) and improves the precision of the hazard ratio.(28) Herein, we presented the results by using 3:1 matching between groups.

In the study, we found very few cases had perioperative death following segmentectomy (0.4%) or lobectomy (0.3%), suggesting either of the procedures is safe. VATS approach rendered less perioperative complications than thoracotomy. The incidences of complications after open segmentectomy were less than open lobectomy after PS matching. However, the incidence of air leak or pulmonary complications were higher in the cases who underwent VATS segmentectomy compared with VATS lobectomy, probably due to learning curve.

Among T1a cases, OS or DFS after segmentectomy or lobectomy seemed to be similar. However, as compared with segmentectomy, lobectomy rendered relatively better OS or DFS in T1b cases.

We acknowledge there are limitations of PFT analysis in the study, e.g., the numbers of group S are quite low, and we have no information about the types of segmentectomy. Therefore, we can't stratify the cases into sides and lobes. However, at the very least, we did not found the significant difference of PFT after segmentectomy or lobectomy. Therefore, a retrospective study with a large sample size and more detailed information, e.g., types of segmentectomy, should be conducted for PFT evaluation with further stratification into lobe and side.

Collectively, we believe both of the surgical types are safe. Therefore, we advocate lobectomy in stage IA cases, especially in T1b cases. Given the disparity of the recurrence rate and survival following the two surgical procedures, we agree with Nomori et al 's view (29) that segmentectomy should be performed with extensive hilar/mediastinal lymph node dissection and a sufficient surgical margin.

Supplementary Material

Supplementary table 1. Demographic and pathological data (Segmentectomy vs. Lobectomy)

Supplementary table 2. Clinical features of the cases to compare postoperative complications (VATS vs. Thoracotomy)

Supplementary table 3. Clinical features of the cases to compare complications via thoracotomy approach (Segmentectomy vs. Lobectomy)

Supplementary table 4. Clinical features of the cases to compare complications via VATS (Segmentectomy vs. Lobectomy)

Supplementary table 5. Clinical features and preoperative PFT of the cases to compare PFT changes (Thoracotomy Approach)

Abbreviations

DLCO

Diffusing capacity of the lung for carbon monoxide

DFS

Disease Free Survival

FEV1

Forced Expiratory Volume in one second

FVC

Forced Vital Capacity

NSCLC

Non–Small-Cell Lung Cancer

PS

Propensity Scores

OS

Overall Survival

PFT

Pulmonary Function Test

VATS

Video Assisted Thoracic Surgery