- Email: [email protected]
Comparison of Segmentectomy and Lobectomy in Stage IA Adenocarcinomas
BRIEF REPORT
Comparison of Segmentectomy and Lobectomy in Stage IA Adenocarcinomas Ze-Rui Zhao, MD,a Dong-Rong Situ, MD,b Rainbow W. H. Lau, MbChB.,a Tony S. K. Mok, MD, FRCPC,c George G. Chen, PhD,a Malcolm J. Underwood, MD,a Calvin S. H. Ng, MDa,* a
Division of Cardiothoracic Surgery, Department of Surgery, The Chinese University of Hong Kong, Prince of Wales Hospital, Hong Kong Special Administrative Region, People’s Republic of China b State Key Laboratory of Oncology in Southern China, Collaborative Innovation Centre for Cancer Medicine, and Department of Thoracic Surgery, Sun Yat-Sen University Cancer Centre, Guangzhou, People’s Republic of China c Department of Clinical Oncology, The Chinese University of Hong Kong, Prince of Wales Hospital, Hong Kong Special Administrative Region, People’s Republic of China Received 16 November 2016; revised 18 December 2016; accepted 9 January 2017 Available online - 19 January 2017
ABSTRACT Introduction: Recent studies have suggested that segmentectomy may be an acceptable alternative treatment to lobectomy for surgical management of smaller lung adenocarcinomas. The objective of this study was to compare survival after lobectomy and segmentectomy among patients with pathological stage IA adenocarcinoma categorized as stage T1b (>0 to 20 mm) according to the new eighth edition of the TNM system. Methods: In total, 7989 patients were identified from the Surveillance, Epidemiology, and End Results registry. Propensity scores generated from logistic regression on preoperative characteristics were used to balance the selection bias of undergoing segmentectomy. Overall and lung cancer–specific survival rates of patients undergoing segmentectomy and lobectomy were compared in propensity score–matched groups. Results: Overall, 564 patients (7.1%) underwent segmentectomy. Lobectomy led to better overall and lung cancer– specific survival than segmentectomy for the entire cohort (log-rank p < 0.01). After 1:2 propensity score matching, segmentectomy (n ¼ 552) was no longer associated with significantly worse overall survival (5-year survival ¼ 74.45% versus 76.67%, hazard ratio ¼ 1.09, 95% confidence interval: 0.90–1.33) or lung cancer–specific survival (5-year survival ¼ 83.89% versus 86.11%, hazard ratio ¼ 1.12, 95% confidence interval: 0.86–1.46) compared with lobectomy (n ¼ 1085) after adjustment for age, sex, lymph node quantity, and histological subtype. Similar negative findings were identified when patients were stratified according to sex, age, histological subtype, and number of evaluated lymph nodes.
Journal of Thoracic Oncology
Vol. 12 No. 5: 890-896
Conclusions: Patients who underwent segmentectomy may have survival outcomes no different than those of some patients who received lobectomy for pathological stage IA adenocarcinomas at least 10 but no larger than 20 mm in size. These results should be further confirmed through prospective randomized trials. 2017 International Association for the Study of Lung Cancer. Published by Elsevier Inc. All rights reserved. Keywords: Adenocarcinoma; Segmentectomy; Lobectomy; Eighth TNM system
Lobectomy is the standard treatment for operable NSCLC, and thus far, sublobar resection has been considered a compromise option for high-risk patients (e.g., those with advanced age or impaired pulmonary function). Growing evidence, mainly from retrospective studies, indicates that sublobar resection, especially for segmentectomy (which is considered superior to wedge resection as a more anatomical approach), may obtain *Corresponding author. Drs. Zhao and Situ equally contributed to this work. Disclosure: The authors declare no conflict of interest. Address for correspondence: Calvin S.H. Ng, MD, Division of Cardiothoracic Surgery, The Chinese University of Hong Kong, Prince of Wales Hospital, Shatin, N.T., Hong Kong SAR, China. E-mail: calvinng@surgery. cuhk.edu.hk ª 2017 International Association for the Study of Lung Cancer. Published by Elsevier Inc. All rights reserved. ISSN: 1556-0864 http://dx.doi.org/10.1016/j.jtho.2017.01.012
May 2017
outcomes similar to those of lobectomy in early-stage NSCLC that measure 2 cm or less.1 Data from the latest NSCLC staging project found that the prognosis differs for tumors sized from 1 to 5 cm, which subsequently led to the reclassification of T descriptors in the revised system.2 Of particular interest, tumors that were previously defined as T1a (2 cm) will be subclassified as T1a (1 cm) and T1b (>1 to 2 cm). In this study, we used the population-based Surveillance, Epidemiology, and End Results (SEER) registry to test the hypothesis that the survival of patients with new pathological T1bN0M0 lung adenocarcinomas who undergo segmentectomy will not be inferior to that of patients undergoing lobectomy.
Methods Study Population The population was selected from the SEER 18 registry and limited to patients in whom stage IA lung adenocarcinomas that measured at least 10 but no more than 20 mm were histologically diagnosed and treated with segmentectomy or lobectomy from 2004 to 2012. Patients’ sociodemographic information and tumor characteristics were collected. Tumor histological subtype was coded according to the 2004 WHO classification.3 Survival was defined as the interval from diagnosis until death or last follow-up. Patients who survived past 31 December 2013 were classified as censored cases. When lung cancer–specific survival (LCSS) was analyzed, causes of death other than lung cancer were treated as censored observations. The completeness of follow-up was measured by a modification of Clark’s method.4 The local institutional review board classified this study as exempt.
Statistical Analysis Overall survival (OS) and LCSS were the primary and secondary outcomes of interest. The Kaplan-Meier method was used to compare unadjusted survival of patients undergoing segmentectomy or lobectomy with a log-rank test. Propensity score matching was used to balance the potential probability of being assigned to a treatment group. A logistic regression model that included age, sex, ethnicity, marital status, tumor laterality, location, and size was used to calculate the propensity scores, and 1: 2 pair matching without replacement was implemented by greedy nearest neighbour matching.5 The covariable balance effect was assessed by a standardized difference with a threshold of 0.10. Multivariable Cox proportional hazards models that included variables with a p value less than 0.2 on univariable analysis were used to decide variables that remained in the final model. A Cox proportional hazards
Segmentectomy in Stage IA Adenocarcinomas
891
model adjusted for age, sex, histological subtype, and number of lymph nodes evaluated was applied to compare outcomes between groups. We also performed secondary subgroup analysis limited to patients with different sex, age (<70 versus 70 years), histological subtype (nonmucinous bronchioloalveolar carcinoma, mucinous adenocarcinoma, and remaining adenocarcinoma subtypes [major adenocarcinoma]), and number of lymph nodes examined (0–10 versus >10).6 The study aim was to test whether segmentectomy was not inferior to lobectomy in treating certain stage IA lung adenocarcinomas. Noninferiority was established at the level of significance if the onesided upper 95% confidence bound was below the prespecified margin of 1.25.7 Statistical power was calculated accordingly.8 Analyses were conducted by using PASW Statistics 18 statistical software (formerly SPSS Statistics [SPSS, Hong Kong Special Administrative Region, People’s Republic of China]).
Results A flow diagram is presented in Figure 1. In total, 7989 patients were identified, of whom 564 (7.1%) underwent segmentectomy. The baseline characteristics are reported in Table 1. Segmentectomy was performed more often in cases involving seniors, left-side tumors, lower lobes, and smaller lesions. After propensity score matching, 552 segmentectomies and 1085 lobectomies were preserved. Preoperative characteristics were well balanced (Fig. 2). However, those in the segmentectomy arm were more likely to have lower lobe cancer, nonmucinous bronchioloalveolar carcinoma (p [ 0.02), and fewer lymph nodes resected (p < 0.01) regardless of matching. A higher incidence of more than 10 lymph nodes resected was observed in the lobectomy group (259 of 1085 [23.9%] versus 56 of 552 [10.1%]). The median follow-up interval was 52.21 months (0– 119 months) before matching. Patients who underwent segmentectomy had worse unadjusted OS and LCSS (both p < 0.01 [Supplementary Fig. 1]). More than 10 lymph nodes resected was associated with better OS (p ¼ 0.03) and LCSS (p < 0.01) (Supplementary Fig. 2). After propensity score matching, the median follow-up intervals were 49.24 months (0–119 months) and 52.06 months (0–119 months) for segmentectomy and lobectomy, respectively. The follow-up modified Clark’s C statistic values were 91.52% and 81.46% for OS and LCSS, respectively. The unadjusted Cox proportional hazard model showed that segmentectomy provided outcomes no different from those of lobectomy in terms of OS (5-year survival ¼ 68.95% versus 70.70%, hazard ratio [HR] ¼ 1.13, 95% confidence interval [CI]: 0.94–1.36) and LCSS
892 Zhao et al
Journal of Thoracic Oncology
Vol. 12 No. 5
Figure 1. Flow diagram of the study.
(5-year survival ¼ 81.30% versus 83.63%, HR ¼ 1.21, 95% CI: 0.93–1.56) (see Supplementary Fig. 1). Age, sex, lymph node quantity, and histological subtype (not for LCSS) were identified as potential predictors for survival function under multivariable analyses (Supplementary
Table 1). The regression model adjusting for these factors showed that segmentectomy was consistently associated with outcomes that were no different in terms of OS (5-year survival ¼ 74.45% versus 76.67%, HR ¼ 1.09, 95% CI: 0.90–1.33) and LCSS (5-year
Table 1. Baseline Characteristics of Patients with Pathological Stage IA Lung Adenocarcinoma (>10 and 20 mm) Treated with Segmentectomy and Lobectomy in the Surveillance, Epidemiology, and End Results Registry Before Propensity Score Matching Characteristic Mean Age ± SD, y Female sex, n (%) Race/Ethnicity, n (%) White Black Asian or Pacific Islander Others Married, n (%) Right laterality, n (%) Tumor location, n (%) Upper lobe Middle lobe Lower lobe Unknown Median tumor size ± SD, mm Histologic subtype, n (%)a BAC, nonmucinous Mucinous adenocarcinoma Major adenocarcinomab Mean lymph nodes ± SD, n a
After Propensity Score Matching
Segmentectomy (n ¼ 564)
Lobectomy (n ¼ 7425)
SD
Segmentectomy (n ¼ 552)
Lobectomy (n ¼ 1085)
SD
69.84 ± 9.17 362 (64.2)
67.31 ± 9.61 4443 (59.8)
0.83 0.09
69.84 ± 9.10 350 (63.4)
69.71 ± 9.08 697 (64.2)
0.04 –0.02
498 (88.3) 46 (8.2) 19 (3.4) 1 (0.2) 309 (54.8) 298 (52.8)
6336 (85.3) 528 (7.1) 519 (7.0) 42 (0.6) 4346 (58.5) 4591 (61.8)
0.09 0.04 –0.16 –0.06 –0.07 –0.18
488 (88.4) 44 (8.0) 19 (3.4) 1 (0.2) 306 (55.4) 296 (53.6)
977 (90.0) 66 (6.1) 40 (3.7) 2 (0.2) 606 (55.9) 604 (55.7)
–0.05 0.07 –0.02 0 –0.01 –0.04
322 (57.1) 16 (2.8) 222 (39.4) 4 (0.8) 15.63 ± 2.93
4629 (62.3) 485 (6.5) 2227 (30.0) 84 (1.2) 16.16 ± 2.87
–0.11 –0.18 0.20 –0.04 –0.31
319 (57.8) 16 (2.9) 214 (38.8) 3 (0.6) 15.69 ± 2.90
634 (58.4) 70 (6.5) 361 (33.3) 20 (1.8) 15.79 ± 2.83
–0.01 –0.17 0.11 –0.11 –0.06
31 (5.5) 27 (4.8) 506 (89.7) 4.19 ± 5.80
197 (2.7) 321 (4.3) 6907 (93.0) 7.66 ± 6.93
0.14 0.02 –0.12 –1.38
31 (5.6) 27 (4.9) 494 (89.5) 4.19 ± 5.76
33 (3.0) 39 (3.6) 1013 (93.4) 7.43 ± 6.88
0.13 0.06 –0.14 –1.29
Variables were not included in the propensity score model, as it is not a preoperative event. Adenocarcinomas other than nonmucinous BAC or mucinous adenocarcinoma. BAC, bronchioloalveolar carcinoma. b
May 2017
Segmentectomy in Stage IA Adenocarcinomas
893
Figure 2. (A) Standardized differences of variables between patients who received lobectomy and segmentectomy. Red crosses symbolize differences before propensity matching and blue circles symbolize differences after propensity matching. Propensity score matching effectively reduced heterogeneity among variables between the two groups in comparison. (B) Mirror histogram of propensity scores for patients who received segmentectomy (below the horizontal line at zero) and lobectomy (above the horizontal line at zero). Matched patients are a subset of the original patient population and their volumes are highlighted in color. Matched groups have similar propensity score distributions.
survival ¼ 83.89% versus 86.11%, HR ¼ 1.12, 95% CI ¼ 0.86–1.46) from those of lobectomy (Fig. 3). On secondary analysis, we did not recognize different OS or LCSS rates when patients were divided according to the number of lymph nodes or pathological subtypes. Demographic characteristics in the subsets of major adenocarcinomas and those with 10 or fewer resected lymph nodes are shown in Supplementary Table 2. Consistent negative results were found when patients were stratified by age, sex, or size (Fig. 4). However, all statistical powers were less than 0.8 for these analyses.
Discussion In the current study of pathological stage IA lung adenocarcinomas larger than 10 and but no larger than 20 mm, we evaluated the survival outcomes of segmentectomy versus those of lobectomy in a noninferiority setting. Our findings showed that segmentectomy has survival no different than that of lobectomy for this specified population. Tailored therapy for patients with early-stage NSCLCs is still an issue of debate.9 A recent report has demonstrated that sublobar resection may achieve outcomes similar to those of lobectomy for NSCLCs
Figure 3. Survival curves adjusted for age, sex, histological subtype, and lymph node quantity compare treatment with segmentectomy versus with lobectomy in patients with pathological stage IA adenocarcinoma (>10 and 20 mm) after propensity score matching for overall (A) and lung cancer–specific (B) survival. Shaded area represents the confidence limits for each group.
894 Zhao et al
Journal of Thoracic Oncology
Vol. 12 No. 5
Figure 4. Multivariable Cox proportional hazards model adjusted for age, sex, histological type, and lymph node quantity for overall (A) and lung cancer–specific (B) survival by subgroup analysis in the matched population. Dashed line indicates the prespecified noninferiority margin of 1.25. Blue shaded band represents the 95% confidence interval (CI) of the hazard ratio (HR) for the overall matched population. BAC, bronchioloalveolar carcinoma; Seg, segmentectomy; Lob, lobectomy.
1 cm or smaller.10 However, limited resection was not equivalent to lobectomy in older patients with lesions 2 cm or smaller.11 Besides tumor size, histological subtype may be another important factor that affects the outcome of sublobar resection. Whether the real therapeutic effect of segmentectomy was underestimated by including patients who underwent wedge resection12 or were classified as having squamous carcinoma11 or overestimated by including lesions smaller than 1 cm remains unclear.10 Thus, we restricted the study population to those who had a
pathological stage IA lung adenocarcinoma at least 1 but no larger than 2 cm in size and underwent segmentectomy and lobectomy only. More than 10 examined lymph nodes was a strong predictor of survival for node-negative early-stage NSCLC6; hence, we integrated this parameter as well as other prognostic confounders into the Cox models for better adjustment. Although lobectomy was associated with higher chance of more than 10 resected lymph nodes, we did not observe a survival difference for segmentectomy versus lobectomy in subgroups
May 2017
divided by number of lymph nodes. However, an intentional segmentectomy with a minimum of 10 lymph nodes resected should be explored in a prospective trial. Histological subtype of lung adenocarcinoma has recently been proved to be a crucial factor that affects the survival of patients who undergo sublobar resection. One pilot study found that lobectomy did not offer an OS advantage over sublobar resection for invasive adenocarcinomas 2 cm or smaller according to the 2011 adenocarcinoma classification,13 although the number of limited resections was quite small (n ¼ 26).1 However, another population-based study found that limited resection was inferior to lobectomy in invasive lung adenocarcinomas 2 cm or smaller, whereas segmentectomy seemed to be equivalent to lobectomy on a subgroup analysis.11 On the basis of the results of our study, we postulate that the discrepancy could be attributed to the high proportion of patients with wedge resections (nearly 80%) who underwent sublobar resection in the latter study, which would potentially override the therapeutic benefit of segmentectomy.11 This study had certain limitations. Preoperative cardiopulmonary function is not available in the SEER registry, which makes it difficult to balance the higher likelihood of choosing patients with poorer cardiopulmonary status to receive sublobar resection. The lack of exact tumor location made it hard to estimate whether the lesion was amenable to segmentectomy. In addition, we were unable to evaluate whether segmentectomy would lead to better preservation of lung function, a greater chance of receiving adjuvant chemotherapy, a higher rate of insufficient margins, or relapse. Additionally, the new classification of adenocarcinoma has not been updated in the SEER registry13; therefore, the study population in this study was still heterogeneous. Patients with even small proportions of micropapillary or solid components have a worse prognosis, which may affect the decision to perform a sublobar procedure.14,15 Furthermore, the lack of clinical staging forced us to select patients solely on the basis of pathological stage, which would inevitably exclude some clinical stage IA cases that were found to be pathological stage IB (e.g., pleural invasion). As excluding such patients would falsely improve survival, the ability to apply the current findings to all clinical stage IA patients could be limited.
Acknowledgments This work was supported by Research Grants Council General Research Fund grant 14117715, Chinese
Segmentectomy in Stage IA Adenocarcinomas
895
University of Hong Kong direct grant 2015.1.087, and Foundation for Science and Technology Research Project of Guangdong grant 2014B020212014. The authors would like to thank the staff members of the National Cancer Institute and their colleagues across the United States who have been involved with the Surveillance, Epidemiology, and End Results program.
Supplementary Data Note: To access the supplementary material accompanying this article, visit the online version of the Journal of Thoracic Oncology at www.jto.org and at http://dx.doi. org/10.1016/j.jtho.2017.01.012.
References 1. Dembitzer FR, Flores RM, Parides MK, Beasley MB. Impact of histologic subtyping on outcome in lobar vs sublobar resections for lung cancer: a pilot study. Chest. 2014;146:175–181. 2. Rami-Porta R, Bolejack V, Crowley J, et al. The IASLC lung cancer staging project: proposals for the revisions of the t descriptors in the forthcoming eighth edition of the TNM classification for lung cancer. J Thorac Oncol. 2015;10:990–1003. 3. Travis WE, Brambilla E, Müller-Hermelink HK, Harris CC, eds. Pathology and Genetics; Tumours of the Lung, Pleura, Thymus and Heart. Lyon, France: IARC Press; 2004. 4. Wu Y, Takkenberg JJ, Grunkemeier GL. Measuring follow-up completeness. Ann Thorac Surg. 2008;85: 1155–1157. 5. Austin PC. The use of propensity score methods with survival or time-to-event outcomes: reporting measures of effect similar to those used in randomized experiments. Stat Med. 2014;33:1242–1258. 6. Ludwig MS, Goodman M, Miller DL, Johnstone PA. Postoperative survival and the number of lymph nodes sampled during resection of node-negative non-small cell lung cancer. Chest. 2005;128:1545–1550. 7. Rothmann M, Li N, Chen G, Chi GY, Temple R, Tsou HH. Design and analysis of non-inferiority mortality trials in oncology. Stat Med. 2003;22:239–264. 8. Chow SC, Wang H, Shao J. Sample Size Calculations in Clinical Research. CRC Press; 2007. 9. Ng CSH, Zhao ZR, Lau RWH. Tailored therapy for stage I non–small-cell lung cancer. J Clin Oncol. 2017;35: 268–270. 10. Kates M, Swanson S, Wisnivesky JP. Survival following lobectomy and limited resection for the treatment of stage I non-small cell lung cancer 1 cm in size: a review of SEER data. Chest. 2011;139:491–496. 11. Veluswamy RR, Ezer N, Mhango G, et al. Limited resection versus lobectomy for older patients with early-stage lung cancer: impact of histology. J Clin Oncol. 2015;33:3447–3453. 12. Smith CB, Swanson SJ, Mhango G, Wisnivesky JP. Survival after segmentectomy and wedge resection in stage
896 Zhao et al I non-small-cell lung cancer. J Thorac Oncol. 2013;8: 73–78. 13. Travis WD, Brambilla E, Noguchi M, et al. International Association for the Study of Lung Cancer/ American Thoracic Society/European Respiratory Society international multidisciplinary classification of lung adenocarcinoma. J Thorac Oncol. 2011;6: 244–285.
Journal of Thoracic Oncology
Vol. 12 No. 5
14. Zhao ZR, Xi SY, Li W, et al. Prognostic impact of patternbased grading system by the new IASLC/ATS/ERS classification in Asian patients with stage I lung adenocarcinoma. Lung Cancer. 2015;90:604–609. 15. Zhao ZR, To KF, Mok TS, Ng CS. Is there significance in identification of non-predominant micropapillary or solid components in early-stage lung adenocarcinoma? Interact Cardiovasc Thorac Surg. 2017;24:121–125.
Comparison of Segmentectomy and Lobectomy in Stage IA Adenocarcinomas Ze-Rui Zhao, MD,a Dong-Rong Situ, MD,b Rainbow W. H. Lau, MbChB.,a Tony S. K. Mok, MD, FRCPC,c George G. Chen, PhD,a Malcolm J. Underwood, MD,a Calvin S. H. Ng, MDa,* a
Division of Cardiothoracic Surgery, Department of Surgery, The Chinese University of Hong Kong, Prince of Wales Hospital, Hong Kong Special Administrative Region, People’s Republic of China b State Key Laboratory of Oncology in Southern China, Collaborative Innovation Centre for Cancer Medicine, and Department of Thoracic Surgery, Sun Yat-Sen University Cancer Centre, Guangzhou, People’s Republic of China c Department of Clinical Oncology, The Chinese University of Hong Kong, Prince of Wales Hospital, Hong Kong Special Administrative Region, People’s Republic of China Received 16 November 2016; revised 18 December 2016; accepted 9 January 2017 Available online - 19 January 2017
ABSTRACT Introduction: Recent studies have suggested that segmentectomy may be an acceptable alternative treatment to lobectomy for surgical management of smaller lung adenocarcinomas. The objective of this study was to compare survival after lobectomy and segmentectomy among patients with pathological stage IA adenocarcinoma categorized as stage T1b (>0 to 20 mm) according to the new eighth edition of the TNM system. Methods: In total, 7989 patients were identified from the Surveillance, Epidemiology, and End Results registry. Propensity scores generated from logistic regression on preoperative characteristics were used to balance the selection bias of undergoing segmentectomy. Overall and lung cancer–specific survival rates of patients undergoing segmentectomy and lobectomy were compared in propensity score–matched groups. Results: Overall, 564 patients (7.1%) underwent segmentectomy. Lobectomy led to better overall and lung cancer– specific survival than segmentectomy for the entire cohort (log-rank p < 0.01). After 1:2 propensity score matching, segmentectomy (n ¼ 552) was no longer associated with significantly worse overall survival (5-year survival ¼ 74.45% versus 76.67%, hazard ratio ¼ 1.09, 95% confidence interval: 0.90–1.33) or lung cancer–specific survival (5-year survival ¼ 83.89% versus 86.11%, hazard ratio ¼ 1.12, 95% confidence interval: 0.86–1.46) compared with lobectomy (n ¼ 1085) after adjustment for age, sex, lymph node quantity, and histological subtype. Similar negative findings were identified when patients were stratified according to sex, age, histological subtype, and number of evaluated lymph nodes.
Journal of Thoracic Oncology
Vol. 12 No. 5: 890-896
Conclusions: Patients who underwent segmentectomy may have survival outcomes no different than those of some patients who received lobectomy for pathological stage IA adenocarcinomas at least 10 but no larger than 20 mm in size. These results should be further confirmed through prospective randomized trials. 2017 International Association for the Study of Lung Cancer. Published by Elsevier Inc. All rights reserved. Keywords: Adenocarcinoma; Segmentectomy; Lobectomy; Eighth TNM system
Lobectomy is the standard treatment for operable NSCLC, and thus far, sublobar resection has been considered a compromise option for high-risk patients (e.g., those with advanced age or impaired pulmonary function). Growing evidence, mainly from retrospective studies, indicates that sublobar resection, especially for segmentectomy (which is considered superior to wedge resection as a more anatomical approach), may obtain *Corresponding author. Drs. Zhao and Situ equally contributed to this work. Disclosure: The authors declare no conflict of interest. Address for correspondence: Calvin S.H. Ng, MD, Division of Cardiothoracic Surgery, The Chinese University of Hong Kong, Prince of Wales Hospital, Shatin, N.T., Hong Kong SAR, China. E-mail: calvinng@surgery. cuhk.edu.hk ª 2017 International Association for the Study of Lung Cancer. Published by Elsevier Inc. All rights reserved. ISSN: 1556-0864 http://dx.doi.org/10.1016/j.jtho.2017.01.012
May 2017
outcomes similar to those of lobectomy in early-stage NSCLC that measure 2 cm or less.1 Data from the latest NSCLC staging project found that the prognosis differs for tumors sized from 1 to 5 cm, which subsequently led to the reclassification of T descriptors in the revised system.2 Of particular interest, tumors that were previously defined as T1a (2 cm) will be subclassified as T1a (1 cm) and T1b (>1 to 2 cm). In this study, we used the population-based Surveillance, Epidemiology, and End Results (SEER) registry to test the hypothesis that the survival of patients with new pathological T1bN0M0 lung adenocarcinomas who undergo segmentectomy will not be inferior to that of patients undergoing lobectomy.
Methods Study Population The population was selected from the SEER 18 registry and limited to patients in whom stage IA lung adenocarcinomas that measured at least 10 but no more than 20 mm were histologically diagnosed and treated with segmentectomy or lobectomy from 2004 to 2012. Patients’ sociodemographic information and tumor characteristics were collected. Tumor histological subtype was coded according to the 2004 WHO classification.3 Survival was defined as the interval from diagnosis until death or last follow-up. Patients who survived past 31 December 2013 were classified as censored cases. When lung cancer–specific survival (LCSS) was analyzed, causes of death other than lung cancer were treated as censored observations. The completeness of follow-up was measured by a modification of Clark’s method.4 The local institutional review board classified this study as exempt.
Statistical Analysis Overall survival (OS) and LCSS were the primary and secondary outcomes of interest. The Kaplan-Meier method was used to compare unadjusted survival of patients undergoing segmentectomy or lobectomy with a log-rank test. Propensity score matching was used to balance the potential probability of being assigned to a treatment group. A logistic regression model that included age, sex, ethnicity, marital status, tumor laterality, location, and size was used to calculate the propensity scores, and 1: 2 pair matching without replacement was implemented by greedy nearest neighbour matching.5 The covariable balance effect was assessed by a standardized difference with a threshold of 0.10. Multivariable Cox proportional hazards models that included variables with a p value less than 0.2 on univariable analysis were used to decide variables that remained in the final model. A Cox proportional hazards
Segmentectomy in Stage IA Adenocarcinomas
891
model adjusted for age, sex, histological subtype, and number of lymph nodes evaluated was applied to compare outcomes between groups. We also performed secondary subgroup analysis limited to patients with different sex, age (<70 versus 70 years), histological subtype (nonmucinous bronchioloalveolar carcinoma, mucinous adenocarcinoma, and remaining adenocarcinoma subtypes [major adenocarcinoma]), and number of lymph nodes examined (0–10 versus >10).6 The study aim was to test whether segmentectomy was not inferior to lobectomy in treating certain stage IA lung adenocarcinomas. Noninferiority was established at the level of significance if the onesided upper 95% confidence bound was below the prespecified margin of 1.25.7 Statistical power was calculated accordingly.8 Analyses were conducted by using PASW Statistics 18 statistical software (formerly SPSS Statistics [SPSS, Hong Kong Special Administrative Region, People’s Republic of China]).
Results A flow diagram is presented in Figure 1. In total, 7989 patients were identified, of whom 564 (7.1%) underwent segmentectomy. The baseline characteristics are reported in Table 1. Segmentectomy was performed more often in cases involving seniors, left-side tumors, lower lobes, and smaller lesions. After propensity score matching, 552 segmentectomies and 1085 lobectomies were preserved. Preoperative characteristics were well balanced (Fig. 2). However, those in the segmentectomy arm were more likely to have lower lobe cancer, nonmucinous bronchioloalveolar carcinoma (p [ 0.02), and fewer lymph nodes resected (p < 0.01) regardless of matching. A higher incidence of more than 10 lymph nodes resected was observed in the lobectomy group (259 of 1085 [23.9%] versus 56 of 552 [10.1%]). The median follow-up interval was 52.21 months (0– 119 months) before matching. Patients who underwent segmentectomy had worse unadjusted OS and LCSS (both p < 0.01 [Supplementary Fig. 1]). More than 10 lymph nodes resected was associated with better OS (p ¼ 0.03) and LCSS (p < 0.01) (Supplementary Fig. 2). After propensity score matching, the median follow-up intervals were 49.24 months (0–119 months) and 52.06 months (0–119 months) for segmentectomy and lobectomy, respectively. The follow-up modified Clark’s C statistic values were 91.52% and 81.46% for OS and LCSS, respectively. The unadjusted Cox proportional hazard model showed that segmentectomy provided outcomes no different from those of lobectomy in terms of OS (5-year survival ¼ 68.95% versus 70.70%, hazard ratio [HR] ¼ 1.13, 95% confidence interval [CI]: 0.94–1.36) and LCSS
892 Zhao et al
Journal of Thoracic Oncology
Vol. 12 No. 5
Figure 1. Flow diagram of the study.
(5-year survival ¼ 81.30% versus 83.63%, HR ¼ 1.21, 95% CI: 0.93–1.56) (see Supplementary Fig. 1). Age, sex, lymph node quantity, and histological subtype (not for LCSS) were identified as potential predictors for survival function under multivariable analyses (Supplementary
Table 1). The regression model adjusting for these factors showed that segmentectomy was consistently associated with outcomes that were no different in terms of OS (5-year survival ¼ 74.45% versus 76.67%, HR ¼ 1.09, 95% CI: 0.90–1.33) and LCSS (5-year
Table 1. Baseline Characteristics of Patients with Pathological Stage IA Lung Adenocarcinoma (>10 and 20 mm) Treated with Segmentectomy and Lobectomy in the Surveillance, Epidemiology, and End Results Registry Before Propensity Score Matching Characteristic Mean Age ± SD, y Female sex, n (%) Race/Ethnicity, n (%) White Black Asian or Pacific Islander Others Married, n (%) Right laterality, n (%) Tumor location, n (%) Upper lobe Middle lobe Lower lobe Unknown Median tumor size ± SD, mm Histologic subtype, n (%)a BAC, nonmucinous Mucinous adenocarcinoma Major adenocarcinomab Mean lymph nodes ± SD, n a
After Propensity Score Matching
Segmentectomy (n ¼ 564)
Lobectomy (n ¼ 7425)
SD
Segmentectomy (n ¼ 552)
Lobectomy (n ¼ 1085)
SD
69.84 ± 9.17 362 (64.2)
67.31 ± 9.61 4443 (59.8)
0.83 0.09
69.84 ± 9.10 350 (63.4)
69.71 ± 9.08 697 (64.2)
0.04 –0.02
498 (88.3) 46 (8.2) 19 (3.4) 1 (0.2) 309 (54.8) 298 (52.8)
6336 (85.3) 528 (7.1) 519 (7.0) 42 (0.6) 4346 (58.5) 4591 (61.8)
0.09 0.04 –0.16 –0.06 –0.07 –0.18
488 (88.4) 44 (8.0) 19 (3.4) 1 (0.2) 306 (55.4) 296 (53.6)
977 (90.0) 66 (6.1) 40 (3.7) 2 (0.2) 606 (55.9) 604 (55.7)
–0.05 0.07 –0.02 0 –0.01 –0.04
322 (57.1) 16 (2.8) 222 (39.4) 4 (0.8) 15.63 ± 2.93
4629 (62.3) 485 (6.5) 2227 (30.0) 84 (1.2) 16.16 ± 2.87
–0.11 –0.18 0.20 –0.04 –0.31
319 (57.8) 16 (2.9) 214 (38.8) 3 (0.6) 15.69 ± 2.90
634 (58.4) 70 (6.5) 361 (33.3) 20 (1.8) 15.79 ± 2.83
–0.01 –0.17 0.11 –0.11 –0.06
31 (5.5) 27 (4.8) 506 (89.7) 4.19 ± 5.80
197 (2.7) 321 (4.3) 6907 (93.0) 7.66 ± 6.93
0.14 0.02 –0.12 –1.38
31 (5.6) 27 (4.9) 494 (89.5) 4.19 ± 5.76
33 (3.0) 39 (3.6) 1013 (93.4) 7.43 ± 6.88
0.13 0.06 –0.14 –1.29
Variables were not included in the propensity score model, as it is not a preoperative event. Adenocarcinomas other than nonmucinous BAC or mucinous adenocarcinoma. BAC, bronchioloalveolar carcinoma. b
May 2017
Segmentectomy in Stage IA Adenocarcinomas
893
Figure 2. (A) Standardized differences of variables between patients who received lobectomy and segmentectomy. Red crosses symbolize differences before propensity matching and blue circles symbolize differences after propensity matching. Propensity score matching effectively reduced heterogeneity among variables between the two groups in comparison. (B) Mirror histogram of propensity scores for patients who received segmentectomy (below the horizontal line at zero) and lobectomy (above the horizontal line at zero). Matched patients are a subset of the original patient population and their volumes are highlighted in color. Matched groups have similar propensity score distributions.
survival ¼ 83.89% versus 86.11%, HR ¼ 1.12, 95% CI ¼ 0.86–1.46) from those of lobectomy (Fig. 3). On secondary analysis, we did not recognize different OS or LCSS rates when patients were divided according to the number of lymph nodes or pathological subtypes. Demographic characteristics in the subsets of major adenocarcinomas and those with 10 or fewer resected lymph nodes are shown in Supplementary Table 2. Consistent negative results were found when patients were stratified by age, sex, or size (Fig. 4). However, all statistical powers were less than 0.8 for these analyses.
Discussion In the current study of pathological stage IA lung adenocarcinomas larger than 10 and but no larger than 20 mm, we evaluated the survival outcomes of segmentectomy versus those of lobectomy in a noninferiority setting. Our findings showed that segmentectomy has survival no different than that of lobectomy for this specified population. Tailored therapy for patients with early-stage NSCLCs is still an issue of debate.9 A recent report has demonstrated that sublobar resection may achieve outcomes similar to those of lobectomy for NSCLCs
Figure 3. Survival curves adjusted for age, sex, histological subtype, and lymph node quantity compare treatment with segmentectomy versus with lobectomy in patients with pathological stage IA adenocarcinoma (>10 and 20 mm) after propensity score matching for overall (A) and lung cancer–specific (B) survival. Shaded area represents the confidence limits for each group.
894 Zhao et al
Journal of Thoracic Oncology
Vol. 12 No. 5
Figure 4. Multivariable Cox proportional hazards model adjusted for age, sex, histological type, and lymph node quantity for overall (A) and lung cancer–specific (B) survival by subgroup analysis in the matched population. Dashed line indicates the prespecified noninferiority margin of 1.25. Blue shaded band represents the 95% confidence interval (CI) of the hazard ratio (HR) for the overall matched population. BAC, bronchioloalveolar carcinoma; Seg, segmentectomy; Lob, lobectomy.
1 cm or smaller.10 However, limited resection was not equivalent to lobectomy in older patients with lesions 2 cm or smaller.11 Besides tumor size, histological subtype may be another important factor that affects the outcome of sublobar resection. Whether the real therapeutic effect of segmentectomy was underestimated by including patients who underwent wedge resection12 or were classified as having squamous carcinoma11 or overestimated by including lesions smaller than 1 cm remains unclear.10 Thus, we restricted the study population to those who had a
pathological stage IA lung adenocarcinoma at least 1 but no larger than 2 cm in size and underwent segmentectomy and lobectomy only. More than 10 examined lymph nodes was a strong predictor of survival for node-negative early-stage NSCLC6; hence, we integrated this parameter as well as other prognostic confounders into the Cox models for better adjustment. Although lobectomy was associated with higher chance of more than 10 resected lymph nodes, we did not observe a survival difference for segmentectomy versus lobectomy in subgroups
May 2017
divided by number of lymph nodes. However, an intentional segmentectomy with a minimum of 10 lymph nodes resected should be explored in a prospective trial. Histological subtype of lung adenocarcinoma has recently been proved to be a crucial factor that affects the survival of patients who undergo sublobar resection. One pilot study found that lobectomy did not offer an OS advantage over sublobar resection for invasive adenocarcinomas 2 cm or smaller according to the 2011 adenocarcinoma classification,13 although the number of limited resections was quite small (n ¼ 26).1 However, another population-based study found that limited resection was inferior to lobectomy in invasive lung adenocarcinomas 2 cm or smaller, whereas segmentectomy seemed to be equivalent to lobectomy on a subgroup analysis.11 On the basis of the results of our study, we postulate that the discrepancy could be attributed to the high proportion of patients with wedge resections (nearly 80%) who underwent sublobar resection in the latter study, which would potentially override the therapeutic benefit of segmentectomy.11 This study had certain limitations. Preoperative cardiopulmonary function is not available in the SEER registry, which makes it difficult to balance the higher likelihood of choosing patients with poorer cardiopulmonary status to receive sublobar resection. The lack of exact tumor location made it hard to estimate whether the lesion was amenable to segmentectomy. In addition, we were unable to evaluate whether segmentectomy would lead to better preservation of lung function, a greater chance of receiving adjuvant chemotherapy, a higher rate of insufficient margins, or relapse. Additionally, the new classification of adenocarcinoma has not been updated in the SEER registry13; therefore, the study population in this study was still heterogeneous. Patients with even small proportions of micropapillary or solid components have a worse prognosis, which may affect the decision to perform a sublobar procedure.14,15 Furthermore, the lack of clinical staging forced us to select patients solely on the basis of pathological stage, which would inevitably exclude some clinical stage IA cases that were found to be pathological stage IB (e.g., pleural invasion). As excluding such patients would falsely improve survival, the ability to apply the current findings to all clinical stage IA patients could be limited.
Acknowledgments This work was supported by Research Grants Council General Research Fund grant 14117715, Chinese
Segmentectomy in Stage IA Adenocarcinomas
895
University of Hong Kong direct grant 2015.1.087, and Foundation for Science and Technology Research Project of Guangdong grant 2014B020212014. The authors would like to thank the staff members of the National Cancer Institute and their colleagues across the United States who have been involved with the Surveillance, Epidemiology, and End Results program.
Supplementary Data Note: To access the supplementary material accompanying this article, visit the online version of the Journal of Thoracic Oncology at www.jto.org and at http://dx.doi. org/10.1016/j.jtho.2017.01.012.
References 1. Dembitzer FR, Flores RM, Parides MK, Beasley MB. Impact of histologic subtyping on outcome in lobar vs sublobar resections for lung cancer: a pilot study. Chest. 2014;146:175–181. 2. Rami-Porta R, Bolejack V, Crowley J, et al. The IASLC lung cancer staging project: proposals for the revisions of the t descriptors in the forthcoming eighth edition of the TNM classification for lung cancer. J Thorac Oncol. 2015;10:990–1003. 3. Travis WE, Brambilla E, Müller-Hermelink HK, Harris CC, eds. Pathology and Genetics; Tumours of the Lung, Pleura, Thymus and Heart. Lyon, France: IARC Press; 2004. 4. Wu Y, Takkenberg JJ, Grunkemeier GL. Measuring follow-up completeness. Ann Thorac Surg. 2008;85: 1155–1157. 5. Austin PC. The use of propensity score methods with survival or time-to-event outcomes: reporting measures of effect similar to those used in randomized experiments. Stat Med. 2014;33:1242–1258. 6. Ludwig MS, Goodman M, Miller DL, Johnstone PA. Postoperative survival and the number of lymph nodes sampled during resection of node-negative non-small cell lung cancer. Chest. 2005;128:1545–1550. 7. Rothmann M, Li N, Chen G, Chi GY, Temple R, Tsou HH. Design and analysis of non-inferiority mortality trials in oncology. Stat Med. 2003;22:239–264. 8. Chow SC, Wang H, Shao J. Sample Size Calculations in Clinical Research. CRC Press; 2007. 9. Ng CSH, Zhao ZR, Lau RWH. Tailored therapy for stage I non–small-cell lung cancer. J Clin Oncol. 2017;35: 268–270. 10. Kates M, Swanson S, Wisnivesky JP. Survival following lobectomy and limited resection for the treatment of stage I non-small cell lung cancer 1 cm in size: a review of SEER data. Chest. 2011;139:491–496. 11. Veluswamy RR, Ezer N, Mhango G, et al. Limited resection versus lobectomy for older patients with early-stage lung cancer: impact of histology. J Clin Oncol. 2015;33:3447–3453. 12. Smith CB, Swanson SJ, Mhango G, Wisnivesky JP. Survival after segmentectomy and wedge resection in stage
896 Zhao et al I non-small-cell lung cancer. J Thorac Oncol. 2013;8: 73–78. 13. Travis WD, Brambilla E, Noguchi M, et al. International Association for the Study of Lung Cancer/ American Thoracic Society/European Respiratory Society international multidisciplinary classification of lung adenocarcinoma. J Thorac Oncol. 2011;6: 244–285.
Journal of Thoracic Oncology
Vol. 12 No. 5
14. Zhao ZR, Xi SY, Li W, et al. Prognostic impact of patternbased grading system by the new IASLC/ATS/ERS classification in Asian patients with stage I lung adenocarcinoma. Lung Cancer. 2015;90:604–609. 15. Zhao ZR, To KF, Mok TS, Ng CS. Is there significance in identification of non-predominant micropapillary or solid components in early-stage lung adenocarcinoma? Interact Cardiovasc Thorac Surg. 2017;24:121–125.