Recurrence and Survival Outcomes After Anatomic Segmentectomy Versus Lobectomy for Clinical Stage I Non–Small-Cell Lung Cancer: A Propensity-Matched Analysis

Abstract

INTRODUCTION

Although surgical resection of clinical stage I non–small-cell lung cancer (NSCLC) remains the standard of care, controversy exists with regard to the extent of parenchymal resection required for local control of the tumor and disease-free survival.^1–5 Intentional use of anatomic segmentectomy for the treatment of small peripheral lung cancers was first performed by Jensik et al.⁶ Lobectomy has been considered the standard of surgical care for early-stage NSCLC identified in the physiologically fit patient⁷; however, anatomic segmentectomy performed by open surgical techniques or through video-assisted thoracic surgical approaches has gained enthusiasm by many surgical groups over the years.^8–10 Additionally, nonsurgical approaches using hypofractionated, stereotactic radiotherapy and image-guided probing of the tumor for delivery of radiofrequency or microwave energy to focally ablate small peripheral lung cancers are also challenging the concept of universal selection of anatomic lobectomy for clinical stage I NSCLC.^11,12

We focused our present analysis on clinical stage I NSCLC, because describing the results of various surgical resection approaches related to pathologic stage I disease necessarily biases outcomes toward surgical interventions with greater objective intraoperative staging of the pulmonary hilum and mediastinum. This is an important point of distinction, in that it is appreciated that pathologic upstaging from the pretreatment clinical stage may be seen in up to 30% of patients, even with modern computed tomography (CT) and positron emission tomography staging techniques being used.^13,14

To accomplish this analysis, we explored our clinical outcomes with anatomic segmentectomy and lobectomy for clinical stage I NSCLC using a propensity-matched risk model and multivariable parameter assessment,^15,16 because results of randomized studies directly comparing these anatomic resection approaches are lacking.

PATIENTS AND METHODS

RESULTS

Recurrence and Survival

At a mean follow-up of 5.4 years, comparing segmentectomy with lobectomy, no differences were noted in either locoregional (5.5% v 5.1%, respectively; P = 1.00) or distant (14.8% v 11.6%, respectively; P = .29) recurrence rates. Kaplan-Meier estimates of freedom from recurrence (70% v 71%, respectively; P = .467) and overall survival (54% v 60%, respectively; P = .258) at 5 years similarly demonstrated no statistically significant difference between groups (Table 3, Fig 1). Importantly, no significant differences were seen in time to recurrence when comparing patients with T1a tumors (≤ 2 cm; P = .585), T1b tumors (2 to 3 cm; P = .395), or T2a tumors (3 to 5 cm; P = .432).

Open in a separate window

Fig 1.

Kaplan-Meier survival estimates for (A) time to recurrence and (B) overall survival between propensity score–matched patients undergoing segmentectomy or lobectomy.

DISCUSSION

How should we approach the peripheral, small, clinically node-negative NSCLC today? Pneumonectomy was the standard of care for resectable lung cancer for nearly two decades.^22–26 However, by the late 1950s, lobectomy had surpassed it as the preferred mode of resection for peripheral lung cancers as a result of observed increases in long-term survival with lobectomy (19% v 12% for pneumonectomy) at 5 years, as well as a reduction in operative morbidity and mortality.²⁷ These generally poor results by today's standards can be primarily attributed to the limitations of clinical staging during that period and to a near totally subjective evaluation of the associated medical conditions (eg, impaired functional lung reserve, cardiac disease) that affect early postsurgical and long-term survival.

The findings of large-population CT screening programs for early lung cancer identification,^28–40 coupled with these clinical staging advances, have led to an increased number of clinically suspicious lung nodules suggestive of early-stage NSCLC in modern practice. The following question has thus arisen: Is lobectomy needed for the small peripheral symptomatic lung nodule serendipitously identified by CT scanning?^41,42

Over the years, our group has been influenced by the insights of breast surgical oncologists considering lesser resections for small breast cancers⁴³ and has explored the potential merits of parenchymal preservation for the patient with lung cancer in the setting of clinical stage I NSCLC.^40–42 Despite a growing number of reports suggesting the utility of a selective application of sublobar resection, and specifically anatomic segmentectomy, for peripheral stage I NSCLC, resection of even the smallest peripheral lung cancer is still considered by many surgeons as a compromised therapy that should be only considered for the physiologically impaired patient who is not a candidate for anatomic lobectomy.^1,4,41 This perception has been primarily established by the findings of the Lung Cancer Study Group, which represents the only prospective, multicenter, randomized trial comparing lobectomy with sublobar resection performed to date. In this study, a three-fold increase in the rate of locoregional recurrence was noted in the sublobar resection arm compared with the lobectomy arm (17.2% v 6.4%, respectively), which included patients undergoing either wedge resection or segmentectomy. Interestingly, we and others have previously found that anatomic segmentectomy may have a reduced risk of locoregional recurrence compared with nonanatomic wedge resection and that anatomic segmentectomy more closely approximates the oncologic results after lobectomy.⁴² The results of the current propensity-matched study suggest that anatomic segmentectomy can achieve perioperative (morbidity and mortality) and oncologic outcomes (recurrence and survival) that are not statistically significantly different compared with those achieved by lobectomy for clinical stage I disease, and thus anatomic segmentectomy should be considered as a valid alternative to lobectomy in properly selected patients.

In this study, we also identified that a similar hilar and mediastinal lymph node station examination was accomplished by segmentectomy and lobectomy. Although a greater number of nodes removed during mediastinal evaluation may lead to greater discovery of surprise node-positive disease, this occurred less than 4% of the time in the recently reported results of the American College of Surgeons Oncology Group Z0030 investigation of mediastinal node dissection versus systematic sampling after resection of stage I NSCLC. Additionally, no improvement in survival could be demonstrated in this trial with the standard use of mediastinal and hilar node dissection over systemic sampling.⁴⁴ In the current study, although lobectomy yielded a greater absolute number of lymph nodes (median, 12 nodes with lobectomy v six nodes with segmentectomy; P < .001), when comparing segmentectomy with lobectomy, there was no significant difference in the number of lymph node stations assessed (median of three lymph node stations in both groups; P = .45) or degree of clinical to pathologic upstaging (36.5% v 29.5%, respectively; P = .073), and no corresponding statistically significant difference in recurrence or overall survival (Table 3, Fig 1).

Approaching the problem of small peripheral lung cancer from a different direction, we see the emergence of enthusiasm with nonsurgical image-guided ablative and focused radiotherapeutic approaches to the small peripheral lung nodule suspicious for lung cancer.^11,12,45 The commentary of advocates of nonsurgical approaches to peripheral lung cancer relate to the avoidance of any morbidity related to surgery with the theoretical possibility of equivalent long-term survival. But as the saying goes, “pay me now or pay me later.” It seems that the risk of local lung injury leading to intermediate and late-term unintended negative consequences, such as progressive radiation pneumonitis, fibrosis, and loss of pulmonary function,^46–48 and the incremental loss of opportunity to identify a significant minority of patients with more advanced cancers who may benefit from adjuvant therapy are concerns.

Today more than ever, we should not be satisfied with the concept of the simple local extirpation of the point of disease presentation manifested in the apparently locally confined NSCLC. Again, an emphasis on improving the overall outcome of clinical stage I NSCLC, which includes patients with occult/surprise lymph node positivity (identified in up to 30% of patients), mandates the pathologic assessment of the disease through surgical resection. Accordingly, tissue still remains the issue in the surgical management of early-stage NSCLC, particularly in this age of individualized therapy for the patient with lung cancer.^49–53 Surgical resection of peripheral lung cancers represents the standard of care. Similar to the arguments with breast cancer surgery, the advantages of surgery are pathologic assessment of surgical margins, the establishment of pathologic regional nodal status, and, in this era of increasing enthusiasm for adjuvant systemic therapy, provision of tissue for pharmacogenomic assessment.

Limitations of this study include the potential introduction of surgical and selection bias associated with the retrospective assessment of prospectively collected data. We have attempted to directly address this fundamental limitation with the use of propensity matching and multivariable analysis. Another limitation of this study is that the presented data are derived from a single institution with expertise in minimally invasive and open lobectomy and segmentectomy techniques. The generalization of the observed outcomes to broad clinical practice must therefore be cautiously scrutinized. Multicenter, prospective, randomized studies currently under way (Cancer and Leukemia Group B 140503; Japan Clinical Oncology Group 0802/West Japan Oncology Group 4607L) will thus be necessary to confirm the results of this study.

In conclusion, the information obtained from this propensity-matched analysis supports this concept of anatomic segmental resection for the small peripheral lung cancer anatomically confined to segmental boundaries. However, confirmation of clear, generous margins of resection and the assurance of accurate intraoperative pathologic nodal staging of the lesion are important considerations that should lead us to favor lobectomy over segmentectomy when an issue. For the small peripheral lung cancer, however, anatomic segmentectomy appears to offer comparable local control and the opportunity for prolonged disease-free and overall survival that is not statistically different when compared with lobectomy.

Footnotes

REFERENCES