The importance of lymph node dissection accompanying wedge resection for clinical stage IA lung cancer†

Abstract

INTRODUCTION

For patient undergoing lobectomy for non-small cell lung cancer (NSCLC), studies have suggested a survival benefit associated with an increased number of lymph nodes (LNs) removed [1, 2]. Due to this relationship, several societies including the International Association for the Study of Lung Cancer (IASLC), the American College of Surgeons (ACOSOG) and the National Comprehensive Cancer Network (NCCN), have proposed guidelines for good surgical resection of early stage NSCLC that include recommendations regarding the extent of LN resection [3–5]. However, most of these recommendations are based primarily upon patients undergoing lobectomy, rather than upon those undergoing sub-lobar resection. As thoracic surgeons increasingly remove smaller tumours and increasingly operate on frailer patients with poor pulmonary reserve, strategies for limited resection have become more attractive, even in the absence of randomized data. It is therefore important to establish whether lymphadenectomy is an important component of sub-lobar resection.

The question of whether lymphadenectomy affects survival after wedge resection may be particularly relevant when comparing wedge resection to other focal treatment strategies such as stereotactic body radiation therapy/stereotactic ablative radiotherapy (SBRT/SABR) or radiofrequency ablation (RFA). SBRT in particular has been progressively used to treat patients with clinical stage I NSCLC, also without the benefit of adequately powered randomized clinical trials [6, 7]. One distinct disadvantage of radiation and ablation techniques compared to surgical resection, is that patient treatment is solely based upon clinical stage of the tumour without pathologic confirmation. It is well known that clinical staging of early NSCLC is often inaccurate and that occult nodal disease may be found in up to 11–14% of patients [8, 9]. Without surgical resection and LN sampling, these patients would be under-staged, potentially undertreated and at higher risk for recurrence. We undertook this study in order to determine rates of wedge resection and LN assessment in our own population of clinical stage I NSCLC surgical patients and to determine whether LN resection and assessment affected rates of recurrence and survival.

PATIENTS AND METHODS

Study design and patients

We conducted a retrospective review of a prospective lung cancer database. The study was approved by the Weill Cornell Medicine, New York-Presbyterian Hospital Institutional Review Board and patient consent waived. The study included patients who underwent wedge resection with curative intent for clinical stage IA NSCLC between February 2000 and May 2014. Patients were excluded from analysis if they had prior lung resection for NSCLC higher than stage I, unless their disease free interval exceeded 2 years and the resected tumour was thought to be a new primary cancer. Patients with multifocal NSCLC, neuroendocrine histology and non- or minimally invasive cancers were also excluded. Demographic, clinical, perioperative and pathological characteristics of selected patients were retrieved from the database. Parenchymal resection margin status was assessed by the pathologist prior to specimen fixation and was measured from the edge of the tumour to the nearest staple line. Complications were graded on a Clavien Dindo scale, with grade III complications classified as those requiring surgical, endoscopic or radiological intervention and Grade IV complications those that are life-threatening or requiring ICU management [10].

Wedge resection operability criteria and surgical technique(s)

In our institution, we performed wedge resection for selected patients with T1N0 peripheral tumours (located in the outer third of lung parenchyma) or for patients who wouldn’t otherwise tolerate the morbidity associated with anatomic lung resection. Type of sub-lobar resection performed (anatomic or non-anatomic) is based on the patient’s general performance status, clinical tumour characteristics and the preference of the surgeon. Small, peripheral tumours are on occasion treated with wedge resection, particularly for patients with significant comorbidities or severe emphysema. Tumours that are located within a resectable anatomical segment, particularly if greater than 1 cm, are usually treated with segmentectomy.

Follow-up, recurrence and survival

Follow-up was accomplished according to established institutional standards. In brief, patients were seen in our follow-up clinic every 3 months during the first postoperative year, every 6 months for 2 additional years, and yearly thereafter. A low dose non-contrast CT scan of the chest and upper abdomen was obtained every 6 months for 2 years and yearly thereafter. Patients from distant geographic locations were followed by direct contact or by contacting their local physician.

Local recurrence was defined as recurrent tumour with similar histology within the same lobe. Regional recurrence was defined as recurrent tumour in the ipsilateral hilar and/or mediastinal LNs. Distant recurrence was defined as evidence of tumour in another lobe within 2 years from surgery, pleural space or elsewhere outside the hemi-thorax.

Overall survival (OS) was defined as the time from surgery until death from any cause. Disease free survival (DFS) was defined as the time from surgery until recurrence or death from any cause.

Statistical analysis

Continuous variables (expressed as median, interquartile range) were compared using Mann–Whitney U test. Frequencies and percentages were compared for categorical variables between the two groups using the two-tailed Pearson’s chi-square test or Fischer’s exact test. Cox proportional hazards regression analysis was done to determine the independent effects of clinical predictors for DFS. Variables included in the model were; age, gender, smoking status, comorbidity index, FEV1 (in %), DLCO (in %), clinical tumour size, clinical T stage, tumour SUVmax, histology, tumour site, LN assessment and surgery year. Univariate predictors with probability values less than 0.20 were included in a multivariate model. Survival probabilities were estimated using the Kaplan–Meier method. The log-rank test was used to determine significance of survival distributions among groups. Propensity score analysis controlling for age, gender, Charlson comorbidity index, patient tolerability of lobectomy, surgery year, tumour size and surgical approach was done (51 patients in each group, caliper 0.2).

Data analysis was performed using IBM SPSS software (IBM Corp. Released 2013. IBM SPSS Statistics for Windows, Version 22.0. Armonk, NY, USA), and PS-matching package version 3.04.

RESULTS

Clinical characteristics

Among 1,358 patients undergoing resection for clinical stage IA lung cancer, we performed 196 wedge resections (14%). Of these, 138 patients (70%) had LNs removed (median = 4 nodes), while the remaining 58 patients (30%) had no lymph nodes (NLNs) resected. The LN and NLN groups were well matched in terms of demographic variables and clinical tumour characteristics (Table 1). Median age was 73 years in each group (P = 0.92), with 61% of the cohort female. Median Charlson Comorbidity Index was 1 (IQ range 0–2) for each group. Median FEV11% was 74% (IQR 54–99%) for NLN patients and 78% (IQR 59–95%) for LN patients (P = 0.86). In the NLN group, 16 patients (28%) had preoperative FEV1% or DLCO% ≤ 50% compared to 31 patients (22%) in the LN group (P = 0.16). Tumours were predominantly located in the upper lobe in both groups and were predominantly adenocarcinoma in both the NLN and LN assessed groups (79 vs. 80%, P = 0.97). Median clinical tumour size was 1.3 cm (IQR 0.9–1.77) in the NLN group, compared to 1.5 cm (IQR 1.2–2) in the LN assessed group. Median PET SUVmax was 2.6 (IQR 1.45–4.9) vs 2.1 (IQR 1–4.5, P = 0.51). The only difference between the groups was in the approach to surgery, with 38% of the NLN group approached by thoracotomy compared to 19% of the LN assessed patient group (P = 0.004). The choice of surgical procedure was related to the time period performed. Prior to 2010, 30 of 73 cases (41%) were performed via thoracotomy, whereas from 2010 on only 18 of 123 cases (15%) were performed via thoracotomy.

Perioperative outcomes

There was no difference in perioperative outcomes between the overall groups of patients with NLN and those with LN assessed in terms of OR time (94 vs 94.5 min, P = 0.36), estimated blood loss (50 vs 50 ml, P = 0.16), chest tube duration (2 vs 2 days, P = 0.73), or hospital or length of stay (4 vs 4 days, P = 0.98). In order to ensure that the increased rate of thoracotomy in the NLN group did not weight the data, we further compared all of these factors in groups of patients with and without LNs removed depending upon surgical approach (Table 2). Again, no differences were apparent in OR time, EBL, chest tube duration or LOS. The overall rate of postoperative complications was 22% in NLN patients versus 24.5% in LN assessed patients (P = 0.74), while the rates of Clavien Dindo III/IV complications were only 3% and 8%, respectively (P = 0.49). We also compared complications rates in both groups of patients depending upon whether resected by VATS or thoracotomy. The overall rate of total complications and of Clavien Dindo III/IV complications was higher in patients undergoing thoracotomy versus VATS. (Table 2).

Pathologiccharacteristics

Pathologic features were also well balanced between the groups of patients (Table 3). Median tumour size was 1.5 cm in each group, with no differences in histologic grade, pleural invasion or lymphovascular invasion between groups. The median parenchymal resection margin was 1 cm in each group (P = 0.42). Slightly higher rates of positive margins were found in NLN patients, 3.4 vs 0.7%, although not statistically significant (P = 0.21). The majority of tumours were pT1 in both the NLN group (83%) and the LN group (83%, P = 0.98). In the LN group, 132 patients (95.7%) were found to be pN0, while 6 patients (4.3%) were found to have occult nodal disease, either pN1 (n = 1) or pN2 (n = 5). This resulted in identifying more pathologic stage IB/II/III patients in the LN group compared to the NLN group (21 vs 17%), although this was not statistically significant (P = 0.55).

Recurrence and survival

For the cohort, median follow up of patients was 35 months (24–78 months). The majority of patients (Table 4) were alive without recurrence in both the NLN group (n = 46) and the LN assessed group (n = 119). Patients in the LN assessed group had higher probability of freedom from loco-regional recurrence compared to the NLN group (5 year: 92 vs 74%, P = 0.025). However, probability of freedom from distant recurrence was not different between the two groups (5 year: 92 vs 85%, P = 0.72). The median time to recurrence for the entire cohort was 18.8 months (95% CI: 11.57–26.05).

OS at 5 years favoured LN patients compared to NLN patients (83 vs 56%, P = 0.042, Fig. 1A). Similarly, there was a trend demonstrating improved 5-year DFS in LN patients compared to NLN patients (65 vs 40%, P = 0.058, Fig. 1B). Univariate and multivariate analyses were performed to determine predictors of recurrence or death in the entire cohort (Table 5). On univariate analysis, failing to assess LNs was associated with recurrence or death (HR 1.627, 95% CI 0.979–2.704), although this was not significant on MVA. Significant predictors of recurrence or death on MVA included male gender (HR 2.89, 95% CI 1.50–5.55) and tumour SUVmax as a continuous variable (HR 1.10, 95% CI 1.02–1.18).

Figure 1

(A) Overall survival between groups of patients with and without LN assessment. (B) Disease free survival between groups of patients with and without LN assessment. LN: lymph node.

Open in new tabDownload slide

Figure 2

Probability of freedom from loco-regional recurrence between propensity matched groups of patients with and without LN assessment. LN: lymph node.

Open in new tabDownload slide

Table 5

Open in new tab

Uni- and multivariable cox-regression analysis of disease free survival  (number of events 64/196)

Independents variables .	Univariate predictors .		Multivariable predictors* .
	HR  (95% CI) .	P-value .	HR  (95% CI) .	P-value .
Age
Age, in years	1.01  (0 .98–1.04)	0 .46
Gender
Female	Reference
Male	2.18  (1.32–3.59)	0 .002	2.89  (1.50–5.55)	0 .001
Smoking status
Never	Reference
Current/Former	0.77  (0 .40–1.48)	0 .43
Pulmonary functions test
FEV1  (in %)	0.99  (0 .98–1.00)	0 .053	0.99  (0 .99–1.01)	0 .75
DLCO  (in %)	0.99  (0 .99–1.01)	0 .61
Co morbidity index  (Charlson)
Co morbidity index	1.05  (0 .91–1.20)	0 .51
Site of the tumour
Upper/Middle lobe	Reference
Lower lobe	1.05  (0 .63–1.74)	0 .85
PET SUV max
PET SUV max	1.08  (1.01–1.15)	0 .015	1.10  (1.02–1.18)	0 .008
Clinic tumour size
Clinic tumour size, in cm	1.25  (0 .82–1.90)	0 .30
Clinical T
cT1a	Reference
cT1b	1.42  (0 .76–2.62)	0 .27
Histology type
Adenocarcinoma	Reference
Squamous/others	1.75  (1.02–3.00)	0 .043	1.01  (0 .51–2.00)	0 .97
LN assessment
Yes	Reference
No	1.63  (0 .98–2.70)	0 .060	1.23  (0 .62–2.41)	0 .55
Year of surgery
< 2010  (median)	Reference
≥ 2010	0.55  (0 .31–0 .99)	0 .045	0.89  (0 .41–1.91)	0 .76

Independents variables .	Univariate predictors .		Multivariable predictors* .
	HR  (95% CI) .	P-value .	HR  (95% CI) .	P-value .
Age
Age, in years	1.01  (0 .98–1.04)	0 .46
Gender
Female	Reference
Male	2.18  (1.32–3.59)	0 .002	2.89  (1.50–5.55)	0 .001
Smoking status
Never	Reference
Current/Former	0.77  (0 .40–1.48)	0 .43
Pulmonary functions test
FEV1  (in %)	0.99  (0 .98–1.00)	0 .053	0.99  (0 .99–1.01)	0 .75
DLCO  (in %)	0.99  (0 .99–1.01)	0 .61
Co morbidity index  (Charlson)
Co morbidity index	1.05  (0 .91–1.20)	0 .51
Site of the tumour
Upper/Middle lobe	Reference
Lower lobe	1.05  (0 .63–1.74)	0 .85
PET SUV max
PET SUV max	1.08  (1.01–1.15)	0 .015	1.10  (1.02–1.18)	0 .008
Clinic tumour size
Clinic tumour size, in cm	1.25  (0 .82–1.90)	0 .30
Clinical T
cT1a	Reference
cT1b	1.42  (0 .76–2.62)	0 .27
Histology type
Adenocarcinoma	Reference
Squamous/others	1.75  (1.02–3.00)	0 .043	1.01  (0 .51–2.00)	0 .97
LN assessment
Yes	Reference
No	1.63  (0 .98–2.70)	0 .060	1.23  (0 .62–2.41)	0 .55
Year of surgery
< 2010  (median)	Reference
≥ 2010	0.55  (0 .31–0 .99)	0 .045	0.89  (0 .41–1.91)	0 .76

CI: confidence interval; DLCO: diffusion capacity of lung for carbon monoxide; FEV: forced expiratory volume; LN: lymph node.

*

Univariate variables with P < 0 .1 were included in the multivariable model.

Table 5

Open in new tab

Uni- and multivariable cox-regression analysis of disease free survival  (number of events 64/196)

Independents variables .	Univariate predictors .		Multivariable predictors* .
	HR  (95% CI) .	P-value .	HR  (95% CI) .	P-value .
Age
Age, in years	1.01  (0 .98–1.04)	0 .46
Gender
Female	Reference
Male	2.18  (1.32–3.59)	0 .002	2.89  (1.50–5.55)	0 .001
Smoking status
Never	Reference
Current/Former	0.77  (0 .40–1.48)	0 .43
Pulmonary functions test
FEV1  (in %)	0.99  (0 .98–1.00)	0 .053	0.99  (0 .99–1.01)	0 .75
DLCO  (in %)	0.99  (0 .99–1.01)	0 .61
Co morbidity index  (Charlson)
Co morbidity index	1.05  (0 .91–1.20)	0 .51
Site of the tumour
Upper/Middle lobe	Reference
Lower lobe	1.05  (0 .63–1.74)	0 .85
PET SUV max
PET SUV max	1.08  (1.01–1.15)	0 .015	1.10  (1.02–1.18)	0 .008
Clinic tumour size
Clinic tumour size, in cm	1.25  (0 .82–1.90)	0 .30
Clinical T
cT1a	Reference
cT1b	1.42  (0 .76–2.62)	0 .27
Histology type
Adenocarcinoma	Reference
Squamous/others	1.75  (1.02–3.00)	0 .043	1.01  (0 .51–2.00)	0 .97
LN assessment
Yes	Reference
No	1.63  (0 .98–2.70)	0 .060	1.23  (0 .62–2.41)	0 .55
Year of surgery
< 2010  (median)	Reference
≥ 2010	0.55  (0 .31–0 .99)	0 .045	0.89  (0 .41–1.91)	0 .76

Independents variables .	Univariate predictors .		Multivariable predictors* .
	HR  (95% CI) .	P-value .	HR  (95% CI) .	P-value .
Age
Age, in years	1.01  (0 .98–1.04)	0 .46
Gender
Female	Reference
Male	2.18  (1.32–3.59)	0 .002	2.89  (1.50–5.55)	0 .001
Smoking status
Never	Reference
Current/Former	0.77  (0 .40–1.48)	0 .43
Pulmonary functions test
FEV1  (in %)	0.99  (0 .98–1.00)	0 .053	0.99  (0 .99–1.01)	0 .75
DLCO  (in %)	0.99  (0 .99–1.01)	0 .61
Co morbidity index  (Charlson)
Co morbidity index	1.05  (0 .91–1.20)	0 .51
Site of the tumour
Upper/Middle lobe	Reference
Lower lobe	1.05  (0 .63–1.74)	0 .85
PET SUV max
PET SUV max	1.08  (1.01–1.15)	0 .015	1.10  (1.02–1.18)	0 .008
Clinic tumour size
Clinic tumour size, in cm	1.25  (0 .82–1.90)	0 .30
Clinical T
cT1a	Reference
cT1b	1.42  (0 .76–2.62)	0 .27
Histology type
Adenocarcinoma	Reference
Squamous/others	1.75  (1.02–3.00)	0 .043	1.01  (0 .51–2.00)	0 .97
LN assessment
Yes	Reference
No	1.63  (0 .98–2.70)	0 .060	1.23  (0 .62–2.41)	0 .55
Year of surgery
< 2010  (median)	Reference
≥ 2010	0.55  (0 .31–0 .99)	0 .045	0.89  (0 .41–1.91)	0 .76

CI: confidence interval; DLCO: diffusion capacity of lung for carbon monoxide; FEV: forced expiratory volume; LN: lymph node.

*

Univariate variables with P < 0 .1 were included in the multivariable model.

Propensity score analysis controlling for age (P = 0.89), gender (P = 1), Charlson comorbidity index (P = 0.67), surgical approach (P = 0.83), tolerability of lobectomy (P = 0.67) and year of surgery (P = 0.84) was performed. There were no differences between the two groups in OS (P = 0.89) or in DFS (P = 0.39). Similar to the whole cohort, patients who underwent LN assessment had significantly higher probability of freedom from loco-regional recurrence (P = 0.024) (2).

DISCUSSION

A previous analysis of a large cancer database demonstrated that an increased number of LNs removed during definitive surgery for stage I NSCLC was associated with a statistically significant increase in survival [1]. In that study using the SEER database, Ludwig et al. [1] found that the survival benefit peaks at approximately 13–16 LNs resected. A more recent analysis of the National Cancer Data Base (NCDB) by Gulack et al. [2] demonstrated similar findings with a significant survival benefit through 11 LNs removed in patients undergoing lobectomy for NSCLC. In that study, even for patients with tumours less than 2 cm, a survival benefit was realized through the incremental resection of each additional LN resected up to four LNs. Despite these findings supporting the benefit of LN removal during surgical resection of lung cancer, neither analysis specifically addressed the question of whether LN resection is beneficial in patients undergoing non-anatomic wedge resection.

Whether or not we as thoracic surgeons should perform a lymphadenectomy with wedge resection is an important question, one with important implications for competing methods of local therapy such as SBRT or RFA. In fact, recent data suggests that thoracic surgeons often do not perform LN sampling with wedge resection. At individual academic institutions, as many as 68% of patients undergoing wedge resection for small (≤2 cm) NSCLC tumours may not have LNs sampled [11]. Furthermore, between 2003–2011, LN evaluation was not performed in 29% of patients undergoing sub-lobar resection for clinical stage IA NSCLC in the NCDB, in which 85% of patients had a wedge resection [2]. Even in well-controlled surgical trials, such as ACOSOG Z4032 designed to specifically evaluate surgical outcomes, 28% of patients undergoing wedge resection had NLNs sampled [12]. Clearly, many thoracic surgeons therefore do not believe lymphadenectomy to be an important part of the surgical therapy of NSCLC treated by wedge resection. However, in addition to potentially understaging patients and failing to identify those eligible for adjuvant therapy, a lack of nodal sampling could potentially contribute to increased local and regional recurrence.

We therefore undertook this study to determine how often we performed LN sampling with wedge resection for NSCLC in our own clinical stage I patient population and to determine whether there is any benefit to lymphadenectomy with regard to staging or oncologic outcome. We also thought it important to establish whether any downsides existed to LN resection, such as longer OR times, more postoperative complications, or need for longer chest tube duration or hospitalization.

Our institutional preference has been for anatomic resection to treat clinical stage IA NSCLC, as evidenced by the fact that 86% of our patients were treated with lobectomy or segmentectomy. Nevertheless, we identified 196 selected patients treated by wedge resection for cIA NSCLC who were reviewed for the current study. The reasons for wedge resection in this population are varied and are not perfectly captured by our database, but include patients enrolled in CALGB 140503, patients with limited pulmonary reserve, patients with previous lung surgery, and patients with very small peripheral nodules without aggressive radiographic features. Among the 196 wedge resections, 138 patients (70%) had LNs sampled while 58 patients (30%) had NLNs removed for pathologic analysis. This rate is similar to that reported in large national databases [1,2]. The LN and NLN groups were remarkably well balanced with regard to demographic and clinical tumour characteristics. Tumours were predominantly adenocarcinoma (80%), less than 2 cm in size, with low-SUVmax and as such are representative of screen detected or incidentally detected asymptomatic tumours.

Importantly, whether approached by open thoracotomy or by VATS resection, sampling or dissecting LNs with wedge resection did not appear to increase OR time, EBL, length of stay or complications. Wedge resection itself appears quite safe. This entire cohort of patients had exceedingly low rates of morbidity with grade III/IV Clavien Dindo complications occurring in only 6.7% of total patients, despite relatively poor pulmonary function (median FEV1 = 76%; median DLCO = 69%; 24% of total cohort with FEV1% or DLCO% ≤50%). Notably, there were no perioperative mortalities in either group. We believe these surgical morbidity and mortality rates are comparable to those reported in modern trials of SBRT and that they are not influenced by lymphadenectomy [13].

We also sought to determine rates of pathologic upstaging in patients undergoing wedge resection. Because these patients had been carefully clinically staged and were thought eligible for wedge resection, we expected low rates of upstaging. Nevertheless, we found a 20% overall rate of upstaging to greater than pIA in the cohort and found a 4.3% rate of nodal disease in patients undergoing wedge resection with LN sampling. Not surprisingly, this is lower than rates of nodal upstaging reported with lobectomy, although in our opinion still clinically significant [9]. Apart from this documented nodal disease, the LN and NLN groups were again remarkably well balanced with regards to pathologic factors. The majority were pT1 tumours with a median size of 1.5 cm, with only a minority of tumours having pleural or lymphovascular invasion. The median distance to the parenchymal margin was 1 cm in both groups.

Despite these clinical and pathologic similarities between the LN and NLN groups, we identified a statistically significant 27% difference in 5-year OS favouring patients who had LNs resected along with wedge resection of the primary tumour. We also demonstrated a halving of local/regional recurrence and a 25% difference in 5-year DFS with LN removal, both of which approached statistical significance. Several possible explanations exist for these findings. By identifying node positive patients, it is possible that further treatment with adjuvant chemotherapy reduced the rate of local and regional recurrence and improved survival. In our series, four out of six N+ patients received adjuvant therapy. Of these, only one recurred. That small number of patients alone would not seem to be able to explain the overall differences in recurrence and survival. Another possibility is that LN removal itself decreases local recurrence. Although not definitely established, there is some evidence of this phenomenon in other solid tumour types. In ACOSOG Z4032, the hazard ratio for local recurrence free survival was 0.88 (CI 0.56–1.37) for LN dissection/sampling versus none [14]. Although not statistically significant (P = 0.57) in this relatively small cohort (n = 112 vs 63 patients), the magnitude of benefit is similar to that seen in our own study.

Finally, it is likely that we as surgeons determine the extent of resection and lymphadenectomy based upon our clinical judgment. In this sense, patients who we may subjectively consider to be frail, at higher risk for death from competing causes, or unlikely to tolerate adjuvant therapy may be more likely to undergo wedge resection without nodal assessment. This could certainly account for some of the differences in OS seen between the NLN and LN groups, despite the fact that the groups appear well balanced for comorbidities and pulmonary function. However, it would seem unlikely that these factors alone could account for the difference in local and regional recurrence seen between the NLN and LN groups, particularly given the similarity between the groups with regard to pathologic factors.

In conclusion, we have demonstrated that NSCLC patients undergoing wedge resection with LN removal appear to have better survival and less local and regional disease recurrence than patients without LN sampling. The study is limited by its retrospective nature, the inherent selection bias that goes into choosing patients for wedge resection, and by the relatively low number of patients in the cohorts. Nevertheless, we believe these findings to be important and oncologically intuitive. LN sampling is obviously more likely to identify patients with occult nodal disease and to identify candidates for adjuvant therapy. We have demonstrated that LN sampling with wedge resection does not increase operative time, perioperative morbidity, length of stay, or mortality. As such, we can see little reason not to perform LN sampling in patients with clinical stage I disease. The potential decrease in local and regional recurrence is yet another compelling reason to remove LNs along with the primary tumour. The ability to do so and to correctly stage NSCLC patients is a distinct advantage of surgical therapy over competing local treatment modalities such as SBRT and RFA.

Conflicts of interest: none declared.

REFERENCES

Author notes