Diagnosis and Management of Lung Cancer, 3rd ed: American College of Chest Physicians Evidence-Based Clinical Practice Guidelines Online Only Articles |

Treatment of Stage I and II Non-small Cell Lung CancerTreatment of Stage I and II Non-small Lung Cancer: Diagnosis and Management of Lung Cancer, 3rd ed: American College of Chest Physicians Evidence-Based Clinical Practice Guidelines FREE TO VIEW

John A. Howington, MD, FCCP; Matthew G. Blum, MD, FCCP; Andrew C. Chang, MD, FCCP; Alex A. Balekian, MD, MSHS; Sudish C. Murthy, MD, PhD, FCCP
Author and Funding Information

From the NorthShore HealthSystem (Dr Howington), University of Chicago Pritzker School of Medicine, Evanston, IL; Penrose Cardiothoracic Surgery (Dr Blum), Memorial Hospital, University of Colorado Health, Colorado Springs, CO; University of Michigan Medical School (Dr Chang), Ann Arbor, MI; Division of Pulmonary, Critical Care, and Sleep Medicine (Dr Balekian), Keck School of Medicine of University of Southern California, Los Angeles, CA; and the Department of Thoracic and Cardiovascular Surgery (Dr Murthy), Cleveland Clinic, Cleveland, OH.

Correspondence to: John A. Howington, MD, FCCP, NorthShore University Health System, 2650 Ridge Ave, 3507 Walgreen Bldg, Evanston, IL 60201; e-mail: jhowington@northshore.org

COI Grids reflecting the conflicts of interest that were current as of the date of the conference and voting are posted in the online supplementary materials.

Funding/Sponsors: The overall process for the development of these guidelines, including matters pertaining to funding and conflicts of interest, are described in the methodology article.1 The development of this guideline was supported primarily by the American College of Chest Physicians. The lung cancer guidelines conference was supported in part by a grant from the Lung Cancer Research Foundation. The publication and dissemination of the guidelines was supported in part by a 2009 independent educational grant from Boehringer Ingelheim Pharmaceuticals, Inc.

Reproduction of this article is prohibited without written permission from the American College of Chest Physicians. See online for more details.

Chest. 2013;143(5_suppl):e278S-e313S. doi:10.1378/chest.12-2359
Text Size: A A A
Published online

Background:  The treatment of stage I and II non-small cell lung cancer (NSCLC) in patients with good or low surgical risk is primarily surgical resection. However, this area is undergoing many changes. With a greater prevalence of CT imaging, many lung cancers are being found that are small or constitute primarily ground-glass opacities. Treatment such as sublobar resection and nonsurgical approaches such as stereotactic body radiotherapy (SBRT) are being explored. With the advent of minimally invasive resections, the criteria to classify a patient as too ill to undergo an anatomic lung resection are being redefined.

Methods:  The writing panel selected topics for review based on clinical relevance to treatment of early-stage lung cancer and the amount and quality of data available for analysis and relative controversy on best approaches in stage I and II NSCLC: general surgical care vs specialist care; sublobar vs lobar surgical approaches to stage I lung cancer; video-assisted thoracic surgery vs open resection; mediastinal lymph node sampling vs lymphadenectomy at the time of surgical resection; the use of radiation therapy, with a focus on SBRT, for primary treatment of early-stage NSCLC in high-risk or medically inoperable patients as well as adjuvant radiation therapy in the sublobar and lobar resection settings; adjuvant chemotherapy for early-stage NSCLC; and the impact of ethnicity, geography, and socioeconomic status on lung cancer survival. Recommendations by the writing committee were based on an evidence-based review of the literature and in accordance with the approach described by the Guidelines Oversight Committee of the American College of Chest Physicians.

Results:  Surgical resection remains the primary and preferred approach to the treatment of stage I and II NSCLC. Lobectomy or greater resection remains the preferred approach to T1b and larger tumors. The use of sublobar resection for T1a tumors and the application of adjuvant radiation therapy in this group are being actively studied in large clinical trials. Every patient should have systematic mediastinal lymph node sampling at the time of curative intent surgical resection, and mediastinal lymphadenectomy can be performed without increased morbidity. Perioperative morbidity and mortality are reduced and long-term survival is improved when surgical resection is performed by a board-certified thoracic surgeon. The use of adjuvant chemotherapy for stage II NSCLC is recommended and has shown benefit. The use of adjuvant radiation or chemotherapy for stage I NSCLC is of unproven benefit. Primary radiation therapy remains the primary curative intent approach for patients who refuse surgical resection or are determined by a multidisciplinary team to be inoperable. There is growing evidence that SBRT provides greater local control than standard radiation therapy for high-risk and medically inoperable patients with NSCLC. The role of ablative therapies in the treatment of high-risk patients with stage I NSCLC is evolving. Radiofrequency ablation, the most studied of the ablative modalities, has been used effectively in medically inoperable patients with small (< 3 cm) peripheral NSCLC that are clinical stage I.

Figures in this Article
General Approach

2.1.1. For patients with clinical stage I and II non-small cell lung cancer (NSCLC) and no medical contraindications to operative intervention, surgical resection is recommended (Grade 1B).

2.1.2. For patients with clinical stage I and II NSCLC, it is suggested that they be evaluated by a thoracic surgical oncologist or a multidisciplinary team even if the patients are considered for nonsurgical therapies such as percutaneous ablation or stereotactic body radiation therapy (Grade 2C).

Remark: At a minimum, we suggest that multidisciplinary teams have representatives from pulmonary medicine, thoracic surgery, medical oncology, radiation oncology, radiology, and pathology. For patients with clinical stage I or II NSCLC and who are medically fit, it is recommended that they be treated by a board certified thoracic surgeon with a focus on lung cancer (Grade 1B).

Remark: Ideally general thoracic surgical procedures would constitute > 75% of the thoracic surgeon’s clinical practice, and involve an average of ≥ 4 anatomic surgical resections performed per month at the center to maintain the experience and smooth function of the care teams.

Lobectomy: Surgical Issues

3.2.1. For patients with clinical stage I NSCLC, a minimally invasive approach such as video-assisted thoracic surgery (thoracoscopy) is preferred over a thoracotomy for anatomic pulmonary resection and is suggested in experienced centers (Grade 2C). For patients with clinical stage I and II NSCLC, systematic mediastinal lymph node sampling or dissection at the time of anatomic resection is recommended over selective or no sampling for accurate pathologic staging (Grade 1B). For patients with clinical stage I NSCLC undergoing anatomic resection who have undergone systematic hilar and mediastinal lymph node staging showing intraoperative N0 status, the addition of a mediastinal lymph node dissection does not provide a survival benefit and is not suggested (Grade 2A). For patients with clinical stage II NSCLC undergoing anatomic resection, mediastinal lymph node dissection may provide additional survival benefit over mediastinal lymph node sampling and is suggested (Grade 2B).

3.5.1. For patients with clinical stage I or II central NSCLC in whom a complete resection can be achieved, a sleeve or bronchoplastic resection is suggested over a pneumonectomy (Grade 2C).

Sublobar Resection

4.3.1. For patients with clinical stage I and II NSCLC who are medically fit for surgical resection, a lobectomy rather than sublobar resection is recommended (Grade1B).

4.8.1. For patients with clinical stage I NSCLC who may tolerate operative intervention but not a lobar resection due to decreased pulmonary function or comorbid disease, sublobar resection is recommended over nonsurgical therapy (Grade 1B).

Remark: Regardless of whether patients undergo wedge or segmentectomy, adequate margins should be achieved.

Remark: Sublobar resection should involve an anatomic segmentectomy whenever possible.

4.8.2. During sublobar resection of solid tumors in compromised patients, it is recommended that margins greater than the maximal tumor diameter for lesions less than 2 cm should be achieved; for tumors larger than 2 cm at least 2 cm gross margins should be sought to minimize the likelihood of a positive margin and/or local recurrence (Grade 1C).

Remark: The data regarding the appropriate gross margin necessary to achieve a pathologically negative margin or minimize local recurrence in larger tumors (> 2 cm) is less well established. It may be that larger margins for larger tumors are required.

Remark: In patients undergoing resection of solid stage cI NSCLC in whom the ability to achieve an adequate margin is compromised, the addition of brachytherapy mesh to sublobar resection may improve local control.

4.10.1. In patients with major increased risk of perioperative mortality or competing causes of death (due to age related or other co-morbidities), an anatomic sublobar resection (segmentectomy) over a lobectomy is suggested (Grade 2C).

4.12.1. For patients with a clinical stage I predominantly ground glass opacity (GGO) lesion ≤ 2 cm, a sublobar resection with negative margins is suggested over lobectomy (Grade 2C).

Nonresectional Treatment Approaches

5.3.1. For patients with clinical stage I NSCLC who cannot tolerate a lobectomy or segmentectomy, stereotactic body radiation therapy (SBRT) and surgical wedge resection are suggested over no therapy (Grade 2C).

Remark: Surgical resection has the potential benefit of definitive histologic analysis (eg, adenocarcinoma subtype) and pathologic nodal information. In compromised patients for whom such information would not change management SBRT is a preferred option. Also, SBRT is favored in patients for whom an adequate margin in unlikely with a surgical wedge resection.

Remark: Radiofrequency ablation may also be considered for peripheral tumors < 3 cm in inoperable patients.

Adjuvant Therapy For patients with completely resected pathologic stage IA,B NSCLC, it is recommended that postoperative chemotherapy not be used (outside of a clinical trial) (Grade 1B). For patients with completely resected pathologic stage IIA,B(N1) NSCLC and good performance status, postoperative platinum-based chemotherapy is recommended (Grade 1A).

Remark: No clear recommendation is possible regarding adjuvant chemotherapy for larger tumors without lymph node involvement. For patients with completely resected pathologic stage I NSCLC, it is recommended that postoperative radiation therapy should not be used (Grade 1A). For patients with completely resected pathologic stage II NSCLC, it is suggested that postoperative radiation therapy should not be used (Grade 2A). For patients with stage I and II NSCLC and a positive bronchial margin (R1 resection), postoperative radiation therapy is suggested (Grade 2C).

Patients with stage I and II non-small cell lung cancer (NSCLC) have early-stage disease and are typically approached with curative intent therapy. Although no patient can be guaranteed a cure, the chance of cure is sufficiently high that no patient with clinical stage I or II NSCLC should be denied treatment until evaluated by a physician or preferably a multidisciplinary team of physicians experienced in the care of patients with lung cancer.

Unfortunately, stage I and II NSCLC combined account for only 25% to 30% of all patients presenting with lung cancer in the United States today. Since the last edition of the lung cancer guidelines was published, the seventh edition of TNM staging in Lung Cancer was published by the International Union Against Cancer and the American Joint Committee on Cancer (AJCC).2 The present staging system is based on analyses of the 67,725 cases of NSCLC within the database. The analysis of this database allowed a detailed study of the impact of tumor size on prognosis. The seventh edition staging system is now recommended for the classification of carcinoid tumors and small cell lung cancer as well as NSCLC. This review and recommendations are focused on the treatment of stage I and II NSCLC only, in a way that is relevant to the current (seventh edition) stage classification system.

Changes in the seventh edition of the lung cancer stage classification system that impacted stage I and II NSCLC are T related and include new size cuts for a refined T stage classification. T1a tumors are ≤ 2 cm, T1b tumors are > 2 to 3 cm in size. T2a tumors are > 3 to 5 cm in size, and T2b tumors are > 5 to 7 cm in size. Tumors > 7 cm in size or tumors with satellite lesions in the same lobe are now classified as T3 tumors. In the new staging system, patients with T2b (> 5-7 cm) with negative nodes and no distant disease are stage IIA. Patients with T2a tumors and N1 lymph nodes are also stage IIA. Patients with N1 lymph nodes are classified as stage IIa if they involve a T2a tumor and stage IIB with a T2b tumor. Patients with T3N0M0 tumors are also classified as stage IIB. In this article we discuss T3 tumors either due to size (> 7 cm) or invasion; T3 tumors due to the presence of same-lobe additional nodules are discussed in the article by Kozower et al,3 “Special Treatment Issues in Non-small Cell Lung Cancer,” in the American College of Chest Physicians (ACCP) Lung Cancer Guidelines. A more detailed discussion of the current stage classification system is provided in the article by Detterbeck et al,4 “The Stage Classification of Lung Cancer,” in the ACCP Lung Cancer Guidelines.

A multidisciplinary writing committee composed of a pulmonologist, a methodologist, and four thoracic surgeons was assembled and approved according to the process for the ACCP Lung Cancer Guidelines as described by Lewis et al,1 “Methodology for Development of Guidelines for Lung Cancer,” in the ACCP Lung Cancer Guidelines.1 To update previous recommendations on the treatment of stages I and II NSCLC, a series of population, intervention, comparator, outcome (PICO) questions (Table S1) were developed and used to structure a literature search of computerized databases (MEDLINE, COCHRANE, Google Scholar, and Embase in some cases). The search included papers published through the end of 2011, overlapping with the previous ACCP lung cancer guidelines. Search terms were chosen according to the questions being addressed (details of the search strategy and results are available on request). Only papers published in English were included in the reviews. Articles were selected if they reported on the outcomes in question in patients with stage I or II NSCLC and included randomized controlled trials (RCTs) as well as cohort and outcomes studies.

A multidisciplinary writing committee reviewed and summarized the available literature and assessed the quality of the supporting evidence using a standardized method (see Lewis et al1 in the ACCP Lung Cancer Guidelines). The writing committee used this to develop recommendations and graded the strength of the recommendations. The evidence and the recommendations were then discussed and refined and finally approved by the entire lung cancer guidelines panel as described elsewhere.1 In accordance with the ACCP Lung Cancer Guidelines (third edition) process, the resulting guideline was reviewed and approved by the ACCP Guidelines Oversight Committee as well as external reviewers.

Surgical resection is widely accepted as the optimal treatment of stage I and II NSCLC over no treatment. There are no RCTs comparing surgery alone to radiotherapy alone, chemotherapy alone, or ablative therapies in otherwise healthy patients with stage I and II NSCLC. The argument for surgery comes primarily from consistent data from retrospective surgical series and databases and registries showing higher survival rates after surgery than after other treatment modalities. Based on such retrospective series of resected stage I and stage II NSCLC, the prognoses for stage I and II NSCLC, expressed in terms of 5-year survival rates, are commonly accepted to be 60% to 80% for stage I and 30% to 50% for stage II NSCLC.5 In contrast, data on the natural history of untreated clinical stage I or II NSCLC consistently shows poor survival (2-year survival of approximately 20% and 15% 5-year survival).6 Although these patients are not comparable, the natural history data are consistent with data from tumor doubling times of NSCLC that were observed for a period of time before resection.6 Although indirect and stemming from uncontrolled comparisons, the large amount of data and the consistency thereof makes a strong argument for surgical resection over no treatment.5,6 This also underscores that surgery should be considered carefully before alternative approaches or no treatment is chosen.

2.1 Recommendation

2.1.1. For patients with clinical stage I and II NSCLC and no medical contraindications to operative intervention, surgical resection is recommended (Grade 1B).

2.1.2. For patients with clinical stage I and II NSCLC, it is suggested that they be evaluated by a thoracic surgical oncologist or a multidisciplinary team even if the patients are considered for nonsurgical therapies such as percutaneous ablation or stereotactic body radiation therapy (Grade 2C).

Remark: At a minimum, we suggest that multidisciplinary teams have representatives from pulmonary medicine, thoracic surgery, medical oncology, radiation oncology, radiology, and pathology.

2.2 Surgical Expertise and Volume

There is a growing body of knowledge showing improved perioperative surgical outcomes and better long-term survival in patients with cancer when they are treated by specialists in cancer surgery and at higher-volume centers. It is certainly plausible that a specific focus, specialized training, and higher volume lead to organization of the process of care in ways that result in better outcomes. Thoracic surgeons have had increased exposure in terms of broader experience and greater operative volume of lung resections during their resident education, particularly in the recent era of a separate pathway focused on general thoracic surgery for certification in thoracic surgery. For the noncardiac thoracic surgeon, thoracic procedures represent a greater proportion of their practice compared with general surgeons or cardiac surgeons. In the United States, general thoracic surgeons are typically defined as American Board of Thoracic Surgery certified surgeons for whom at least 50%, and for most individuals 75%, of their clinical practice is composed of thoracic procedures.

2.2.1 Specialization:

Our systematic literature search identified a number of papers reporting both short-term (perioperative mortality) and long-term outcomes (4- or 5-year survival) outcomes relative to surgeon specialty. In the United States, about 30% of pulmonary resections are performed by general surgeons, 45% by board-certified (BC) cardiothoracic surgeons who practice primarily cardiac surgery, and 25% by BC cardiothoracic surgeons who practice primarily thoracic surgery.7 We selected papers reporting an adjusted OR, summarized in Figure 1.811 We report the OR relative to a consistent denominator of the general surgeon as the reference (or the cardiac surgeon when comparing only among cardiac and thoracic surgeons).

Figure Jump LinkFigure 1. Outcomes of surgical resection for lung cancer according to specialization.

Inclusion criteria: multicenter database studies reporting adjusted outcomes for thoracic surgery relative to specialization up to 2012. Lobe = lobectomy; MC = Medicare; NIS = Nationwide Inpatient Sample; NS = not significant; Pneum = pneumonectomy; S = Surgeon; SEER = Surveillance Epidemiology and End Results.

aFor long-term survival, rates are an adjusted hazard ratio.

Grahic Jump Location

Figure 1 illustrates the consistent finding in every study of lower mortality in patients treated by thoracic or cardiac surgeons compared with general surgeons. However, the differences were sometimes not statistically significant. Only a few studies have reported a comparison between cardiac and thoracic surgeons, with some showing no difference and others showing lower perioperative mortality for thoracic surgeons. This is corroborated by a recent high-quality meta-analysis, with a pooled OR for perioperative mortality of 0.78 (95% CI, 0.7-0.88; P < .0001) for thoracic surgeons vs general surgeons, and an OR of 0.82 (95% CI, 0.69-0.96; P = .016) for cardiothoracic surgeons vs general surgeons.12 Furthermore, the rate of complications (or a surrogate for complications, eg, prolonged length of stay [LOS]) was lower for thoracic and cardiac surgeons as compared with the general surgeons in most studies.79 This lower incidence of complications occurred despite the finding that the patients treated by thoracic surgeons had more comorbidities.13

The studies have also consistently noted better long-term survival after surgery performed by a thoracic or cardiac surgeon vs a general surgeon. The rate of mediastinal node dissection is twofold to threefold higher among specialty-trained surgeons. Other measures, such as the use of PET scanning and video-assisted thoracic surgery (VATS) resections, are more commonly done by thoracic and cardiac surgeons, whereas the use of neoadjuvant or adjuvant therapy is not different.7

The data shown in Figure 1 are OR that are adjusted for other factors, such as patient characteristics or facility characteristics (eg, high or low volume). However, which factors and how many were considered in the adjustment is inconsistent.12 Different databases were used for these analyses, each of which has strengths and weaknesses. Furthermore, although the definition of a general surgeon vs one who is BC in cardiothoracic surgery is clear, the criteria used to differentiate a cardiac from a thoracic surgeon are somewhat varied. The studies in Figure 1 all come from the United States with the exception of one from The Netherlands.11

The two largest studies used the National Inpatient Sample (NIS) database, which is specifically designed to represent a US cross-section of hospitals, involving about 20% of all hospital discharges regardless of insurance type.8,9 Schipper et al8 observed a raw mortality of 11.5%, 10.1%, and 6.4% for pneumonectomy and 4.1%, 3.0%, and 2.3% for lobectomy among general surgeons, cardiac surgeons, and thoracic surgeons, respectively. After adjustment for age, sex, type of hospital, case volume, insurance type, and 24 comorbidities, there was little difference between general and cardiac surgeons, but there remained a strong trend, albeit not statistically significant, for thoracic surgeons vs general surgeons (OR, 0.70; 95% CI, 0.47-1.06; P = .093). In this study, surgeons were not defined by board certification but by caseload in their practice. In a similar study involving largely overlapping cohorts from the NIS database, Ellis et al9 found statistically significant differences between thoracic and cardiac surgeons vs general surgeons. Exploratory sensitivity analyses suggested that the threshold for defining a surgeon as one or another type affected the results, with the best results seen in those very specifically focused on thoracic surgery (95% thoracic surgical cases).9

2.2.2 Case Volume and Structural Characteristics:

The search identified a large number of studies investigating whether outcomes are affected by the case-volume of centers performing surgery for lung cancer; we selected studies reporting on adjusted ORs, taking into account other potential factors. These are summarized in Figure 2.1433 The studies have demonstrated lower hospital mortality rates in higher-volume centers, with only one exception. The difference in perioperative mortality was statistically significant in most of the studies. This is corroborated by a recent good-quality meta-analysis (Fig 3A), which found a pooled OR of 0.71 (95% CI, 0.62-0.81; P < .0001).12 The few studies addressing individual surgeon volumes for lung cancer resections have found lower perioperative mortality is associated with high-volume surgeons.29,30

Figure Jump LinkFigure 2. Outcomes of surgical resection for lung cancer according to case volume.

Inclusion Criteria: Multicenter database studies reporting adjusted outcomes for thoracic surgery relative to case volume up to 2012. CA = California database; CI = 95% confidence interval; FL = Florida database; High-vol = high volume; Low-Vol = low volume; JACS = Japanese Association of Chest Surgery; JATS = Japanese Association of Thoracic Surgery; NCDB = National Cancer Database; NSQIP = National Surgical Quality Improvement Program; NY = New York database; Reg = registry; Seg = segmentectomy; Surg-Vol = surgeon volume. See Figure 1 legend for expansion of other abbreviations.

aFor long-term survival rates are an adjusted hazard ratio.

bN is total number of patients.

cRelative to high volume teaching hospital.

Grahic Jump Location
Figure Jump LinkFigure 3. [Section 2.2.2] Meta-analysis of relationship between outcomes and hospital case volume of lung cancer resections. A, Perioperative mortality. B, Long-term survival. Higher limit = lowest annual volume in the high-volume category; lower limit = highest annual volume in the lower limit category. (Reproduced with permission from von Meyenfeldt et al.12)Grahic Jump Location

A large number of studies have also found a benefit to long-term survival in patients having undergone resection at a higher-volume institution (Fig 2), although two smaller studies found the opposite effect. The recent meta-analysis found a pooled OR of 0.93 (95% CI, 0.84-1.03; P = .16) (Fig 3B).12

These studies have involved adjusted ORs, taking into account other factors that might have an impact. However, which factors and how many were included is variable. The thresholds for volume that were used were also quite variable. Although most of the studies were conducted in the United States, the results are consistent with studies from Asia and Europe. The types of databases used also varied.

The largest study examined 90,088 patients undergoing pneumonectomy. The unadjusted perioperative (in-hospital) mortality was 2.7% vs 4.9% in the highest-volume vs the lowest-volume deciles of hospitals, and a rate of prolonged LOS, respectively, was 14% vs 8%.14 A large US National Cancer Database study estimated that if perioperative mortality and survival rates equal to that in the best-performing quintile of hospitals could be realized across the board, > 500 perioperative deaths and nearly 1,500 subsequent deaths annually could be avoided in the United States among patients undergoing lung cancer surgery.15

The search also identified several studies addressing the impact of teaching hospitals (those with a training program or university medical center) that met the inclusion criteria. Most of these have also found better outcomes (both short term and long term) at teaching hospitals, even after adjustment for other factors such as case-volume (Fig 4).34 A significant benefit was found for surgery performed at a thoracic surgery teaching hospital vs a non-thoracic surgery teaching hospital (perioperative mortality of 2.7% vs 3.7%, P < .05).34

Figure Jump LinkFigure 4. Outcomes of surgical resection for lung cancer according to teaching facility status.

Inclusion criteria: multicenter database studies reporting adjusted outcomes for thoracic surgery relative to teaching facility status up to 2012. See Figure 1 and 2 legends for expansion of abbreviations.

aFor long-term survival, rates are an adjusted hazard ratio.

bRelative to medium-sized nonteaching hospital.

Grahic Jump Location
2.2.3 Summary:

The data in aggregate (Figs 1, 2, 4) present a strong body of evidence for referring patients with known or suspected early-stage NSCLC to physicians and centers with a focus on lung cancer care. In general, the benefit amounts to an approximately 15% to 20% improvement in surgical outcomes, which is far greater that attributable to interventions such as adjuvant therapy or newer chemotherapy drugs. A focus on high-quality care is necessary to improve the outcomes of patients with lung cancer, a disease that requires a multidisciplinary approach, in which approximately 30% of patients with early-stage disease receive no definitive treatment and which carries an overall cure rate < 20%.

It is difficult, however, to define exactly which factors account for the improved outcomes. Although the relationship between volume and outcomes may be fairly consistent within a large database, it is less consistent as a predictor of outcomes when applied to individual institutions. Furthermore, low perioperative mortality in a given year is a poor predictor of overall good outcomes.35 Furthermore, at least a portion of the impression that low-volume centers have worse outcomes may be attributed to the way the data are analyzed and portrayed36,37; and the contribution of volume to the variation in perioperative mortality rates is relatively low.37

Furthermore, there are difficulties with implementation of case volume as a tool to improve outcomes. A sophisticated meta-analysis was unable to identify a potential threshold as a target.12 Travel distance to an appropriate hospital is probably less important: studies have found that 75% of patients could reach a high-volume hospital by traveling < 25 additional miles, and 25% actually travel farther to a more distant low-volume hospital.38 However, if patients were redirected to the centers currently in the upper 2 quintiles of volume (top 40%), their volume would increase by 250%, and 91% of hospitals would stop providing thoracic surgery.17 The wisdom and practicality of such redirection needs thoughtful consideration.

Furthermore, there are data that having effective systems of care may be a key factor, and volume may merely be associated with this aspect. One large study has demonstrated that the incidence of complications was not correlated with perioperative mortality and appears to be attributable to patient-related factors and not to surgical quality. Instead, decreased perioperative mortality correlates primarily with a better response once complications occur.39 It appears that a shift to higher-volume institutions alone is not adequate,25 and that organization of care plays an important role.40

In summary, the setting in which pulmonary resection is accomplished has a major impact on short-term and long-term outcomes. Treatment at a center with higher volume, specialty trained individuals, and an organized system of care is likely to result in significant benefits. However, the relationship of these factors to outcomes is complex and requires further study.

2.2.4 Recommendation For patients with clinical stage I or II NSCLC and who are medically fit, it is recommended that they be treated by a board certified thoracic surgeon with a focus on lung cancer (Grade 1B).

Remark: Ideally general thoracic surgical procedures would constitute > 75% of the thoracic surgeon’s clinical practice, and involve an average of ≥ 4 anatomic surgical resections performed per month at the center to maintain the experience and smooth function of the care teams.

3.1 Minimally Invasive vs Open Surgical Resection

Several recent single-institution retrospective series have demonstrated that VATS lobectomy is associated with fewer complications,4147 lower estimated blood loss or transfusion rate,42,48 and shorter hospital LOS,41,4448 even for older patients (Figs 5A, 5B).41,42,47,4951 Two meta-analyses50,52 (quality assessed as fair and good) and two systematic reviews53,54 (both poor quality) of VATS vs open lobectomy have been reported. These studies all found VATS is associated with a short-term benefit (reduced complications, perioperative mortality, pain) and long-term recurrence and survival rates that are at least as high as after thoracotomy. In addition, outcomes studies using US databases5557 confirmed lower complications, mortality, and length of hospital stay, even when hospital volume is taken into account, as well as equivalent long-term survival. Nevertheless, most of the data come from nonrandomized comparisons; only three small RCTs have been conducted,5860 which demonstrated equivalent outcomes but were too small to demonstrate this in a statistically robust manner. Most of these studies involved a mixture of stages, but the large majority of patients were stage I and II.

Figure Jump LinkFigure 5. [Section 3.1] Outcomes of thoracoscopic vs open lung cancer resection. Results of a meta-analysis of comparative studies. Most of the included studies were cohort studies, not randomized studies. A, Perioperative outcomes. B. Long-term outcomes. Card = cardiac; Compl = complication; F/U = follow-up; Hosp LOS = hospital length of stay; Periop = perioperative; Pulm = pulmonary; rand = randomized studies; Resp dysfctn = respiratory dysfunction; S at max f/u = survival at maximal follow-up; Surv = survival; VATS = video-assisted thoracic surgery. (Data from Cheng et al.50 Figures reproduced with permission from Detterbeck.51)Grahic Jump Location

Although our search identified many papers comparing VATS to open lobectomy, few of them specifically addressed stage I or II NSCLC. Analyzing data obtained from the Society of Thoracic Surgeons General Thoracic Database, a prospective surgical specialty-specific registry, Paul et al56 identified subjects undergoing lobectomy from 2002 to 2007. Two matched cohorts of 1,281 patients each were defined by surgical approach: VATS or thoracotomy. Subjects were matched by propensity score using baseline preoperative variables, including patient demographics, comorbidities, and pulmonary function; however, data were missing for clinical stage in nearly 40% (1,021 of 2,562) and pathologic stage in 21% (546 of 2,562). Patients undergoing VATS were more likely to have clinical stage IA/IB disease, although the final (pathologic) staging distribution was not significantly different between the two groups. Patients undergoing VATS lobectomy experienced fewer postoperative complications overall: (26% vs 35%, P < .0001), fewer pulmonary complications (8% vs 12%, P < .0001), less atrial fibrillation requiring medical treatment (7% vs 12%, P < .0004), and fewer blood transfusions (2% vs 5%, P < .003). In addition, VATS lobectomy was associated with earlier chest tube removal (median 3 days vs 4 days, P < .0001) and shorter hospital LOS (median 4 days vs 6 days, P < .0001). This study identified higher operative times (median 173 min vs 143 min, P < .0001) for subjects having VATS resection. In addition, a small randomized study found no difference in long-term survival in 100 patients with stage cIa disease.60

Using the 2008 NIS transactional database, Park and colleagues57 evaluated the outcomes of VATS lobectomy compared with open pulmonary resection. Unlike the Society of Thoracic Surgeons General Thoracic Database, the NIS is an all-payer inpatient care database that captures the discharge records of 20% of the population receiving care at nonfederal institutions within the United States. This study identified 6,292 patients undergoing lobectomy, of whom 1,523 (24%) were treated by VATS. VATS lobectomy was associated with fewer total complications and shorter hospital LOS in both univariate and multivariate analysis.

A systematic review and meta-analysis of the number of lymph nodes dissected or biopsied from both RCT and cohort studies demonstrates no difference between VATS vs open resection, although a few studies have suggested that VATS resections may be associated with fewer mediastinal44,47 or N1 nodes assessed.61 This may be an issue that is less an inherent limitation of the VATS technique and more of a reflection on the surgeon’s commitment to accurate nodal staging.

In summary, consistent data from small RCTs, matched case-control studies, cohort studies, meta-analyses, and outcomes studies demonstrate lower operative mortality, complications, and LOS. These studies also demonstrate at least equivalent long-term survival either overall or when matched for stage.

3.2 Recommendation

3.2.1. For patients with clinical stage I NSCLC, a minimally invasive approach such as video-assisted thoracic surgery (thoracoscopy) is preferred over a thoracotomy for anatomic pulmonary resection and is suggested in experienced centers (Grade 2C).

3.3 Mediastinal Lymph Node Dissection vs Systematic Sampling

Principles and expectations for acceptable surgical management have been established for the good-performance-status patient with early-stage NSCLC.6264 A definition of “complete resection” has been developed by the International Staging Committee of the International Association for the Study of Lung Cancer. This is the only data-derived definition and serves as the international standard. This definition mandates that a negative microscopic margin be obtained and systematic lymph node dissection be performed before a resection can be declared “complete.”62 Moreover, extracapsular nodal extension cannot be present, and the highest (most distant) mediastinal lymph node must be examined and found to be free of cancer to fulfill the definition. Anything less (positive margin, extracapsular lymph node spread, or positive mediastinal lymph nodes left in situ) renders the resection incomplete.

Formal definitions have also been developed by the European Society of Thoracic Surgeons, the largest organization worldwide of non-cardiac thoracic surgeons,65 for terms describing the extent of mediastinal staging technique, and this serves as the most widely used international standard nomenclature. One should distinguish between a selective biopsy or sampling (involving only selected suspicious or representative nodes), a systematic sampling (exploration and biopsy of a standard set of lymph node stations in each case), and a formal mediastinal lymph node dissection (MLND), which involves removal of all node-bearing tissue within defined landmarks for a standard set of lymph node stations. However, despite the formal definitions, many have used the terms systematic sampling and node dissection very loosely,66 especially in the past. There is clearly a continuum between MLND and systematic sampling, and the overlap between the two complicates clinical trials comparing them.67 Finally, definitions of levels of thoroughness of intraoperative (and preoperative) mediastinal staging exist.68

The last issue that confounds assessing the impact of MLND for patients with early-stage NSCLC is inaccuracy of clinical staging. Studies providing the most reliable pre-resection clinicopathologic information (radiographic and tissue confirmation) are intertwined with those using only imaging to define the clinical stage; these differences cause difficulties in interpretation of the data.

3.3.1 Stage I NSCLC:

MLND does not add a survival benefit for patients with stage I NSCLC compared with systematic sampling. More specifically, MLND offers no survival advantage for patients who have undergone a negative systematic mediastinal and hilar lymph node sampling procedure. A randomized, multi-institutional trial in this unique cohort of patients with clinicopathologic stage I disease clearly demonstrates equality between MLND and systematic sampling.69 As entry criteria for this trial, patients with clinical stage I NSCLC were subjected to meticulous sampling of the hilum and mediastinum prior to randomization. Only patients in whom all systematically sampled lymph nodes were found to be free of cancer were eligible and randomized. Those patients subsequently undergoing anatomic resection and MLND had similar survival, and no important difference in local, regional, or distant lung cancer recurrence, compared with those randomized to resection alone. It has been argued that the two study groups were not disparate enough to identify a potential small impact of MLND,67 despite far more lymph tissue being resected in the MLND group.70 Moreover, in retrospect, given the rigorous entry criteria, the study might have been slightly underpowered despite its impressive accrual because of the low incidence of occult mediastinal disease in the study population.71,72 Regardless, this landmark study has addressed the issue far more critically and scientifically than any previous study and serves as the foundation upon which guidelines must be rendered.

Other less well-constructed randomized and retrospective trials generally corroborate the conclusion that MLND does not offer a clear survival advantage for resected early-stage NSCLC. A two-center randomized, but smaller, trial performed 10 years prior demonstrated no evident differences in disease-free or overall survival when MLND was added to anatomic lung cancer resection compared with conventional systematic sampling.73 Local recurrence rates were slightly lower in the MLND group, although not statistically dissimilar from those observed in systematic sampling patients. In a subset analysis from this study, patients with very-low-burden lymph node involvement did demonstrate a relapse-free survival benefit from MLND.73 A retrospective propensity analysis of patients undergoing either MLND or a modified “selective” MLND (where the lymphadenectomy extent was influence by cancer presentation) also demonstrated no benefit of MLND and suggested that the additional small morbidity associated with MLDN is not justified.74 Finally, in a retrospective analysis, despite removal of nearly twice the number of mediastinal lymph nodes in a MLND arm, no difference in detection of occult pN2 disease or survival advantage was bestowed upon stage I patients when compared with systematic sampling.75

A third randomized trial comparing MLND and selective sampling did show some therapeutic benefit of MLND, but it is confounded by the inaccuracies of clinical staging used for entrance criteria.76 Almost one-half of the patients with clinical stage I disease undergoing MLND migrated to stage IIIA on pathologic staging (and became excluded from subset analysis), whereas far fewer patients with stage IIIA disease were found in the cohort undergoing selective sampling. Consequently, it must be assumed that a large number of patients with clinical stage I but pathologic stage IIIA remained undiscovered in the selective sampling group, and their presence skewed the subset analysis of patients with stage I disease in favor of the MLND cohort. This source of bias, perhaps best known as the Will Rogers phenomenon, likely confounds the vast majority of studies on this topic (randomized or not) and cannot be ignored when interpreting these data. Perhaps the only study exempt from this criticism is that by Darling et al,69 highlighting its impact in guideline construction.

Despite rigorous randomized trials suggesting the contrary, a few retrospective studies do favor some benefit of MLND for stage I disease. These studies are, as expected, plagued by the standard biases and inaccuracies associated with nonrandomized data. Nonetheless, there does appear to be less local recurrence in patients undergoing MLND in these studies.7779 The question of whether these patients truly had stage I disease, however, confounds all but one of these studies.78 Consequently, it is difficult to use their results to address this question.

In summary, in the setting of very careful and extensive hilar and mediastinal sampling and intraoperative confirmation of lack of nodal involvement, there is no benefit to proceeding with a complete mediastinal node dissection. The practicality of extensive intraoperative sampling to confirm this is a potential issue. There is also little detriment to proceeding with MLND (see “3.3.3 Safety of Lymphadenectomy for Stage I/II NSCLC”).

3.3.2 Stage II NSCLC:

No specific studies address this question, in part because stage II disease is a relatively uncommon presentation of NSCLC. Hilar lymph node disease presenting in the context of some mediastinal (N2) burden (stage IIIA), a far more common presentation of N1 disease, may benefit from MLND.80 Nonetheless, several studies have had large enough cohorts of patients with stage II disease to provide some insight.73,76 One randomized study suggests trends favoring overall survival, relapse-free survival, and reduction of distant and local recurrences for pN1-positive patients (analyzed jointly with one-station pN2-positive patients).73 The findings are supported by nonrandomized data demonstrating a threefold difference in local control favoring MLND for patients with stage II disease.77

3.3.3 Safety of Lymphadenectomy for Stage I/II NSCLC:

As might be expected, MLND dissection adds operative time, blood loss, and postoperative chest drainage when compared with sampling, but the effect is minor, of questionable clinical significance, and does not translate into increased morbidity. In the American College of Surgeons Oncology Group (ACOSOG) Z0030 trial, the two randomized arms had very similar morbidity profiles and were equivalently safe.81 Moreover, patients from both arms had similar hospital LOS. There is consistency among studies regarding safety, although rare examples of unique complications from MLND74,82 have been observed. As noted earlier, there are numerous variations of MLND. “Extended” mediastinal lymphadenectomy may be associated with excessive morbidity, although some claim that the procedure can be justified because of a survival benefit.63,83

3.3.4 Recommendation