Treatment of stage III non–small cell lung cancer (NSCLC) traditionally has involved combinations of chemotherapy, radiation, and surgical resection. Although the multimodality approach remains standard, only a fraction of patients with stage III lung cancer can undergo complete resection, and long-term prognosis remains poor. The PACIFIC trial generated significant enthusiasm when it demonstrated that the programmed death ligand-1 inhibitor, durvalumab, improved survival in patients with unresectable stage III NSCLC after completion of definitive concurrent chemoradiation. This article reviews the indications for traditional therapies in stage III NSCLC and highlights ongoing advances that have led to the incorporation of novel therapeutic agents.
The prognosis of stage III non–small cell lung cancer (NSCLC) remains poor, despite the important roles of surgical resection and concurrent chemoradiation.
Encouragingly, consolidation durvalumab after concurrent chemoradiotherapy improves survival in unresectable stage III NSCLC, suggesting a new role for immunotherapy in locally advanced disease.
Optimizing the use of checkpoint inhibitors and targeted therapies in stage III NSCLC remains an ongoing objective of clinical research.
Stage III non–small cell lung cancer (NSCLC) is a heterogeneous disease whose treatment incorporates multiple therapeutic modalities delivered in collaboration by thoracic surgeons, medical oncologists, and radiation oncologists. Although the long-term prognosis of patients with stage III disease receiving definitive therapy has traditionally been poor, immunotherapy is beginning to change the treatment paradigm. As a result, it is critical that providers not only understand the historical basis of traditional therapies but, also recognize the new treatment efforts and advances being made within the field. This article reviews the management of stage III NSCLC in detail, focusing on (1) the current role of surgery, chemotherapy, and radiation; (2) the newest indications for immunotherapy; and (3) the ongoing challenges and areas of unmet need.
Epidemiology of stage III non–small cell lung cancer
Current Population Trends
Despite significant therapeutic advancements over the past decade, lung cancer remains the leading cause of cancer-related death within the United States. At the same time, longitudinal data suggest that the landscape of lung cancer is changing, with declining incidence rates over the past several years, particularly among men. Within the context of these larger trends, the proportion of new lung cancer cases presenting as stage III has decreased steadily, from 28.6% in 1998 to 26.6% in 2006.
Despite curative intent treatment, the median 5-year relative survival rate of stage III NSCLC remains only 15.8%. However, a recent study demonstrated an early overall survival (OS) benefit for the programmed death ligand-1 (PD-L1) inhibitor, durvalumab, in patients with unresectable stage III disease. Although it remains to be seen if this benefit persists, long-term survival with immunotherapy in the metastatic setting has been encouraging and offers hope that similar improvement can be achieved in locally advanced disease.
Role of Staging in Prognosis
Implemented in January 2018, the eighth edition of the TNM staging system defines stage III lung cancer as locally advanced tumor and/or nodal disease without distant metastases. Although studies defining current standard of care for stage III NSCLC used older editions of the TNM staging system, understanding the newest staging criteria is necessary in order to recognize and prognosticate stage III disease within the current clinical context.
A comprehensive summary of the eighth edition of the TNM staging system is found in Lynn T. Tanoue’s article, “ Lung Cancer Staging ,” elsewhere in this issue. Here, it is important to at least highlight two changes made from the seventh to the eighth edition that have had a direct impact on the definition of stage III disease. First, greater emphasis is now placed on increasing tumor size as a poor prognostic feature. Specifically, tumors measuring greater than 5 cm have been upstaged from T2b to T3 and tumors measuring greater than 7 cm have been upstaged from T3 to T4. Second, tumors invading the diaphragm are now similarly upstaged from T3 to T4. These modifications are based on retrospective analysis of a large data set from the International Association for the Study of Lung Cancer, which showed that increasing tumor size and pathologic involvement of the diaphragm are each independently associated with a worse prognosis. By comparison, lymph node descriptors did not change in the newest edition of the TNM staging system, although the prognostic significance of increasing nodal involvement from ipsilateral hilar nodes to contralateral mediastinal nodes has been demonstrated clearly in prior studies.
The eighth edition of the TNM staging system also expands stage III NSCLC into 3 subgroups: IIIA, IIIB, and IIIC. When applied retrospectively to patients with NSCLC, each successive stage III subgroup is associated with incrementally shorter survival, ranging from 36% at 5 years for clinical stage IIIA disease to 13% for clinical stage IIIC disease. This suggests that the expanded classification groups better capture the heterogeneity of stage III NSCLC and should help oncologists define prognosis in newly diagnosed patients more accurately.
Demographic and Molecular Prognostic Factors
Demographic characteristics associated with worse prognosis in stage III NSCLC include older age. System level factors also have an impact on care, with 1 study demonstrating a lower risk of death for patients with stage III NSCLC who received definitive concurrent chemoradiotherapy (CCRT) at a high-volume center.
The prognostic effect of PD-L1 expression also has been a topic of recent interest, given the emerging use of immunotherapy in the treatment of stage III disease. Results to date, however, have been mixed, with some studies suggesting a positive association between PD-L1 expression and survival and others suggesting the opposite. Although the reason for these conflicting results is not known, methodological differences in PD-L1 staining may produce inconsistencies that do not necessarily reflect the underlying tumor biology. Alternatively, other potential immune biomarkers, such as tumor mutation burden, may be the true drivers of prognosis.
Although demographic and molecular characteristics can aid oncologists as they consider the prognosis of individual patients, they should not be used routinely to escalate or de-escalate treatment at this time. Determining precisely how these additional risk factors should be best considered in combination with staging and how their prognostic effects might be mitigated by novel therapies remain important areas of ongoing research.
Surgery as cure: the role of resection
Given its curative potential, complete surgical resection continues to play a key role in the management of stage III NSCLC. Complete resection (R0) is defined as removal of the primary tumor with negative microscopic margins, systematic nodal dissection, absence of extracapsular nodal extension, and absence of tumor in the highest sampled mediastinal lymph node. Compared with resection that leaves microscopic (R1) or macroscopic (R2) positive margins, complete resection results in significant improvements in 5-year OS (33% [R0] vs 19% [R1] vs 12% [R2]; P <.0001 for R0 vs R1).
The ability to achieve complete resection in stage III NSCLC may be limited by adverse tumor characteristics and/or patient-related factors. Adverse tumor characteristics that deem stage III disease unresectable include (1) large or invasive T4 tumors that cannot be resected to negative margins and (2) involvement of supraclavicular or contralateral hilar or mediastinal lymph nodes (N3) that extend beyond the procedural field established by the primary tumor. , Patient-related factors, such as impaired pulmonary function (eg, predicted postoperative forced expiratory volume in 1 second <60% or predicted postoperative diffusing capacity for carbon monoxide <60%) or cardiovascular comorbidities also may preclude safe surgical resection.
Compared with stage IIIB/C NSCLC, which generally is unresectable by virtue of T4N2 or TxN3 disease, stage IIIA T3N1 NSCLC with negative preoperative mediastinal lymph node biopsies often is anatomically amenable to complete resection. , , The presence of ipsilateral hilar nodal metastases (N1), however, is a risk factor for locoregional and distant recurrence. , In addition, there is considerable prognostic variability within the subgroup of N1 disease, with involvement of multiple N1 lymph nodes conferring a worse prognosis. Therefore, even when surgery is pursued for stage IIIA T3N1 NSCLC, it is combined with adjuvant chemotherapy with or without sequential postoperative radiation therapy (PORT).
Defining the resectability of stage IIIA(N2) disease, by comparison, is more complicated. As lymph node involvement progresses from ipsilateral hilar (N1) to mediastinal (N2) stations, 5-year OS rates decline from 29% to 16% ( P <.0001). As might be expected, involvement of multiple versus single N2 lymph nodes confers a worse prognosis. , In addition, some patients who have clinically negative N2 nodes based on preoperative endobronchial ultrasound–guided or mediastinoscopy-guided biopsy may be found to have pathologic N2 disease on frozen section or surgical pathology at the time of resection. These patients are defined as having occult or incidental pathologic N2 involvement and are felt to have a better prognosis with complete resection. , ,
In comparison, patients identified by preoperative biopsies to have clinical stage IIIA NSCLC with N2 involvement (stage IIIA[N2]) have not benefited from the addition of surgery in randomized trials ( Table 1 ). Although 3 of these trials closed early with incomplete accrual, a meta-analysis of all 5 trials demonstrated no effect of surgery on survival (hazard ratio [HR] 0.92; 95% CI, 0.82–1.04; P = .19). Despite these results, it continues to be debated whether subgroups of patients with clinical stage IIIA(N2) NSCLC might benefit from induction chemoradiation followed by complete resection. Partly, this stems from the randomized Intergroup 0139 trial, which found in subgroup analyses that lobectomy versus pneumonectomy and achievement of a complete versus incomplete pathologic nodal response after induction chemoradiation were associated with improved survival in the surgical arm. In an unplanned subgroup of patients matched by age, sex, performance status, and clinical T stage, investigators further found that median OS (33.6 versus 22.7 months, p = .002) was longer with lobectomy than chemotherapy plus radiation alone. Additionally, more patients in the surgery arm died without progression (18% vs 10%; P = .02), contributing to a significant improvement in progression-free survival with surgery (12.8 mo vs 10.5 mo; HR 0.77; 95% CI, 0.62–0.96; P = .017).
|Trial||Treatment Arms||Overall Survival|
|Shepherd et al, 1998 a||PV + surgery vs RT 60 Gy||OS: 18.7 mo vs 16.2 mo ( P = NS)|
|Johnstone et al, 2002 a||(M)VP + surgery vs (M)VP + RT 64 Gy||OS: 19.4 mo vs 17.4 mo ( P = .46)|
|van Meerbeeck et al, 2007||Platinum-based chemotherapy + surgery vs platinum-based chemotherapy + RT 60–62.5 Gy||OS: 16.4 mo vs 17.5 mo ( P = .60)|
|Albain et al, 2009||EP + RT 45 Gy + surgery vs EP + RT 61 Gy||OS: 23.6 mo vs 22.2 mo ( P = .24)|
|Eberhardt et al, 20,151||PC + cisplatin/vinorelbine/RT 45 Gy + surgery vs PC + cisplatin/vinorelbine/RT 65–71 Gy||OS: 44% vs 40% at 5 y ( P = .34)|
Higher mortality with pneumonectomy and improved survival with N2 clearance after induction therapy in patients undergoing resection also have been demonstrated in retrospective studies. However, retrospective studies and subgroup analyses from clinical trials, however, are hypothesis-generating. Although they suggest that patients achieving a complete pathologic N2 response could benefit from resection, there are no randomized trials in this subgroup of patients specifically. For these reasons, CCRT followed by consolidation durvalumab still is considered standard for most patients presenting with stage IIIA(N2) NSCLC. The National Comprehensive Cancer Network (NCCN) guidelines do suggest that induction therapy followed by lobectomy can be considered in select patients who present with involvement of a single, nonbulky (<3 cm), N2 lymph node. If this strategy is pursued, patients should be informed of the goals, risks, and benefits, and resection should be carried out by surgeons having expertise in this area.
The continued role of chemotherapy and radiation
Neoadjuvant and Adjuvant Chemotherapy in Resectable Disease
The risk of recurrence after complete resection of locally advanced NSCLC led to several randomized trials evaluating the efficacy of neoadjuvant or induction chemotherapy prior to surgical resection ( Table 2 ). Although results have not always been consistent, many of these studies, in particular those enrolling patients with stage IIIA disease alone, have demonstrated significantly improved survival with the addition of neoadjuvant chemotherapy. , In addition, a meta-analysis incorporating these trials demonstrated significantly improved survival with neoadjuvant chemotherapy, further supporting the bimodality approach in patients with completely resectable NSCLC (HR 0.87; 95% CI, 0.78–0.96; P = .007).
|Trial||Stage||Treatment Arms||Overall Survival||Complete Resection|
|Pass et al, 1992||IIIA(N2)||EP + surgery ± RT vs surgery + RT||28.7 mo vs 15.6 mo ( P = .095)||—|
|Rosell et al, 1994||IIIA||MIP + surgery vs surgery alone||26 mo vs 8 mo ( P <.001)||85% vs 90%|
|Roth et al, 1994||IIIA||CEP + surgery vs surgery alone||64 mo vs 11 mo ( P = .008)||39% vs 31%|
|Depierre et al, 2002||I–IIIA||MIP + surgery ± RT vs surgery ± RT alone||37 mo vs 26 mo ( P = .15)||92% vs 86%|
|Nagai et al, 2003||IIIA(N2)||Cisplatin/vindesine + surgery vs surgery alone||17 mo vs 16 mo ( P = .5274)||65% vs 77%|
|Mattson et al, 2003||IIIA/B||Docetaxel + surgery or RT vs surgery or RT alone||14.8 mo vs 12.6 mo ( P = NS)||77% vs 76%|
|Gilligan et al, 2007||I–III||Platinum-based chemotherapy + surgery vs surgery alone||54 mo vs 55 mo ( P = .86)||81% vs 79%|
|Pisters et al, 2010||IB–IIIA||PC + surgery vs surgery alone||62 mo vs 41 mo ( P = .11)||84% vs 87%|
|Scagloiotti et al, 2012||IB–IIIA||GC + surgery vs surgery alone||7.8 y vs 4.8 y ( P = .04)||88% vs 84%|
The optimal sequence of surgery and chemotherapy, however, remains a topic of debate. Among the arguments in favor of neoadjuvant chemotherapy is its potential to downstage disease, thereby improving the feasibility of complete resection and reducing the need for more invasive surgeries. As shown in Table 2 , however, complete resection rates generally were similar in patients receiving neoadjuvant chemotherapy versus upfront surgery in randomized trials. In addition, adjuvant chemotherapy, by comparison, avoids treatment-induced complications that could delay resection. As a result, several randomized clinical trials have been conducted to evaluate the efficacy of adjuvant platinum-based chemotherapy regimens in stage III NSCLC. A pooled analysis by the Lung Adjuvant Cisplatin Evaluation (LACE) group, in particular, incorporated 5 trials of cisplatin-based chemotherapy in patients with completely resected stages IA–III NSCLC. Among the cohort of 4584 patients, adjuvant chemotherapy was associated with a 5.4% absolute benefit in survival at 5 years (HR 0.89; 95% CI, 0.82–0.96; P = .005). This benefit was pronounced particularly among patients with stage II or stage III disease (stage III: HR 0.83; 95% CI, 0.72–0.94). Based on these results, adjuvant cisplatin-based chemotherapy after complete resection of stage IIIA NSCLC is now recommended by clinical practice guidelines. Neoadjuvant chemotherapy remains an alternative if tumor reduction prior to surgery is necessary.
Postoperative Radiation Therapy
Compared with adjuvant chemotherapy, the benefit of sequential PORT in patients with completely resected stage III NSCLC has been less clear. In an early randomized trial of patients with stages I–III NSCLC, PORT was associated with a lower OS rate at 5 years compared with observation (30% vs 43; relative risk 1.33; 95% CI, 1.11–1.59; P = .002), although this was driven primarily by nonmalignant intercurrent deaths, such as cardiorespiratory failure. Subgroup analyses further suggested that the harmful effect of PORT was most significant in patients with stages I–II disease, whereas the effect on patients with stage III disease did not reach statistical significance (relative risk 1.14; 95% CI, 0.89–1.46; P = .30). Similarly, data from large retrospective cohorts have suggested that PORT may improve survival in patients with pathologic N2 involvement after surgical resection. ,
In recent years, radiation techniques have significantly improved to limit off-target toxicity, and the use of adjuvant chemotherapy has become standard. As a result, there is a strong rationale supporting randomized trials of PORT in the modern setting. We are currently awaiting the results of the Lung ART trial, a phase 3 randomized study of PORT versus observation in patients with completely resected stage III NSCLC found to have pathologic N2 disease at the time of surgery. In the meantime, a subset analysis was conducted of patients enrolled in the ANITA trial, a randomized phase 3 study evaluating cisplatin plus vinorelbine after complete resection of stages IB–IIIA NSCLC. Receipt of PORT, at 45 Gy to 60 Gy, off-protocol at the discretion of the primary treatment center was associated with worse OS (HR 1.34; 95% CI, 1.10–1.63; P = .003). Those patients with pathologic N2 involvement, however, had a longer median OS with chemotherapy plus PORT compared with chemotherapy alone (47.4 mo vs 23.8 mo; P = not reported). Based on these results, the NCCN guidelines allow consideration of PORT in patients with clinical stages I–III T1-3N0-1 NSCLC who are incidentally found to have occult pathologic N2 involvement at the time of complete surgical resection. Due to its potential risks, PORT currently is not recommended for lesser stage disease.
Concurrent Chemoradiation in Unresectable Disease
In patients with unresectable stage III NSCLC, the combination of chemotherapy plus radiotherapy is associated with a better response and longer survival compared with radiation alone. The more recent development in the past decade, however, has been the confirmation that CCRT is more effective than the sequential approach. In the Radiation Therapy Oncology Group (RTOG) 9410 trial, the concurrent use of cisplatin plus vinblastine with thoracic radiotherapy was associated with a median survival of 17.0 months compared with 14.6 months for the same therapies given sequentially. Nonetheless, the 5-year OS rate of the concurrent therapy arm still was only 16%, suggesting a need for additional progress.
One attempt at improving survival further has been the use of consolidation chemotherapy after completion of CCRT. Initial promise for this approach was suggested by the phase 2 SWOG S9504 trial, which showed that consolidation docetaxel after definitive CCRT in patients with stage IIIB NSCLC achieved a 3-year survival rate of 37%, comparing favorably to historical controls. , Unfortunately, a phase 3 trial by the Hoosier Oncology Group evaluating consolidation docetaxel versus observation after definitive CCRT showed no benefit in survival (21.2 mo vs 23.2 mo; P = .883) with the increased cost of toxicity. A recent meta-analysis evaluating consolidation chemotherapy after CCRT similarly found that OS was not improved (odds ratio 1.24; 95% CI, 0.89–1.72; P = .21) whereas rates of infection ( P = .02), pneumonitis ( P = .003), and treatment-related death ( P = .04) were increased.
Another area of controversy pertains to the optimal chemotherapy regimen to be used with concurrent radiation therapy. Multiple platinum-based regimens are currently are approved, with cisplatin plus etoposide and carboplatin plus paclitaxel being most commonly used within the United States. , Real-world studies, however, suggest that outcomes with several different platinum-doublet regimens may be similar. In addition, platinum chemotherapy plus pemetrexed now also is approved in combination with radiation for patients with unresectable stage III nonsquamous NSCLC. The PROCLAIM trial demonstrated no difference in survival between patients receiving cisplatin plus pemetrexed versus cisplatin plus etoposide (26.8 mo vs 25.0 mo; HR 0.98; 95% CI, 0.79–1.20; P = .831), although rates of grade 3 to grade 4 adverse events were significantly lower in the pemetrexed arm (64.0% vs 76.8%; P = .001).
The emergence of immunotherapy in stage III non–small cell lung cancer
Despite optimizing the use of chemoradiotherapy, the prognosis of stage III NSCLC remains poor, and attempts to incorporate novel therapies such as cetuximab and bevacizumab have been unsuccessful. , Fortunately, the success of immunotherapy in metastatic disease is now beginning to trickle down to the locally advanced setting. Two recent trials—PACIFIC and KEYNOTE-042—that have led to the approval of checkpoint inhibitors in specific subsets of patients with stage III NSCLC are discussed.
The PACIFIC study was a landmark randomized, phase 3 trial evaluating durvalumab, a monoclonal IgG1 PD-L1 inhibitor, in patients with unresectable stage IIIA/B NSCLC. , Following platinum-doublet chemotherapy plus concurrent radiation at 54 Gy to 66 Gy, patients received either consolidation durvalumab, 10 mg/kg, or placebo, every 2 weeks for up to 12 months. The randomized cohort of 713 patients consisted predominantly of individuals who were older (median age 64), male (70%), and white (70%) and had a positive smoking history (90%). Baseline PD-L1 expression prior to CCRT was greater than or equal to 25% in 22.3% of patients and less than 25% in 41.0% of patients. At the initial interim analysis, the study met the first of its 2 coprimary endpoints, with a median progression-free survival lasting 16.8 months versus 5.6 months for patients receiving durvalumab versus placebo (HR 0.52; 95% CI, 0.42–0.65; P <.001). This survival benefit extended to all patients regardless of PD-L1 expression greater than or equal to 25% or less than 25%. In an updated analysis, the second primary endpoint of OS also was significantly prolonged in patients receiving durvalumab, with a 24-month OS rate of 66.3% versus 55.6% (HR 0.68; CI 0.47–0.997; P = .0025).
In terms of safety, serious adverse events (28.6% vs 22.6%), drug discontinuation (15.4% vs 9.8%), and immune-related adverse events (24.2% vs 8.1%) were more common with durvalumab than placebo. Pneumonitis and radiation pneumonitis (33.9% vs 24.8%) also were more common, which is notable given that symptomatic pneumonitis already occurs in up to 30% of patients after CCRT alone. , Although pneumonitis of grade 3 to 4 severity was rare in both treatment arms (3.4% vs. 2.6%), patients with risk factors such as autoimmune disease, grade 2 or greater pneumonitis secondary to prior chemoradiotherapy, or poor performance status were excluded. As a result, in real-world patients who may not match the relatively fit, low-risk cohort of the PACIFIC trial, the potential for durvalumab to exacerbate pneumonitis after radiation should be considered carefully.
Based on these results from the PACIFIC trial, consolidation durvalumab after definitive chemotherapy is now standard of care in unresectable stage III NSCLC. It will be important, however, going forward to clarify whether the survival benefit of durvalumab extends to patients whose tumors do not express PD-L1. In post hoc analyses, the HR for death for durvalumab versus placebo in patients whose tumors expressed PD-L1 less than 1% was 1.36 (95% CI, 0.79–2.34). Although caution is warranted when interpreting post hoc analyses, these results at the very least suggest that additional studies are needed to better characterize the predictive role of PD-L1 expression. Until more definitive results are available, guidelines recommend consolidation durvalumab after definitive CCRT for all patients with unresectable stage III NSCLC based on the survival benefit in the intention-to-treat population.
The KEYNOTE-042 study was a randomized phase 3 trial in which patients with locally advanced or metastatic NSCLC with PD-L1 expression greater than or equal to 1% patients were randomized to either pembrolizumab or chemotherapy with carboplatin (area under the curve 5 to 6) plus paclitaxel (200 mg/m 2 ) or pemetrexed (500 mg/m 2 every 3 weeks). Among patients with PD-L1 expression greater than or equal to 1%, median OS lasted 16.7 versus 12.1 months (HR 0.81; 95% CI, 0.71–0.93; P = .0018) for pembrolizumab versus chemotherapy. Subgroup analyses showed a trend toward benefit in patients presenting with locally advanced NSCLC in particular (HR 0.74; 95% CI, 0.49–1.13), although the absolute number of patients in this subgroup was small (n = 160). Based on these data, pembrolizumab was approved in April 2019 for patients with untreated, EGFR and ALK wild-type, stage III NSCLC with PD-L1 expression greater than or equal to 1% who are not candidates for surgery or chemoradiation.
The significance of this approval is that it adds an additional treatment option for a subgroup of patients with stage III NSCLC who may not qualify for definitive therapy. The proportion of patients with stage IIIA disease who did not receive any treatment in clinical practice increased significantly by 0.21% ( P = .003) from 1998 to 2012, with factors, such as comorbidity, contributing. In the KEYNOTE-042 trial, pembrolizumab was associated with more immune-mediated adverse events and infusion reactions (28% vs 7%) but fewer grade 3 to grade 5 adverse events overall (18% vs 41%) compared with chemotherapy. Therefore, for patients with stage III NSCLC who are unable to undergo surgery or receive chemoradiation, pembrolizumab may have a better toxicity profile and thereby represent a favorable treatment alternative, especially compared with best supportive care alone.
Following on the heels of the PACIFIC trial, 2 early phase trials have evaluated neoadjuvant immunotherapy in patients with surgically resectable NSCLC. In a study of 22 patients with stages I–IIIA disease (stage IIIA, 33%), 2 doses of neoadjuvant nivolumab, 3 mg/kg, every 2 weeks were administered prior to resection. Major pathologic response, defined as less than or equal to 10% viable tumor cells in the resected primary tumor, was achieved in 9 of 20 evaluable patients (45%) and was associated with a higher mean tumor mutational burden. On the other hand, it was not consistently associated with radiographic response, suggesting that nivolumab-mediated inflammation at the primary tumor site may obscure radiographic interpretation of disease. In a second study, neoadjuvant ipilimumab, a CTLA-4 inhibitor, was administered at 10 mg/kg in combination with 2 cycles of platinum-doublet chemotherapy prior to resection in 13 patients with stages II–IIIA NSCLC (stage IIIA 77%). Downstaging of mediastinal lymph nodes occurred in 4 of 13 patients (31%), and rates of adverse perioperative complications were similar to those of historical controls. Together, these early trials suggest that checkpoint inhibitors may be safe and clinically active in the perioperative setting. Although comparative survival data are lacking, clinical trials are ongoing that hopefully should clarify the expanding role of immunotherapy in the management of stage III NSCLC ( Table 3 ).