1 INTRODUCTION

The 19th Lung Cancer Summit was successfully held in Guangzhou by the Chinese Association of Lung Cancer, Guangdong Association of Clinical Trials (GACT)/(CTONG) on March 5, 2022. After in-depth communications and discussions, experts on lung cancer clinical research, transformational research, and basic research achieved consensuses on the multidisciplinary comprehensive treatment and biomarkers for resectable and unresectable stage III Non-small cell lung cancer (NSCLC) at the Summit.

NSCLC accounts for 80%–85% of all lung cancer patients, of which one-third of patients are initially diagnosed as stage III [1, 2]. The treatment for stage III NSCLC is the most complex and requires the dedication of a multiple disciplinary team. Some stage III patients can be treated by surgery but need comprehensive perioperative therapy. For unresectable stage III patients, the major treatments include radiotherapy, chemotherapy, and immunotherapy [3]. With the advent of precision therapy, targeted therapy and immunotherapy provide more novel options for patients with stage III NSCLC; however, they raise new questions, such as “What type of patients are suitable for surgery?”, “Do we need to redefine ‘resectable’?”, “How can we optimally apply targeted therapy and immunotherapy to perioperative care?”, “Can we choose perioperative treatment based on biomarkers?”, “What type of patients should receive radiotherapy or chemotherapy?”, “What is the best combination of chemoradiotherapy and immunotherapy?”, and “What biomarkers can guide the immunotherapy or targeted therapy for stage III lung cancer?” Focusing on these questions, the Summit provided six expert consensuses after in-depth discussions and communications by combining the current diagnosis and treatment strategies for stage III NSCLC in China and recent worldwide research data. All the panel members contributed to revising and finalizing this document.

The level of evidence and strength of recommendations for this consensus are in Table 1.

2 CONSENSUS 1: REDEFINING THE DEFINITION OF “RESECTABILITY”

Stage III NSCLC encompasses a group of diseases with high heterogeneity. According to the eighth edition of The American Joint Committee on Cancer/Union for International Cancer Control Tumor–Node–Metastasis staging system [4], stage III NSCLC can be further staged as IIIA, IIIB, or IIIC with 5-year survival rates of 36%, 26%, and 13%, respectively [5]. Only some stage IIIA and a few stage IIIB NSCLC patients have resectable cases, whereas most stage IIIB and stage IIIC NSCLC patients have unresectable cases and require chemotherapy-based treatments. As stage III NSCLC patients are highly heterogeneous clinically and pathologically, their clinical treatment remains controversial. However, such patients are still curable and therefore should actively receive comprehensive treatments. It is imperative for all stage III NSCLC patients to receive consultation with a multiple disciplinary team, which allows the development of a proposal for clinical diagnosis and treatment.

Stage III NSCLC can be categorized as completely resectable, unresectable, or potentially resectable according to the possibility of complete surgical resection. There are three criteria for resectable stage III lung cancer, including (1) primary tumor and hilar and mediastinal lymph nodes that can be completely resected (R0 resection), but a pneumonectomy is given cautiously; (2) there is no metastasis toward N3 lymph nodes; and (3) the expected perioperative mortality rate is less than 5%. Resectable stage III NSCLC mainly includes stage IIIA N0–1, N2 with a single mediastinal lymph node metastasis and a short diameter <2 cm and some T4N1 (with solitary carcinomatous nodules in different lobes of the same lung). In addition, unresectable stage III NSCLC encompasses some phase IIIA/IIIB and all phase IIIC, usually including a single N2 mediastinal lymph node with a short diameter ≥3 cm, multiple N2 lymph nodes with or without fused clusters (lymph node short diameter ≥2 cm on a computed tomography scan), T4, and all N3 with invasion into the esophagus, heart, and great vessels. Moreover, consultation with the multiple disciplinary team is recommended to make the decision regarding surgical resection for some stage III NSCLC with a high surgical difficulty, such as stage IIIA with a single N2 mediastinal lymph node with a short diameter <3 cm, a superior sulcus tumor, and a T3 or T4 central tumor [6]. Potential resectability is defined as resectable after neoadjuvant therapy. Furthermore, the number of patients for whom resection is possible has been increased by the improved surgical techniques and the advancement of neoadjuvant therapy.

Neoadjuvant therapy consists of chemotherapy, radiotherapy, targeted therapy, and an immune checkpoint blockade-based regimen. In recent years, neoadjuvant therapy has shown good perioperative efficacy in resectable NSCLC patients with negative epidermal growth factor receptor (EGFR) and negative anaplastic lymphoma kinase expression. As shown in the CheckMate-816 study with stage IB (≥4 cm) to stage IIIA resectable NSCLC patients, nivolumab combined with chemotherapy significantly improved the pathologic complete remission rate compared to chemotherapy alone (24% vs. 2.2%); the pathologic complete remission rates of stage IIIA resectable NSCLC were 23% and 0.9%, respectively [7]. Moreover, targeted neoadjuvant therapy is promising for the treatment of resectable EGFR-mutated NSCLC. In the CTONG1103 trial led by Professors Yi-Long Wu and Wenzhao Zhong, erlotinib significantly improved the major pathological response in stage III N2-positive NSCLC patients with EGFR-sensitive mutations compared with gemcitabine and cisplatin combined therapy (9.7% vs. 0%) [8]. In addition, at the recent European Lung Cancer Conference, the prospective multicenter single-arm phase II NEOS trial updated that neoadjuvant osimertinib induced an objective response rate of 71.1% and an R0 resection rate of 94% in patients with resectable stage II–IIIB (T3-4N2) and EGFR-mutated lung adenocarcinoma [9]. These findings suggest that stage III NSCLC may benefit from immunotherapy or targeted therapy. Currently, multiple studies analyzing targeted therapy or immunotherapy are ongoing. For instance, the NeoADAURA study (NCT04351555), a global multicenter stage III trial, aimed to evaluate the efficacy and safety of osimertinib alone or in combination with chemotherapy as well as chemotherapy alone as neoadjuvant therapy in patients with resectable EGFR-mutated stage II–IIIB NSCLC. This study verified the efficiency of neoadjuvant therapy [10]. Therefore, neoadjuvant therapy should be recommended to these types of NSCLC patients.

3 CONSENSUS 2: ADJUVANT TARGETED THERAPY FOR STAGE III NSCLC PATIENTS

Since the U.S. Food and Drug Administration (FDA) approved osimertinib for the postoperative adjuvant treatment in early stage NSCLC patients with the EGFR mutation in December 2020, osimertinib has gained approval as a postoperative adjuvant therapy in more than 50 countries and regions globally. In June 2021, China's National Medical Products Administration approved icotinib for postoperative adjuvant therapy of stage II–IIIA NSCLC patients with EGFR-sensitive mutations. In October 2021, the FDA approved atezolizumab for adjuvant therapy in post-surgery or platinum-treated stage II–IIIA NSCLC patients with ≥1% PD-L1 positivity in tumor cells. Moreover, SP263 was approved as a companion diagnostic assay. As the postoperative adjuvant therapy for NSCLC has evolved from chemotherapy to precision therapy, routine detection of EGFR and PD-L1 should be recommended for stage III NSCLC patients prior to treatment, allowing the possibility for precision therapy.

The ADJUVANT/CTONG1104 trial [11, 12] showed that adjuvant gefitinib therapy significantly prolonged the disease-free survival (DFS) of patients with EGFR-mutated stage II–IIIA NSCLC compared to standard platinum-based chemotherapy consisting of vinorelbine and cisplatin (median DFS: 30.8 vs. 19.8 months); and postoperative gefitinib treatment for 2 years reduced the risk of relapse or death by 44% (hazard ratio [HR] = 0.56, 95% confidence interval [CI] = 0.40–0.79, p = 0.001). Subsequently, clinical trials, such as EVAN [13], SELECT [14], and EVIDENCE [15], showed that different adjuvant drugs among the first generation of EGFR-TKIs resulted in similar DFS times. For example, Liang et al. [16] compared the efficacy and safety of the first-generation EGFR-TKIs (gefitinib, erlotinib, and icotinib) in stage II–III patients after definitive operation, and the median DFS times were 36.1 months, 42.8 months, and 32.5 months for gefitinib, erlotinib, and icotinib, respectively, without statistical significance (p = 0.223). In contrast to the head-to-head comparisons of the first-generation EGFR-TKIs and chemotherapy, ADAURA, a phase 3, global, multicenter, randomized controlled trial [17], enrolled patients into adjuvant chemotherapy based on the doctor's judgement and expanded treatment to patients with stage IB–IIIA NSCLC. In patients with stage II–IIIA EGFR-mutant NSCLC, the main endpoint of the study was that adjuvant osimertinib treatment significantly reduced the 3-year rate of relapse or death by 83% compared with the placebo (HR = 0.17, 99.06% CI = 0.11–0.26; 2-year DFS rate: 90% vs. 44%). In stage IB–IIIA patients, the key secondary endpoint was that adjuvant therapy with osimertinib significantly reduced the rate of relapse or death by 80% (HR = 0.20, 99.12% CI = 0.14–0.30; 2-year DFS rate: 89% vs. 52%). Notably, stage III NSCLC patients have a high risk of brain metastases. The ADJUVANT trial [18] showed that the recurrence rate of intracranial metastases in the gefitinib group was greater than that in the vinorelbine plus cisplatin group (27.4% vs. 24.1%). In addition, Liang et al. [16] demonstrated that there was no significant difference in the cause of treatment failure of the first-generation EGFR-TKIs. The treatment failure caused by intracranial metastasis was 6.1%, 7.5%, and 3.9% for gefitinib, erlotinib, and icotinib, respectively. These findings may be related to poor penetration of the first-generation EGFR-TKI drugs through the blood–brain barrier. As a third-generation EGFR-TKI, osimertinib can easily pass through the blood–brain barrier. The ADAURA trial [17] showed that adjuvant osimertinib therapy significantly reduced the rate of metastases toward the central nervous system (including brain metastases) or death by 82% compared with placebo (HR = 0.18, 95% CI = 0.10–0.33). Thus, for stage III NSCLC postoperative patients with EGFR-sensitive mutations, adjuvant therapy with osimertinib is preferably recommended.

ICAN/CTONG1804, a real-world noninterventional trial with 486 patients from 26 Chinese research institutes found that adjuvant chemotherapy had no apparent benefit for patients with EGFR mutations [19]. Moreover, the ADAURA trial enrolled patients based on physician assessment and the patient's choice of adjuvant chemotherapy, and the subgroup analysis showed [20] that adjuvant osimertinib therapy significantly improved the DFS compared to placebo, regardless of the history of adjuvant chemotherapy use. Compared to placebo, adjuvant osimertinib reduced the risk of relapse or death by 84% (HR = 0.16, 95% CI = 0.10–0.26) or 77% (HR = 0.23, 95% CI = 0.13–0.40) in patients with or without prior adjuvant chemotherapy, respectively [21]. For EGFR-mutated stage IB–IIIA NSCLC patients, adjuvant osimertinib can provide a significant benefit after complete tumor resection, regardless of previous adjuvant chemotherapy, implying that adjuvant chemotherapy is dispensable.

The decision-making of adjuvant therapy should focus on the exploration of a more precise treatment model, that is, precisely screening the patients who will benefit and exploring the most appropriate treatment duration to improve the cure rate of resectable NSCLC [22]. Because EGFR-mutated NSCLC at the early stage has a high molecular heterogeneity, biomarker prediction is needed to determine the individualized adjuvant therapy. Wu et al. [23] extracted genomic DNA from 171 tumor tissues and performed next-generation sequencing for 422 cancer-related genes from the ADJUVANT/CTONG1104 trial. They discovered five key genes that could predict the efficacy of adjuvant therapy in those with EGFR mutations, including missense mutations in exons 4 and 5 of the tumor protein P53 gene (TP53), RB1 mutation, and amplification of the NK2 homeobox 1, cyclin-dependent kinase 4, and MYC genes. Based on this discovery, they further established the Multigene Index to Evaluate the Relative benefit of Various Adjuvant therapies (MINERVA) scoring model. Analysis of the MINERVA scoring model showed that patients with the RB1 mutation with or without missense mutation in exons 4 and 5 of TP53 or MYC amplification had a longer DFS upon adjuvant chemotherapy than those with adjuvant gefitinib therapy (median DFS: 34.2 vs. 19.3 months), suggesting that patients with both EGFR and RB1 mutations may benefit from adjuvant chemotherapy rather than gefitinib; therefore, adjuvant chemotherapy is preferably recommended.

4 CONSENSUS 3: ADJUVANT IMMUNOTHERAPY FOR WILD TYPE STAGE III NSCLC

Impower-010 [24], the world's first phase III, randomized, controlled trial of postoperative adjuvant immunosuppressive agents on early stage NSCLC patients, recruited 1005 NSCLC patients at stage IB (≥4 cm) to IIIA who received platinum-containing chemotherapy. Adjuvant atezolizumab therapy was randomly administered to these patients and reduced the risk of disease progression by 34% in stage II–IIIA patients with PD-L1 ≥ 1% compared with the optimal supportive intervention (HR = 0.66, 95% CI = 0.50–0.88, p = 0.0039). In addition, the KEYNOTE-091 trial [25] on stage IB (≥4 cm) to IIIA NSCLC patients who underwent complete resection compared adjuvant pembrolizumab therapy and placebo. Interim analysis showed that 85.8% of patients who previously received adjuvant chemotherapy tended to benefit from adjuvant pembrolizumab therapy compared to those who received placebo (HR = 0.73, 95% CI = 0.60–0.89), especially for the subgroup with PD-L1 expression of 1%–49%. Thus, adjuvant immunotherapy is recommended after completion of postoperative chemotherapy.

Further subgroup analysis of the IMpower-010 trial [24] showed that stage II–IIIA patients with PD-L1 expression of 1%–49% had a 13% reduced risk of relapse or death upon adjuvant atezolizumab therapy compared with the optimal supportive intervention (HR = 0.87, 95% CI = 0.60–1.26). However, adjuvant atezolizumab therapy reduced the risk of relapse or death by 57% in patients with PD-L1≥50% (HR = 0.43, 95% CI = 0.27–0.68), suggesting that the benefit for the total population mainly lied in the subgroup with high PD-L1 expression. Although atezolizumab is approved by the FDA to treat stage II–IIIA NSCLC patients with PD-L1 ≥ 1%, it is preferably recommended to patients with PD-L1 > 50% in clinical practice.

For patients unwilling to receive adjuvant chemotherapy, can we directly provide adjuvant immunotherapy? The IMpower-110 trial [26] showed that atezolizumab prolonged the survival of patients with PD-L1≥50% compared to chemotherapy (20.2 months vs. 13.1 months, HR = 0.76, 95% CI = 0.54–1.09, p = 0.01), suggesting that adjuvant atezolizumab therapy is beneficial for patients with PD-L1 ≥ 50%. Therefore, stage III NSCLC patients who refuse adjuvant chemotherapy but have PD-L1 > 50% can consider receiving adjuvant immunotherapy.

5 CONSENSUS 4: SELECTION OF NEOADJUVANT OR ADJUVANT THERAPY FOR STAGE III NSCLC PATIENTS

ADAURA and numerous other studies have demonstrated that adjuvant EGFR-TKI targeted therapy could improve the DFS of resectable EGFR-mutated NSCLC patients. Meanwhile, neoadjuvant targeted therapy is also suggested to be beneficial. The CTONG1103 trial [8] found that neoadjuvant erlotinib therapy significantly improved the median PFS compared to gemcitabine and cisplatin-combined therapy (21.5 vs. 11.4 months, HR = 0.39, 95% CI = 0.23–0.67, p < 0.001). The NEOS trial [9] also showed that stage II–IIIB resectable lung adenocarcinoma patients with EGFR mutations acquired an objective response rate of 71.1% and a disease control rate of 100% after neoadjuvant therapy with osimertinib for 6 weeks. In 32 patients who underwent surgery, 94% achieved R0 resection. Of the patients evaluable pathologically, 11% (3/28) achieved a major pathological response, including 1 case (4%) with a pathologic complete remission rate and 46% (13/28) who exhibited pathological remission ≥50%. The phase III NeoADAURA trial [10] (NCT04351555) aims to compare osimertinib alone or in combination with chemotherapy and standard chemotherapy for the neoadjuvant treatment of EGFR-mutated NSCLC, but this trial is still ongoing.

The CheckMate-816 trial [7] on stage IB (≥4 cm) to IIIA resectable NSCLC patients showed that neoadjuvant therapy combining nivolumab and chemotherapy could significantly improve the pathologic complete remission rate compared with traditional chemotherapy (24.0% vs. 2.2%, odds ratio = 13.94, p < 0.0001). This study updated that nivolumab combined with chemotherapy significantly improved the event-free survival (EFS) of patients compared to chemotherapy alone at the 2022 American Association for Cancer Research conference (median EFS: 31.6 vs. 20.8 months, HR = 0.63, 97.38% CI = 0.43–0.91, p = 0.0052). Therefore, neoadjuvant immunotherapy may be considered for the treatment of stage III NSCLC patients without driver genes.

There are three categories of current clinical trials of perioperative immunotherapy for lung cancer, including single neoadjuvant clinical studies, single adjuvant clinical studies, and neoadjuvant-surgical-adjuvant clinical studies. Most of them are focused on perioperative immunochemotherapy, such as CheckMate-77T, KEYNOTE-671, IMpower-030, AEGEAN, and so on. Furthermore, perioperative immunochemotherapy is the future direction for the treatment of driver gene-negative stage III NSCLC patients.

6 CONSENSUS 5: TREATMENT OPTIONS FOR UNRESECTABLE STAGE III NSCLC PATIENTS

The PACIFIC trial showed that dorvalizumab consolidation immunotherapy for 1 year following concurrent chemoradiotherapy significantly improved the survival of stage III NSCLC unresectable patients as indicated by the significantly longer PFS than that in the placebo group (median PFS: 16.8 vs. 5.6 months, HR = 0.52, p < 0.001) [27]. The latest 5-year follow-up data showed [28, 29] that durvalumab improved the median overall survival by 18.4 months (47.5 vs. 29.1 months, HR = 0.72), 42.9% of the patients had an overall survival of more than 5 years, and 33.1% of the patients presented with a PFS of more than 5 years.

Although concurrent chemoradiotherapy had been increasingly accepted in the real world before the publication of the PACIFIC study, 40%–70% of patients still received sequential chemoradiotherapy in China as well as in some countries and regions in Europe [30, 31]. The GEMSTONE-301 trial [32, 33] of stage III NSCLC patients who made no progress after concurrent or sequential chemoradiotherapy showed that sugemalimab therapy for up to 2 years significantly improved the PFS compared to placebo (median PFS: 9.0 vs. 5.8 months, HR = 0.64, 95% CI = 0.48–0.85, p = 0.0026), which was indifferent in both the concurrent and sequential chemoradiotherapy subgroups. This study further confirmed the efficacy of anti-PD-L1 monoclonal antibody consolidation therapy following concurrent chemoradiotherapy and provided the option of consolidation immunotherapy for patients after sequential chemoradiotherapy. This study also provided data regarding the safety and efficacy of consolidation immunotherapy in stage III NSCLC patients after chemoradiotherapy in China via large, randomized, double-blinded, phase III clinical research. PACIFIC-6, a phase II clinical trial, recruited stage III NSCLC patients who failed to progress after sequential chemoradiotherapy. Among these patients, 114 cases with a performance status of 0–1 were treated with durvalumab (1500 mg once every 4 weeks for up to 2 years), and their median PFS was 13.1 months [34]. The 4-week durvalumab regimen (1500 mg once every 4 weeks) has been approved by the FDA and European Medicines Agency for the treatment of patients with a body weight of more than 30 kg [35]. This regimen is convenient due to the reduced numbers of injections and visits. In the real-world PACIFIC-R trial, 1399 patients with stage III NSCLC who failed to progress after concurrent or sequential chemoradiotherapy received durvalumab and achieved a median PFS of 21.3 months [36]. Moreover, the phase III PACIFIC-5 trial will further confirm the efficacy and safety of durvalumab after concurrent and sequential chemoradiotherapy.

Consolidation immunotherapy post chemoradiotherapy has succeeded in the PACIFIC and GEMSTONE-301 trials. Nevertheless, researchers still attempt to explore the new combination modality of immunotherapy and chemoradiotherapy in stage III unresectable NSCLC. Can immune checkpoint inhibitors be administered prior to or concurrently with chemoradiotherapy? Such questions should be answered when investigating new combination modalities by considering the safety. There are several common exploratory modalities, that is, (1) chemoradiotherapy with concurrent immunosuppressant therapy followed by immunosuppressant consolidation therapy [37]; (2) after induction chemotherapy, chemoradiotherapy with concurrent immunosuppressant therapy followed by immunosuppressant consolidation therapy [38]; (3) immunosuppressant induction therapy followed by chemoradiotherapy [39]; and (4) after induction immunosuppressant therapy or induction chemotherapy, surgical intervention, or chemoradiotherapy suggested by multiple disciplinary team consultation. Patients with stage III unresectable NSCLC should also participate in the appropriate clinical trials.

7 CONSENSUS 6: TRANSLATIONAL RESEARCH ON MOLECULAR RESIDUAL DISEASE (MRD)

Tissue, blood, and gut microbiota can be sampled for the detection of biomarkers, which are widely used predictors of efficacy and prognosis in clinical practice. ctDNA, as a valuable biomarker, can be used to evaluate the possible benefit that postoperative patients with MRD may gain from adjuvant therapy [40, 41]. The phase II LCMC3 trial evaluating atezolizumab as a neoadjuvant therapy in untreated resectable stage IB–IIIB NSCLC patients found that MRD-negative patients by ctDNA analysis had a longer 2-year DFS rate after atezolizumab treatment or surgery than MRD-positive patients (after atezolizumab treatment, HR = 0.34, 95% CI = 0.1–1.17; after surgery, HR = 0.25, 95% CI = 0.06–1.01) [42]. Therefore, ctDNA analysis of MRD can serve as a prognostic factor. Patients testing positive for MRD by ctDNA analysis after definitive treatment have a poorer prognosis than those testing negative.

The phase III Impower-010 trial evaluating adjuvant atezolizumab therapy in stage II–IIIA NSCLC-resected patients after adjuvant chemoradiotherapy showed that MRD-positive patients by ctDNA analysis could benefit from adjuvant atezolizumab therapy (HR = 0.61, 95% CI = 0.29–0.94) [43]. Accordingly, postoperative adjuvant therapy is recommended for MRD-positive patients by ctDNA analysis after definitive treatment. For MRD-negative patients, there is insufficient evidence to support the postoperative benefit of adjuvant therapy.

A retrospective study assessing consolidation immunotherapy in stage IIB–IIIB NSCLC unresectable patients after definitive chemoradiotherapy showed that MRD-positive patients by ctDNA analysis exhibited a better outcome after receiving consolidation immunotherapy (p = 0.04), whereas MRD-negative patients had a good prognosis with or without consolidation immunotherapy. This retrospective study should be further verified due to the relatively small sample size, but the presence of MRD should be explored to determine the best clinical therapeutic modality for unresectable stage III NSCLC [44].

The detection of MRD by ctDNA analysis warrants further improvement. Its clinical value also needs to be confirmed by large-scale prospective clinical studies. Despite the fact that clinical studies on lung cancers have reported the prognostic value and predictive assistance for the relapse time of MRD in several countries, the guiding value of ctDNA analysis for clinical decision-making needs merits further investigation [45-51]. Patients are encouraged to participate in prospective clinical trials on MRD detection by ctDNA analysis. With more clinical data, this method will be able to better guide clinical decision-making.

AUTHOR CONTRIBUTIONS

Conception and design: Yi-Long Wu. Manuscript writing: All authors. Final approval of manuscript: All authors. Accountable for all aspects of the work: All authors.

CONFLICT OF INTEREST STATEMENT

Yi-Long Wu reports grants and personal fees from AstraZeneca and Boehringer Ingelheim, grants from Bristol Myers Squibb, and personal fees from Beigen, Eli Lilly, MSD, Hengrui, Pfizer, Roche, and Sanofi outside the submitted work. Wenzhao Zhong reports other support from AstraZeneca, Bristol Myers Squibb, MSD, Roche, and Innovent outside the submitted work. Qing Zhou reports lecture and presentations fees from AstraZeneca, Boehringer Ingelheim, BMS, Eli Lilly, MSD, Pfizer, Roche, and Sanofi outside the submitted work.

ETHICS STATEMENT

Not applicable.

INFORMED CONSENT

Not applicable.