Treatment benefits and outcome determinants in (neo)adjuvant trials

ADAURA: study update

The adjuvant use of the third-generation EGFR tyrosine kinase inhibitor (TKI) osimertinib is being explored in the phase III ADAURA study that includes almost 700 patients with completely resected stage IB, II, and IIIA, EGFR-mutant non-small-cell lung cancer (NSCLC), with or without adjuvant chemotherapy. The primary analysis has revealed benefits of osimertinib compared to placebo with respect to disease-free survival (DFS) in patients with stage II/IIIA disease (36-month DFS rates, 78 % vs. 28 %; HR, 0.17; p < 0.001) and in the overall population (36-month DFS rates, 79 % vs. 40 %; HR, 0.20; p < 0.001) [1, 2]. At ESMO 2022, Tsuboi presented an updated DFS analysis at the protocol-specified maturity of 50 % as well as other findings after 2 more years of follow-up [3]. All patients had had the opportunity to complete 3 years of planned study treatment.

These updated data corroborated the previous results. For the primary endpoint, the analysis yielded median DFS of 65.8 vs. 21.9 months in the stage II/IIIA population, with 48-month DFS rates of 70 % vs. 29 % (Figure 1). This translated into a 77 % risk reduction (HR, 0.23). In the overall population, median DFS was 65.8 vs. 28.1 months, and 73 % vs. 38 % of patients were disease-free at 48 months (HR, 0.27). DFS benefits with osimertinib were observed across all predefined subgroups. In the groups with and without adjuvant chemotherapy, the HRs for DFS were 0.29 and 0.36, respectively. During the study, the patients had been restaged according to the AJCC/IUCC 8th edition staging manual; however, the pro­portions of stages according to the 7th and 8th editions were similar, and the DFS results were consistent across all stages.

The authors also reported a prespecified exploratory analysis of recurrence patterns in the overall population. Fewer patients treated in the experimental arm had disease recurrence (27 % vs. 60 %), as osimertinib treatment reduced the risk of both locoregional and distant relapses. The most common first sites of recurrence were identical in the two arms, with lower rates for osimertinib in the lung (12 % vs. 26 %), the lymph nodes (6 % vs. 17 %), and the CNS (6 % vs. 11 %). Updated results on DFS in the CNS in 63 patients with stage II/IIIA disease showed that 14 % were on osimertinib treatment at the time of CNS recurrence, vs. 71 % in the placebo group. Median CNS DFS had not been reached yet in either arm, and the 48-month rates were 90 % vs. 75 %. Here, the osimertinib-mediated relative risk reduction was as high as 76 % (HR, 0.24), with the cumulative incidence of CNS relapses being consistently lower in the experimental arm than in the placebo arm. The estimated probability of observing CNS recurrence at 36 months was 2 % vs. 13 %. No new safety concerns emerged with the extended treatment duration. In their entirety, the updated results reinforce adjuvant osimertinib as the standard of care for patients with EGFR-mutant, stage IB-IIIA NSCLC after complete tumor resection, with or without adjuvant chemotherapy.

Figure 1: Updated disease-free survival results obtained with adjuvant osimertinib vs. placebo in the ADAURA study

Figure 1: Updated disease-free survival results obtained with adjuvant osimertinib vs. placebo in the ADAURA study

ctDNA monitoring after resection of EGFR-mutant NSCLC

Personalization of adjuvant strategies such as anti-EGFR treatment with os­imertinib requires the selection of patients who are most likely to benefit from adjuvant EGFR TKI therapy, with the aim to reduce costs and avoid unnecessary side effects as well as the psychological burden of long-term treatment. Circulating tumor DNA (ctDNA) provides a potentially valuable biomarker for early diagnosis, prognostic stratification, detection of minimal residual disease (MRD), and recurrence. Ahn et al. investigated the role of longitudinal monitoring of ctDNA in 278 patients with curatively resected stage IA-IIIA, ­EGFR-mutant NSCLC. The radiological follow-up was accompanied by serial ctDNA monitoring using droplet digital PCR. Samples were taken preoperatively, 4 weeks after surgery and in increasing intervals for 5 years (every 3 months in year 1, every 4 months in year 2, every 6 months in year 3 and annually thereafter) or until clinically definitive recurrence.

At the time of the analysis reported at ESMO 2022, the median follow-up was 62.0 months [4]. Baseline ctDNA was detected in 67 patients (24.1 %). ctDNA detection rates trended to increase with the disease stage (IA, 23.4 %; IB, 17.6 %; IIA, 17.9 %; IIB, 50.0 %; IIIA, 42.3 %). The same observation was made for baseline ctDNA copy numbers, with the difference being significant (p < 0.0001). Overall, 51 of the 67 patients who had baseline ctDNA showed ctDNA clearance 4 weeks after surgery (76.1 %). This was fairly comparable across the stages. With respect to the type of EGFR mutation, the analysis did not reveal any differences between exon 19 deletion and L858R mutation regarding ctDNA detection rates (24 % vs. 24.3 %) or clearance rates (75 % vs. 77.8 %).

Relapses occurred in 28.1 %. The 3-year DFS rate was highest in patients with ctDNA negativity at baseline (83.3 %), while it was slightly lower for those with ctDNA positivity but MRD negativity 4 weeks after surgery (78.0 %) and considerably lower for those who were positive for both markers (50.0 %). These differences were significant (p = 0.02). According to a multivariate analysis including various clinicopathologic variables, the ctDNA group remained an independent determinant of DFS regardless of stage (HR, 1.27; p = 0.03). These results suggest that baseline ctDNA positivity and MRD positivity are associated with poor DFS in curatively resected stage IA-IIIA, ­EGFR-mutant NSCLC.

CheckMate 816: significance of pathological features

The randomized phase III CheckMate 816 study investigated neoadjuvant nivolumab plus platinum-doublet chemotherapy compared to chemotherapy alone in 358 patients with newly diagnosed, resectable, stage IB(≥ 4 cm)-IIIA NSCLC. Surgery was performed within 6 weeks post-treatment and was followed by optional adjuvant chemo­therapy with or without radiotherapy. Compared to chemotherapy, the combination significantly improved event-free survival (EFS) and pathologic complete response (pCR) [5]. In an additional analysis, EFS was improved with both strategies in patients with pCR or major pathologic response (MPR) in the primary tumor relative to patients without [6]. Here, lower percentages of residual viable tumor (% RVT) were associated with improved EFS in the experimental arm. A post-hoc analysis relating to the results in patients with or without pathologic evidence of lymph node involvement from CheckMate 816 was presented at ESMO 2022 [7].

The addition of nivolumab to chemotherapy led to similar EFS and pCR benefits in patients with and without lymph node involvement (Table 1). Closer inspection of the group with lymph node involvement who had received the combination showed that 19 % had achieved 0 % RVT in both the primary tumor and the lymph nodes, which equals pCR. This cohort derived the best 24-month EFS rate of 92 %, ­followed by 76 % in those who had 0 % RVT in either the primary tumor or the lymph nodes but not in both. Not surprisingly, the lowest rate (49 %) resulted in patients with residual viable tumor (> 0 % RVT) in both the primary site and the nodes. This observation adds to the growing body of evidence demonstrating that pathologic response might predict long-term outcomes in resectable NSCLC. However, clinically relevant % RVT thresholds beyond pCR warrant further investigation.

The assessment of the % RVT was achieved by pathologic recognition of tumor regression, i.e., the area occupied by the tumor prior to immune-mediated clearance [8, 9]. Nivolumab plus chemotherapy led to lower % RVT and higher % regression compared to chemotherapy only, while the extent of necrosis did not differ between the treatment arms. According to a time-dependent ROC curve analysis, % regression and % RVT in the primary tumor predicted EFS at 2 years in a similar manner in the experimental arm. Taken together, these findings further support the neoadjuvant use of nivolumab plus chemotherapy as a novel treatment option for patients with resectable NSCLC irrespective of lymph node involvement.

Table 1 Efficacy of nivolumab plus chemotherapy vs. chemotherapy only in patients with or without pathologic evidence of lymph node involvement

Health-related quality of life in CheckMate 816

Another piece of evidence underscoring the benefits obtained with the CheckMate 816 regimen is the health-related quality of life (HRQoL) analysis reported by Felip et al. [10]. Prespecified patient-reported outcome exploratory endpoints in CheckMate 816 were investigated using the EQ-5D visual analog scale (VAS) and the EQ-5D health status utility index (EQ-5D UI). Assessments were performed before the start of the neoadjuvant treatment, before cycles 2 and 3, at post-neoadjuvant visit 1 (approximately 30 days after the last dose) and post-neoadjuvant visit 2 (approximately 70 days after the first post-neoadjuvant visit).

The EQ-5D VAS and UI scores from baseline to post-neoadjuvant visit 1 in all randomized patients yielded no differences across the treatment arms, with scores being similar to those obtained in the UK population norm. EQ-5D VAS and UI scores during the neoadjuvant period and post-surgery (i.e., until post-neoadjuvant visit 2) showed slight declines after surgery, although these were similar in both arms and are consistent with previous reports [11-13]. No ­notable differences in EQ-5D VAS and UI scores were seen between the treatment arms across patient subgroups in the post-operative population. Overall, HRQoL was preserved from baseline during the neoadjuvant treatment period with nivolumab plus chemotherapy similar to chemotherapy alone, and the combination did not impact post-operative patient-reported outcomes compared to chemotherapy only.

NEOpredict-Lung: nivolumab plus relatlimab

The randomized, phase II NEOpredict-Lung trial examined the combined neoadjuvant administration of nivolu­mab and relatlimab on days 1 and 15 in patients with stage IB, II or IIIA NSCLC for whom curative resectability had been established by the multidisciplinary lung cancer board. Relatlimab is a monoclonal antibody targeting the immune checkpoint LAG-3; the efficacy of nivolumab plus relatlimab in melanoma has been demonstrated by the RELATIVITY-047 trial [14]. In the control arm of the NEOpredict-Lung study, nivolumab monotherapy was administered on day 1 and 15. Each treatment group contained 30 patients. After surgery, the patients received standard-of-care adjuvant therapy. Schuler et al. reported first results for the primary endpoint, which was ­feasibility of curatively intended surgery within 43 days, and other outcomes [15].

As the analysis showed, preoperative combined immune-checkpoint inhibitor therapy with 2 courses of nivolumab and relatlimab is safe and feasible. Treatment-related adverse events (TRAEs) were mostly grade 1 and 2 and consistent with the known profiles ­(Table 2). No patient experienced delay of surgery due to toxicity, and all patients in both arms were able to undergo surgery within 43 days. Overall response rates according to RECIST 1.1 were 27 % and 10 % for the combination and nivolumab monotherapy, respectively. Thirty-one patients received sequential FDG PET/CT at baseline and prior to surgery at one institution. This allowed for the assessment of metabolic response, which was 38 % in both arms. Complete and major pathological responses occurred in 30 % vs. 27 %. R0 resection was performed in 97 % vs. 100 %. At 12 months, 100 % vs. 92 % of patients were alive, and 91 % vs. 92 % were disease-free.

All PD-L1 expression groups experienced histopathological responses, ­although a trend towards deeper remissions emerged in patients with PD-L1–positive tumors. Also, complete remissions were restricted to this group. Patients who had ≤ 50 % vital tumor cells in their resected tissue tended to show increased effector T cell counts in the peripheral blood. Comprehensive correlative studies and biomarker analyses are ongoing. In addition, the protocol has been amended to add a third treatment arm that explores a higher dose of relatlimab for increased LAG-3 target occupancy.

Table 2 NEOpredict-Lung: treatment-related adverse events observed with nivolumab/relatlimab vs. nivolumab

Ipi/nivo in addition to CRT: INCREASE

A segment of patients with stage III ­NSCLC who have high T-stage and low N-stage (e.g., superior sulcus tumors) is borderline resectable and might become resectable after induction with chemoradiotherapy (CRT). Based on the assumption that the combination of pre-operative CRT with dual checkpoint inhibition could enhance pathological and immunological response rates through activation of antigen-presenting cells and effector T cells, the single-arm, prospective, phase II INCREASE study assessed the following regimen: in patients with cT3-4 N0-2 M0 NSCLC eligible for post-induction resection, nivolumab plus ipilimumab was administered on day 1 of CRT, and nivolumab alone was administered on day 22. The patients underwent surgery 6 weeks after the end of treatment. Pathological responses (i.e., pCR, MPR) and safety constituted the primary endpoint. Twenty-seven individuals received induction therapy, and 24 proceeded to surgery.

In the group of operated patients, the treatment resulted in substantial responses, with pCR and MPR rates of 63 % and 79 %, respectively [16]. Compared to the historical reference of 30 %, the difference for pCR was significant (p < 0.001). In terms of radiological response, 12.5 % and 87.5 % of the resected population achieved partial remissions (PR) and stable disease (SD), respectively. There were no differences between the groups who had achieved pCR vs. MPR regarding PR (13 % vs. 11 %) or SD (87 % vs. 89 %).

The toxicity rates were acceptable, with grade 3/4 TRAEs in 67 % and grade 3/4 immune-related AEs in 19 %. No patient failed to undergo surgery or died due to TRAEs. Among immune-related AEs, dermatitis was observed most commonly (any grade, 41 %), followed by thyroid disorders (33 %), pneumonitis (11 %), and hepatitis (7 %). According to a translational analysis, induction treatment led to higher levels of CD8-positive effector T cells in the peripheral blood. The CD39-positive T cell subset mainly increased in patients who had achieved pCR. A comparison with historical controls (i.e. matched patients who underwent CRT and surgery without immunotherapy) showed increased CD8-positive T cell proliferation in tumor-draining lymph nodes.

PEARLS/KEYNOTE-091 results according to PD-L1 expression

In the randomized, triple-blind, phase III PEARLS/KEYNOTE-091 study, adjuvant pembrolizumab for ≤ 18 administrations (n = 590) was tested against placebo (n = 587) in patients with confirmed stage IB (T ≥ 4 cm), II, or IIIA NSCLC ­after complete surgical resection with negative margins (i.e., R0 resection). Adjuvant chemotherapy for a maximum of 4 cycles was optional. The dual primary endpoints included DFS in the overall population and in the group with PD-L1 TPS > 50 %. At the time of the ­second interim analysis after a median follow-up of 35.6 months, DFS prolongation with pembrolizumab compared to placebo was clinically meaningful and statistically significant in the overall population (HR, 0.76; p = 0.0014) but, surprisingly, not in the population with PD-L1 expression > 50 % (HR, 0.82; p = 0.14) [17]. In search of an explanation for this discrepancy, Peters et al. evaluated the trial outcomes according to PD-L1 expression status [18].

The baseline characteristics were generally similar across the overall population and the PD-L1 ≥ 50 % population and were also balanced between the treatment arms within both groups. All-cause AEs as well as immune-mediated AEs and infusion reactions were similar for the overall and PD-L1-high populations, which also applied to treatment exposure. The DFS analysis showed significant risk reductions with pembrolizumab vs. placebo for the groups with TPS 1-49 % (HR, 0.67) and TPS < 1 % (HR, 0.78). As expected, median and long-term DFS estimates were numerically improved with pembrolizumab in the TPS ≥ 50 % population versus the TPS 1-49 % and < 1 % groups; however, the difference within the TPS ≥ 50 % group was offset by the fact that the placebo arm also performed better than the placebo arms in the other groups (Figure 2). Median DFS had not been reached yet in either arm of the TPS ≥ 50 % population. Therefore, the absence of a statistically significant DFS benefit in this group is likely to have resulted from placebo overperformance.

Overall, these data from the PEARLS/KEYNOTE-091 trial confirm the benefit of pembrolizumab in patients with completely resected stage IB-IIIA ­NSCLC and, if recommended, prior adjuvant chemotherapy, regardless of PD-L1 expression. DFS in the TPS ≥ 50 % population will be tested again at the next interim analysis.

Figure 2: Disease-free survival performance of pembrolizumab (left) and placebo (right) according to the three PD-L1 categories in PEARLS/KEYNOTE-091

Figure 2: Disease-free survival performance of pembrolizumab (left) and placebo (right) according to the three PD-L1 categories in PEARLS/KEYNOTE-091

No benefits of adjuvant canakinumab

Negative results were obtained for the adjuvant use of the anti-IL-1β antibody canakinumab in the placebo-controlled phase III CANOPY-A trial [19]. IL-1β is a key mediator of inflammation, and canakinumab has been approved in many countries for the management of inflammatory conditions. In the phase III CANTOS study evaluating the role of canakinumab in cardiovascular disease, an exploratory analysis had shown reductions in the incidence and mortality of NSCLC [20], which led to the launch of CANOPY-A. Almost 1,400 patients with completely resected, stage IIA-IIIA and IIIB (T > 5 cm, N2) NSCLC who had received cisplatin-based chemotherapy and/or radiation therapy participated in the trial.

DFS, which was the primary endpoint, did not differ between the canakinumab-treated arm and the placebo arm (35.0 vs. 29.7 months; HR, 0.94; p = 0.258) [19]. The subgroup analysis based on demographics, disease characteristics, and biomarkers of interest did not show any meaningful differences across the treatment arms. Likewise, lung-cancer–specific survival was similar for canakinumab and placebo (HR, 0.90). OS was not formally tested due to the lack of significance regarding DFS. No new safety concerns arose; infections constituted the most common AEs of special interest and were reported with comparable incidences. Also, the assessment yielded similar rates of AEs, serious AEs, and fatal serious AEs.

Exploratory biomarker data will continue to be analyzed, including tumor molecular profiling and longitudinal analyses of circulating inflammatory markers. The collective data gained throughout the CANOPY study program that further includes a neoadjuvant trial and two studies conducted in the advanced disease setting will continue to elucidate the role of IL-1β in NSCLC and inform ongoing ­research.


  1. Wu YL et al., Osimertinib in resected EGFR-mutated non-small-cell lung cancer. N Engl J Med 2020; 383(18): 1711-1723
  2. Herbst R et al., Osimertinib as adjuvant therapy in patients with stage IB–IIIA EGFR mutation positive (EGFRm) NSCLC after complete tumor resection: ADAURA. J Clin Oncol 2020; 38: 18_suppl.LBA5
  3. Tsuboi M et al., Osimertinib as adjuvant therapy in patients with resected EGFRm stage IB-IIIA NSCLC: updated results from ADAURA. ESMO 2022, LBA47
  4. Ahn MJ et al., Longitudinal monitoring of circulating tumor DNA from plasma in patients with curative resected stage IA-IIIA EGFR mutant non-small cell lung cancer. ESMO 2022, abstract 933MO
  5. Forde PM et al., Neoadjuvant nivolumab plus chemotherapy in resectable lung cancer. N Engl J Med 2022; 386(21): 1973-1985
  6. Provencio-Pulla M et al., Neoadjuvant nivolumab + platinum-doublet chemotherapy versus chemo for resectable (IB–IIIA) non-small cell lung cancer: Association of pathological regression with event-free survival in CheckMate 816. J Clin Oncol 2022; 40:17_suppl.LBA8511
  7. Stein Deutsch J et al., Analysis of pathological features and efficacy outcomes with neoadjuvant nivolumab plus platinum-doublet chemotherapy for resectable non-small cell lung cancer in CheckMate 816. ESMO 2022, abstract LBA50
  8. Cottrell TR et al., Pathologic features of response to neoadjuvant anti-PD-1 in resected non-small-cell lung carcinoma: a proposal for quantitative immune-related pathologic response criteria (irPRC). Ann Oncol 2018; 29(8): 1853-1860
  9. Stein JE et al., Pan-tumor pathologic scoring of response to PD-(L)1 blockade. Clin Cancer Res 2020; 26(3): 545-551
  10. Felip E et al., Nivolumab plus platinum-doublet chemotherapy versus chemotherapy as neoadjuvant treatment for resectable non-small cell lung cancer: health-related quality of life outcomes from CheckMate 816. ESMO 2022, abstract 932MO
  11. Balduyck B et al., Quality of life evolution after lung cancer surgery: a prospective study in 100 patients. Lung Cancer 2007; 56(3): 423-431
  12. Ichimura H et al., Preoperative predictors of restoration in quality of life after surgery for lung cancer. Thoracic Cancer 2021; 12(6): 835-844
  13. Poghosyan H et al., Health-related quality of life after surgical treatment in patients with non-small cell lung cancer: a systematic review. Lung Cancer 2013; 81(1): 11-26
  14. Tawbi HA et al., Relatlimab and nivolumab versus nivolumab in untreated advanced melanoma. N Engl J Med 2022; 386: 24-34
  15. Schuler M et al., A randomized phase II study of preoperative nivolumab plus relatlimab over nivolumab in patients with resectable non-small-cell lung cancer. ESMO 2022, abstract LBA37
  16. Bahce I et al., Ipilimumab plus nivolumab and chemoradiotherapy followed by surgery in patients with resectable and borderline resectable lung cancer: the INCREASE trial. ESMO 2022, abstract 950O
  17. O’Brien M et al., Pembrolizumab versus placebo as adjuvant therapy for completely resected stage IB–IIIA non-small-cell lung cancer (PEARLS/KEYNOTE-091): an interim analysis of a randomised, triple-blind, phase 3 trial. Lancet Oncol 2022;
  18. Peters S et al., PD-L1 expression and outcomes of pembrolizumab and placebo in completely resected stage IB-IIIA NSCLC: subgroup analysis of PEARLS/KEYNOTE-091. ESMO 2022, abstract 930MO
  19. Garon EB et al., CANOPY-A: phase III study of canakinumab as adjuvant therapy in patients with completely resected non-small cell lung cancer. ESMO 2022, LBA49
  20. Ridker PM et al., Effect of interleukin-1β inhibition with canakinumab on incident lung cancer in patients with atherosclerosis: exploratory results from a randomised, double-blind, placebo-controlled trial. Lancet 2017; 390(10105): 1833-1842

© 2022 Springer-Verlag GmbH, Impressum