Adjuvant Pembrolizumab in Patients with Stage IIIA/N2 Non–Small Cell Lung Cancer Completely Resected after Neoadjuvant Concurrent Chemoradiation: A Prospective, Open-Label, Single-Arm, Phase 2 Trial
Article information
Abstract
Purpose
Optimal treatment for stage IIIA/N2 non–small cell lung cancer (NSCLC) is controversial. We aimed to assess the efficacy and safety of adjuvant pembrolizumab for stage IIIA/N2 NSCLC completely resected after neoadjuvant concurrent chemoradiation therapy (CCRT).
Materials and Methods
In this open-label, single-center, single-arm phase 2 trial, patients with stage IIIA/N2 NSCLC received adjuvant pembrolizumab for up to 2 years after complete resection following neoadjuvant CCRT. The primary endpoint was disease-free survival (DFS). Secondary endpoints included overall survival (OS) and safety. As an exploratory biomarker analysis, we evaluated the proliferative response of blood CD39+PD-1+CD8+ T cells using fold changes in the percentage of proliferating Ki-67+ cells from days 1 to 7 of cycle 1 (Ki-67D7/D1).
Results
Between October 2017 and October 2018, 37 patients were enrolled. Twelve (32%) and three (8%) patients harbored EGFR and ALK alterations, respectively. Of 34 patients with programmed cell death ligand 1 assessment, 21 (62%), nine (26%), and four (12%) had a tumor proportion score of < 1%, 1%-50%, and ≥ 50%, respectively. The median follow-up was 71 months. The median DFS was 22.4 months in the overall population, with a 5-year DFS rate of 29%. The OS rate was 86% at 2 years and 76% at 5 years. Patients with tumor recurrence within 6 months had a significantly lower Ki-67D7/D1 among CD39+PD-1+CD8+ T cells than those without (p=0.036). No new safety signals were identified.
Conclusion
Adjuvant pembrolizumab may offer durable disease control in a subset of stage IIIA/N2 NSCLC patients after neoadjuvant CCRT and surgery.
Introduction
Approximately 20% of patients with non–small cell lung cancer (NSCLC) are diagnosed at stage IIIA, with N2 disease accounting for around 40%-50% of those cases [1]. Optimal treatment for stage IIIA/N2 NSCLC remains controversial. Neoadjuvant or adjuvant chemotherapy provides only modest benefit, with a 4%-5% absolute improvement in 5-year overall survival (OS) rate compared to surgery alone for operable NSCLC, regardless of chemotherapy regimen, the platinum agent used, and whether adjuvant radiotherapy is given or not [2,3]. As a result, the prognosis of stage IIIA NSCLC is poor, with a 5-year OS rate of approximately 36% [4].
Compared to neoadjuvant chemotherapy alone, neoadjuvant chemoradiation improves rates of mediastinal downstaging, pathological complete response (pCR) of mediastinal lymph nodes, and R0 resection for operable stage IIIA/N2 NSCLC, although survival benefit remains unclear [5]. The Intergroup 0139 trial showed potential benefits of neoadjuvant concurrent chemoradiation therapy (CCRT) followed by surgery in stage IIIA/N2 NSCLC compared to definitive CCRT, including improved progression-free survival and, in patients eligible for lobectomy, improved OS [6]. Real-world data also support the efficacy of neoadjuvant CCRT for N2 disease [7]. Therefore, trimodal therapy with neoadjuvant CCRT and surgery is a viable and potentially beneficial therapeutic option for stage IIIA/N2 NSCLC [8].
Multiple pivotal trials have recently shown that adding immune checkpoint inhibitors (ICIs) to standard neoadjuvant or adjuvant systemic therapy, whether pre- [9], post- [10,11], or perioperatively [12-15], improves survival outcomes in resectable NSCLC patients, especially those with programmed cell death ligand 1 (PD-L1) tumor proportion score (TPS) of 1% or greater. Neoadjuvant programmed cell death protein 1 (PD-1) inhibitor plus chemotherapy followed by surgery and adjuvant immunotherapy is currently one of the standard-of-care options for locally advanced NSCLC patients without contraindications to immunotherapy [16]. However, none of these trials have investigated whether ICI has a role in patients receiving trimodal therapy including neoadjuvant CCRT and surgery. Given the effective local control with neoadjuvant CCRT and durable disease control with adjuvant ICI observed across studies, we hypothesized that combining these therapies may improve the outcome of patients with stage IIIA/N2 NSCLC. This study aims to (1) evaluate the efficacy and safety of adjuvant pembrolizumab in stage IIIA/N2 NSCLC patients receiving complete tumor resection after neoadjuvant CCRT and (2) explore tissue- and blood-based biomarkers that can predict durable disease control with adjuvant pembrolizumab.
Materials and Methods
1. Study design and participants
This is an investigator-initiated, open-label, single-center, single-arm phase 2 trial to evaluate the efficacy of adjuvant pembrolizumab in patients with completely resected stage IIIA/N2 NSCLC after neoadjuvant CCRT. Eligible patients were aged 18 years or older, had pathologically confirmed NSCLC of stage IIIA/N2 according to the American Joint Committee on Cancer Staging Manual 8th edition, and received the institutional standard protocol of neoadjuvant CCRT with 44-Gy thoracic radiotherapy in 2-Gy fractions beginning on day 1 plus paclitaxel (50 mg/m2) and cisplatin (25 mg/m2) on days 1, 8, 15, 22, and 29, followed by complete surgical resection including negative margins. Key eligibility criteria included no evidence of disease on clinical examination and radiographic assessment per Response Evaluation Criteria in Solid Tumors ver. 1.1 after surgery, Eastern Cooperative Oncology Group performance status of 0 or 1, and adequate organ functions confirmed within 14 days before enrollment. Patients were excluded if they had previously received any treatment for NSCLC or a PD-1 or PD-L1 inhibitor for any indications; received an immunosuppressant or a systemic corticosteroid within 28 days before study entry; had a history of other malignancy within 2 years before enrollment except for cutaneous squamous or basal cell carcinoma, early gastric cancer, well-differentiated thyroid carcinoma, and cervical intraepithelial neoplasia; or had a history of autoimmune disease requiring systemic immunosuppression within 2 years before study entry. Complete eligibility criteria are outlined in the Supplementary Methods.
2. Procedures and assessments
After complete recovery from surgery and within six weeks, patients commenced adjuvant systemic therapy with intravenous pembrolizumab at a fixed dose of 200 mg on day 1 of each 21-day cycle for 2 years or until disease recurrence, occurrence of unacceptable toxicity, or death. Pembrolizumab therapy could be continued beyond the date of disease recurrence if the investigator deemed the patient likely to benefit. Pembrolizumab dose reduction was not permitted. Pembrolizumab therapy could be interrupted or delayed at the occurrence of grade 2 or higher treatment-related adverse events (TRAEs). The study treatment was permanently discontinued if (1) a grade 2 or higher TRAE occurred and did not resolve to grade 0 or 1 after a maximum 12-week treatment interruption; (2) grade 3 or higher TRAEs involving increased alanine or aspartate aminotransferase, hyperbilirubinemia, infusion-related reaction, pneumonitis, or acute kidney injury occurred; (3) grade 3 or serious TRAEs recurred after pembrolizumab resumption; or (4) any grade 4 TRAE occurred.
Screening assessment included clinical examination, complete blood count with differential, comprehensive metabolic panel, thyroid function test, electrocardiogram, chest X-ray, and chest computed tomography (CT). After each treatment cycle, patients received clinical examinations, routine blood tests, and chest X-rays to assess adverse events. For patients without documented disease recurrence, chest CT was done every 12 weeks for the first year, every 16 weeks for the second year, every 6 months for the third year, and annually thereafter. Patients who stopped study treatment continued to be followed up as part of the study unless they withdrew consent to be followed up. Local pathologists assessed pathological response to neoadjuvant CCRT. Adverse events were assessed at every visit and graded according to the Common Terminology Criteria for Adverse Events ver. 4.0.
3. Assessment of tumor PD-L1 expression
Formalin-fixed paraffin-embedded tissues from pre-CCRT (diagnosis) and post-CCRT (surgical) archival samples were analyzed for tumor PD-L1 expression using the PD-L1 immunohistochemistry (IHC) 22C3 pharmDx assay (Agilent Technologies, Santa Clara, CA). The percentage of tumor cells with membranous PD-L1 staining was assessed for each sample. Patients were classified as PD-L1–positive if TPS was ≥ 1% in initial biopsy samples (prioritized) or surgical specimens (used when biopsies were unsuitable for PD-L1 assessment due to limited tissue or technical constraints). Otherwise, patients with TPS lower than 1% in either sample were classified as PD-L1–negative.
4. Multi-color flow cytometry
Peripheral blood was collected immediately before treatment (cycle 1 day 1) and six days after the first dose of pembrolizumab (cycle 1 day 7). Peripheral blood mononuclear cells (PBMCs) were isolated from whole blood using standard Ficoll-Paque (GE Healthcare, Pittsburgh, PA) density gradient centrifugation. The cells were counted and cryopreserved for further use. PBMCs were stained with the indicated antibodies for immune cell analysis. The following fluorochrome-conjugated monoclonal antibodies were used in multi-color flow cytometry: anti-CD8 (RPA-T8), anti-CD3 (HIT3a), and anti-CD4 (SK3) from BD Biosciences (Beford, MA); anti–PD-1 (EH12.2H7) and anti-Ki-67 (Ki-67) from BioLegend (San Diego, CA); anti-CD14 (61D3), anti-CD19 (HIB19), and anti-CD39 (eBioA1) from eBioscience (San Diego, CA). Surface-stained cells were fixed and permeabilized for intracellular staining with the Foxp3/Transcription Factor Staining Buffer Set (eBiosciences). Stained PBMCs were acquired on the LSR II flow cytometer (BD Biosciences). Data were analyzed using the FlowJo software (TreeStar, San Carlos, CA). We assessed the proportion of proliferating Ki-67+ cells among peripheral blood CD8+ T lymphocyte subsets at each timepoint. The 1-week proliferative response of each T cell subset was assessed using fold changes in the percentage of Ki-67+ cells from day 1 to day 7 of the first treatment cycle (Ki-67D7/D1).
5. Outcomes
The primary endpoint was disease-free survival (DFS). Secondary endpoints included OS and safety. All analyses were based on the intention-to-treat population. DFS was defined as the time from surgery to lung cancer recurrence or death from any cause, whichever occurred first. OS was defined as the time from surgery to death from any cause. Additional prespecified exploratory outcomes included DFS and OS in patient subgroups stratified by PD-L1 TPS, epidermal growth factor receptor (EGFR)/anaplastic lymphoma kinase (ALK) alteration status, and other clinicopathologic characteristics, and the association of the first-week proliferative response of circulating CD8+ T cell subsets upon pembrolizumab treatment and durable clinical benefit (DCB; defined as the absence of disease recurrence within 6 months after surgery), DFS, and OS.
6. Statistical analysis
We calculated the sample size based on the primary endpoint, DFS. This study was designed to have a power of 85% to reject the null hypothesis that the median DFS was 20 months or less at a one-sided α level of 0.15, assuming that the true median DFS was 30 months. This assumption was based on the internally analyzed DFS data of patients with stage IIIA/N2 NSCLC treated with neoadjuvant CCRT and curative surgery at Samsung Medical Center before this study was designed. The one-sample test assuming an exponential distribution of survival times, an accrual period of 15 months, an additional 36 months of follow-up, and a 5% dropout rate indicated that 37 patients were required. The final analysis was planned to be conducted when at least 22 patients experienced disease recurrence or death.
We summarized the data as medians and ranges for continuous variables and numbers and percentages for discrete variables. Fisher’s exact tests were used to assess associations between categorical variables. PD-L1 TPS was compared between the pre- and post-CCRT samples using the Wilcoxon signed-rank test. Differences in Ki-67D7/D1 for each CD8+ T cell subset were compared between groups using the Mann-Whitney test. The optimal cutoff of Ki-67D7/D1 was determined as the point at which Youden’s index was maximized and used to divide patients into two groups. We used the Kaplan-Meier method to estimate DFS and OS. Patients without a documented event had their data censored on the last tumor assessment date for DFS and on the last follow-up date for OS. We used the Cox proportional hazards regression analysis to estimate hazard ratios (HRs) and test for differences in DFS and OS between groups. The reverse Kaplan-Meier method was used to estimate the follow-up time. All tests were two-tailed. p-values smaller than 0.05 were considered statistically significant. R statistical software ver. 4.3.2 (The R Foundation for Statistical Computing, Vienna, Austria) was used for computation.
Results
1. Patient characteristics
From October 18, 2017, to October 24, 2018, 40 patients were screened for eligibility, and 37 were enrolled (Table 1, S1 Table). Approximately two-thirds were male, with a median age at diagnosis of 63 years. The majority (73.0%) of patients presented with adenocarcinoma, while 27.0% had squamous cell carcinoma. Following neoadjuvant CCRT, all but three patients underwent lobectomy; two received bilobectomy, and one underwent pneumonectomy. pCR was achieved in two patients (5.4%). Among non-pCR cases, 17 (45.9%) and 10 (27.0%) patients had involvement of single and multiple N2 nodal stations, respectively, while the remaining eight patients (21.6%) had complete N2 nodal clearance. EGFR-activating mutations were identified in 12 patients (32%), with L858R substitution (n=6), exon 19 deletion (n=3), exon 20 insertion (n=1), and S768I substitution (n=2) constituting the major alterations. One patient with S768I also harbored a G719X mutation. ALK gene rearrangements were detected in three patients (8.1%), including two (5.4%) with ALK IHC score of 3+ determined by an Food and Drug Administration (FDA)–approved companion diagnostic test and one (2.7%) confirmed with fluorescence in situ hybridization.
Patient characteristics
PD-L1 TPS was assessed in 21 patients (56.8%) before neoadjuvant CCRT using initial biopsy samples and in 30 patients (81.1%) after CCRT using surgical specimens. Of the 17 patients with both pre-CCRT and post-CCRT PD-L1 expression measurements, eight (47.1%) experienced an increase in PD-L1 TPS, two (11.8%) experienced a decrease, and the remaining seven (41.2%) had PD-L1 TPS of < 1% at both timepoints (S2 Fig.). There was no statistically significant difference in PD-L1 TPS between the pre-CCRT and post-CCRT assessments (p=0.091). One patient who achieved a pCR to neoadjuvant CCRT had a pre-CCRT PD-L1 TPS of 0%. Integrating results from both assessments (with baseline assay prioritized), we classified 21 (56.8%), nine (24.3%), and four (10.8%) patients into the PD-L1 TPS categories of < 1%, 1%-50%, and ≥ 50%, respectively. Three patients had no available tumor tissue for PD-L1 expression assessment.
All enrolled patients received at least one dose of adjuvant pembrolizumab, and 12 (32.4%) completed the protocol-specified 2-year treatment course (Fig. 1). The median number of pembrolizumab cycles was 19, with a median duration of therapy of 55 weeks. Among the 25 patients who discontinued study treatment, 18 (72.0%) had disease recurrence, four (16.0%) discontinued due to TRAEs, and three (12%) withdrew consent before completing the adjuvant therapy.
Clinicopathologic profiles and disease-free survival follow-up of individual patients. Each horizontal color bar indicates each patient’s disease-free survival follow-up period. Patients who had disease recurrence, died without recurrence, and were disease-free and alive at the end of follow-up are denoted with different bar colors. The patient indicated with an asterisk received four doses of pembrolizumab beyond disease recurrence. ALK, anaplastic lymphoma kinase; EGFR, epidermal growth factor receptor; EOT, end of treatment; PD-L1, programmed cell death ligand 1; TPS, tumor proportion score; TRAE, treatment-related adverse event.
2. Efficacy
As of the data cutoff date (December 22, 2023), 11 patients (29.7%) in the intention-to-treat population remained disease-free and alive. With a median follow-up of 71 months, 24 patients (64.9%) had disease recurrence, and 11 patients (30%) died. The median DFS was 22.4 months, with 2-year and 5-year DFS rates of 46% (95% confidence interval [CI], 30 to 61) and 29% (95% CI, 16 to 44), respectively (Fig. 2A). The median OS was not reached in the intention-to-treat population. The OS rate was 86% (95% CI, 71 to 94) at 2 years and 76% (95% CI, 58 to 87) at 5 years (Fig. 2B).
Disease-free survival (A) and overall survival (B) in the intention-to-treat population. Rates of disease-free survival and overall survival at 24 and 60 months and their 95% confidence intervals are shown in the inset tables.
Among patients who had a PD-L1 TPS of 1% or higher, the median DFS was 38.4 months, and the 2-year and 5-year DFS rates were 62% (95% CI, 31 to 82) and 37% (95% CI, 13 to 62), respectively (Fig. 3A). Patients with a PD-L1 TPS of less than 1% had a median DFS of 14.1 months and 2-year and 5-year DFS rates of 29% (95% CI, 12 to 48) and 19% (95% CI, 6 to 38), respectively (Fig. 3A). Patients with PD-L1 TPS ≥ 1% had a trend toward improved DFS over those with PD-L1 TPS < 1%, although the difference in DFS between these two groups was not statistically significant (HR for PD-L1 TPS ≥ 1% vs. < 1%, 0.52; 95% CI, 0.22 to 1.22; p=0.134). For patients with EGFR-activating mutations or ALK translocations, the median DFS was 14.3 months, with 2-year and 5-year DFS rates of 40% (95% CI, 16 to 63) and 20% (95% CI, 4.9 to 42), respectively (Fig. 3B). The median DFS for patients without EGFR or ALK alterations was 25.7 months, with 2-year and 5-year DFS rates of 50% (95% CI, 28 to 68) and 36% (95% CI, 17 to 55), respectively (Fig. 3B). There was a trend toward improved DFS in patients without EGFR/ALK alterations, although it did not reach statistical significance (HR for patients with vs. without EGFR/ALK alterations, 1.47; 95% CI, 0.67 to 3.19; p=0.334).
Disease-free survival according to the programmed cell death ligand 1 (PD-L1) tumor proportion score (TPS) (A) and epidermal growth factor receptor (EGFR)/anaplastic lymphoma kinase (ALK) alteration status (B) and overall survival according to the PD-L1 tumor proportion score (C) and EGFR/ALK alteration status (D). Rates of disease-free and overall survival at 24 and 60 months and their 95% confidence intervals (CIs) are shown in the inset tables. HR, hazard ratio.
In contrast, subgroups defined by PD-L1 TPS (≥ 1% vs. < 1%) showed an opposite trend in OS compared to DFS, although the difference in OS between these groups was not statistically significant (HR for PD-L1 TPS ≥ 1% vs. < 1%, 1.5; 95% CI, 0.46 to 4.94; p=0.501) (Fig. 3C). Similarly, EGFR/ALK alteration status was not significantly associated with OS, but a trend toward an inverse association compared to DFS was observed (HR for patients with vs. without EGFR/ALK alterations, 0.29; 95% CI, 0.06 to 1.32; p=0.110) (Fig. 3D). There was no association between PD-L1 expression and EGFR/ALK alterations (odds ratio, 1.02; p > 0.99). All tested clinicopathological variables showed no significant association with DFS (S3A Fig.) and OS (S3B Fig.).
3. Proliferative response of circulating CD39+PD-1+CD8+ T cells after pembrolizumab therapy as a surrogate biomarker for durable disease control
Of the 37 enrolled patients, 34 had pretreatment (cycle 1 day 1) and on-treatment (cycle 1 day 7) blood samples. Patients with DCB had a significantly higher Ki-67D7/D1 among CD39+PD-1+CD8+ T cells than those without DCB (p=0.036) (Fig. 4A). In contrast, PD-1+CD8+ T cells (Fig. 4B) and CD8+ T cells (Fig. 4C) showed no significant difference in Ki-67D7/D1 between patients who achieved DCB and who did not (p=0.857 and p=0.817, respectively). Based on the optimal cutoff of 2.055 for Ki-67D7/D1 among CD39+PD-1+CD8+ T cells, the sensitivity and specificity for predicting DCB were 80% and 75%, respectively (S4 Fig.). Twenty-four of 25 patients (96%) with Ki-67D7/D1 ≥ 2.055 achieved DCB, while three of nine patients (33.3%) with Ki-67D7/D1 < 2.055 progressed within 6 months after surgery (p=0.048) (Fig. 4D). Among patients with Ki-67D7/D1 ≥ 2.055, the median DFS was 22.5 months, and the DFS rate was 48% (95% CI, 28 to 66) at 2 years and 27% (95% CI, 12 to 46) at 5 years (Fig. 4E). Patients with Ki-67D7/D1 < 2.055 had a median DFS of 12.8 months and 2-year and 5-year DFS rates of 44% (95% CI, 14 to 72) and 33% (95% CI, 8 to 62), respectively (Fig. 4E). The DFS did not significantly differ between these two groups (HR for patients with Ki-67D7/D1 < 2.055 vs. Ki-67D7/D1 ≥ 2.055, 1.18; 95% CI, 0.47 to 2.99; p=0.721). Similarly, patients with Ki-67D7/D1 ≥ 2.055 tended to live longer than those with Ki-67D7/D1 < 2.055, but the difference in OS was not significant (HR for patients with Ki-67D7/D1 < 2.055 vs. Ki-67D7/D1 ≥ 2.055, 1.47; 95% CI, 0.38 to 5.7; p=0.576) (Fig. 4F).
Early proliferative response of CD39+PD-1+CD8+ T cells after anti–PD-1 therapy as a predictor of durable clinical benefit. (A-C) Ki-67D7/D1 among peripheral blood CD39+PD-1+CD8+ T cells (A), PD-1+CD8+ T cells (B), and CD8+ T cells (C). One patient with durable clinical benefit had a baseline (day 1) percentage of Ki-67+ cells among CD39+PD-1+CD8+ T cells of 0, hence an infinite Ki-67D7/D1 value (shown as a cropped dot in panel A). Error bars indicate standard errors. (D) The rate of recurrence within 6 months after surgery in patients with Ki-67D7/D1 ≥ 2.055 and those with Ki-67D7/D1 < 2.055 among CD39+PD-1+CD8+ T cells. (E, F) Disease-free survival (E) and overall survival (F) according to Ki-67D7/D1 among CD39+PD-1+CD8+ T cells. CI, confidence interval; HR, hazard ratio; PD-1, programmed cell death protein 1.
4. Safety
Most TRAEs were of grade 1 or 2 (Table 2). Common TRAEs included pneumonitis, hypothyroidism, hyperthyroidism, and skin rash, most of which were of low grade. One patient each experienced grade 4 pneumonitis and grade 3 elevation of aspartate and alanine aminotransferases, which led to permanent discontinuation of pembrolizumab. Two additional patients had to discontinue adjuvant therapy due to persistent grade 2 pneumonitis and grade 2 fatigue, respectively. Five serious TRAEs occurred, including four that led to treatment discontinuation and one grade 2 pneumonia. In most other cases, TRAEs were manageable with supportive care with or without transient treatment interruption. No new safety signals were identified.
Treatment-related adverse events
Discussion
This study is the first clinical trial that evaluated the feasibility, efficacy, and safety of adjuvant ICI in patients with stage IIIA/N2 NSCLC completely resected after neoadjuvant CCRT. Adjuvant pembrolizumab for up to 2 years resulted in a median DFS of 22.4 months and 2-year DFS and OS rates of 46% and 86%, respectively. Stratifying patients by PD-L1 TPS, EGFR/ALK alterations, and other clinicopathologic factors did not reveal any significant association with DFS, suggesting restricting patient selection based on these features for this multimodal therapy might not be necessary. However, the small sample size necessitates further investigation to establish the predictive role of these tumor-intrinsic biomarkers. Most TRAEs were of low grade, and their frequency was similar to or slightly lower than those reported previously with (neo)adjuvant pembrolizumab despite the longer duration of therapy in our study [11,14].
The IMPOWER-010 and PEARLS/KEYNOTE-091 trials evaluated adjuvant atezolizumab and pembrolizumab, respectively, in patients with completely resected stage IB-IIIA NSCLC [10,11]. They reported an almost identical 2-year DFS rate of 58% in the overall intention-to-treat population. While the observed efficacy of adjuvant pembrolizumab in our study does not compare favorably with results from these phase 3 trials, important differences in study designs and populations preclude valid comparisons. All patients in IMPOWER-010 and most (86%) in PEARLS/KEYNOTE-091 received adjuvant platinum-based chemotherapy before randomization, whereas all patients in our study received neoadjuvant CCRT and no adjuvant chemotherapy. The higher prevalence of EGFR/ALK alterations in our study (41%) than in IMPOWER-010 (15%) and PEARLS/KEYNOTE-091 (7%) might also have contributed to differing results [17]. Finally, and most importantly, neither of the two phase 3 trials reported the outcome of the stage IIIA/N2 patient subgroup separately, which is likely inferior to the reported overall study population outcomes.
More recently, three randomized trials (NADIM II, KEYNOTE-671, and AEGEAN) demonstrated that neoadjuvant chemoimmunotherapy followed by surgery and adjuvant immunotherapy for 6 to 12 months improves the pCR rate and event-free survival compared to neoadjuvant chemotherapy followed by surgery alone in patients with resectable NSCLC [13-15]. Among these trials, NADIM II enrolled only stage IIIA or IIIB patients, and only AEGEAN reported the median DFS of stage IIIA patients separately. The 2-year progression-free survival rate was 67.2% among patients in the experimental arm of the NADIM II trial who received perioperative nivolumab [13]. In the phase 3 AEGEAN study, the 2-year event-free survival rate in the modified intentionto-treat population was 63.3%, and the median event-free survival was not reached both in the modified intention-totreat population and the subgroup of patients with stage IIIA NSCLC [15]. The single-arm NADIM trial enrolled only stage III NSCLC patients, among whom less than 46% had stage IIIA/N2 disease, and showed that neoadjuvant nivolumab plus chemotherapy followed by surgery and 1-year adjuvant nivolumab could achieve a 2-year progression-free survival rate of 77.1% [12]. Taken together, the event-free survival outcomes of stage III NSCLC patients in perioperative immunotherapy trials appear more favorable when compared to the DFS observed in our study. However, again, fundamental differences in patient characteristics and endpoint definitions render valid comparisons across these studies infeasible. The primary endpoint of our study, DFS, was calculated from the date of surgery, while the event-free survival was calculated from the date of randomization in NADIM II, KEYNOTE-671, and AEGEAN studies and from the date of diagnosis in the NADIM trial. In addition, stage IIIA/N2 disease accounted for less than half of patients in all perioperative immunotherapy trials. Lastly, NADIM, NADIM II, and AEGEAN excluded patients with actionable driver gene alterations, and only 7% of patients in KEYNOTE-671 had EGFR or ALK alterations [12-15].
Multiple randomized trials have explored neoadjuvant chemoradiation in stage IIIA/N2 NSCLC [6,18,19]. The reported median progression-free (or event-free) survival in the intervention arms (patients who received preoperative chemoradiation) ranged from 12.4 to 12.8 months, and OS from 24 to 40 months. Therefore, survival outcomes observed in our study appear numerically superior to those of historical cohorts who received neoadjuvant chemoradiation without adjuvant ICI, suggesting a potential benefit from adjuvant pembrolizumab. This is particularly noteworthy given the differing definitions of DFS and progression-free survival as described above. Importantly, our study’s eligibility required complete tumor resection, whereas previous chemoradiation trials did not impose a surgical outcome criterion.
While pCR is a surrogate for long-term survival in completely resected NSCLC [20], the 5% pCR rate observed in our study using neoadjuvant CCRT falls within the historical range of 4%-16% achieved with neoadjuvant chemotherapy alone [21-23]. Moreover, existing evidence suggests that the survival benefit of adding radiotherapy to neoadjuvant chemotherapy is minimal, despite potential improvements in pCR rates [19,24]. This implicates the limited impact of radiotherapy beyond local disease control [25]. In contrast, enhanced pathological response with neoadjuvant chemoimmunotherapy reflects systemic treatment effect and likely micrometastasis eradication, as evidenced by improved pCR rate and event-free survival in all four randomized trials [9,13-15]. However, few studies, if at all, have investigated the role of radiotherapy in the era of immunotherapy for operable lung cancer, and none directly compared neoadjuvant CCRT with neoadjuvant chemoimmunotherapy. Given that only three patients experienced recurrence after completing 2-year adjuvant pembrolizumab in this study with a median follow-up of 71 months, identifying predictive biomarkers for early recurrence may further optimize patient selection.
Studies have shown that dynamic blood-based immune cell biomarkers can predict anti–PD-1 therapy response in advanced NSCLC, avoiding invasive tissue biopsy [26-28]. Notably, the proliferative response (increase in the proportion of Ki-67+ cells) of PD-1+CD8+ T cells 1 to 4 weeks after ICI treatment consistently correlated with efficacy [26,27]. As anti-tumor T cell response to anti–PD-1 therapy is primarily driven by tumor-reactive CD8+ T cells marked by CD39 expression [29], we hypothesized that the proliferative response of CD39+PD-1+CD8+ T cells could robustly predict DCB. Indeed, we observed that the early proliferative response of circulating CD39+PD-1+CD8+ T cells, but not that of PD-1+CD8+ or CD8+ T cells, was associated with DCB, indicating that failure to reinvigorate tumor-reactive, exhausted CD8+ T cells promptly after pembrolizumab treatment may predict early recurrence. Given that nearly one-fifth of patients who received neoadjuvant pembrolizumab in KEYNOTE-671 failed to receive surgery [14], our findings imply a potential adaptive treatment strategy that uses blood-based biomarkers to identify patients who are unlikely to benefit from pembrolizumab, facilitating early employment of alternative treatments such as neoadjuvant CCRT. This encouraging and biologically plausible finding merits confirmation in a larger prospective study.
Limitations of our study include the small sample size, incomplete tumor PD-L1 assessment due to the use of archival tissues, a heterogeneous patient population with a high prevalence of EGFR/ALK alterations, and the absence of a control group. The debatable role of neoadjuvant CCRT for stage IIIA/N2 NSCLC also limits the interpretation of the data. However, since immunotherapy just paved its path into the multimodal treatment for operable lung cancer, research into different sequencing and combinations of surgery, radiotherapy, chemotherapy, targeted therapy, and immunotherapy is crucial to determine optimal ICI use in operable NSCLC. The phase 3 ADAURA trial demonstrated significant improvements in DFS and OS in completely resected, EGFR mutation–positive, stage IB-IIIA NSCLC with adjuvant osimertinib [30]. Combining anti-angiogenic agents with ICI and chemotherapy for EGFR-mutant tumors, given its efficacy demonstrated in advanced disease [31], may also be worth investigating in the neoadjuvant setting.
In conclusion, our study suggests that quadrimodal therapy with neoadjuvant CCRT, surgery, and adjuvant pembrolizumab may offer durable disease control in a subset of stage IIIA/N2 NSCLC patients. However, larger, randomized controlled trials are vital to confirm its feasibility and efficacy.
Electronic Supplementary Material
Supplementary materials are available at Cancer Research and Treatment website (https://www.e-crt.org).
Notes
Ethical Statement
The corresponding author conceived and designed the protocol. It was approved by the institutional review board (IRB) of Samsung Medical Center (IRB approval number, 2016-09-018) and the Ministry of Food and Drug Safety, along with all subsequent amendments. All patients provided written informed consent. The study adhered to the ethical principles of the Declaration of Helsinki and the Good Clinical Practice guidelines of the International Conference on Harmonization.
Author Contributions
Conceived and designed the analysis: Shin J, Park S, Ahn MJ.
Collected the data: Shin J, Park S, Kim KH, Shin EC, Jung HA, Cho JH, Sun JM, Lee SH, Choi YS, Ahn JS, Kim J, Park K, Shim YM, Kim HK, Noh JM, Ahn YC, Pyo H, Ahn MJ.
Contributed data or analysis tools: Shin J, Park S, Kim KH, Shin EC, Jung HA, Cho JH, Sun JM, Lee SH, Choi YS, Ahn JS, Kim J, Park K, Shim YM, Kim HK, Noh JM, Ahn YC, Pyo H, Ahn MJ.
Performed the analysis: Shin J, Park S, Kim KH, Ahn MJ.
Wrote the paper: Shin J, Kim KH, Ahn MJ.
Supervised the study, acquired funding, and provided resources: Ahn MJ.
Conflicts of Interest
Yong Chan Ahn, the editor-in-chief of the Cancer Research and Treatment, was not involved in the editorial evaluation or decision to publish this article. All remaining authors have declared no conflicts of interest.
This study is partly supported by a research grant from the Investigator-Initiated Studies Program of Merck Sharp & Dohme Corp. The opinions expressed in this paper are those of the authors and do not necessarily represent those of Merck Sharp & Dohme Crop.
MSD provided pembrolizumab and partial funding but was not involved in data collection, analysis, manuscript writing, or the decision to submit the manuscript for publication.