Neoadjuvant atezolizumab for resectable non-small cell lung cancer: an open-label, single-arm phase II trial

Abstract

In an ongoing, open-label, single-arm phase II study ( NCT02927301 ), 181 patients with untreated, resectable, stage IB-IIIB non-small cell lung cancer received two doses of neoadjuvant atezolizumab monotherapy. The primary end point was major pathological response (MPR; ≤10% viable malignant cells) in resected tumors without EGFR or ALK alterations. Of the 143 patients in the primary end point analysis, the MPR was 20% (95% confidence interval, 14-28%). With a minimum duration of follow-up of 3 years, the 3-year survival rate of 80% was encouraging. The most common adverse events during the neoadjuvant phase were fatigue (39%, 71 of 181) and procedural pain (29%, 53 of 181), along with expected immune-related toxicities; there were no unexpected safety signals. In exploratory analyses, MPR was predicted using the pre-treatment peripheral blood immunophenotype based on 14 immune cell subsets. Immune cell subsets predictive of MPR in the peripheral blood were also identified in the tumor microenvironment and were associated with MPR. This study of neoadjuvant atezolizumab in a large cohort of patients with resectable non-small cell lung cancer was safe and met its primary end point of MPR ≥ 15%. Data from this single-arm, non-randomized trial suggest that profiles of innate immune cells in pre-treatment peripheral blood may predict pathological response after neoadjuvant atezolizumab, but additional studies are needed to determine whether these profiles can inform patient selection and new therapeutic approaches.

Fig. 1. Patient disposition.

Primary efficacy population is bolded. ^aTwo patients were determined to have hemangioma and solitary fibrous tumor at resection despite initial pathology consistent with NSCLC. ^bIncludes one EGFR-positive patient. ^cThe reasons were clinical progression (n = 3), physician did not want to delay patient surgery (n = 1), physician did not consider the patient a good surgical candidate (n = 1) and physician discontinued patient from the study because of an AE (n = 1). ^dOne patient was determined to have pre-existing congestive heart failure, one declined surgery and one was lost to follow-up.

Fig. 2. Clinical outcomes in patients who had surgical resection and whose tumors did not have known EGFR or ALK alterations.

a, Pathological response (n = 143). Pathological regression is defined as percentage viable tumor cells – 100%. b, DFS by MPR status in patients with R0 resections (n = 137). c, OS by MPR status in patients with R0 resections (n = 137). HR, hazard ratio.

Fig. 3. Performance of GAM–LASSO MPR predictive models.

a, Use of the GAM–LASSO model to predict MPR in test set 2, which consisted of patients within LCMC3 who were not included in either the training set (n = 57) or test set 1 (n = 54). MPR was not assessed in these nine patients because of no resection. The MPR and non-MPR cohorts derived from the merge of the model’s training set and test set 1. The maximum and minimum values of the boxes denote the IQR. The line within the IQR denotes the median. The extremities of the dashed lines represent the minimum and maximum values of the data, which are 1.5× below the first quartile and 1.5× above the third quartile. The parameters for null hypothesis testing via analysis of variance (ANOVA) were as follows: d.f. = 2, total sum of squares = 1.976, mean squares = 0.988, F-value = 32.799 and Pr(>F) = 4.914 × 10⁻¹². The statistical details for the comparison of MPR and non-MPR were t = −5.47, d.f. = 27.02, two-sided P = 8.6 × 10⁻⁶ and 95% CI = −0.439 to −0.200. The statistical details for the comparison of MPR and PD were t = −3.18, d.f. = 28.45, two-sided P = 0.0035 and 95% CI = −0.383 to −0.083. The statistical details for the comparison of non-MPR and PD were t = −1.77, d.f. = 9.52, two-sided P = 0.11 and 95% CI = −0.195 to 0.023. No adjustment was made for multiplicity. b, ROC curves for the training set and test set 1. The dashed y = x line, which represents random assignment, is included for reference. ^aImmunophenotyping via flow cytometry. IQR, interquartile range; ROC, receiver operating characteristic.

Fig. 4. scRNA-seq analysis of selected genes expressed in tumor tissue from 15 patients following treatment with neoadjuvant atezolizumab.

a, The expression of different NK cell surface receptors was determined by scRNA-seq. b, Tumor samples collected at resection were classified into three groups of five samples each on the basis of the percentage of viable tumor cells by pathological analysis: low (≤25% viable tumor cells), middle (26–50%) and high (>50%). Dot size represents the percentage of NK cells in the group expressing the gene. Color represents the scaled average normalized expression. NK cells were downsampled to have the same number of cells in each group. ILT2 is also known as LILRB1, ILT4 as LILRB2, KIR2DL1 as CD158a, NKG2D as CD314 and KLRK1 and PD-L1 as CD274. CD, cluster of differentiation; CMP, common myeloid progenitor; ILT, immunoglobulin-like transcript; KIR2DL1, killer cell immunoglobulin-like receptor DL1; LILRB, leukocyte immunoglobulin-like receptor subfamily B; NKG2, natural killer group protein 2.

Extended Data Fig. 1. Study design.

*Mandatory. CT, computed tomography; NSCLC, non-small cell lung cancer; PET, positron emission tomography; q3m, every 3 months; SOC, standard of care.

Extended Data Fig. 2. CONSORT diagram for the various biomarker analyses.

^aTwo patients were determined to have hemangioma and solitary fibrous tumor at resection despite initial pathology consistent with NSCLC. ^bIncludes 1 EGFR-positive patient. ^cThe reasons were as follows: clinical progression (n = 3), physician did not want to delay patient surgery (n = 1), physician did not consider the patient a good surgical candidate (n = 1), and physician discontinued patient from the study because of an adverse event (n = 1). ^dOne patient was determined to have preexisting congestive heart failure, 1 declined surgery, and 1 was lost to follow-up. ^eEighty patients provided a total of 98 samples for the bulk RNA-seq analyses (54 at baseline and 44 at surgery). ALK, anaplastic lymphoma kinase; EGFR, epidermal growth factor receptor; ES, exome sequencing; MPR, major pathologic response; NSCLC, non-small cell lung cancer; PD, progressive disease; RECIST, Response Evaluation Criteria in Solid Tumors; RNA-seq, ribonucleic acid sequencing; scRNA-seq, single-cell ribonucleic acid sequencing; TMB, tumor mutational burden.

Extended Data Fig. 3. Relationship between MPR, histology, and TMB.

(a) Pathologic response by mutation status (n = 85). TMB showed a positive but not significant trend with pathologic response in non-squamous (R = 0.28; P = 0.05) and squamous (R = 0.23; P = 0.22) tumors. KEAP1 mutations were not significantly associated with pathologic response in either non-squamous (P = 0.46) or squamous tumors (P = 0.98). KRAS and STK11 mutations were found only in non-squamous tumors, where mutations in STK11 were found to significantly associate with pathologic response (P = 0.01), and mutations in KRAS were not (P = 0.87). (b) Pathologic response by the overlapping mutational status of KRAS and STK11 in patients with non-squamous NSCLC (n = 78). The maximum and minimum values of the boxes denote the IQR. The line within the IQR denotes the median. The extremities of the dashed lines represent the 5^th to 95^th percentiles. *TMB was only determined for the subset of patients with ES data from baseline and/or surgery with tumor purity ≥15%. P values for TMB in Panel a were determined via linear correlation test (Pearson). P values for mutation status in Panels a and b were determined via two-sided Wilcoxon rank sum test. ES, exome sequencing; IQR, interquartile range; KEAP1, Kelch-like ECH-associated protein 1; KRAS, Kirsten rat sarcoma viral oncogene homolog; MPR, major pathologic response; NS, non-significant; NSCLC, non-small cell lung cancer; STK11, serine/threonine kinase 11; TMB, tumor mutational burden.

Extended Data Fig. 4. DFS (a) and OS (b) by disease stage, use of adjuvant atezolizumab, and lymph node status (n = 143).

The MPR rate in patients who did and did not receive adjuvant atezolizumab was 41% (22/54) and 8% (7/89), respectively. DFS, disease-free survival; OS, overall survival.

Extended Data Fig. 5. Gating strategy for the IMMUNOME tubes described in Supplementary Table 5.

(a) Tubes 1–11 were gated on the basis of CD45 and side scatter. The immune cell populations of interest were those falling within the region ‘lymphocytes,’ which was defined by CD45^bright, a low side scatter, and a lymphosum (CD3⁺ CD19⁺ CD56⁺ CD16⁺) totaling 95%–100%, with <5% myeloid contamination. Tube 1 (lymphosum) contained lymphosum markers, as well as myeloid markers (CD13/CD14). The gate established for Tube 1 was applied to Tubes 2–11. (b) Tube 12 (myeloid cells) used LIN (CD3⁺ CD19⁺ CD56⁺) as an exclusion gate. Plots were gated using three different strategies: LIN⁻ total, LIN⁻ CD11b⁺, and LIN⁻ CD33⁺. (c) Tube 13 (senescent cells) used both lymphocyte and sequential gating to find the immune cell population of interest. Lymphocyte gating isolated CD28⁻ CD16⁻ CD56⁻ CD3⁺ cells, which were sequentially gated on CD57. Events falling with the CD57⁺ region were considered senescent cells. Senescent cells were further classified into subsets defined by positivity and negativity for CD4, CD8, KLRG1, and CD127. (d) Tube 14 (dendritic cells) used LIN as an exclusion gate. Plots were gated using three different strategies: LIN⁻ total, LIN⁻ CD1c⁺, and LIN⁻ CD141⁺. CD, cluster of differentiation; KLRG1, killer cell lectin-like receptor subfamily G member 1; LIN, lineage.

Extended Data Fig. 6. ROC curves for test set 1 (n = 54), which considered only peripheral blood IMMUNOME, as well as for test set 1 in combination with histology (non-squamous [n = 32] vs. squamous [n = 22]), nodal status (N1/N2 [n = 35] vs. N0 [n = 19]), PD-L1 expression (n = 54), sex (female [n = 26] vs. male [n = 28]), or smoking status/history (never [n = 6] vs. current/former [n = 48]).

The dashed y = x line, which represents random assignment, is included for reference. ^aImmunophenotyping via flow cytometry. PD-L1, programmed death-ligand 1; ROC, receiver operating characteristic.

Extended Data Fig. 7. Immune cell subsets with significant (two-sided p < 0.05) changes in abundance before (baseline) and after (at surgery) neoadjuvant treatment with atezolizumab. No adjustment was made for multiplicity.

A, CD45⁺ CD62L⁺ CD27⁺ CD56/16⁻ CD45RO⁺ CCR7⁻ CD45RA⁻ CD4⁻ CD3⁺ CD8⁺ (p = 0.0081); B, CD45⁺ CD33⁺ HLA-DR⁺ CD124⁻ CD14⁺ CD11b⁺ CD66b⁻ CD16⁻ CD33⁺ CD15⁺ (p = 0.0042); C, CD45⁺ HLA-DR⁺ CD124⁻ CD14⁺ CD11b⁺ CD66b⁻ CD16⁻ CD33⁺ CD15⁺ (p = 0.0042); D, CD45⁺ CD94⁺ NKG2D⁺ CD3⁻ CD56⁺ CD117⁻ NKG2A⁺ CD127⁻ CD161⁻ CD16⁺ (p = 0.0036); E, CD45⁺ LIN⁺ HLA-DR⁺ CD33⁺ CD16⁺ CD11b⁺ CD15⁺ (p = 0.0092); F, CD45⁺ HLA-DR⁻ CD69⁺ CD19⁻ CD56⁺ CD16⁻ CD134⁻ CD4⁻ CD3⁺ CD8⁺ (p = 0.0079); G, CD45⁺ CD11b⁺ HLA-DR⁺ CD124⁻ CD14⁺ CD11b⁺ CD66b⁻ CD16⁻ CD33⁺ CD15⁺ (p = 0.0157); H, CD45⁺ γ/δ^- α/β⁺ CD19⁻ CD56⁺ CD16⁻ CD13/14⁻ CD4⁺ CD3⁺ CD8⁻ (p = 0.0113); I, CD45⁺ HLA-DR⁻ CD69⁺ CD3⁺ KIR3DL1⁻ KIR2DL2⁻ NKG2A⁻ CD56⁺ KIR2DL1⁻ CD16⁻ (p < 0.0001); J, CD45⁺ CD107a/b^- CD159c⁺ CD3⁻ KIR3DL1⁻ KIR2DL2⁻ NKp80⁺ CD56⁺ KIR2DL1⁻ CD16⁺ (p = 0.0025); K, CD45⁺ HLA-DR⁻ CD69⁺ CD19⁻ CD56⁺ CD16⁻ CD134⁻ CD4⁻ CD3⁺ CD8⁻ (p = 0.0004); L, CD45⁺ CD16⁺ CD336⁻ CD3⁻ CD244⁺ CD335⁺ NKG2D⁺ CD56⁻ CD161⁺ CD337⁻ (p = 0.0002); M, CD45⁺ γ/δ^- α/β⁺ CD19⁻ CD56⁺ CD16⁻ CD13/14⁻ CD4⁻ CD3⁺ CD8⁻ (p = 0.0191); N, CD45⁺ HLA-DR⁻ CD69⁺ CD3⁺ KIR3DL1⁻ KIR2DL2⁻ NKG2A⁺ CD56⁺ KIR2DL1⁻ CD16⁻ (p = 0.0307); O, CD45⁺ CD62L⁻ CD27⁺ CD56/16⁺ CD45RO⁻ CCR7⁻ CD45RA⁺ CD4⁻ CD3⁺ CD8⁺ (p = 0.0400); P, CD45⁺ CD94⁻ NKG2D⁻ CD3⁺ CD56⁺ CD117⁻ NKG2A⁻ CD127⁺ CD161⁺ CD16⁺ (p = 0.0480); Q, CD45⁺ HLA-DR⁺ CD69⁺ CD3⁻ KIR3DL1⁻ KIR2DL2⁻ NKG2A⁺ CD56⁻ KIR2DL1⁻ CD16⁺ (p = 0.0014); R, CD45⁺ LIN⁻ HLA-DR⁻ CD33⁻ CD16⁺ CD11b⁺ CD15⁺ (p = 0.0030). *Pre-treatment predictor of MPR. LIN included CD19, CD3, and CD56. NKG2A is also known as CD159a, NKG2D as CD314 and KLRK1, KIR2DL1 as CD158a, KIR2DL2 as CD158b, CD335 as NKp46, and CD337 as NKp30. α/β, α/β chains of the T cell receptor; γ/δ, γ/δ chains of the T cell receptor; CCR7, C-C motif chemokine receptor 7; CD, cluster of differentiation; HLA, human leukocyte antigen; KIR, killer cell immunoglobulin-like receptor; KLR, killer cell lectin-like receptor; LIN, lineage; MPR, major pathologic response; NK, natural killer; NKG2, natural killer group protein 2.

Extended Data Fig. 8. Association between gene expression in NK cells and the percentage of viable tumor cells.

The expression of different NK cell surface receptors was determined by scRNA-seq. Tumor samples collected at resection were classified into 3 groups of 5 samples each based on the percentage of viable tumor cells by pathologic analysis: low (≤25% viable tumor cells), middle (26%–50%), and high (>50%). UMAP plots of NK cells show normalized expression of several receptors in the 3 groups of viable tumor cell numbers. NK cells were down-sampled to have the same number of cells in each group. ILT2 is also known as LILRB1, ILT4 as LILRB2, NKG2A as CD159a, NKG2D as CD314 and KLRK1, KIR2DL1 as CD158a, KIR2DL3 as CD158b, KIR2DL4 as CD158d, KIR3DL1 as CD158e1, and KIR3DL2 as CD158k. CD, cluster of differentiation; ILT, immunoglobulin-like transcript; KIR, killer cell immunoglobulin-like receptor; KLR, killer cell lectin-like receptor; LILRB, leukocyte immunoglobulin-like receptor subfamily B; max, maximum; min, minimum; NK, natural killer; NKG2, natural killer group protein 2; scRNA-seq, single-cell ribonucleic acid sequencing; UMAP, uniform manifold approximation and projection.

Extended Data Fig. 9. Association of (a) ILT2 and (b) PD⁻L1 transcripts with the cell type enrichment scores of specific immune cell subsets derived from bulk RNA-seq at baseline and at surgery by tumor histology.

Relative immune cell abundance was estimated from RNA-seq data using the cell type enrichment analysis tool xCell and correlated with the abundance (log₂ cpm) of ILT2 or PD-L1. Results are shown for all baseline and surgical samples with RNA-Seq, split by histology. A positive correlation coefficient (blue) indicates samples with an increased abundance of a given immune cell signature also having an increased abundance of either ILT2 or PD-L1. Asterisks indicate significant correlations (unadjusted Pearson correlation two-sided P < 0.05). ILT2 is also known as LILRB1 and PD-L1 as CD274. CD, cluster of differentiation; cpm, counts per million reads mapped; DC, dendritic cell; ILT2, immunoglobulin-like transcript 2; LILRB, leukocyte immunoglobulin-like receptor subfamily B; NK, natural killer; PD-L1, programmed death-ligand 1; RNA-seq, ribonucleic acid sequencing.

Extended Data Fig. 10. Association of ILT2 and PD-L1 transcripts with pathologic response at baseline and at surgery by tumor histology.

Expression values for each transcript are presented as log₂ (cpm + 1). Two-sided P-values for Pearson (R) and Spearman (ρ [Rho]) correlation are shown. The grey band represents the 95% confidence interval for the linear regression line. No adjustment was made for multiplicity. Of the 54 patients at baseline with RNA-seq data, 52 also had pathologic response data available. ILT2 is also known as LILRB1 and PD-L1 as CD274. CD, cluster of differentiation; cpm, counts per million reads mapped; ILT2, immunoglobulin-like transcript 2; LILR1, leukocyte immunoglobulin-like receptor 1; MPR, major pathologic response; PD-L1, programmed death-ligand 1; RNA-seq, ribonucleic acid sequencing.