The treatment gap Immune checkpoint inhibitors (ICIs) have transformed lung cancer treatment, but many patients do not benefit. Tumors with few immune cells - called immunosuppressed or 'cold' tumors - tend to respond poorly to immunotherapy. Finding ways to warm up these cold tumors is a major research priority.
IL-1beta as a target Interleukin-1beta (IL-1beta) is a pro-inflammatory cytokine that promotes tumor development and creates an immunosuppressive tumor microenvironment (TME). Preclinical studies showed that blocking IL-1beta with canakinumab could shift the TME toward immune activation and improve the effect of ICIs like pembrolizumab.
The CANOPY trials Two randomized clinical trials evaluated canakinumab plus pembrolizumab in NSCLC: CANOPY-1 (first-line advanced disease, 643 patients) and CANOPY-N (neoadjuvant resectable disease). Neither trial showed a benefit from adding canakinumab in the overall patient population. This study reports exploratory biomarker analyses to understand which patients might benefit.
Key finding In CANOPY-1, patients with low CD8+ T-cell infiltration (an immunosuppressive TME) did benefit from canakinumab addition: progression-free survival hazard ratio was 0.53 in the low CD8 group, suggesting a 47% reduction in progression risk. High CD8 patients showed no benefit (HR 1.01). In CANOPY-N, canakinumab treatment reduced immunosuppressive cell populations in the TME.
T-cell infiltration and response The tumor immune microenvironment is not uniform across patients. Tumors with high CD8+ cytotoxic T-cell infiltration (inflamed phenotype) generally respond better to immunotherapy than tumors with few T cells (desert phenotype) or T cells trapped at the tumor border (excluded phenotype).
IL-1beta and immunosuppression High IL-1beta in tumors has been linked to poor prognosis in NSCLC. IL-1beta promotes tumor-associated inflammation that recruits immunosuppressive cells and inhibits effective antitumor immunity. Blocking it could theoretically convert cold, immunosuppressed tumors into immunotherapy-responsive ones.
Prior evidence for canakinumab Data from the CANTOS cardiovascular trial showed a dose-dependent reduction in lung cancer incidence and mortality with canakinumab, providing a rationale for the CANOPY program. The CANOPY-1 and CANOPY-N trials were among four CANOPY trials designed to test canakinumab in NSCLC.
Why negative trials still provide insight Although CANOPY-1 and CANOPY-N did not meet their primary endpoints in the overall population, they collected extensive baseline and post-treatment tumor samples with matched outcome data - creating a unique dataset to discover which TME features predict benefit from IL-1beta blockade.
IHC phenotype classification In CANOPY-1, CD8+ T-cell infiltration was measured using a duplex chromogenic IHC assay with anti-CD8 and anti-panCK antibodies. Tumors were classified into three T-cell phenotypes: inflamed (CD8 in carcinoma greater than 1.3%), excluded (CD8 stroma:tumor ratio greater than 4), or desert (neither criterion met).
Multiplex immunofluorescence in CANOPY-N Pre-treatment and post-surgery tumor samples from CANOPY-N were analyzed by multiplex immunofluorescence using panels targeting CD3, CD11b, CD19, CD66b (neutrophils), CD163 (macrophages), CD56, FOXP3 (regulatory T cells), and granzyme B. This allowed comprehensive characterization of immune cell populations and their changes with treatment.
Transcriptomic analyses For 298 CANOPY-1 patients with available tissue, RNA sequencing was performed. An 18-gene T-cell-inflamed signature was used to classify patients as T-cell high or T-cell low. Additional TME gene signatures from published and in-house datasets were analyzed by principal component analysis to characterize the overall immune environment.
Statistical framework Cox proportional hazards models estimated hazard ratios for PFS and OS within each CD8 subgroup and T-cell phenotype group, adjusting for stratification factors (PD-L1, histology, geographic region). All subgroup analyses were pre-specified exploratory analyses, not primary endpoints.
CD8 level as a prognostic factor In CANOPY-1, high CD8+ T-cell infiltration was independently prognostic: high CD8 patients had longer PFS (HR 0.73) and OS (HR 0.76) compared to low CD8 patients, regardless of treatment arm. Patients with the inflamed T-cell phenotype had the longest median OS (23.08 months), compared to desert and excluded phenotypes.
Canakinumab benefit in CD8-low patients Among patients with low CD8+ T-cell infiltration (n=250), adding canakinumab to pembrolizumab-based chemotherapy significantly improved PFS (HR 0.53, 95% CI 0.36-0.78). A numerical OS benefit trend was also observed (HR 0.72). In contrast, CD8-high patients showed no benefit (PFS HR 1.01).
Desert phenotype subgroup Patients with the immune desert T-cell phenotype - the most immunosuppressed group - showed the strongest numerical trends toward benefit with canakinumab (PFS HR 0.47, OS HR 0.58), though with wide confidence intervals due to the smaller sample size. The excluded phenotype group showed no benefit.
PD-L1 does not modify the CD8 benefit Both PD-L1-negative and PD-L1-positive patients in the CD8-low subgroup showed longer PFS with canakinumab, suggesting the CD8 level is a more specific predictor of canakinumab benefit than PD-L1 expression alone. Patients with PD-L1-negative and low CD8 tumors showed an OS HR of 0.59 with canakinumab addition.
RNA-seq confirms T-cell phenotypes For patients with available transcriptomic data, the T-cell-inflamed gene signature was significantly enriched in CD8-high patients, while a fibroblast gene signature was enriched in CD8-excluded phenotype patients - validating the IHC-based phenotype classification using independent molecular data.
Immunosuppressive TME in CD8-low tumors Principal component analysis of TME gene signatures revealed that CD8-low tumors had higher scores for immunosuppressive signatures, including those associated with tumor-associated macrophages (TAMs), regulatory T cells, myeloid-derived suppressor cells, and TGF-beta signaling - defining a broader immunosuppressive environment.
CD8-high tumors are immune-inflamed Conversely, CD8-high tumors were enriched for signatures of T-cell activation, cytotoxic T-cell activity, NK cell presence, and interferon-gamma signaling - consistent with an immune-activated TME that already responds well to ICI without needing IL-1beta blockade.
IL-1beta target biology supported These findings support the hypothesis that IL-1beta-driven immunosuppression disproportionately affects cold, CD8-low tumors, and that blocking IL-1beta may specifically help convert these immunosuppressive microenvironments into states more permissive to ICI therapy.
Pre-to-post treatment changes In CANOPY-N, tumor samples were collected before treatment and at the time of surgery after neoadjuvant canakinumab with or without pembrolizumab. This design allowed direct measurement of how canakinumab changes the TME immune landscape in real patient tumors.
Reduction in immunosuppressive populations Treatment with canakinumab plus pembrolizumab was associated with a reduction in immunosuppressive cell populations in the TME, including CD163+ tumor-associated macrophages, regulatory T cells (FOXP3+), and CD66b+ neutrophils - cells that suppress antitumor immune responses.
Changes in CD8+ T cells While the CANOPY-N trial did not show a significant increase in major pathologic response with canakinumab addition, the biomarker data suggest that canakinumab treatment can modulate the cellular composition of the TME, reducing immunosuppressive barriers that might limit ICI efficacy.
Mechanistic support The CANOPY-N findings provide direct mechanistic evidence that canakinumab can shift the TME in NSCLC patients - reducing barriers to immune activation. This supports the hypothesis that in the right patient population (CD8-low, immunosuppressed tumors), this shift may translate into clinical benefit.
Exploratory analyses only All biomarker analyses reported here were pre-specified exploratory analyses, not primary endpoints of the trials. The interactions between treatment benefit and CD8 subgroup were not statistically significant in formal interaction testing, limiting the strength of conclusions that can be drawn.
Sample size constraints The subgroup analyses were limited by sample size, particularly for the desert phenotype group (n=81). The confidence intervals on hazard ratios in key subgroups are wide, meaning the findings require confirmation in larger, prospectively powered biomarker studies.
Optimal patient selection still undefined While low CD8+ T-cell infiltration identified a subgroup with potential canakinumab benefit, the exact combination of biomarkers (CD8, PD-L1, TME gene signature score) that would optimally select patients for IL-1beta blockade is not yet defined. Composite biomarker panels may be needed.
Future directions Future trials should prospectively test canakinumab or other IL-1beta inhibitors in NSCLC patients selected for low CD8 or immunosuppressive TME features as the primary biomarker eligibility criterion. Integration of IHC, RNA-seq, and TME profiling into clinical selection algorithms represents the translational path forward.