Functional Engagement of the PD-1/PD-L1 Complex But Not PD-L1 Expression Is Highly Predictive of Patient Response to Immunotherapy in Non–Small-Cell Lung Cancer
What This Study Shows
This landmark study of 135 lung cancer patients demonstrates that measuring whether immune checkpoint proteins are actually interacting—rather than just counting how much protein is present—dramatically improves our ability to predict who will benefit from immunotherapy.
Patients whose tumors showed high PD-1/PD-L1 interaction lived three times longer than those with low interaction. Meanwhile, the standard test (PD-L1 expression) failed to predict survival at all.
Key Findings
Survival with High Interaction
Patients with high PD-1/PD-L1 interaction had a median overall survival of 31 months when treated with immunotherapy.
Survival with Low Interaction
Patients with low interaction had only 10 months median survival—a 3-fold difference.
Missed by Standard Testing
Nearly a quarter of patients had high interaction but low PD-L1 expression—they would be excluded from therapy by current companion diagnostics.
Why This Matters
Current tests measure how much PD-L1 protein is present on tumor cells. But presence doesn't mean function—a protein can be highly expressed but not actually doing anything.
The functional biomarker approach measures actual molecular interaction at nanometer scale, revealing which tumors are truly using the PD-1/PD-L1 checkpoint to evade the immune system.
Clinical Implications
- 🎯 Better patient selection: Identify patients most likely to respond to checkpoint inhibitors, avoiding unnecessary treatment and toxicity for non-responders.
- 🔓 Rescue missed patients: The 24% of patients with high interaction but low expression could benefit from immunotherapy but are currently excluded.
- 💰 Healthcare economics: Checkpoint inhibitors cost $150,000+ per year. Better selection improves cost-effectiveness and reduces wasted treatment.
Learn More
Explore the glossary to understand the technology behind these measurements:
Abstract
Immune checkpoint inhibitors (ICIs) targeting the PD-1/PD-L1 axis have transformed NSCLC treatment, yet only 20-30% of patients achieve durable responses. Current companion diagnostics based on PD-L1 immunohistochemistry (IHC) show inconsistent predictive value. This retrospective study evaluated whether quantitative measurement of PD-1/PD-L1 molecular interaction using QF-Pro could better predict clinical outcomes in NSCLC patients treated with anti-PD-1/PD-L1 therapy.
Methods
Study Design: Retrospective cohort study of 135 NSCLC patients treated with anti-PD-1/PD-L1 immunotherapy at multiple centers in Spain, France, and the Netherlands. FFPE tumor samples were analyzed using the QF-Pro platform.
iFRET Methodology: Two-site cell-cell amplified FRET was performed using donor-conjugated anti-PD-1 and acceptor-conjugated anti-PD-L1 antibodies. FLIM measurements were acquired with time-correlated single photon counting (TCSPC) at 256×256 pixel resolution. FRET efficiency was calculated from donor lifetime quenching: E = 1 - (τDA/τD).
Comparator: PD-L1 expression was assessed by IHC (SP263 Ventana, Roche) with TPS ≥50% as the positive threshold per EMA criteria.
Endpoints: Primary endpoint was overall survival (OS). Secondary endpoints included progression-free survival (PFS) and objective response rate (ORR).
Key Findings
Median OS: High iFRET
Patients stratified as iFRET-high demonstrated median OS of 31 months (95% CI: 24.2-37.8).
Median OS: Low iFRET
iFRET-low patients had median OS of 10 months (95% CI: 7.1-12.9), HR = 0.38 (95% CI: 0.25-0.58).
Statistical Significance
Log-rank test for OS stratified by iFRET status achieved P < 0.0001. Multivariate Cox regression confirmed independent prognostic value.
"PD-L1 expression (TPS ≥50%) showed no significant association with overall survival (P = 0.162), while iFRET-determined PD-1/PD-L1 interaction strongly predicted outcomes (P < 0.0001)."
Discordance Analysis
Analysis of concordance between iFRET status and PD-L1 TPS revealed significant discordance:
This 42% total discordance rate underscores the fundamental limitation of expression-based biomarkers: protein quantity does not equal protein function.
Molecular Rationale
The superior predictive performance of iFRET reflects its mechanistic basis:
- 📐 Distance constraint: FRET occurs only when PD-1 and PD-L1 are within 1-10nm (Förster radius ~5nm for the fluorophore pair), confirming direct molecular engagement.
- 🔬 Two-site specificity: The coincidence detection requirement (both antibodies must bind) eliminates false positives from non-specific binding or autofluorescence.
- ⚡ Functional readout: iFRET measures the immune evasion mechanism directly—actual checkpoint engagement—rather than potential for engagement (expression).
Subgroup Analyses
First-line vs. subsequent therapy: The predictive value of iFRET was strongest in the first-line setting (HR = 0.31, P < 0.001), consistent with less pre-treated, more immunotherapy-naive tumor microenvironments.
Histology: Predictive value was maintained across adenocarcinoma and squamous cell carcinoma subtypes.
Specificity: In a chemotherapy-treated control cohort, iFRET status showed no association with outcomes, confirming specificity for ICI response prediction.