Predicting Immunotherapy Response in Non-Small Cell Lung Cancer

The Leading Cause of Cancer Death—With an Imperfect Biomarker

Lung cancer remains the leading cause of cancer death worldwide, claiming nearly 1.8 million lives annually. Non-small cell lung cancer (NSCLC) accounts for approximately 85% of lung cancer cases. For patients with advanced disease, the outlook was historically grim.

Checkpoint inhibitors changed the landscape. Drugs targeting PD-1 and PD-L1 can produce durable responses, with some patients achieving long-term survival that would have been unthinkable a decade ago. These treatments have become standard of care for advanced NSCLC.

For the 70-80% of patients who don't respond, this means months of treatment with immune-related adverse events, without the therapeutic benefit. For healthcare systems, it means billions spent on ineffective therapy. There has to be a better way.

From Expression to Interaction

The logic is straightforward: if anti-PD-1/PD-L1 drugs work by blocking the interaction between PD-1 and PD-L1, the biomarker should measure whether that interaction is actually occurring.

PD-L1 expression testing counts protein molecules on tumor cells. But expression doesn't guarantee function. The protein might be in the wrong location, improperly folded, or simply not engaged with its binding partner on T cells. Counting molecules tells you about potential, not activity.

The study was published in the Journal of Clinical Oncology in 2023, analyzing FFPE tissue samples from 188 NSCLC patients treated with anti-PD-1/PD-L1 immunotherapy. Each sample was assessed for both traditional PD-L1 expression (22C3 pharmDx assay) and PD-1/PD-L1 engagement by iFRET.

A Three-Fold Difference in Survival

The results were unambiguous. Patients stratified by checkpoint engagement showed dramatically different outcomes:

42% of Patients Miscategorized by Standard Testing

Perhaps the most striking finding was the discordance between iFRET status and PD-L1 expression. When comparing the two biomarkers, the tests gave different answers in 42% of patients:

This means nearly one-quarter of patients who could benefit from immunotherapy would be excluded based on current companion diagnostics. Meanwhile, 18% of patients are receiving treatment they're unlikely to benefit from, based on a test that measures the wrong thing.

The biological explanation is clear: expression without engagement means the checkpoint isn't active. High PD-L1 expression with low interaction may indicate protein localized away from the tumor-T cell interface, or PD-L1 present but not engaging infiltrating T cells. Either way, blocking an interaction that isn't happening won't help.

Specificity for Immunotherapy Response

A critical question: does iFRET predict better outcomes in general, or specifically predict response to checkpoint blockade? The study addressed this with a chemotherapy control cohort.

In patients treated with chemotherapy alone, iFRET status showed no association with outcomes. This confirms that the biomarker specifically predicts response to drugs that block PD-1/PD-L1 interaction—not just general prognosis. The mechanism matches the measurement.

Additional subgroup analyses confirmed:

• First-line therapy: Predictive value was strongest (HR = 0.31, P < 0.001)
• Histology: Value maintained across adenocarcinoma and squamous cell subtypes
• Multivariate analysis: iFRET remained significant after adjusting for clinical factors

Beyond the Statistics

The 21-month difference in median survival isn't just a number. It's the difference between seeing a child graduate, meeting a grandchild, having time to put affairs in order. For the 24% of patients with favorable iFRET status but low PD-L1 expression, current testing could mean being denied access to a treatment that might give them years rather than months.

Building the Evidence