The longstanding clinical puzzle of when to add chemotherapy to immunotherapy in advanced lung cancer may have a more precise molecular answer. For the roughly 30% of non-small cell lung cancer patients with high PD-L1 expression, current guidelines support immunotherapy alone — yet a substantial subset still progresses rapidly. A liquid biopsy biomarker may now help identify that at-risk group before treatment begins.
Using a nationwide U.S. clinicogenomic database of 965 patients with advanced non-small cell lung cancer, researchers evaluated whether circulating tumor DNA (ctDNA) tumor fraction (TF) could predict differential benefit from immunotherapy alone versus chemo-immunotherapy combinations. A TF threshold of ≥5% emerged as clinically meaningful: patients above this cutoff showed a hazard ratio of 0.58 for real-world progression-free survival when chemotherapy was added to immune checkpoint blockade (95% CI 0.41–0.82, p=0.002), indicating substantially greater benefit from the combined regimen. This 5% cutoff was independently validated in a separate cohort from Gustave Roussy. Among 283 patients with paired PET imaging, ctDNA TF correlated meaningfully with total metabolic tumor volume (rho=0.46, p<0.001), and TF levels were modulated by TP53 and RB1 mutations — genes associated with more aggressive tumor biology.
This work is significant because it positions ctDNA TF as a dual-function biomarker: simultaneously capturing disease burden (a surrogate for tumor volume) and underlying tumor biology (genomic instability markers). PD-L1 immunohistochemistry, the current standard for treatment selection, notoriously misclassifies patients due to spatial and temporal heterogeneity. A blood-based, quantitative biomarker that is agnostic to biopsy site addresses several of those shortcomings. That said, this remains a retrospective analysis using real-world data, which introduces confounding and selection biases that a prospective randomized trial would need to resolve. The correlation between ctDNA TF and metabolic tumor volume is biologically coherent but does not establish causality or define which measure adds more independent predictive value. If validated prospectively, ctDNA TF at a 5% threshold could meaningfully refine treatment escalation decisions, particularly for high PD-L1 patients where oncologists currently rely on clinical intuition over molecular evidence.