Fluid Evidence

Of the more than 200 different types of cancer we have identified, those that affect the skin and lung are among the most common. Together, they represent about one-fifth of all new cancer diagnoses (1)(2)(3)(4). Immune checkpoint inhibitors, drugs that boost the body’s T-cell anti-tumor response by removing the brakes that typically prevent the immune system from killing tumor cells, have greatly improved clinical outcomes for patients with melanoma and non-small cell lung cancer (NSCLC). One commonly targeted molecular brake is programmed cell death protein-1 (PD-1), a receptor found on the surface of T cells. By preventing PD-1 from binding to its target on cancer cells, programmed death ligand-1 (PD-L1) allows T cells to attack the tumor – which is why anti-PD-1 antibodies, such as pembrolizumab and nivolumab, are now approved first-line therapies for patients with advanced melanoma and NSCLC (5)(6). Clinical pathologists measure the expression of PD-L1 in tumor tissue to identify patients who score high for PD-L1 expression and therefore might derive the most benefit from immunotherapy.

Using PD-L1 as a predictive biomarker for patient selection is challenging because the immunohistochemistry (IHC) tests used to determine the presence of PD-L1 in tumor biopsies are not standardized – in fact, the four FDA-approved IHC tests for PD-L1 expression use different antibodies, detection systems, scoring systems, and thresholds (7)(8)(9). But is there another way? Droplet digital PCR (ddPCR) may help. This form of PCR is highly reproducible across different labs (10), is optimized for rapid, minimally invasive liquid biopsy (11), and allows for both increased sensitivity and precision due to its ability to partition a sample into thousands of droplets (12).

Nevertheless, PD-L1’s reliability as a predictive biomarker requires further characterization. Its efficacy as a biomarker may be limited to patients with specific disease characteristics that are not yet well understood. PD-L1 levels may also change over time or as a result of prior treatments, suggesting that a single assessment from a tissue biopsy at diagnosis might be insufficient to inform ongoing therapy. To investigate how well PD-L1 levels correlate with treatment outcomes, researchers from the University of Pisa used ddPCR to measure PD-L1 mRNA levels in liquid biopsies (13). They analyzed plasma-derived exosomes, a source of intact mRNA involved in cancer cell signaling and immunity.

The researchers evaluated changes in PD-L1 expression at baseline and at two months in patients with advanced cancer treated with nivolumab and pembrolizumab. They found that PD-L1 levels correlated significantly with treatment response. Complete and partial responders had the highest levels of PD-L1 expression at baseline and a significant reduction of PD-L1 levels after treatment. Patients with stable disease exhibited lower levels of baseline PD-L1 expression and did not show significant changes in levels after treatment, whereas patients with progressive disease had the lowest baseline levels of PD-L1 and a significant increase after treatment.

It seems clear that patients with elevated levels of PD-L1, decreasing soon after treatment, might derive the most benefit from anti-PD-L1/ PD-1 immunotherapies. By dynamically evaluating PD-L1 mRNA from exosomes via ddPCR, we may be able to obtain useful information on clinical outcomes in cancer patients – information that can be continuously updated as patients undergo treatment.