Contrary to its role in lung cancer, PD-L1 expression does not predict benefit from immunotherapy in mesothelioma. Data from major trials shows similar outcomes regardless of PD-L1 status, leading clinicians to omit this test entirely and streamline treatment decisions.

While the need for prospective trials dominates the ctDNA discussion, a more fundamental obstacle is the lack of standardization between assay types (e.g., tumor-informed vs. agnostic). Without a common measurement approach, data from disparate trials cannot be pooled to create a universally accepted surrogate endpoint for regulatory approval.

An analysis of over 17,000 oncology drug development trajectories revealed that trials incorporating biomarkers had almost twice the overall success probability (10%) compared to those without (5%). This success boost is most significant in early-phase (Phase 1 and 2) trials.

Given that PD-L1 scores for gastroesophageal cancers can be exceptionally variable between labs, some clinicians prefer a simple CPS cutoff of 1. This 'some expression versus no expression' approach is considered more reproducible and practical for decision-making than relying on specific higher scores that may not be consistent across different testing sites.

Beyond clinical validation, the adoption of novel biomarkers like microRNA is hindered by practical lab issues. Disagreements over sample type (serum vs. plasma), establishing universal cutoffs, and achieving high concordance between different testing centers are critical, non-clinical hurdles that must be overcome for widespread clinical use.

Clinicians must recognize that liquid and solid biopsies show significant discordance. ESR1 mutations are more frequently detected in liquid assays, while PIK3CA mutations are more often found in solid tissue. This variability by gene directly impacts the optimal testing strategy for patients.

An expert argues forcefully that the PD-L1 biomarker should be "ditched" in bladder cancer. Citing its repeated failure to predict overall survival benefit across multiple major trials, it is deemed an oversimplified and unreliable tool that leads to both over- and under-treatment of patients.

Despite stratifying patients by PD-L1 status, the AGO-OV-229 trial found it was not a predictive marker. Hazard ratios for survival were similar for both PD-L1 positive and negative tumors, challenging its utility for patient selection.

Even when trials like LITESPARK 022 and Keynote 564 use identical eligibility criteria, outdated staging systems result in patient populations with different underlying risks. This makes direct comparison of outcomes between trials, even for the same drug, an unfair and statistically flawed analysis that ignores the function of a control arm.