Biomarker-Driven Screening

The screening engine prioritizes patients based on the presence, absence, or expression level of a specific biomarker rather than tumor site of origin. This enables tumor-agnostic trial matching where a patient with a rare cancer harboring an NTRK fusion can be matched to a pan-tumor trial.

• Evaluates structured genomic reports for variants like EGFR exon 19 deletions, ALK rearrangements, or BRAF V600E
• Parses immunohistochemistry results for protein expression levels such as PD-L1 TPS ≥ 50%
• Interprets quantitative lab values including MSI-H status or TMB-H thresholds
• Cross-references against trial criteria specifying exact biomarker cutoff values

Advanced screening systems correlate biomarker evidence across disparate data modalities within the patient record. A single eligibility criterion may require simultaneous validation from genomic sequencing, pathology reports, and laboratory results.

• Links next-generation sequencing reports to corresponding pathology specimens
• Validates that a liquid biopsy finding is corroborated by tissue-based testing
• Reconciles conflicting biomarker results from tests performed at different timepoints
• Applies temporal reasoning to ensure the biomarker assessment falls within the trial's required testing window

Raw genomic data requires computational interpretation before it can be used for trial matching. The system must resolve variant nomenclature inconsistencies and classify the clinical significance of detected alterations.

• Normalizes variants to standard HGVS nomenclature across different sequencing platforms
• Distinguishes between pathogenic mutations, variants of unknown significance, and benign polymorphisms
• Maps proprietary gene panel results to standard ontologies like HGNC gene symbols
• Identifies actionable mutations with tiered evidence levels per AMP/ASCO/CAP guidelines

Many biomarker-driven trials specify quantitative or semi-quantitative cutoffs that must be precisely evaluated. The screening engine extracts numeric scores and applies comparison logic against trial-defined thresholds.

• Extracts PD-L1 Tumor Proportion Score with exact percentage values from pathology reports
• Evaluates HER2 immunohistochemistry scores (0, 1+, 2+, 3+) with reflex FISH testing logic
• Processes hormone receptor status with Allred score or H-score quantification
• Applies trial-specific cutoff logic such as "ER-positive defined as ≥ 1% nuclear staining"

Screening must identify not only qualifying biomarkers but also exclusionary resistance markers that would disqualify a patient. This bidirectional biomarker logic prevents enrollment of patients unlikely to benefit.

• Detects KRAS mutations that confer resistance to anti-EGFR therapies
• Identifies ALK resistance mutations like G1202R in patients previously treated with crizotinib
• Flags BRCA reversion mutations that restore homologous recombination function
• Cross-references prior therapy history with known resistance-conferring alterations

Many biomarker-driven trials require testing through a specific FDA-approved companion diagnostic or trial-specified assay. The screening engine validates the testing methodology against protocol requirements.

• Verifies that PD-L1 testing used the SP142 assay when specified by the trial protocol
• Confirms sequencing was performed on a trial-approved NGS platform such as FoundationOne CDx
• Checks that the testing laboratory holds appropriate CLIA certification or CAP accreditation
• Flags patients tested with non-approved methodologies for potential re-testing workflows