Presentation (CTI)

OBJECTIVES: Real-world evidence (RWE) studies in lung cancer rely on accurate biomarker documentation to support treatment evaluation, outcomes research, and precision medicine decisions. Incomplete biomarker capture — particularly for complex genomic panels — can compromise data integrity and undermine study validity. This study examined abstraction error patterns in lung cancer cases to determine whether missed biomarker documentation is a primary driver of quality assurance (QA) discrepancies.
METHODS: A retrospective QA review of Q1 2026 lung cancer abstraction cases was conducted across seven abstractors using internal quality metrics. Error rates were assessed across domains including biomarkers, treatment, disease status, staging, and ECOG performance status. NGS panel abstraction was evaluated for completeness across all reported results. Descriptive statistical analyses were performed in R to quantify error frequency and identify dominant patterns.
RESULTS: A total of 134 errors were identified across seven abstractors reviewing 690 cases. Biomarker-related discrepancies were the dominant error domain, accounting for 82.1% of all errors (110/134). The most frequent error was incomplete NGS panel abstraction, representing 74 errors (55.2% of total). This highlights a broader challenge in RWD oncology abstraction: NGS completeness standards vary across organizations, and inconsistent capture of full panel results — including non-actionable findings — may introduce systematic bias into downstream analyses. Additional biomarker errors included missed biomarkers (18 occurrences, 13.4%) and incorrect result interpretation (7 occurrences, 5.2%). Treatment-related discrepancies accounted for 10.4% of errors; remaining errors spanned disease status, staging, and ECOG documentation.
CONCLUSIONS: Incomplete NGS panel abstraction represents the most critical vulnerability in real-world lung cancer data quality, with implications for biomarker prevalence estimates, treatment eligibility studies, and comparative effectiveness research. Variation in NGS completeness standards across RWD organizations may introduce systematic bias difficult to detect without rigorous QA protocols. Standardized NGS abstraction guidance and targeted abstractor training are warranted to strengthen RWE reliability in oncology.