Comparison of Annotation Services for Next-Generation Sequencing in a Large-Scale Precision Oncology Program

Area of Science:

• Genomic Medicine
• Oncology
• Bioinformatics

Background:

• Next-generation sequencing (NGS) multigene panel testing is increasingly utilized in precision oncology programs, such as the Veterans Affairs National Precision Oncology Program (NPOP).
• Interpreting the vast number of genomic alterations identified by NGS remains a significant challenge for implementing precision medicine effectively.
• Standardized and reliable annotation of these variants is essential for clinical decision-making.

Purpose of the Study:

• To evaluate the concordance of pathogenicity determination and clinical actionability assessments provided by different annotation services within the NPOP.
• To identify potential bottlenecks in the interpretation of genomic alterations from NGS data.

Main Methods:

• Comparison of annotations for unique gene variants from NGS panel results across three services: N-of-One (NoO), IBM Watson for Genomics (WfG), and OncoKB.
• Assessment of concordance for pathogenicity (all three services) and Level 1 drug actionability (WfG and OncoKB) using Cohen's kappa statistic.

Main Results:

• Analysis of 1,388 unique variants from 1,227 NGS results revealed low agreement in pathogenicity assessment between services (e.g., WfG and NoO: 30% agreement, kappa -0.26).
• Moderate agreement was observed for Level 1 therapeutic actionability between WfG and OncoKB (96.9% agreement, kappa 0.44), with significant discrepancies in identified actionable targets.
• TP53, STK11, and CDKN2A were among the most frequently mutated genes, with lung adenocarcinoma being the predominant cancer type.

Conclusions:

• Substantial variability exists in the pathogenicity assessment of NGS variants in solid tumors across different annotation services.
• Agreement on Level 1 therapeutic actionability recommendations between WfG and OncoKB was only moderate, highlighting inconsistencies in clinical guidance.
• There is a critical need to enhance the precision and consistency of NGS multigene panel variant annotation to support effective precision oncology.