Abstract
Companion diagnostic (CDx) tests play important roles in identifying oncogenic driver genes and tailoring effective molecularly targeted therapies for lung cancer patients. In Japan, the Oncomine Dx target test (ODxTT) and the AmoyDx pan lung cancer PCR panel (AmoyDx) are prominent CDx tests and only one of these tests is covered by the domestic insurance system. However, these CDx tests cover different target regions and apply different technologies (ODxTT is amplicon-based next-generation sequencing and AmoyDx is multiplex PCR-based assay), which may lead to missing of actionable mutations affecting patient prognosis. Here, we performed a direct comparison analysis of 1059 genetic alterations of eight driver genes from 131 samples and evaluated the concordance between two CDx tests for detecting actionable variants and fusions. When excluding the eight uncovered variants (ODxTT: two variants, AmoyDx: six variants), the overall percent agreement was 97.6% (1026/1051) with 89.0% of overall positive percent agreement (89/100) and 98.5% of overall negative percent agreement (937/951). Of the 25 discordant genetic alterations, two were undetected despite being covered in the AmoyDx (one EGFR variant and one ROS1 fusion). Furthermore, there were potential false positives in the ODxTT (nine MET exon 14 skippings) and in the AmoyDx (five variants, six ROS1 and three RET fusions). These potential false positives in the AmoyDx likely due to non-specific amplification, which was validated by the unique molecular barcoding sequencing. The ODxTT missed two uncovered EGFR rare variants, which was visually confirmed in the raw sequencing data. Our study provides insights into real-world performance of CDx tests for lung cancer and ensures reliability to advance precision medicine.