Comparison of in Situ and Extraction-Based Methods for the Detection of ROS1 Rearrangements in Solid Tumors

Abstract

Clinical data confirmed that patients with ROS1 rearrangement are sensitive to specific inhibitors. Therefore, reliable detection of ROS1 rearrangements is essential. Several diagnostic techniques are currently available. However, previous studies were hampered by the low number of ROS1-positive samples. Thirty-five samples, including 32 ROS1 fluorescent in situ hybridization (FISH)-positive and three ROS1 FISH-negative samples were evaluated by ROS1 chromogenic in situ hybridization, ROS proto-oncogene 1, receptor tyrosine kinase (ROS1) immunohistochemistry (IHC), an Agilent SureSelect^{XT HS} custom panel, the Archer FusionPlex Comprehensive Thyroid and Lung panel, and a custom NanoString fusion panel. Some samples were additionally analyzed with the Illumina TruSight Tumor 170 assay. Eleven samples were ROS1 FISH positive by a break-apart signal pattern. In all 11 samples, a ROS1 fusion was confirmed by at least one other method. The other 21 samples tested ROS1 FISH positive by an isolated 3' green signal pattern. Ten of 21 samples could be confirmed by at least two other methods. The other 11 samples tested negative by ROS1 IHC and at least one other method, indicating a false-positive ROS1 FISH result. Our study found that all ROS1 FISH-positive samples with isolated 3' green signals should be confirmed by another method. When sufficient material is available, extraction-based parallel sequencing approaches for the verification of these cases might be preferable.