Histologic and Molecular Characterization of Non−Small Cell Lung Carcinoma With Discordant ROS1 Immunohistochemistry and Fluorescence In Situ Hybridization

Supplemental Digital Content is available in the text. Introduction: ROS1 immunohistochemical (IHC) positivity requires follow-up with confirmatory testing such as fluorescence in situ hybridization (FISH). Identifying predictive characteristics of false positive ROS1 IHC cases could aid in optimizing testing algorithms, decrease testing costs and preserve tissue. Materials and Methods: Retrospective results were retrieved for 2054 patients with non−small cell lung carcinoma submitted to our laboratory for molecular testing. Reflex ROS1 FISH was done on all ROS1 immunoreactive cases using ROS1 D4D6 antibody. Staining intensity and histo-score was recorded for all ROS1 immunoreactive cases. Results of any additional molecular testing (KRAS, BRAF, EGFR, ALK FISH, RET FISH, MET FISH) were also tabulated. Results: ROS1 immunoreactivity was seen in 305/2054 (14.8%) cases. Immunoreactivity was weak in majority of the cases with only 4.6% cases having an histo-score >100 and 5.9% of cases had moderate staining intensity. FISH was negative in 99% (302/305) cases with any degree of IHC expression (discordant cases) while 3 cases were positive by FISH. Diffuse strong IHC staining in greater than 90% of the tumor was noted in 6 cases, 3 (0.98%) of which were confirmed to have ROS1 rearrangement by FISH. The discordant cases had significantly higher rates of EGFR mutations (P<0.0005) in comparison to ROS1 IHC negative cases, were seen more often in adenocarcinoma and adenosquamous cell carcinoma (P<0.0005) with lepidic and acinar patterns, and more likely to occur in primary lung carcinomas (P<0.0005). Conclusions: False positive ROS1 immunoreactivity was very frequent, occurred more commonly in primary NSCLC cases with acinar and/or lepidic histologies and was more likely in EGFR mutated cases. Using higher positivity thresholds for ROS1 IHC and incorporating the histologic and molecular correlates into algorithmic strategies could result in increased specificity and clinical utility of ROS1 IHC assay.