Abstract 3127: Multicenter study evaluating the ROS1 status in lung adenocarcinoma using the novel SP384 immunohistochemistry clone. Towards a new algorithm for ROS1 status assessment in routine

Background: The detection of a ROS1rearrangement in advanced or metastatic lung adenocarcinoma (LUAD) lead to a targeted treatment with tyrosine kinase inhibitors with improved progression free survival (PFS) and overall survival (OS) of the patients. Thus, it is mandatory to screen systematically for the ROS1rearrangement in this patient population. ROS1rearrangements can be detected using fluorescence in situ hybridization (FISH), however ROS1immunohistochemistry (IHC) can be used as a screening test since is largely available, easy, rapid to perform, and cost-effective. However, some false positive and negative IHC results are observed when using the D4D6 cloneleading to confirmatory ROS1FISH in IHC positive samples. Patients and methods: We evaluated the sensitivity and specificity of anti- ROS1SP384 (Ventana, Tucson, AZ) and D4D6 (Cell Signaling, Danvers, MA) antibodies in a multicenter population of 336 LUAD cases enriched for ROS1FISH positive cases (n=51) provided from 6 French molecular pathologyplatforms. Three senior lung pathologists independently scored the SP384 slides as positive or negative around a cutoff of staining in >30% tumor cells at a ≥2+ intensity level, and for the D4D6 clonearound a cutoff of ≥2+ intensity level in any tumor cells. Inter-reader precision between pathologists was assessed. Results were correlated to the PFS and the OS of patients treated with crizotinib. Results: Sensitivity and specificity rates were 100% (95%CI 93.2-100%) and 99.31% (95%CI 97.51-99.92%) for the SP384 clone, and 90.6% (95%CI 79.34-96.87%) and 99.65% (95%CI 98.05-99.99%) for the D4D6 clone, respectively. Inter-reader agreement was 97.5% (95%CI 94.8-99.7) for the SP384 cloneand 86.3% (95%CI 91.3-95.6) for the D4D6 clone. Overall, when compared to ROS1FISH analysis, the SP384 clonehad an accuracy of 99.4%, while D4D6 cloneof 98.2%. Using log-rank test, we observed that LUAD with positive ROS1SP384 status had a longer PFS than those characterized by the D4D6 clone(P=0.021). Conclusions: Interpretation above a cutoff of >30% tumor cells with staining at a ≥2+ intensity level, the ROS1SP384 clonedemonstrates superior sensitivity to D4D6 clone, while preserving similar specificity rate for the detection of ROS1rearrangements. The presented data provide evidence that the SP384 clonemay be used for effective stratification prior to confirmation with orthogonal methods. Citation Format: Veronique Hofman, Isabelle Rouquette, Sandra Lassalle, Simon Heeke, Jean-Christophe Sabourin, Nicolas Piton, Julien Mazieres, Jean-Michel Vignaud, Clemence Yguel, Anne Laure Lepage, Frederic Bibeau, Elodie Long-Mira, Katia Zahaf, Hugues Begueret, Jonathan Benzaquen, Michel Poudenx, Charles-Hugo Marquette, Marius Ilie, Paul Hofman. Multicenter study evaluating the ROS1status in lung adenocarcinoma using the novel SP384 immunohistochemistry clone. Towards a new algorithm for ROS1status assessment in routine [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2019; 2019 Mar 29-Apr 3; Atlanta, GA. Philadelphia (PA): AACR; Cancer Res 2019;79(13 Suppl):Abstract nr 3127.