Start Your Search
MA26 - New Therapies and Emerging Data in ALK, EGFR and ROS1 (ID 930)
- Event: WCLC 2018
- Type: Mini Oral Abstract Session
- Track: Targeted Therapy
- Presentations: 1
- Now Available
- Coordinates: 9/26/2018, 13:30 - 15:00, Room 201 BD
MA26.07 - ROS1 (SP384) Immunohistochemistry Inter-Reader Precision Between 12 Pathologists (Now Available) (ID 12387)
14:10 - 14:15 | Author(s): Mark Kockx
ROS1 positive non-small cell lung cancer (NSCLC) patients can be treated with specific tyrosine kinase inhibitors including crizotinib. ROS1 positivity is often clinically detected by fluorescence in situ hybridization (FISH), however ROS1 IHC can be used to screen samples prior to FISH confirmation of ROS1 status. The ROS1 (SP384) antibody detects ROS1 with high sensitivity, specificity, and consistency. Consistent interpretation of a ROS1 IHC assay between pathologists is important patient evaluation. Here we present inter-reader precision of 12 pathologists across 60 FFPE cases stained with ROS1 (SP384).Method
A retrospective cohort of 60 FFPE NSCLC cases stained with H&E, Rabbit Monoclonal Negative Control Ig, and ROS1 (SP384) were selected to represent positive, negative, and borderline ROS1 IHC status. Twelve practicing lung pathologists independently scored the cases as positive or negative around a cutoff of cytoplasm staining in > 30% tumor cells at a ≥2+ intensity level using Pathotrainer software (Pathomation bvba). Scoring was blinded to other readers and ROS1 status of the cases. Overall percent agreement (OPA), negative percent agreement (NPA), and positive percent agreement (PPA) were calculated in comparison to the group mode. Average overall percent agreement (AOPA), average positive agreement (APA), and average negative agreement (ANA) were calculated pairwise for each reader pair. Following independent assessment, participating pathologists conducted a discordant case review establishing consensus reads for all 60 cases and compared 44 cases to available FISH results.Result
OPA of each of the 12 readers to the mode was 96.4% (95% CI 93.9-98.6) with PPA of 96.3% (95% CI 92.7-99.4) and NPA of 96.5% (95% CI 92.8-99.5). Pairwise AOPA between each of the 12 readers was 94.5% (95%CI 91.2-97.7) with APA 94.0% (95% CI 89.5-97.6) and ANA 95.0% (95%CI 91.2-97.9).
Consensus IHC scores were concordant with FISH 90.0% (40/44 cases).Conclusion
Inter-reader precision around a cutoff of >30% tumor cells with cytoplasmic staining at a ≥2+ intensity level was high in interpreting ROS1 (SP384) in NSCLC samples. Case review highlighted confirmation with FISH in questionable cases and staining patterns to be considered when interpreting ROS1 (SP384) IHC.