Abstract 1406P
Background
The underlying objective of this external quality assessment (EQA) conducted by the German Quality Assurance Initiative (QuIP) was to evaluate the accuracy of different ROS1 testing methods in a multicenter approach. We evaluated the diagnostic performance of immunohistochemistry (IHC), in situ hybridization (ISH) and RNA/DNA sequencing (NGS) on non-small cell lung cancer (NSCLC) tissue samples.
Methods
Samples were thoroughly pretested using IHC, ISH and RNA/DNA NGS, before starting the external EQA. The test cohort consisted of ten specimens each, including five ROS1 positive and five ROS1 negative samples for IHC/ISH and four ROS1 positive and six ROS1 negative samples for NGS. Twenty test sets were available for IHC and ISH, twenty-one test sets for NGS. The participants were provided with two unstained tumor slides for IHC/ISH and three unstained tumor slides for NGS and the results were submitted within 14 days via an online questionnaire.
Results
IHC has the highest passing rate with 90% (18/20 participants) while five different antibodies (Cell Signaling, DCS, Epitomics, Medac, Ventana) were used. The sensitivity for the IHC part was 96.9% and the specificity was 93.5%. ISH had the second highest passing rate (85%) with probes from Abbott Molecular and Zytovision and the sensitivity was 94% and the specificity was 97%. From 17 participants who used fluorescent ISH (FISH) 15 passed the ring trial and two out of three participants who used chromogenic ISH (CISH). The NGS samples were mainly processed with kits from ThermoFisher, Illumina and Archer and reached a sensitivity of 93.8% and a specificity of 99.1%. The NGS part had a passing rate of 76%.
Conclusions
All three methods (IHC, ISH, NGS) produced reliable results for formalin-fixed samples within the different institutions and IHC achieved the highest passing rate. The highest sensitivity was observed for IHC whereas the highest specificity was seen for NGS. Most institutions used more than one method to confirm ROS1 positive samples. The German institutions seem to be well-positioned with regard to ROS1 testing in NSCLC patients, however the limited tissue availability may hamper the application of NGS for routine diagnostics.
Clinical trial identification
Editorial acknowledgement
Legal entity responsible for the study
The authors.
Funding
Has not received any funding.
Disclosure
All authors have declared no conflicts of interest.