Multicenter evaluation of a novel ROS1 immunohistochemistry assay (SP384) for detection of ROS1 rearrangements in a large cohort of lung adenocarcinoma patients

Abstract Introduction The detection of a ROS1 rearrangement in advanced and metastatic lung adenocarcinoma (LUAD) lead to a targeted treatment with tyrosine kinase inhibitors, with favorable progression-free survival and overall survival of the patients. Thus, it is mandatory to screen for the ROS1 rearrangement in all these patients. ROS1 rearrangements can be detected using break-apart fluorescence in situ hybridization (FISH), which is the "gold standard"; however, ROS1 immunohistochemistry (IHC) can be used as a screening test since it is widely available, easy and rapid to perform, and cost-effective. Methods We evaluated the diagnostic accuracy and inter-pathologist agreement of two anti-ROS1 IHC clones, SP384 (Ventana, Tucson, AZ) and D4D6 (Cell Signaling, Danvers, MA), in a training cohort of 51 positive ROS1 FISH LUAD cases, and then in a large validation cohort of 714 consecutive cases of LUAD from six routine molecular pathology platforms. Results In the two cohorts, the SP384 and D4D6 clones demonstrate variable sensitivity and specificity rates on the basis of two cutoff points ≥1+ (all % tumor cells and ≥2+ (>30% stained tumor cells). In the validation cohort, the D4D6 yielded the best accuracy for the presence of a ROS1 rearrangement by FISH. Inter-pathologist agreement was moderate to good (ICC, 0.722–0.874) for the D4D6 clone and good to excellent (ICC, 0.830–0.956) for the SP384 clone. Conclusions ROS1 IHC is an effective screening tool for the presence of ROS1 rearrangements. However, users must be acutely aware of the variable diagnostic performance of different anti-ROS1 antibodies prior to implementation into routine clinical practice.