Interlocal reproducibility was performed in 3 external testing laboratories in the United States and Europe. Each laboratory performed automated dyeing operations on each of the five non-consecutive days. The same set of samples was taken 5 times, each set consisting of tissue parts from 36 FFPE-HUMAN-NSCLC samples. Efforts were made to select a balanced number of positive and negative samples, with about 25% of the samples near the cutoff. The coloring was performed by 1 operator (laboratory technician) over a period of at least 20 days, with reagents of 1 PD-L1 22C3 phARMDX Lot, on 1 or 2 autostainer Link 48 instruments at any location. The dye results were analyzed with a 50% GST (as described above) by 1 observer (pathologist) in each of the 3 test laboratories. Figure 2A.2A provides an overview of the design of the study. The statistical analysis was carried out using a percentile bootstrap method. Pair comparisons revealed 2,700 consistent total results (1477 PD-L1 negative and 907 PD-L1 positive), with a total of 316 conflicting results. The following agreements were concluded: ANA reached 90.3% with a low of 84.4%, reached 85.2% with a low of 75.6% and OA reached 88.3% with a floor of 81.4% (Table 33). All precision studies have resulted in a 100% agreement. Therefore, the Wilson Score method was used to calculate the CL for Negative Agreement (NPA), Positive Agreement (AAE) and OA. In particular, CTs were calculated based on the number of comparisons by independent pairs in each study.
Note that the bootstrap method is only suitable for results with a certain method of disordance. Resitene reproducibility was also assessed over the course of 5 days by resistance tests at each site. Pair comparisons resulted in 1,080 consistent total results (601 negative and 373 positive), with a total of 106 conflicting results. The results of the analysis were 91.9% with a floor of 88.8%, APA of 87.6% with a floor of 82.5% and OA of 90.2% with a floor of 86.3%. The prevalence of the PD-L1 protein in tissue segments of FFPE with dakos PD-L1 IHC 22C3 pharmDx-Assay was evaluated. One hundred and twenty-seven unique archived NSCLC samples were set up for PD-L1 with 22C3 monoclonal cloned body. All samples were presented as whole tissue segments and primary, metastatic, stage III and iv samples were included. The expression PD-L1 was observed on tumor and immune cells. For the evaluation, the expression PD-L1 was considered positive only on the membranes of tumor cells.
Partial coloration of the tumor cell membrane was sufficient and at least 100 viable tumor cells had to be present for an evaluable interpretation. The term PD-L1 was quantified as GST: average negative negative agreement in negative percentage (ANA), average positive agreement (APA) and total percentage agreement (OA) were calculated for intersite and intrasite reproducibility and interobserver using confidence intervals of 95% percentile-boot-traps (IC) at 2 terroirs.