144 Chapter 5 Figure 1: Two-dimensional scoring system that characterizes four staining patterns: uniform negative, gradient toward the periphery, uniform positive, and gradient toward the center. A. Schematic representation of the two-dimensional scoring system designed to describe the four defined staining patterns. B. Photographs representing the defined staining patterns observed using D1C2 (x20 objective). For D1C2, the gradient toward the center staining pattern was not observed (indicated by a gray square). The percentage of cancer cells subjective to shedding was calculated for each staining pattern by subtracting the percentage of N-terminal MET immunoreactivity from the percentage of C-terminal MET immunoreactivity. Assessment of association between C-terminal MET, ECD shedding, loss of E-cadherin, and survival To assess which staining pattern for D1C2, A2H2-3, and NCH-38 is most informative with respect to survival for cancers positive for TM C-terminal MET, receiver operating characteristic (ROC) curve analyses were performed for both overall survival (OS) and disease-free survival (DFS) using the area under the curve as a performance measure (Supplementary tables 1 and 2). An identical approach was taken for ECD shedding across the gradient toward the periphery and uniform positive staining pattern (Supplementary tables 3 and 4). The optimal cutoff values for uniform positive D1C2 immunoreactivity, uniform negative NCH-38 immunoreactivity, and ECD shedding within the D1C2 uniform positive staining pattern in view of survival analysis for both OS and DFS, were determined using the maximal value of the Youden index (Supplementary tables 5 through 8 for uniform positive D1C2, Supplementary tables

RkJQdWJsaXNoZXIy MTk4NDMw