145 MET, ECD shedding, and loss of E-cadherin 9 through 12 for uniform negative NCH-38 and Supplementary tables 13 through 16 for ECD shedding within the D1C2 uniform positive staining pattern). Survival analyses The D1C2 uniform positive staining pattern, MET ECD shedding, and E-cadherin uniform negative staining pattern OS and DFS curves were calculated by means of the Kaplan-Meier (KM) method. The log-rank test was used to assess significance of differences in survival times. Univariable and multivariable Cox proportional hazards regression models were used to assess the prognostic value of the D1C2 uniform positive staining pattern, MET ECD shedding, the E-cadherin uniform negative staining pattern, and demographical, clinical, and histopathological characteristics. The median test and independent-samples t-test were used to confirm that inclusion of tissues fixed using nonbuffered formalin (surgically removed before 1995) had no effect on the medians and averages of the prognostically relevant staining patterns (D1C2 uniform positivity, A2H2-3 uniform negativity, and NCH-38 uniform negativity) for tissues fixed using buffered formalin. Therefore, all samples (1984 – 2010) were included in view of the sample size for multivariable analyses. Calculations were performed using SPSS Statistics (version 25; IBM; Armonk, NY, USA). Unless otherwise mentioned, statistical significance was set at a P-value <0.05. Definitions for OS and DFS can be found in Supplementary information. Results Performance of the novel two-dimensional scoring system To study C- and N-terminal MET and E-cadherin immunoreactivity separately and with respect to one another in OSCC, parallel WTSs of 203 cancers were stained with D1C2, A2H2-3 and/or NCH-38 and evaluated using the developed two-dimensional scoring system. The baseline characteristics are presented in Table 1. The defined staining patterns occur in combinations within a cancer section (Supplementary figure 2, Supplementary table 17). In general, the patterns of gradient toward the periphery and toward the center are mutually exclusive for the MET antibodies and NCH-38 (Figure 1B, Supplementary figure 1). 5
RkJQdWJsaXNoZXIy MTk4NDMw