Table 2.

Multicovariate analysis of podoplanin, ΔNp63, clusters of EIC, and all three biomarkers for oral cancer development

Analysis of maximum likelihood estimates
VariablePHR (95% CI)
A. Model 1 (E/N = 35/150)
    Age0.051.031 (1.000-1.064)
    BC + RP vs 13cRA0.790.889 (0.378-2.090)
    RP only vs 13cRA0.411.426 (0.617-3.297)
    Histologic status at baseline: dysplasia vs hyperplasia0.111.776 (0.881-3.581)
    Podoplanin: 2-4 vs 0-10.0023.094 (1.505-6.363)
B. Model 2 (E/N = 36/152)
    Age0.021.035 (1.004-1.066)
    BC + RP vs 13cRA0.951.027 (0.452-2.336)
    RP only vs 13cRA0.521.311 (0.576-2.981)
    Histologic status at baseline: dysplasia vs hyperplasia0.161.632 (0.827-3.222)
    ΔNp63: 6-9 vs 0-50.00073.308 (1.663-6.580)
C. Model 3 (E/N = 32/145)
    Age0.0061.047 (1.013-1.082)
    BC + RP vs 13cRA0.490.727 (0.294-1.795)
    RP only vs 13cRA0.241.647 (0.714-3.796)
    Histologic status at baseline: dysplasia vs hyperplasia0.0462.054 (1.013-4.164)
    Clusters of EIC: 1 vs 00.0572.032 (0.980-4.213)
D. Model 4 (E/N = 37/155)
    Age0.121.024 (0.994-1.056)
    BC + RP vs     13cRA0.860.929 (0.413-2.093)
    RP only vs 13cRA0.191.757 (0.759-4.070)
    Histologic status at baseline: dysplasia vs hyperplasia0.361.385 (0.689-2.783)
    No. positive biomarkers: 3 vs 0-20.00054.372 (1.912-9.992)

NOTE: For multicovariate analysis, the Cox proportional model was fitted for the three markers separately because they are highly associated with each other. For each model, age, treatment arm, and histologic status and a specific biomarker or group of biomarkers were entered.

Abbreviation: E/N, number of events and number of patients.