75 MET immunoreactivity and poor prognosis Supplementary table S7: Evaluation of possible confounders and significant statistical interaction for 5-year DFS. Explanatory variable 5-year DFS Confounder Effect modifier b1 eq. 1 b1 eq. 2 Variation b1’s (%) exp{b3} eq. 3 p-value Sex 0.68 0.66 2.65 0.52 0.11 Age at diagnosis (years) 0.68 0.65 5.15 0.98 0.31 Smoking 0.68 0.67 1.47 1.54 0.35 Alcohol 0.68 0.70 -1.90 1.20 0.65 Site 0.68 0.68 0.44 0.41 0.27 cT-stage 0.68 0.60 11.91 0.89 0.78 cN-stage 0.68 0.67 1.47 0.91 0.81 pT-stage 0.68 0.62 8.68 0.89 0.77 pN-stage 0.68 0.70 -2.65 0.98 0.96 Infiltration depth 0.68 0.67 0.88 0.26 0.28 Differentiation grade 0.68 0.67 2.21 1.14 0.78 Vasoinvasive growth 0.70 0.66 6.27 0.37 0.03 Bone invasion 0.68 0.66 3.53 1.01 0.98 Perineural invasion 0.71 0.79 -11.13 1.93 0.12 Extranodal growth 0.67 0.52 23.02 0.61 0.22 Growth pattern 0.68 0.72 -6.16 1.45 0.48 Postoperative (C)RT 0.69 0.72 -4.64 0.50 0.16 Abbreviations: DFS, Disease Free Survival; (C)RT, (chemo)radiotherapy; eq., equation. Bold values highlight statistical significance. SUPPOSE logistic regression equations li(t) = l0(t)exp{b1x1} [1] li(t) = l0(t)exp{b1x1 + b2x2} [2] li(t) = l0(t)exp{b1x1 + b2x2 + b3x1.x2} [3] WITH x1: MET staining pattern x2: to be investigated baseline variable THEN x1 and x2 are confounders when there is more than 10% variation between the b1’s from equation 1 and 2 Statistical interaction occurs between x1 and x2 if exp{b3} significantly contributes to equation 3 3
RkJQdWJsaXNoZXIy MTk4NDMw