TABLE 2

Multivariable Cox Proportional Hazards Regression Model for OS

Parameter	Coefficient	Hazard ratio	P
Sex (male vs. female)	−0.05 (−0.39 to 0.28)	0.95 (0.68–1.32)	0.75
Age (y)	0.01 (−0.01 to 0.03)	1.01 (0.99–1.03)	0.19
Performance status (1 vs. 0)	1.13 (0.53 to 1.73)	—^*	<0.001
Performance status × time (mo)	−0.04 (−0.07 to −0.01)	—^*	0.005
Clinical stage (IIIB vs. IIIA/IIB)	0.34 (0.01 to 0.66)	1.40 (1.01–1.94)	0.04
Log MTV (centered)^†	0.18 (0.05 to 0.31)	1.20 (1.05–1.36)	0.007
100 × SumMean (centered)^‡	−0.44 (−1.06 to 0.18)	0.64 (0.35–1.20)	0.16
Log MTV (centered) × 100 × SumMean (centered)	−0.57 (−1.05 to −0.08)	—	0.02

↵* Single hazard ratio is not reported as the specified covariate is time-varying, thus implying that the hazard ratio varies over time.
↵† Log MTV (centered) corresponds to Log MTV − mean of Log MTV. Because both covariates involved in statistical interaction are centered, hazard ratio for Log MTV can be interpreted at mean value of 100 × SumMean. Hazard ratio corresponds to 1-unit increase in MTV on natural log scale.
↵‡ 100 × SumMean (centered) is 100 × SumMean − mean of 100 × SumMean. Because both covariates involved in statistical interaction are centered, hazard ratio for SumMean can be interpreted at mean value of Log MTV. Hazard ratio corresponds to 0.01 increase in SumMean.
Data in parentheses are 95% confidence intervals.