Statistical Analysis

Distribution differences in clinical characteristics were compared using χ² test or Fisher exact test for categorical variables and t test for continuous variables. Data for plasma EBV DNA level were obtained using a previously reported method.¹⁵ Univariate analysis was performed with the Kaplan-Meier method and compared using the log-rank test. Multivariate analyses were conducted using Cox regression models to calculate hazard ratios (HRs) with 95% CIs and adjusted P values. Association rules with the parallel coordinates plot were determined using the “apriori” function from the “arules” package of R software. The interaction plot, created using the “stats” package of R, was used to determine whether HBsAg and plasma EBV DNA level were associated with DMFS. Kaplan-Meier survival curves and restricted cubic spline (RCS) curves were plotted using the “ggplot2” package of R. Maximally selected rank statistics were used to independently determine the plasma EBV DNA level cutoff value. We stratified analyses based on HBsAg status and different plasma EBV DNA levels, ranging from 0.5 to 50 × 1000 copies/mL. The stats, survival, Hmisc, rms, maxstat, ggplot2, apriori, and survminer packages from R (R Foundation for Statistical Computing) were used to conduct all statistical analyses. Detailed summary of the statistical considerations is presented in eMethods in Supplement 1.³⁴

For cytological experiments, statistical analysis was performed using GraphPad Prism software, version 9.0 (GraphPad Prism Inc). Data are expressed as mean (SD). Comparisons between 2 groups were performed with 2-tailed unpaired t test. P < .05 was considered statistically significant.