The case-cohort design is a valuable two-phase sampling scheme often used in situations with low disease incidence and challenges in obtaining covariate data due to cost constraints, and many methods have been proposed for its analysis. However, most of the existing approaches, particularly for interval-censored data, ignore the presence of non-expensive covariates that are typically observed for the entire cohort and thus can result in the significant loss of efficiency. To address this, under the framework of the semi-parametric transformation hazards model, we propose a supersampling approach that augments the case-cohort samples by introducing an additional sub-cohort, allowing us to incorporate non-expensive covariates. In the proposed method, the multiple imputation procedure is employed based on rejection sampling, and the asymptotic properties of the resulting estimator of regression parameters are established. To evaluate the finite sample performance of the proposed method, a simulation study is conducted, and it demonstrates its effectiveness in practical situations. Finally, the proposed methodology is applied to an HIV vaccine trial that motivated this study.