The paper introduces a pathwise algorithm for fitting the Cox proportional hazards model with convex combinations of $\ell_1$ and $\ell_2$ penalties (elastic net). The algorithm uses cyclical coordinate descent and warm starts to find solutions along a regularization path efficiently. The authors demonstrate the effectiveness of their algorithm on both real and simulated data sets, showing significant speed improvements over competing methods. The algorithm is implemented in the R package *glmnet* and is available on CRAN. The paper also discusses the stability and efficiency of the algorithm, including its performance in handling large datasets and correlated predictors.The paper introduces a pathwise algorithm for fitting the Cox proportional hazards model with convex combinations of $\ell_1$ and $\ell_2$ penalties (elastic net). The algorithm uses cyclical coordinate descent and warm starts to find solutions along a regularization path efficiently. The authors demonstrate the effectiveness of their algorithm on both real and simulated data sets, showing significant speed improvements over competing methods. The algorithm is implemented in the R package *glmnet* and is available on CRAN. The paper also discusses the stability and efficiency of the algorithm, including its performance in handling large datasets and correlated predictors.