In high-dimensional survival analysis, complex relationships involving both linear and nonlinear effects are prevalent. Traditional linear Cox proportional hazards models often fail to capture these complex nonlinearities, while purely neural network-based methods struggle with overfitting and a lack of interpretability in high-dimensional settings. This paper proposes a Deep Partial Linear Cox Model with adaptive Minimax Concave Penalty (MCP) regularization, termed DMCOX. While preserving the interpretability of the classic Cox model, the proposed method integrates deep neural networks into the Partial Linear Cox Model (PLCM) framework. It leverages the universal approximation capability of neural networks to flexibly capture the non-linear effects of low-dimensional covariates, while simultaneously introducing MCP regularization to achieve unbiased estimation and precise feature selection for high-dimensional linear covariates. A hybrid optimization objective combining MCP regularization and neural network approximation is constructed, and an alternating optimization algorithm based on coordinate descent and gradient-based updates is designed for model solving. Extensive numerical simulation experiments demonstrate that DMCOX outperforms traditional Cox models, simple deep learning-based models, and methods using Lasso, SCAD, or L0 penalties. Under various censoring rates and non-linear complexity scenarios, DMCOX exhibits superior predictive accuracy (C-index) and variable selection performance (Recall, F1-score), effectively overcoming overfitting and underfitting issues. Furthermore, the application of the model to real-world high-grade serous ovarian cancer (HGSOC) data, combined with SP-LIME for feature screening and interpretability analysis, successfully identified key prognostic gene features such as TAP1, CXCL9, and COL11A1. The model achieved predictive performance superior to existing methods, validating its effectiveness and clinical potential in precision medicine and biomarker discovery.
WU Weiyan
,
ZHANG Sanguo
. A deep partially linear Cox model with MCP penalty and its application in ovarian cancer prognosis[J]. Journal of University of Chinese Academy of Sciences, 0
: 20260510
.
DOI: 10.7523/j.ucas.2026.030
[1] LeCun Y, Bengio Y, Hinton G. Deep learning[J]. Nature, 2015, 521(7553): 436-444. DOI:10.1038/nature14539.
[2] Krizhevsky A, Sutskever I, Hinton G E.ImageNet classification with deep convolutional neural networks[J]. Communications of the ACM, 2012, 60: 84-90. DOI:10.1145/3065386.
[3] Faraggi D, Simon R.A neural network model for survival data[J]. Statistics in Medicine, 1995, 14(1): 73-82. DOI:10.1002/sim.4780140108.
[4] Katzman J L, Shaham U, Cloninger A, et al.DeepSurv: personalized treatment recommender system using a Cox proportional hazards deep neural network[J]. BMC Medical Research Methodology, 2018, 18(1): 24. DOI:10.1186/s12874-018-0482-1.
[5] Ching T, Zhu X, Garmire L X.Cox-nnet: An artificial neural network method for prognosis prediction of high-throughput omics data[J]. PLoS Computational Biology, 2018, 14(4): e1006076. DOI:10.1371/journal.pcbi.1006076.
[6] Cox D R.Regression models and life-tables[J]. Journal of the Royal Statistical Society.Series B (Methodological), 1972, 34(2): 187-220.
[7] Cybenko G.Approximation by superpositions of a sigmoidal function[J]. Mathematics of Control, Signals and Systems, 1989, 2(4): 303-314. DOI:10.1007/BF02551274.
[8] Hornik K, Stinchcombe M, White H.Multilayer feedforward networks are universal approximators[J]. Neural Networks, 1989, 2(5): 359-366. DOI:10.1016/0893-6080(89)90020-8.
[9] Yarotsky D.Error bounds for approximations with deep ReLU networks[J]. Neural Networks,2017,94:103-114.DOI:10.1016/j.neunet.2017.07.002.
[10] Mhaskar H, Liao Q L, Poggio T.When and why are deep networks better than shallow ones?[C]//Proceedings of the Thirty-First AAAI Conference on Artificial Intelligence. February 4 - 9, 2017, San Francisco, California, USA. ACM, 2017: 2343-2349. DOI:10.5555/3298483.3298577.
[11] Schmidt-Hieber J.Nonparametric regression using deep neural networks with ReLU activation function[J]. The Annals of Statistics, 2020, 48(4): 1875-1897. DOI:10.1214/19-aos1875.
[12] Bauer B, Kohler M.On deep learning as a remedy for the curse of dimensionality in nonparametric regression[J]. The Annals of Statistics, 2019,47(4): 2261-2285. DOI:10.1214/18-aos1747.
[13] Zhong Q X, Mueller J, Wang J L.Deep learning for the partially linear Cox model[J]. The Annals of Statistics, 2022, 50(3): 1348-1375.DOI:10.1214/21-aos2153.
[14] Sun Y M, Kang J, Haridas C, et al. Penalized deep partially linear Cox models with application to CT scans of lung cancer patients[J]. Biometrics, 2024, 80(1): ujad024. DOI:10.1093/biomtc/ujad024.
[15] Cai T T, Xie M Q, Hu T, et al.Simultaneous variable selection and estimation for a partially linear Cox model[J]. Statistical Methods in Medical Research, 2025, 34(4): 783-795. DOI:10.1177/09622802251322988.
[16] Ribeiro M T, Singh S, Guestrin C.“Why should I trust you?”Explaining the predictions of any classifier[C]//Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining.San Francisco California USA. ACM, 2016: 1135-1144.DOI:10.1145/2939672.2939778.
[17] Sasieni P.Information bounds for the conditional hazard ratio in a nested family of regression models[J]. Journal of the Royal Statistical Society Series B:Statistical Methodology,1992,54(2):617-635.DOI:10.1111/j.2517-6161.1992.tb01901.x
[18] Huang J.Efficient estimation of the partly linear additive Cox model[J]. The Annals of Statistics, 1999, 27(5): 1536-1563.
[19] Breheny P, Huang J.Coordinate descent algorithms for nonconvex penalized regression, with applications to biological feature selection[J]. The Annals of Applied Statistics, 2011, 5(1): 232-253.DOI:10.1214/10-AOAS388.
[20] Zhang C H.Nearly unbiased variable selection under minimax concave penalty[J]. The Annals of Statistics, 2010, 38(2): 894-942.
[21] Harrell F E Jr, Califf R M, Pryor D B, et al. Evaluating the yield of medical tests[J]. JAMA, 1982, 247(18): 2543-2546. DOI:10.1001/jama.1982.03320430047030.
[22] Harrell F E Jr, Lee K L, Mark D B. Multivariable prognostic models: Issues in developing models, evaluating assumptions and adequacy, and measuring and reducing errors[J]. Statistics in Medicine, 1996, 15(4): 361-387. DOI:10.1002/(SICI)1097-0258(19960229)15:4<361::AID-SIM168>3.0.CO;2-4.
[23] Fan J C C, Gijbels R. Local Polynomial Modelling and Its Applications[M]. London: Chapman & Hall, 1996.
[24] Zhai R, Dan C, Suggala A, et al.Boosted CVaR classification[C]//Advances in Neural Information Processing Systems 34: Proceedings of the 35th Annual Conference on Neural Information Processing Systems 2021. December 6-14, 2021, Virtual. Curran Associates, Inc., 2021: 21860-21871.
[25] Wang R Z,Malladi S,Wang T H,et al. The marginal value of momentum for small learning rate SGD[EB/OL].2023:arXiv:2307.15196.(2023-07-23)[2026-03-28].https://arxiv.org/abs/2307.15196
[26] Talhouk A, George J, Wang C, et al.Prognostic gene expression signature for high-grade serous ovarian cancer[J]. Nature Communications, 2020, 11: 2678. DOI:10.1038/s41467-020-16575-1.
[27] Hao J, Kim Y, Mallavarapu T, et al.Interpretable deep neural network for cancer survival analysis by integrating genomic and clinical data[J]. BMC Medical Genomics, 2019, 12(10): 189. DOI:10.1186/s12920-019-0624-2.
