Survival Prediction of Lung Cancer Using Small-Size Clinical Data with a Multiple Task Variational Autoencoder

被引:9
作者
Thanh-Hung Vo [1 ]
Lee, Guee-Sang [1 ]
Yang, Hyung-Jeong [1 ]
Oh, In-Jae [2 ,3 ]
Kim, Soo-Hyung [1 ]
Kang, Sae-Ryung [3 ,4 ]
机构
[1] Chonnam Natl Univ, Dept Artificial Intelligence Convergence, Gwangju 61186, South Korea
[2] Chonnam Natl Univ Med Sch, Dept Internal Med, Jeonnam 58128, South Korea
[3] Hwasun Hosp, Jeonnam 58128, South Korea
[4] Chonnam Natl Univ Med Sch, Dept Nucl Med, Jeonnam 58128, South Korea
基金
新加坡国家研究基金会;
关键词
survival analysis; lung cancer; variational autoencoder; multiple tasks; prognosis;
D O I
10.3390/electronics10121396
中图分类号
TP [自动化技术、计算机技术];
学科分类号
0812 ;
摘要
Due to the increase of lung cancer globally, and particularly in Korea, survival analysis for this type of cancer has gained prominence in recent years. For this task, mathematical and traditional machine learning approaches are commonly used by medical doctors. While the deep learning approach has had proven success in computer vision tasks, natural language processing and other AI techniques are also adopted for this task. Due to the privacy issues and management process, data in medicine are difficult to collect, which leads to a paucity of samples. The small number of samples makes it difficult to use deep learning and renders this approach unusable. In this investigation, we propose a network architecture that combines a variational autoencoder (VAE) with the typical DNN architecture to solve the survival analysis task. With a training size of n = 4107, MVAESA achieves a C-index of 0.722 while CoxCC, CoxPH, and CoxTime achieved scores of 0.713, 0.703, and 0.710, respectively. With a small training size of n = 379, MVAESA achieves a C-index of 0.707, compared with 0.689, 0.688 and 0.690 for CoxCC, CoxPH, and CoxTime, respectively. The results show that the combination of a VAE with a target task makes the network more stable and that the network could be trained using a small-sized sample.
引用
收藏
页数:13
相关论文
共 25 条
[1]  
[Anonymous], 2016, ARXIV160406737
[2]   A time-dependent discrimination index for survival data [J].
Antolini, L ;
Boracchi, P ;
Biganzoli, E .
STATISTICS IN MEDICINE, 2005, 24 (24) :3927-3944
[3]   Report of the Korean Association of Lung Cancer Registry (KALC-R), 2014 [J].
Choi, Chang-Min ;
Kim, Ho Cheol ;
Jung, Chi Young ;
Cho, Deog Gon ;
Jeon, Jae Hyun ;
Lee, Jeong Eun ;
Ahn, Jin Seok ;
Kim, Seung Joon ;
Kim, Yeongdae ;
Choi, Yoo-Duk ;
Suh, Yang-Gun ;
Kim, Jung-Eun ;
Lee, Boram ;
Won, Young-Joo ;
Kim, Young-Chul .
CANCER RESEARCH AND TREATMENT, 2019, 51 (04) :1400-1410
[4]  
Collobert R., 2008, MACHINE LEARNING, P160, DOI DOI 10.1145/1390156.1390177
[5]  
Deng L, 2013, INT CONF ACOUST SPEE, P8599, DOI 10.1109/ICASSP.2013.6639344
[6]   Fast R-CNN [J].
Girshick, Ross .
2015 IEEE INTERNATIONAL CONFERENCE ON COMPUTER VISION (ICCV), 2015, :1440-1448
[7]  
Haarburger C, 2019, I S BIOMED IMAGING, P1197, DOI [10.1109/isbi.2019.8759499, 10.1109/ISBI.2019.8759499]
[8]   SALMON: Survival Analysis Learning With Multi-Omics Neural Networks on Breast Cancer [J].
Huang, Zhi ;
Zhan, Xiaohui ;
Xiang, Shunian ;
Johnson, Travis S. ;
Helm, Bryan ;
Yu, Christina Y. ;
Zhang, Jie ;
Salama, Paul ;
Rizkalla, Maher ;
Han, Zhi ;
Huang, Kun .
FRONTIERS IN GENETICS, 2019, 10
[9]   DeepSurv: personalized treatment recommender system using a Cox proportional hazards deep neural network [J].
Katzman, Jared L. ;
Shaham, Uri ;
Cloninger, Alexander ;
Bates, Jonathan ;
Jiang, Tingting ;
Kluger, Yuval .
BMC MEDICAL RESEARCH METHODOLOGY, 2018, 18
[10]  
Kim DW, 2019, SCI REP-UK, V9, DOI [10.1038/s41598-018-36760-y, 10.1038/s41598-019-43372-7]