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

Abstract

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.

A disease network‐based deep learning approach for characterizing melanoma

“…In the future, we can modify the model by linking to survival prediction 63 or considering multiple tasks. 64 The upgraded model may serve as a novel, powerful tool for dermato-oncologists to identify patients with a high risk for progressive disease and poor survival who will benefit from modified, cluster-adapted follow-up schemes or more aggressive treatment strategies. Vice versa, patients with a low risk according to their gene expression and DNA methylation profiles may also take advantage of the model by saving them from unnecessary treatment-related toxicities.…”

“…In the future, we can modify the model by linking to survival prediction 63 or considering multiple tasks 64 . The upgraded model may serve as a novel, powerful tool for dermato‐oncologists to identify patients with a high risk for progressive disease and poor survival who will benefit from modified, cluster‐adapted follow‐up schemes or more aggressive treatment strategies.…”

Multi-task deep latent spaces for cancer survival and drug sensitivity prediction

“…MTL can be used to learn multiple tasks simultaneously instead of building separate models for each task or learning in stages. The combination of autoencoders for feature extraction and predictive models as part of MTL has been previously utilized for drug-sensitivity prediction (Chen et al, 2022) as well as survival prediction (Vo et al, 2021) separately in the context of cancer omics. In this paper we implement a novel end-to-end deep learning architecture extending the CODE-AE-ADV to include patient survival prediction and a joint learning scheme to learn a latent space that is suitable for prioritizing drugs for groups of patients based on their omics profiles.…”

Deep learning-based personalised outcome prediction after acute ischaemic stroke

“…Recently, the application of deep learning in the field of survival analysis has been increasing in popularity 17–21. Representatively, there are models, such as DeepSurv, that expand CoxPH to learn nonlinear relationships between clinical factors and risk, and Deep Survival Machines (DeepSM), which estimates survival function as a mixture of individual parametric survival distributions without assumptions of proportional hazards 22 23.…”