1
|
Al-Tashi Q, Saad MB, Sheshadri A, Wu CC, Chang JY, Al-Lazikani B, Gibbons C, Vokes NI, Zhang J, Lee JJ, Heymach JV, Jaffray D, Mirjalili S, Wu J. SwarmDeepSurv: swarm intelligence advances deep survival network for prognostic radiomics signatures in four solid cancers. Patterns (N Y) 2023; 4:100777. [PMID: 37602223 PMCID: PMC10435962 DOI: 10.1016/j.patter.2023.100777] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [What about the content of this article? (0)] [Affiliation(s)] [Abstract] [Key Words] [Grants] [Track Full Text] [Subscribe] [Scholar Register] [Received: 12/16/2022] [Revised: 04/18/2023] [Accepted: 05/26/2023] [Indexed: 08/22/2023]
Abstract
Survival models exist to study relationships between biomarkers and treatment effects. Deep learning-powered survival models supersede the classical Cox proportional hazards (CoxPH) model, but substantial performance drops were observed on high-dimensional features because of irrelevant/redundant information. To fill this gap, we proposed SwarmDeepSurv by integrating swarm intelligence algorithms with the deep survival model. Furthermore, four objective functions were designed to optimize prognostic prediction while regularizing selected feature numbers. When testing on multicenter sets (n = 1,058) of four different cancer types, SwarmDeepSurv was less prone to overfitting and achieved optimal patient risk stratification compared with popular survival modeling algorithms. Strikingly, SwarmDeepSurv selected different features compared with classical feature selection algorithms, including the least absolute shrinkage and selection operator (LASSO), with nearly no feature overlapping across these models. Taken together, SwarmDeepSurv offers an alternative approach to model relationships between radiomics features and survival endpoints, which can further extend to study other input data types including genomics.
Collapse
Affiliation(s)
- Qasem Al-Tashi
- Department of Imaging Physics, The University of Texas MD Anderson Cancer Center, Houston, TX 77030, USA
| | - Maliazurina B. Saad
- Department of Imaging Physics, The University of Texas MD Anderson Cancer Center, Houston, TX 77030, USA
| | - Ajay Sheshadri
- Department of Pulmonary Medicine, The University of Texas MD Anderson Cancer Center, Houston, TX 77030, USA
| | - Carol C. Wu
- Department of Thoracic Imaging, The University of Texas MD Anderson Cancer Center, Houston, TX 77030, USA
| | - Joe Y. Chang
- Department of Radiation Oncology, The University of Texas MD Anderson Cancer Center, Houston, TX 77030, USA
| | - Bissan Al-Lazikani
- Department of Genomic Medicine, The University of Texas MD Anderson Cancer Center, Houston, TX 77030, USA
| | - Christopher Gibbons
- Section of Patient-Centered Analytics, The University of Texas MD Anderson Cancer Center, Houston, TX 77030, USA
| | - Natalie I. Vokes
- Department of Thoracic/Head and Neck Medical Oncology, The University of Texas MD Anderson Cancer Center, Houston, TX 77030, USA
| | - Jianjun Zhang
- Department of Thoracic/Head and Neck Medical Oncology, The University of Texas MD Anderson Cancer Center, Houston, TX 77030, USA
| | - J. Jack Lee
- Department of Biostatistics, The University of Texas MD Anderson Cancer Center, Houston, TX 77030, USA
| | - John V. Heymach
- Department of Thoracic/Head and Neck Medical Oncology, The University of Texas MD Anderson Cancer Center, Houston, TX 77030, USA
| | - David Jaffray
- Office of the Chief Technology and Digital Officer, The University of Texas MD Anderson Cancer Center, Houston, TX 77030, USA
| | - Seyedali Mirjalili
- Centre for Artificial Intelligence Research and Optimization, Torrens University Australia, Fortitude Valley, Brisbane, QLD 4006, Australia
- Yonsei Frontier Lab, Yonsei University, Seoul 03722, Korea
- University Research and Innovation Center, Obuda University, 1034 Budapest, Hungary
| | - Jia Wu
- Department of Imaging Physics, The University of Texas MD Anderson Cancer Center, Houston, TX 77030, USA
- Department of Thoracic/Head and Neck Medical Oncology, The University of Texas MD Anderson Cancer Center, Houston, TX 77030, USA
| |
Collapse
|
2
|
Yang CH, Chen WC, Chen JB, Huang HC, Chuang LY. Overall mortality risk analysis for rectal cancer using deep learning-based fuzzy systems. Comput Biol Med 2023; 157:106706. [PMID: 36965323 DOI: 10.1016/j.compbiomed.2023.106706] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [What about the content of this article? (0)] [Affiliation(s)] [Abstract] [Key Words] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 08/03/2022] [Revised: 02/01/2023] [Accepted: 02/19/2023] [Indexed: 03/19/2023]
Abstract
Colorectal cancer is a leading cause of cancer mortality worldwide, with an increasing incidence rate in developing countries. Integration of genetic information with cancer therapy guidance has shown promise in cancer treatment, indicating its potential as an essential tool in translation oncology. However, the high-throughput analysis and variability of genomic data poses a major challenge to conventional analytic approaches. In this study, we propose an advanced analytic approach, named "Fuzzy-based RNNCoxPH," incorporated fuzzy logic, recurrent neural networks (RNNs), and Cox proportional hazards regression (CoxPH) for detecting missense variants associated with high-risk of all-cause mortality in rectum adenocarcinoma. The test data set was downloaded from "Rectum adenocarcinoma, TCGA-READ" the Genomic Data Commons (GDC) portal. In this study, four model-based risk score models were derived using RNN, CoxPH, RNNCoxPHAddition, and RNNCoxPHMultiplication. The RNNCoxPHAddition and RNNCoxPHMultiplication models were obtained as the sum and product of the RNN risk degree matrix and the CoxPH risk degree matrix, respectively. Moreover, the fuzzy logic system was used to calculate the survival risk values of missense variants and classified their membership grade to improve the identification of high-risk gene variation locations associated with cancer mortality. The four models were integrated to develop an advanced risk estimation model. There were 20 028 variants associated with survival status, amongst 17 638 variants were associated with survival and 2390 variants associated with mortality. The proposed Fuzzy-based RNNCoxPH model obtained a balanced accuracy of 93.7%, which was significantly higher than that of the other four test methods. In particular, the CoxPH model is commonly used in medical researches and the XGBoost model is famous for its high accuracy in machine learning. The results suggest that the Fuzzy-based RNNCoxPH model exhibits a higher efficacy in identifying and classifying the missense variants related to mortality risk in rectum adenocarcinoma.
Collapse
Affiliation(s)
- Cheng-Hong Yang
- Department of Information Management, Tainan University of Technology, Tainan, Taiwan; Department of Electronic Engineering, National Kaohsiung University of Science and Technology, Kaohsiung, Taiwan; Ph. D. Program in Biomedical Engineering, Kaohsiung Medical University, Kaohsiung, Taiwan; School of Dentistry, Kaohsiung Medical University, Kaohsiung, Taiwan; Drug Development and Value Creation Research Center, Kaohsiung Medical University, Kaohsiung, Taiwan.
| | - Wen-Ching Chen
- Department of Electronic Engineering, National Kaohsiung University of Science and Technology, Kaohsiung, Taiwan.
| | - Jin-Bor Chen
- Department of Internal Medicine, Kaohsiung Chang Gung Memorial Hospital and Chang Gung University College of Medicine, Kaohsiung, Taiwan.
| | - Hsiu-Chen Huang
- Department of Community Health, Chia-Yi Christian Hospital, Chia-Yi City, Taiwan.
| | - Li-Yeh Chuang
- Department of Chemical Engineering & Institute of Biotechnology and Chemical Engineering, I-Shou University, Kaohsiung, Taiwan.
| |
Collapse
|