Multi-Label Nonlinear Matrix Completion with Transductive Multi-Task Feature Selection for Joint MGMT and IDH1 Status Prediction of Patient with High-Grade Gliomas

Lei Chen, Han Zhang, Junfeng Lu, Kimhan Thung, Abudumijiti Aibaidula, Luyan Liu, Songcan Chen, Lei Jin, Jinsong Wu, Qian Wang, Liangfu Zhou, Dinggang Shen

Research output: Contribution to journalArticlepeer-review

21 Citations (Scopus)

Abstract

The O 6 -methylguanine-DNA methyltransferase (MGMT) promoter methylation and isocitrate dehydrogenase 1 (IDH1) mutation in high-grade gliomas (HGG) have proven to be the two important molecular indicators associated with better prognosis. Traditionally, the statuses of MGMT and IDH1 are obtained via surgical biopsy, which has limited their wider clinical implementation. Accurate presurgical prediction of their statuses based on preoperative multimodal neuroimaging is of great clinical value for a better treatment plan. Currently, the available data set associated with this study has several challenges, such as small sample size and complex, nonlinear (image) feature-to-(molecular) label relationship. To address these issues, we propose a novel multi-label nonlinear matrix completion (MNMC) model to jointly predict both MGMT and IDH1 statuses in a multi-task framework. Specifically, we first employ a nonlinear random Fourier feature mapping to improve the linear separability of the data, and then use transductive multi-task feature selection (performed in a nonlinearly transformed feature space) to refine the imputed soft labels, thus alleviating the overfitting problem caused by small sample size. We further design an optimization algorithm with a guaranteed convergence ability based on a block prox-linear method to solve the proposed MNMC model. Finally, by using a single-center, multimodal brain imaging and molecular pathology data set of HGG, we derive brain functional and structural connectomics features to jointly predict MGMT and IDH1 statuses. Results demonstrate that our proposed method outperforms the previously widely used single- A nd multi-task machine learning methods. This paper also shows the promise of utilizing brain connectomics for HGG prognosis in a non-invasive manner.

Original languageEnglish
Article number8294228
Pages (from-to)1775-1787
Number of pages13
JournalIEEE Transactions on Medical Imaging
Volume37
Issue number8
DOIs
Publication statusPublished - 2018 Aug

Bibliographical note

Funding Information:
Manuscript received December 12, 2017; revised February 1, 2018 and February 12, 2018; accepted February 12, 2018. Date of publication February 19, 2018; date of current version July 31, 2018. This work was supported in part by the National Institutes of Health under Grant EB006733, Grant EB008374, Grant MH100217, Grant MH108914, Grant AG041721, Grant AG049371, Grant AG042599, Grant AG053867, and Grant EB022880, in part by the National Natural Science Foundation of China under Grant 61732006, Grant 61672281, and Grant 61572263, in part by the Natural Science Foundation of Jiangsu Province under Grant BK20161516 and Grant BK20151511, in part by the Post-Doctoral Science Foundation of China under Grant 2015M581794, in part by the National Key Technology Research and Development Program of China under Grant 2014BAI04B05, in part by the Science and Technology Commission of Shanghai Municipality under Grant 16410722400, and in part by the Post-Doctoral Science Foundation of Jiangsu Province under Grant 1501023C. (Lei Chen, Han Zhang, and Junfeng Lu contributed equally to this work.) (Corresponding authors: Liangfu Zhou; Dinggang Shen.) L. Chen is with the Jiangsu Key Laboratory of Big Data Security and Intelligent Processing, Nanjing University of Posts and Telecommunications, Nanjing 210023, China, and also with the Department of Radiology and BRIC, University of North Carolina, Chapel Hill, NC 27599 USA (e-mail: chenlei@njupt.edu.cn).

Publisher Copyright:
© 2012 IEEE.

Keywords

  • Brain tumor
  • connectomics
  • functional connectivity
  • high-grade glioma
  • matrix completion
  • molecular biomarker
  • prognosis
  • structural connectivity

ASJC Scopus subject areas

  • Software
  • Radiological and Ultrasound Technology
  • Computer Science Applications
  • Electrical and Electronic Engineering

Fingerprint

Dive into the research topics of 'Multi-Label Nonlinear Matrix Completion with Transductive Multi-Task Feature Selection for Joint MGMT and IDH1 Status Prediction of Patient with High-Grade Gliomas'. Together they form a unique fingerprint.

Cite this