Share：
Share this content in WeChat
X
[Chinese][PDF]851
• Clinical Article •
Prediction of adult-type diffuse gliomas IDH phenotype through an ensemble machine learning model with integrating of MRI visual and DTI histogram
HAN Xin  LI Xiaoran  LU Jie 

Cite this article as: HAN X, LI X R, LU J. Prediction of adult-type diffuse gliomas IDH phenotype through an ensemble machine learning model with integrating of MRI visual and DTI histogram[J]. Chin J Magn Reson Imaging, 2024, 15(11): 51-59, 89. DOI:10.12015/issn.1674-8034.2024.11.009.


[Abstract] Objective To investigate the value of ensemble learning model constructed based on MRI visual and diffusion tensor imaging (DTI) histogram for predicting isocitrate dehydrogenase (IDH) phenotypes in adult-type diffuse gliomas.Materials and Methods A retrospective analysis was conducted on conventional MRI and DTI images of 106 adult diffuse gliomas identified by the 2021 edition of the WHO Classification of Central Nervous System Tumors. Visually accessible Rembrandt images (VASARI) features were evaluated on conventional MRI. The absolute and relative values of fractional anisotropy (FA), relative anisotropy (RA), volume ratio anisotropy (VR), and mean diffusivity (MD) of DTI images were measured, as well as the histogram features. Recursive feature elimination (RFE) and Boruta algorithms were used for feature screening in the training set, and Gaussian Naive Bayes (GNB) models of VASARI features, DTI clinical parameters and DTI histograms were ensembled with a support vector machine (SVM) based on the stacking method. The ensemble machine learning model was then used to predict the IDH phenotype of adult diffuse gliomas. The performance of each model was evaluated by measuring the area under the curve (AUC) of receiver operating characteristic curves.Results A total of 106 glioma patients (50.05±15.17 years old, 54 males) were enrolled in the study, comprising 55 patients with IDH-mutant and 51 patients with IDH wildtype. The cascade recursive dimension reduction of RFE and Boruta, respectively, identified six VASARI features, eight DTI clinical parameters features, and eight DTI histogram features as the primary layer classifier. The model constructed based on histogram features had the highest AUC (0.90/0.87, training set/test set), which was superior to the model constructed from DTI clinical parameters (AUC: 0.83/0.78, training dataset/testing dataset) and the model constructed from conventional MRI visual features (AUC: 0.84/0.66, training dataset/testing dataset). The ensemble learning model based on stacking generalization achieved the highest AUC for predicting IDH phenotype (0.92/0.89 for the training dataset/testing dataset).Conclusions The ensemble learning model based on the combined conventional MRI features and DTI features can effectively predict the IDH genotype of adult-type diffuse gliomas before surgery and assists in the rapid clinical assessment of the prognosis.
[Keywords] glioma；magnetic resonance imaging；diffusion tensor imaging；ensemble machine learning；isocitrate dehydrogenase

HAN Xin1, 2   LI Xiaoran1, 2   LU Jie1, 2*  

1 Department of Radiology and Nuclear Medicine, Xuanwu Hospital, Capital Medical University, Beijing100053, China

2 Beijing Key Laboratory of Magnetic Resonance Imaging and Brain Informatics, Beijing100053, China

Corresponding author: LU J, E-mail: imaginglu@hotmail.com

Conflicts of interest   None.

Received  2024-06-26
Accepted  2024-11-10
DOI: 10.12015/issn.1674-8034.2024.11.009
Cite this article as: HAN X, LI X R, LU J. Prediction of adult-type diffuse gliomas IDH phenotype through an ensemble machine learning model with integrating of MRI visual and DTI histogram[J]. Chin J Magn Reson Imaging, 2024, 15(11): 51-59, 89. DOI:10.12015/issn.1674-8034.2024.11.009.

[1]
MILLER K D, OSTROM Q T, KRUCHKO C, et al. Brain and other central nervous system tumor statistics, 2021[J]. CA: Cancer J Clin, 2021, 71(5): 381-406. DOI: 10.3322/caac.21693.
[2]
OSTROM Q T, CIOFFI G, WAITE K, et al. CBTRUS Statistical Report: Primary Brain and Other Central Nervous System Tumors Diagnosed in the United States in 2014–2018[J]. Neuro Oncol, 2021, 23(Supplement_3): iii1-iii105. DOI: 10.1093/neuonc/noab200.
[3]
LOUIS D N, PERRY A, WESSELING P, et al. The 2021 WHO Classification of Tumors of the Central Nervous System: a summary[J]. Neuro Oncol, 2021, 23(8): 1231-1251. DOI: 10.1093/neuonc/noab106.
[4]
WONG Q H W, LI K K W, WANG W W, et al. Molecular landscape of IDH-mutant primary astrocytoma grade Ⅳ/glioblastomas[J]. Mod Pathol, 2021, 34(7): 1245-1260. DOI: 10.1038/s41379-021-00778-x.
[5]
YOUSSEF G, MILLER J J. Lower grade gliomas[J/OL]. Curr Neurol Neurosci Rep, 2020, 20(7): 21 [2024-06-09]. https://pmc.ncbi.nlm.nih.gov/articles/PMC7244462/. DOI: 10.1007/s11910-020-01040-8.
[6]
YOU W, MAO Y, JIAO X, et al. The combination of radiomics features and VASARI standard to predict glioma grade[J/OL]. Front Oncol, 2023, 13: 1083216 [2024-06-09]. https://pmc.ncbi.nlm.nih.gov/articles/PMC10073533/. DOI: 10.3389/fonc.2023.1083216.
[7]
SETYAWAN N H, CHORIDAH L, NUGROHO H A, et al. Beyond invasive biopsies: using VASARI MRI features to predict grade and molecular parameters in gliomas[J/OL]. Cancer Imaging, 2024, 24(1): 3 [2024-06-09]. https://pmc.ncbi.nlm.nih.gov/articles/PMC10759759/. DOI: 10.1186/s40644-023-00638-8.
[8]
KURT N, BINBOGA KURT B, GULSARAN U, et al. Diffusion tensor imaging and diffusion-weighted imaging on axillary lymph node status in breast cancer patients[J]. Diagn Interv Radiol, 2022, 28(4): 329-336. DOI: 10.5152/dir.2022.21460.
[9]
HALILIBRAHIMOĞLU H, POLAT K, KESKIN S, et al. Associating IDH and TERT mutations in glioma with diffusion anisotropy in normal-appearing white matter[J]. Am J Neuroradiol, 2023, 44(5): 553-561. DOI: 10.3174/ajnr.A7855.
[10]
AUGELLI R, CICERI E, GHIMENTON C, et al. Magnetic resonance diffusion-tensor imaging metrics in high grade gliomas: Correlation with IDH1 gene status in WHO 2016 era[J]. Eur J Radiol, 2019, 116: 174-179. DOI: 10.1016/j.ejrad.2019.04.020.
[11]
GAO A, ZHANG H, YAN X, et al. Whole-tumor histogram analysis of multiple diffusion metrics for glioma genotyping[J]. Radiology, 2022, 302(3): 652-661. DOI: 10.1148/radiol.210820.
[12]
ALIOTTA E, NOURZADEH H, BATCHALA P P, et al. Molecular subtype classification in lower-grade glioma with accelerated DTI[J]. AJNR Am J Neuroradiol, 2019, 40(9): 1458-1463. DOI: 10.3174/ajnr.A6162.
[13]
LIN K, CIDAN W, QI Y, et al. Glioma grading prediction using multiparametric magnetic resonance imaging-based radiomics combined with proton magnetic resonance spectroscopy and diffusion tensor imaging[J]. Med Phys, 2022, 49(7): 4419-4429. DOI: 10.1002/mp.15648.
[14]
YAN J, ZHAO Y, CHEN Y, et al. Deep learning features from diffusion tensor imaging improve glioma stratification and identify risk groups with distinct molecular pathway activities[J/OL]. EBioMedicine, 2021, 72: 103583 [2024-06-09]. https://pmc.ncbi.nlm.nih.gov/articles/PMC8479635/ DOI: 10.1016/j.ebiom.2021.103583.
[15]
ZHANG M, GENG R W, BAI Y, et al. Value of arterial spin labeling and diffusion tensor imaging in evaluating IDH1 gene phenotype in gliomas[J]. Chin J Magn Reson Imaging, 2023, 14(10): 58-64. DOI: 10.12015/issn.1674-8034.2023.10.011.
[16]
HYARE H, RICE L, THUST S, et al. Modelling MR and clinical features in grade Ⅱ/Ⅲ astrocytomas to predict IDH mutation status[J]. Eur J Radiol, 2019, 114: 120-127. DOI: 10.1016/j.ejrad.2019.03.003.
[17]
SAHU A, PATNAM N G, GODA J S, et al. Multiparametric magnetic resonance imaging correlates of isocitrate dehydrogenase mutation in WHO high-grade astrocytomas[J/OL]. J Pers Med, 2022, 13(1): 72 [2024-06-09]. https://pmc.ncbi.nlm.nih.gov/articles/PMC9865247/. DOI: 10.3390/jpm13010072.
[18]
EISENBARTH D, WANG Y A. Glioblastoma heterogeneity at single cell resolution[J]. Oncogene, 2023, 42(27): 2155-2165. DOI: 10.1038/s41388-023-02738-y.
[19]
ZHOU H, VALLIÈRES M, BAI H X, et al. MRI features predict survival and molecular markers in diffuse lower-grade gliomas[J]. Neuro Oncol, 2017, 19(6): 862-870. DOI: 10.1093/neuonc/now256.
[20]
CARRILLO J A, LAI A, NGHIEMPHU P L, et al. Relationship between tumor enhancement, edema, IDH1 mutational status, MGMT promoter methylation, and survival in glioblastoma[J]. AJNR Am J Neuroradiol, 2012, 33(7): 1349-1355. DOI: 10.3174/ajnr.A2950.
[21]
HE J L, GAO Y, WU Q, et al. Predicting IDH1 gene mutation of gliomas by combining clinical and imaging features with multiple sequence radiomics[J]. Chin J Magn Reson Imaging, 2023, 14(8): 27-33. DOI: 10.12015/issn.1674-8034.2021.06.017.
[22]
VERDUIN M, PRIMAKOV S, COMPTER I, et al. Prognostic and predictive value of integrated qualitative and quantitative magnetic resonance imaging analysis in glioblastoma[J/OL]. Cancers, 2021, 13(4): 722 [2024-06-09]. https://pmc.ncbi.nlm.nih.gov/articles/PMC7916478/. DOI: 10.3390/cancers13040722.
[23]
LI Y, ZHANG W. Quantitative evaluation of diffusion tensor imaging for clinical management of glioma[J]. Neurosurg Rev, 2020, 43(3): 881-891. DOI: 10.1007/s10143-018-1050-1.
[24]
XIONG J, TAN W, WEN J, et al. Combination of diffusion tensor imaging and conventional MRI correlates with isocitrate dehydrogenase 1/2 mutations but not 1p/19q genotyping in oligodendroglial tumours[J]. Eur Radiol, 2016, 26(6): 1705-1715. DOI: 10.1007/s00330-015-4025-4.
[25]
KICKINGEREDER P, SAHM F, RADBRUCH A, et al. IDH mutation status is associated with a distinct hypoxia/angiogenesis transcriptome signature which is non-invasively predictable with rCBV imaging in human glioma[J/OL]. Sci Rep, 2015, 5(1): 16238 [2024-06-09]. https://pubmed.ncbi.nlm.nih.gov/26538165/. DOI: 10.1038/srep16238.
[26]
HUANG Z, LU C, LI G, et al. Prediction of lower grade insular glioma molecular pathology using diffusion tensor imaging metric-based histogram parameters[J/OL]. Front Oncol, 2021, 11: 627202 [2024-06-09]. https://pmc.ncbi.nlm.nih.gov/articles/PMC7988075/. DOI: 10.3389/fonc.2021.627202.
[27]
PARK C J, CHOI Y S, PARK Y W, et al. Diffusion tensor imaging radiomics in lower-grade glioma: improving subtyping of isocitrate dehydrogenase mutation status[J]. Neuroradiology, 2020, 62(3): 319-326. DOI: 10.1007/s00234-019-02312-y.
[28]
QIAN Z, LI Y, WANG Y, et al. Differentiation of glioblastoma from solitary brain metastases using radiomic machine-learning classifiers[J]. Cancer Lett, 2019, 451: 128-135. DOI: 10.1016/j.canlet.2019.02.054.
[29]
KIM M, JUNG S Y, PARK J E, et al. Diffusion- and perfusion-weighted MRI radiomics model may predict isocitrate dehydrogenase (IDH) mutation and tumor aggressiveness in diffuse lower grade glioma[J]. Eur Radiol, 2020, 30(4): 2142-2151. DOI: 10.1007/s00330-019-06548-3.
[30]
TAN Y, MU W, WANG X C, et al. Improving survival prediction of high-grade glioma via machine learning techniques based on MRI radiomic, genetic and clinical risk factors[J/OL]. Eur J Radiol, 2019, 120: 108609 [2024-06-09]. https://pubmed.ncbi.nlm.nih.gov/31606714/. DOI: 10.1016/j.ejrad.2019.07.010.
[31]
NEVES B J, AGNES J P, GOMES M D N, et al. Efficient identification of novel anti-glioma lead compounds by machine learning models[J/OL]. Eur J Med Chem, 2020, 189: 111981 [2024-06-09]. https://pubmed.ncbi.nlm.nih.gov/31978780/. DOI: 10.1016/j.ejmech.2019.111981.
[32]
VAN KEMPEN E J, POST M, MANNIL M, et al. Accuracy of machine learning algorithms for the classification of molecular features of gliomas on MRI: A systematic literature review and meta-analysis[J/OL]. Cancers, 2021, 13(11): 2606 [2024-06-09]. https://pmc.ncbi.nlm.nih.gov/articles/PMC8198025/. DOI: 10.3390/cancers13112606.

PREV Research on the white matter microstructural changes in boys with enuresis using diffusion kurtosis imaging
NEXT The value of synthetic MRI combined with amide proton transfer imaging in projecting IDH1 gene state in gliomas
  


LINK


Tel & Fax: +8610-67113815    E-mail: editor@cjmri.cn