陆军军医大学(第三军医大学)军事预防医学系军队卫生统计学教研室,重庆
WU Yazhou, E-mail: asiawu@tmmu.edu.cn
Received:27 November 2025,
Revised:2026-01-13,
Accepted:13 January 2026,
Published:15 February 2026
移动端阅览
郭泓麟,秦茂洋,宋秋月,等.EGWO-GAT:一种用于膀胱尿路上皮癌分期诊断的图注意力网络与增强型灰狼优化算法多组学整合模型[J].陆军军医大学学报,2026,48(03):366-377.DOI:10.16016/j.2097-0927.202511119.
GUO H L,QIN M Y,SONG Q Y,et al.EGWO-GAT: A graph attention network and enhanced grey wolf optimizer-based multi-omics integration model for bladder urothelial carcinoma staging diagnosis[J].J Army Med Univ,2026,48(03):366-377.DOI:10.16016/j.2097-0927.202511119.
郭泓麟,秦茂洋,宋秋月,等.EGWO-GAT:一种用于膀胱尿路上皮癌分期诊断的图注意力网络与增强型灰狼优化算法多组学整合模型[J].陆军军医大学学报,2026,48(03):366-377.DOI:10.16016/j.2097-0927.202511119. DOI:
GUO H L,QIN M Y,SONG Q Y,et al.EGWO-GAT: A graph attention network and enhanced grey wolf optimizer-based multi-omics integration model for bladder urothelial carcinoma staging diagnosis[J].J Army Med Univ,2026,48(03):366-377.DOI:10.16016/j.2097-0927.202511119. DOI:
目的
2
本研究提出了一种基于图注意力网络(graph attention network,GAT)与增强型灰狼优化算法(enhanced-grey wolf optimizer, EGWO)的多组学整合模型EGWO-GAT,实现对膀胱尿路上皮癌(bladder urothelial carcinoma
BLCA)肿瘤样本的分期预测。
方法
2
基于在加州大学圣克鲁斯分校Xena功能基因组学探索器(University of California, Santa Cruz Xena,UCSC Xena)网站收集的404例癌症基因组图谱(The Cancer Genome Atlas, TCGA)BLCA样本,包含mRNA、DNA甲基化和微小RNA(microRNA,miRNA)数据,将3种组学数据分别进行预处理和差异分析之后得到其节点特征与边特征,以GAT为基础,引入EGWO进行超参数优化,采用多层感知机(multilayer perceptron,MLP)进行后续癌症分期预测。经5折交叉验证分析,将本研究创建的EGWO-GAT模型与多种经典机器学习分期模型的性能进行对比,并进行组学贡献分析和保留不同相似边条数的模型性能比较,采用准确率、精确率、召回率、F1分数及曲线下面积(area under the curve, AUC)作为性能评估的核心指标。
结果
2
差异特征筛选结果显示,mRNA组学获得534个差异基因,DNA甲基化组学获得3 108个差异探针,miRNA组学获得114个差异miRNA。多模型对比结果表明,当整合所有组学数据类型且保留相似性排名前3的其他患者作为边时,EGWO-GAT模型性能最佳,其AUC值达到0.744,准确率达到0.711,精确率达到0.792,召回率达到0.782,F1分数达到0.785,其综合分类性能显著优于其余经典机器学习方法,且较GS-GAT模型在各项指标上均有明显提升。组学贡献分析显示,全组学(mRNA+DNA甲基化+ miRNA)整合的性能显著优于其他6种组学组合方式。相似性边数性能比较结果表明,保留前3条相似边时模型的AUC、准确率、精确率及F1分数均高于保留前5条或7条边的情况,综合性能最优。
结论
2
本研究构建的EGWO-GAT多组学整合模型在BLCA分期中性能优异,可为精准分期提供技术支撑,解决因样本异质性引发的临床分期难题,对辅助个体化治疗及改善患者预后意义重大。
Objective
2
This study proposes a multi-omics integration model
EGWO-GAT
based on graph attention network (GAT) and enhanced-grey wolf optimizer (EGWO)
for staging prediction of bladder urothelial carcinoma (BLCA) samples.
Methods
2
A total of 404 BLCA samples from The Cancer Genome Atlas (TCGA) were collected from the University of California
Santa Cruz Xena (UCSC Xena) functional genomics explorer
including mRNA
DNA methylation
and microRNA (miRNA) data. After separate preprocessing and differential analysis of the 3 omics datasets
node features and edge features were extracted. Using GAT as the backbone framework
EGWO was employed for hyperparameter optimization
and multilayer perceptron (MLP) was applied for cancer staging prediction. Through 5-fold cross-validation
the performance of EGWO-GAT was compared with classical machine learning models
along with omics contribution analysis and performance conparison of models retaining different numbers of similar edges. Accuracy
precision
recall
F1 score
and area under the curve (AUC) were used as core evaluation metrics.
Results
2
Differential feature screening identified 534 differentially expressed genes in mRNA
3 108 differential probes in DNA methylation
and 114 differential miRNAs. Model comparisons showed that EGWO-GAT achieved optimal performance when integrating all omics data and retaining edges from the top 3 similar patients
with an AUC value of 0.744
an accuracy of 0.711
a precision of 0.792
a recall of 0.782
and an F1 score of 0.785. Its performance significantly surpassed other classical methods and outperformed the GS-GAT model across all metrics. Omics contribution analysis confirmed that full integration (mRNA+DNA methylation+miRNA) outperformed all other combinations. Result of the similarity edge number performance comparison demonstrated that retaining the top 3 similar edges yielded the highest AUC
accuracy
precision
and F1 score compared to the top 5 or 7 edges.
Conclusion
2
The EGWO-GAT model exhibits excellent performance in BLCA staging
providing reliable technical support for precise diagnosis. It addresses clinical challenges caused by sample heterogeneity and holds significant potential for guiding individualized treatment and improving patient prognosis.
王海涛 , 罗伟铭 , 陈剑 , 等 . 中高危非肌层浸润性膀胱癌卡介苗灌注治疗后复发的临床预测模型构建与验证 [J]. 陆军军医大学学报 , 2025 , 47 ( 9 ): 959 - 968 . DOI: 10.16016/j.2097-0927.202501061 http://dx.doi.org/10.16016/j.2097-0927.202501061 .
WANG H T , LUO W M , CHEN J , et al . Development and validation of clinical prediction model for post-treatment recurrence in high-risk non-muscle invasive bladder cancer after BCG intravesical instillation [J]. J Army Med Univ , 2025 , 47 ( 9 ): 959 - 968 . DOI: 10.16016/j.2097-0927.202501061 http://dx.doi.org/10.16016/j.2097-0927.202501061 .
HUANG X , DU G T , YANG Y , et al . Advancing bladder cancer management: development of a prognostic model and personalized therapy [J]. Front Immunol , 2024 , 15 : 1430792 . DOI: 10.3389/fimmu.2024.1430792 http://dx.doi.org/10.3389/fimmu.2024.1430792 .
ABDEL-RAHMAN O . Validation of the eighth AJCC new substages for bladder cancer among different staging contexts [J]. Clin Genitourin Cancer , 2017 , 15 ( 6 ): e1095 - e1106 . DOI: 10.1016/j.clgc.2017.07.025 http://dx.doi.org/10.1016/j.clgc.2017.07.025 .
AMIN M B , GREENE F L , EDGE S B , et al . The Eighth Edition AJCC Cancer Staging Manual: Continuing to build a bridge from a population-based to a more “personalized” approach to cancer staging [J]. CA Cancer J Clin , 2017 , 67 ( 2 ): 93 - 99 . DOI: 10.3322/caac.21388 http://dx.doi.org/10.3322/caac.21388 .
LINEHAN W M , RICKETTS C J . The Cancer Genome Atlas of renal cell carcinoma: findings and clinical implications [J]. Nat Rev Urol , 2019 , 16 ( 9 ): 539 - 552 . DOI: 10.1038/s41585-019-0211-5 http://dx.doi.org/10.1038/s41585-019-0211-5 .
DONG Y Y , YANG W K , WANG J W , et al . MLW-gcForest: a multi-weighted gcForest model towards the staging of lung adenocarcinoma based on multi-modal genetic data [J]. BMC Bioinformatics , 2019 , 20 ( 1 ): 578 . DOI: 10.1186/s12859-019-3172-z http://dx.doi.org/10.1186/s12859-019-3172-z .
DAI H J , YU Z J , ZHAO Y , et al . Integrating machine learning models with multi-omics analysis to decipher the prognostic significance of mitotic catastrophe heterogeneity in bladder cancer [J]. Biol Direct , 2025 , 20 ( 1 ): 56 . DOI: 10.1186/s13062-025-00650-x http://dx.doi.org/10.1186/s13062-025-00650-x .
SUBRAMANIAN I , VERMA S , KUMAR S , et al . Multi-omics data integration, interpretation, and its application [J]. Bioinform Biol Insights , 2020 , 14 : 1177932219899051 . DOI: 10.1177/1177932219899051 http://dx.doi.org/10.1177/1177932219899051 .
LI B J , WANG T Y , NABAVI S . Cancer molecular subtype classification by graph convolutional networks on multi-omics data [C]// Proceedings of the 12th ACM International Conference on Bioinformatics, Computational Biology, and Health Informatics . Gainesvillem, FL, USA : 2021 : 1 - 9 . DOI: 10.1145/3459930.3469542 http://dx.doi.org/10.1145/3459930.3469542 .
GUO H , LV X , LI Y Z , et al . Attention-based GCN integrates multi-omics data for breast cancer subtype classification and patient-specific gene marker identification [J]. Brief Funct Genomics , 2023 , 22 ( 5 ): 463 - 474 . DOI: 10.1093/bfgp/elad013 http://dx.doi.org/10.1093/bfgp/elad013 .
KESIMOGLU Z N , BOZDAG S . SUPREME: multiomics data integration using graph convolutional networks [J]. NAR Genom Bioinform , 2023 , 5 ( 2 ): lqad063 . DOI: 10.1093/nargab/lqad063 http://dx.doi.org/10.1093/nargab/lqad063 .
WANG T X , SHAO W , HUANG Z , et al . MOGONET integrates multi-omics data using graph convolutional networks allowing patient classification and biomarker identification [J]. Nat Commun , 2021 , 12 ( 1 ): 3445 . DOI: 10.1038/s41467-021-23774-w http://dx.doi.org/10.1038/s41467-021-23774-w .
VELIČKOVIĆ P , CUCURULL G , CASANOVA A , et al . Graph attention networks [EB/OL]. 2017 : arXiv : 1710 . 10903 . https://arxiv.org/abs/1710.10903 https://arxiv.org/abs/1710.10903
TANVIR R B , ISLAM M M , SOBHAN M , et al . MOGAT: a multi-omics integration framework using graph attention networks for cancer subtype prediction [J]. Int J Mol Sci , 2024 , 25 ( 5 ): 2788 . DOI: 10.3390/ijms25052788 http://dx.doi.org/10.3390/ijms25052788 .
ZHAO W Y , GU X , CHEN S Q , et al . MODIG integrating multi-omics and multi-dimensional gene network for cancer driver gene identification based on graph attention network model [J]. Bioinformatics , 2022 , 38 ( 21 ): 4901 - 4907 . DOI: 10.1093/bioinformatics/btac622 http://dx.doi.org/10.1093/bioinformatics/btac622 .
MIRJALILI S , MIRJALILI S M , LEWIS A . Grey wolf optimizer [J]. Adv Eng Softw , 2014 , 69 : 46 - 61 . DOI: 10.1016/j.advengsoft.2013.12.007 http://dx.doi.org/10.1016/j.advengsoft.2013.12.007 .
TOMCZAK K , CZERWIŃSKA P , WIZNEROWICZ M . The Cancer Genome Atlas (TCGA): an immeasurable source of knowledge [J]. Contemp Oncol , 2015 , 19 ( 1A ): A68 - A77 . DOI: 10.5114/wo.2014.47136 http://dx.doi.org/10.5114/wo.2014.47136 .
XU N , YAO Z M , SHANG G G , et al . Integrated proteogenomic characterization of urothelial carcinoma of the bladder [J]. J Hematol Oncol , 2022 , 15 : 76 . DOI: 10.1186/s13045-022-01291-7 http://dx.doi.org/10.1186/s13045-022-01291-7 .
ZHANG C J , SHEN K J , ZHENG Y X , et al . Genome-wide screening of abberant methylated drivers combined with relative risk loci in bladder cancer [J]. Cancer Med , 2020 , 9 ( 2 ): 768 - 782 . DOI: 10.1002/cam4.2665 http://dx.doi.org/10.1002/cam4.2665 .
STEMPOR P A , AVNI D , LEIBOWITZ R , et al . Comprehensive analysis of correlations in the expression of miRNA genes and immune checkpoint genes in bladder cancer cells [J]. Int J Mol Sci , 2021 , 22 ( 5 ): 2553 . DOI: 10.3390/ijms22052553 http://dx.doi.org/10.3390/ijms22052553 .
李佳芮 , 钱力 , 沈俊杰 , 等 . 基于转录组学和变分自编码器的癌症分期诊断研究 [J]. 陆军军医大学学报 , 2025 , 47 ( 6 ): 613 - 622 . DOI: 10.16016/j.2097-0927.202412079 http://dx.doi.org/10.16016/j.2097-0927.202412079 .
LI J R , QIAN L , SHEN J J , et al . Cancer staging diagnosis based on transcriptomics and variational autoencoder [J]. J Army Med Univ , 2025 , 47 ( 6 ): 613 - 622 . DOI: 10.16016/j.2097-0927.202412079 http://dx.doi.org/10.16016/j.2097-0927.202412079 .
WANG B , MEZLINI A M , DEMIR F , et al . Similarity network fusion for aggregating data types on a genomic scale [J]. Nat Methods , 2014 , 11 ( 3 ): 333 - 337 . DOI: 10.1038/nmeth.2810 http://dx.doi.org/10.1038/nmeth.2810 .
PAI S , BADER G D . Patient similarity networks for precision medicine [J]. J Mol Biol , 2018 , 430 ( 18 ): 2924 - 2938 . DOI: 10.1016/j.jmb.2018.05.037 http://dx.doi.org/10.1016/j.jmb.2018.05.037 .
POPESCU M , BALAS V E , PERESCU-POPESCU L , et al . Multilayer perceptron and neural networks [J]. WSEAS Trans Circuits Syst Arch , 2009 , 8 : 579 - 588 . DOI: 10.5555/1639537.1639542 http://dx.doi.org/10.5555/1639537.1639542 .
ZHANG Z , WANG C , LIU T T , et al . MiRNA-182-5p promotes human bladder cancer proliferation and migration through the FOXF2/SHH axis [J]. Neoplasma , 2022 , 69 ( 2 ): 321 - 330 . DOI: 10.4149/neo_2021_210903N1266 http://dx.doi.org/10.4149/neo_2021_210903N1266 .
ZOU Z H , CHENG Q , ZHANG S C , et al . MiR-205-3p suppresses bladder cancer progression via GLO1 mediated P38/ERK activation [J]. BMC Cancer , 2023 , 23 : 956 . DOI: 10.1186/s12885-023-11175-9 http://dx.doi.org/10.1186/s12885-023-11175-9 .
NARWAL A , KUMARI K , KAUSHAL S , et al . The study of miRNA-200c expression and epithelial-to-mesenchymal transition-related transcription factors in the primary bladder urothelial carcinoma [J]. Urol Ann , 2023 , 15 ( 1 ): 35 - 42 . DOI: 10.4103/ua.ua_72_22 http://dx.doi.org/10.4103/ua.ua_72_22 .
ZAGHLOOL S B , MOOK-KANAMORI D O , KADER S , et al . Deep molecular phenotypes link complex disorders and physiological insult to CpG methylation [J]. Hum Mol Genet , 2018 , 27 ( 6 ): 1106 - 1121 . DOI: 10.1093/hmg/ddy006 http://dx.doi.org/10.1093/hmg/ddy006 .
0
Views
1
下载量
0
CSCD
Publicity Resources
Related Articles
Related Author
Related Institution
京公网安备11010802024621
