ABSTRACT
Noninvasive prediction of glioma CDKN2A/B gene homozygous deletion can assist doctors in preoperative diagnosis of glioma patients, which is crucial for guiding prognosis and the best therapeutic regime. At present, the task of glioma gene classification has the problem of too many parameters and high computational complexity. It can not capture the spatial information of different scale channels, and the introduced attention can only capture local information. Aiming at this problem, a ConvNeXt glioma gene classification method integrating pyramid channel attention is proposed. Firstly, so as to reduce network parameters, group convolution is used to extract information of different scale channel. Secondly, in order to learn important features, attention mechanism is introduced. Finally, the obtained weights are weighted to each channel to recalibrate the feature map and stablish a long-term dependence. Multiparametric brain magnetic resonance imaging (MRI) data and corresponding genomic information of 127 subjects (68 positives and 57 negatives for CDKN2A/B homozygous deletion) were obtained from The Cancer Imaging Archive (TCIA) and The Cancer Genome Atlas (TCGA) respectively. The experimental results certify that proposed method achieves 83.34%, 81.12% and 84.62% on ACC, AUC and F1, respectively. Compared with the baseline model ConvNeXt, the classification performance is improved, which proves the effectiveness and efficiency of the proposed model in the CDKN2A/B gene classification task of glioma.
- Ostrom Q T, Gittleman H, Fulop J, CBTRUS statistical report: primary brain and central nervous system tumors diagnosed in the United States in 2008-2012[J]. Neuro-oncology, 2015, 17(suppl_4): iv1-iv62.Google ScholarCross Ref
- Lu V M, O'Connor K P, Shah A H, The prognostic significance of CDKN2A homozygous deletion in IDH-mutant lower-grade glioma and glioblastoma: a systematic review of the contemporary literature [J]. Journal of neuro-oncology, 2020, 148: 221-229.Google Scholar
- Chang K, Bai H X, Zhou H, Residual Convolutional Neural Network for the Determination of IDH Status in Low-and High-Grade Gliomas from MR ImagingNeural Network for Determination of IDH Status in Gliomas[J]. Clinical Cancer Research, 2018, 24(5): 1073-1081.Google ScholarCross Ref
- Korfiatis P, Kline T L, Lachance D H, Residual deep convolutional neural network predicts MGMT methylation status[J]. Journal of digital imaging, 2017, 30: 622-628.Google Scholar
- Chang P, Grinband J, Weinberg B D, Deep-learning convolutional neural networks accurately classify genetic mutations in gliomas[J]. American Journal of Neuroradiology, 2018, 39(7): 1201-1207.Google ScholarCross Ref
- Akkus Z, Ali I, Sedlar J, Predicting 1p19q chromosomal deletion of low-grade gliomas from MR images using deep learning[J]. arXiv preprint arXiv:1611.06939, 2016.Google Scholar
- Tang Z, Xu Y, Jin L, Deep learning of imaging phenotype and genotype for predicting overall survival time of glioblastoma patients[J]. IEEE transactions on medical imaging, 2020, 39(6): 2100-2109.Google ScholarCross Ref
- Bangalore Yogananda C G, Shah B R, Vejdani-Jahromi M, A novel fully automated MRI-based deep-learning method for classification of IDH mutation status in brain gliomas[J]. Neuro-oncology, 2020, 22(3): 402-411.Google ScholarCross Ref
- Lu Z, Bai Y, Chen Y, The classification of gliomas based on a pyramid dilated convolution resnet model[J]. Pattern Recognition Letters, 2020, 133: 173-179.Google ScholarCross Ref
- Sievers P, Hielscher T, Schrimpf D, CDKN2A/B homozygous deletion is associated with early recurrence in meningiomas[J]. Acta neuropathologica, 2020, 140: 409-413.Google Scholar
- He K, Zhang X, Ren S, Deep residual learning for image recognition[C]//Proceedings of the IEEE conference on computer vision and pattern recognition. 2016: 770-778.Google Scholar
- Liu Z, Mao H, Wu C Y, A convnet for the 2020s[C]//Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. 2022: 11976-11986.Google Scholar
- Hu J, Shen L, Sun G. Squeeze-and-excitation networks[C]//Proceedings of the IEEE conference on computer vision and pattern recognition. 2018: 7132-7141.Google Scholar
- Woo S, Park J, Lee J Y, Cbam: Convolutional block attention module[C]//Proceedings of the European conference on computer vision (ECCV). 2018: 3-19.Google Scholar
Recommendations
A statistical method for identifying differential gene--gene co-expression patterns
Motivation: To understand cancer etiology, it is important to explore molecular changes in cellular processes from normal state to cancerous state. Because genes interact with each other during cellular processes, carcinogenesis related genes may ...
Gene pathways and subnetworks distinguish between major glioma subtypes and elucidate potential underlying biology
Molecular diagnostic tools are increasingly being used in an attempt to classify primary human brain tumors more accurately. While methods that are based on the analysis of individual gene expression prove to be useful for diagnostic purposes, they are ...
Differences in gene expression between B-cell chronic lymphocytic leukemia and normal B cells: a meta-analysis of three microarray studies
Motivation: A major focus of current cancer research is to identify genes that can be used as markers for prognosis and diagnosis, and as targets for therapy. Microarray technology has been applied extensively for this purpose, even though it has ...
Comments