Regional radiomics similarity networks (R2SNs) in the human brain: reproducibility, small-world properties and a biological basis

BackgroundStructural covariance network (SCN) has been applied successfully to structural magnetic resonance imaging (MRI) study. However, most SCNs were constructed by the unitary marker, which was insensitive for the different disease phases. The aim of this study is to devise a novel regional radiomics similarity network (R2SN) that could provide more comprehensive information in morphological network analysis. MethodsRegional radiomics similarity network (R2SN) was constructed by computing the Pearson correlations between the radiomics features extracted from any pair of regions for each subject. We further assessed the small-world property of R2SN using the graph theory method, as well as the reproducibility in the different datasets and the reliability with test-retest analysis. The relationship between the R2SN and inter-regional co-expression of gene enriched was also explored, as well as the relationship with general intelligence. ResultsThe R2SN can be replicated in different datasets, also regardless of using different feature subsets. The R2SN showed high reliability with the test-retest analysis (ICC>0.7). Besides, the small-word property ({sigma}>2) and the high correlation with the gene expression (R=0.24, P<0.001) and the general intelligence was found by R2SN. ConclusionR2SN provides a novel, reliable, and biologically plausible method to understand human morphological covariance based on structural MRI. Impact StatementImaging biomarkers are the cornerstone of modern radiology, and the development of valid biomarkers is crucial for optimizing individualized prediction in neurological disorders like AD. Thus, the development of the data mining method from neuroimaging is crucial for adding the biomarkers of disease. This study confirmed that R2SN provides a novel, robust and biologically plausible model and a new perspective for understanding the human brain, therefore. Thus, the R2SN has great promise in further study.