Fig. 3: investigating overlap of metabolomics and radiomics data structures. | npj Precision Oncology

Fig. 3: investigating overlap of metabolomics and radiomics data structures.

From: Deep representation learning of tissue metabolome and computed tomography annotates NSCLC classification and prognosis

Fig. 3

a Two-dimensional PCA of metabolomic features from UHRS dataset. b Two-dimensional PCA of radiomic features from UHRS dataset. c Two-dimensional PCA of radiomic features from a larger TCIA dataset. d Pearson correlation heatmap between CT radiomic features and six metabolomic features important for classification of histology subtypes from UHRS dataset.

Back to article page