The image texture features obtained from ¹⁸F-fluorodeoxyglucose positron emission tomography/computed tomography (¹⁸F-FDG PET/CT) images of non-small cell lung cancer (NSCLC) have revealed tumor heterogeneity. A combination of genomic data and radiomics may improve the prediction of tumor prognosis. This study aimed to predict NSCLC metastasis using a graph neural network (GNN) obtained by combining a protein–protein interaction (PPI) network based on gene expression data and image texture features. ¹⁸F-FDG PET/CT images and RNA sequencing data of 93 patients with NSCLC were acquired from The Cancer Imaging Archive. Image texture features were extracted from ¹⁸F-FDG PET/CT images and area under the curve receiver operating characteristic curve (AUC) of each image feature was calculated. Weighted gene co-expression network analysis (WGCNA) was used to construct gene modules, followed by functional enrichment analysis and identification of differentially expressed genes. The PPI of each gene module and genes belonging to metastasis-related processes were converted via a graph attention network. Images and genomic features were concatenated. The GNN model using PPI modules from WGCNA and metastasis-related functions combined with image texture features was evaluated quantitatively. Fifty-five image texture features were extracted from ¹⁸F-FDG PET/CT, and radiomic features were selected based on AUC (n = 10). Eighty-six gene modules were clustered by WGCNA. Genes (n = 19) enriched in the metastasis-related pathways were filtered using DEG analysis. The accuracy of the PPI network, derived from WGCNA modules and metastasis-related genes, improved from 0.4795 to 0.5830 (p < 2.75 × 10⁻¹²). Integrating PPI of four metastasis-related genes with ¹⁸F-FDG PET/CT image features in a GNN model elevated its accuracy over a without image feature model to 0.8545 (95% CI = 0.8401–0.8689, p-value < 0.02). This model demonstrated significant enhancement compared to the model using PPI and ¹⁸F-FDG PET/CT derived from WGCNA (p-value < 0.02), underscoring the critical role of metastasis-related genes in prediction model. The enhanced predictive capability of the lymph node metastasis prediction GNN model for NSCLC, achieved through the integration of comprehensive image features with genomic data, demonstrates promise for clinical implementation.

This study aimed to identify a distant-recurrence image biomarker in NSCLC by investigating correlations between heterogeneity functional gene expression and fluorine-18-2-fluoro-2-deoxy-D-glucose positron emission tomography (¹⁸F-FDG PET) image features of NSCLC patients. RNA-sequencing data and ¹⁸F-FDG PET images of 53 patients with NSCLC (19 with distant recurrence and 34 without recurrence) from The Cancer Imaging Archive and The Cancer Genome Atlas Program databases were used in a combined analysis. Weighted correlation network analysis was performed to identify gene groups related to distant recurrence. Genes were selected for functions related to distant recurrence. In total, 47 image features were extracted from PET images as radiomics. The relationship between gene expression and image features was estimated using a hypergeometric distribution test with the Pearson correlation method. The distant recurrence prediction model was validated by a random forest (RF) algorithm using image texture features and related gene expression. In total, 37 gene modules were identified by gene-expression pattern with weighted gene co-expression network analysis. The gene modules with the highest significance were selected (p-value < 0.05). Nine genes with high protein–protein interaction and area under the curve (AUC) were identified as hub genes involved in the proliferation function, which plays an important role in distant recurrence of cancer. Four image features (GLRLM_SRHGE, GLRLM_HGRE, SUVmean, and GLZLM_GLNU) and six genes were identified to be correlated (p-value < 0.1). AUCs (accuracy: 0.59, AUC: 0.729) from the 47 image texture features and AUCs (accuracy: 0.767, AUC: 0.808) from hub genes were calculated using the RF algorithm. AUCs (accuracy: 0.783, AUC: 0.912) from the four image texture features and six correlated genes and AUCs (accuracy: 0.738, AUC: 0.779) from only the four image texture features were calculated using the RF algorithm. The four image texture features validated by heterogeneity group gene expression were found to be related to cancer heterogeneity. The identification of these image texture features demonstrated that advanced prediction of NSCLC distant recurrence is possible using the image biomarker.

We developed machine and deep learning models to predict chemoradiotherapy in rectal cancer using ¹⁸F-FDG PET images and harmonized image features extracted from ¹⁸F-FDG PET/CT images. Patients diagnosed with pathologic T-stage III rectal cancer with a tumor size > 2 cm were treated with neoadjuvant chemoradiotherapy. Patients with rectal cancer were divided into an internal dataset (n = 116) and an external dataset obtained from a separate institution (n = 40), which were used in the model. AUC was calculated to select image features associated with radiochemotherapy response. In the external test, the machine-learning signature extracted from ¹⁸F-FDG PET image features achieved the highest accuracy and AUC value of 0.875 and 0.896. The harmonized first-order radiomics model had a higher efficiency with accuracy and an AUC of 0.771 than the second-order model in the external test. The deep learning model using the balanced dataset showed an accuracy of 0.867 in the internal test but an accuracy of 0.557 in the external test. Deep-learning models using ¹⁸F-FDG PET images must be harmonized to demonstrate reproducibility with external data. Harmonized ¹⁸F-FDG PET image features as an element of machine learning could help predict chemoradiotherapy responses in external tests with reproducibility.