Cancer management remains fragmented across its continuum, from late-stage diagnosis and salvage therapies to non-personalized surveillance. Here, we present Oncoformer, a unified multimodal transformer model trained on the China Oncology Multimodal Prediction and Surveillance Study (COMPASS) cohort (3.67 million individuals, 17.7 million clinical visits) and validated on independent external cohorts, including the UK Biobank. Oncoformer integrates longitudinal electronic health records with chest X-ray imaging to address multiple clinical tasks: pan-cancer diagnosis (area under the receiver operating characteristic curve [AUROC] = 0.956), future cancer prediction up to 1 year before diagnosis (AUROC = 0.869), tumor stage inference (mean AUROC > 0.90), patient-specific treatment-response forecasting, and recurrence-free survival stratification across ten cancer types (all p < 0.01). Staging predictions were independently validated against postoperative pathological endpoints and shown to converge on core cancer genomic pathways. By translating routine clinical data into a dynamic view of cancer evolution, Oncoformer provides a framework for risk-informed cancer prediction and treatment stratification using routine clinical data.
Fei Liu, Kai Wang, Hui Xu et al.· Cell· 1 citation
Accurate evaluation of adrenal masses remains a significant challenge in endocrinology and radiology, as differential diagnosis involves a wide spectrum of benign and malignant lesions. Radiomics and deep learning (DL) have emerged as promising tools to enhance the precision of adrenal mass assessment by extracting high-dimensional imaging features and enabling automated, data-driven analysis. This review summarizes the latest advancements in the application of radiomics and DL techniques for adrenal mass evaluation. We systematically describe the workflow of radiomic feature extraction and model development, emphasizing their roles in differentiating key lesions such as pheochromocytomas/paragangliomas (PPGLs), adrenal cortical adenomas, and adrenal cortical carcinomas. Additionally, the utility of these approaches in genotype prediction and prognostic evaluation is highlighted. The review further explores the advantages and potential of DL, particularly convolutional neural networks (CNNs), in automated segmentation, feature learning, and end-to-end diagnostic frameworks. Finally, current challenges including technical limitations, clinical translation barriers, and future research directions are discussed, aiming to provide a theoretical foundation for constructing intelligent and precise adrenal mass evaluation systems.
Chaohui Liang, Hao Zhu· Frontiers in Endocrinology· 0 citations