What is radiomics?

Abstract

Radiomics is an emerging field that combines medical imaging techniques with data science to extract a vast array of quantitative features from images for clinical or research applications. These features, often imperceptible to the human eye, hold the potential to enhance healthcare and improve patient outcomes 1 through the identification of novel imaging markers that enable precise diagnosis, prognosis and treatment planning for a variety of childhood diseases. Integration of radiomics with other data types, such as genomics,2 offers opportunities for further multimodal insights, while longitudinal studies can establish radiomics’ role in monitoring disease progression or treatment response. Interpretability, accountability and reliability are essential within a healthcare setting. In contrast to ‘black-box’ artificial intelligence approaches, radiomics typically integrates interpretable algorithms, ensuring that predictions and insights can be understood, validated and trusted by clinicians, offering auditable workflows that are clear and reproducible. While traditional radiological reporting is invaluable, it may not fully exploit all the complex information embedded within modern imaging data. Radiomics can provide additional value to conventional radiological reporting, enhancing our ability to characterise pathology quantitatively and equipping clinicians with meaningful insights to reliably inform decision-making.