The term “omics” refers to a broad field of biological science dedicated to understanding and cataloging the full range of information within cells, tissues, or entire organisms. Each omics discipline examines a different layer of this biological information. For example, genomics focuses on the complete set of DNA in an organism, identifying genes and their variations, while proteomics studies the proteins these genes produce and how they interact.
By collectively studying these systems, omics fields provide a comprehensive view of biological functions and disease processes. In medical research, the power of omics lies in its ability to analyze massive datasets, uncover complex molecular signatures, predict disease outcomes, identify drug targets, and personalize treatments.
Extending this approach into medical imaging, “radiomics” extracts quantitative data from scans to reveal insights about tumor biology, treatment responses, and disease progression. This brings us closer to a truly data-driven understanding of cancer. To explore this further, I spoke with Dr. Andrea Corsi from Radiomics.bio, an innovative company based in Belgium and Canada specializing in advanced imaging analysis services for the pharmaceutical and biotech industries. These services include volumetric and radiomic analyses, predictive modeling, and prognostic tools, offering valuable insights for clinical trials and post-market investigations. Dr. Corsi, as the Medical Director, oversees all medical and segmentation work within the company.
What is Radiomics and How Does it Work?
As mentioned before, each “omics” field delves into a specific layer of biology, generating vast datasets that help us understand diseases more comprehensively. Radiomics, a new member of the omics family, applies this concept to medical imaging, particularly in cancer research.
Radiomics focuses on extracting detailed, quantitative information from medical images—such as CT scans, MRIs, or PET scans—transforming subjective visual evaluations into rich datasets that uncover hidden patterns and insights. Like other “omics “approaches, radiomics quantifies previously invisible aspects of diseases, offering a new level of precision in cancer diagnosis and treatment. Here, we examine how radiomics works, its reliance on artificial intelligence (AI) for data analysis, and its role in advancing personalized cancer care.
According to Dr. Corsi, radiomics addresses critical challenges in oncology: “For the last decades, the oncological field has been predominantly guided by RECIST v1.1 criteria. Such criteria are universally applied across different clinical trials and they had the great merit of equalizing the field with the technological capabilities available at the time.”
“However, they are known to be affected by inter- and intra-reader variability. But most importantly, it is recognized today that this technique is unable to capture the complexity of responses emerging from novel treatment approaches, particularly in the context of immunotherapy. This limitation becomes evident in cases of mixed responses, where some lesions may progress (or new ones appear) while on the other hand some may disappear or respond to treatment. Total tumor burden and volumetric analyses can help in better depicting different response profiles, while radiomic features may contribute to differentiate lesions that are likely to respond from those refractory to treatment.”
Radiomics employs algorithms to analyze textures, shapes, and other features in medical images, providing insights into the tumor’s molecular and genetic characteristics. While traditional imaging reveals basic traits like size and shape, radiomics dives deeper, extracting complex data from image pixels and voxels. This enables researchers and clinicians to create detailed “fingerprints” of tumors, correlating them with cancer aggressiveness or treatment responses.
The Role of AI in Radiomics
AI and machine learning are integral to radiomics, particularly for analyzing large-scale imaging data. AI algorithms can process thousands of images rapidly, identifying subtle patterns linked to treatment outcomes, recurrence, or disease progression.
By learning from extensive datasets, AI enhances radiomics’ predictive power, making it a vital tool for tailoring cancer treatments. Further, AI streamlines the radiomics workflow, from image segmentation to data extraction, improving both efficiency and accuracy.
Applications and Advantages of Radiomics in Cancer Care
Radiomics offers transformative benefits across cancer care, from diagnosis to treatment. By analyzing subtle characteristics within medical images, it can distinguish malignant from benign tumors with high accuracy, supporting early and precise diagnosis.
Radiomics also provides insights into tumor-specific treatment responses, enabling personalized therapies tailored to each patient. Unlike traditional biopsies, radiomics delivers this data non-invasively, benefiting patients with hard-to-access tumors or those at higher surgical risk.
Additionally, radiomics allows real-time tracking of tumor changes, helping clinicians monitor treatment effectiveness and adjust strategies accordingly. These capabilities make radiomics a powerful tool in adaptive and personalized cancer care.
Dr. Corsi emphasized the practical advantages of radiomics: “This approach has significant potential and can help guide treatment planning. I would like to provide a concrete example on this. According to RECIST v1.1, the appearance of new lesions often results in patients being removed from clinical trials. However, with immunotherapeutic agents, newly emerging response patterns, such as pseudoprogression, have been identified. In such cases, progressive or newly appearing lesions may coexist with responsive ones. By using RECIST, patients who are subjectively improving would still be considered progressive and excluded from trials.”
And he adds “However, thanks to advancements in quantitative total tumor burden analysis, tumor growth kinetics, and radiomics investigations, this paradigm is shifting. The field is increasingly recognizing that response profiles are highly complex, and in some cases, patients kept on trials beyond progression experience measurable benefits in overall survival. Another example is the ability to recognize, from a medical image, and without recurring to physical biopsies, whether a given patient will benefit from a specific treatment or predict the treatment impact, being response or adverse event, on a patient. This will be of significant help for physicians when defining personalized treatment plans.”
The Future of Radiomics in Personalized Cancer Treatment
Radiomics is poised to play a crucial role in personalized medicine. As technology and analytical methods advance, radiomics may become a standard part of cancer diagnosis and treatment planning, offering data-driven insights throughout the care continuum. However, challenges such as standardizing practices and establishing data-sharing protocols remain.
Dr. Corsi highlighted ongoing efforts to address these issues: “There has been a lot of buzz around radiomics as a scientific field over the last decade, with incredible results published in the literature that often failed to be validated in real-life scenarios. However, in recent years, I’ve seen the field eat humble pie, focusing on the reproducibility and explainability of results. I believe this shift has been appreciated amongst others by regulatory agencies, which are now considering the inclusion of radiomics as part of imaging biomarkers for use in clinical trials. These biomarkers, when added to standard ones such as RECIST v1.1, can help drive precision medicine and give oncologic patients better treatments.”
Radiomics integrates the power of omics with cancer imaging, unlocking new possibilities for understanding and treating the disease. With continued innovation, it promises a future where cancer care is guided by comprehensive, individualized insights, transforming patient outcomes.
