Researchers from Imperial College London have demonstrated how combining medical imaging with AI can be used to provide a ‘virtual biopsy’ for cancer patients. 

Their non-invasive method has been shown to classify the type of lung cancer a patient has, which can be crucial in selecting the right treatment and can predict if the cancer is likely to progress. 

According to the researchers, the technique could be used by doctors when it’s not possible or suitable to obtain a physical tissue biopsy from a patient. 

The study, published in the journal NPJ Precision Oncology, was led by Imperial College London, Imperial College Healthcare NHS Trust and funded by the National Institute for Health and Care Research (NIHR) Imperial Biomedical Research Centre (BRC), alongside collaborators in Córdoba, Spain. 

The researchers found a significant and powerful correlation between patients’ metabolomic profiles and ‘deep features’ of their CT scans, which appear as brighter or darker areas in the image. 

Using this method, the researchers theorised they could bypass the need for physical tissue samples and infer tumour metabolic characteristics from the CT scan alone. 

To test this, they used their TMR-CT model in a separate group of 723 lung cancer patients who were treated at Royal Marsden Hospital, Guy’s and St Thomas’ Hospital or Imperial College NHS Healthcare Trust. All patients had a CT scan but there was no available metabolomics data. 

The results showed that TMR-CT adeptly classified lung cancer and, importantly, gave dependable predictions about patient outcomes, surpassing the performance of traditional CT-based methods and clinical assessments.