CD8 Radiomics

At the ASCO Congress, the results of a study on a non-invasive tool to assess tumour-infiltrating CD8 T cells from contrast-enhanced CTs using AI approaches were presented. This tool shows promise for conducting multiple virtual biopsies to assess and characterise the spatial heterogeneity of non-small cell lung cancer in patients treated with durvalumab. The virtual biopsies can evaluate the tumour microenvironment of each lesion directly from the imaging data, assessing the risk of progression at the lesion level.

The significance of this tool lies in its potential to address the limitations of conventional biomarkers, which currently perform only moderately in predicting immunotherapy responses. Conventional biomarkers require biopsies and thus only consider a small sample of the entire disease. By applying AI to imaging, the tool maps the tumour microenvironment of each lesion, identifying which lesions are most likely to progress. This capability allows for targeted treatment, including ablative therapy or combination treatments.

The study focused particularly on liver metastases, which are associated with poor prognosis. The tool effectively identified “cold” and “hot” liver metastases, impacting the prognosis of the patient. Identifying patients with cold liver metastases is particularly crucial, as they have a higher risk of not responding to immunotherapy. This approach of using non-invasive biomarkers could lead to new strategies by better characterising the patient’s disease and guiding targeted treatments for pejorative lesions. This method may pave the way for a new era of AI-guided, ultra-precision medicine.