Image, image on the wall, can I see a biomarker change after all?

Cancer biomarker testing is essential for tailoring treatments to individual patients and has now added pathologists and molecular pathologists as key stakeholders in the treatment pathway.

There are multiple initiatives to improve access to testing, increase testing rates, and provide more comprehensive testing, such as Next-Generation Sequencing (NGS). Yet, according to recent publications, approximately 64% of patients do not benefit from tailored treatments due to various barriers along the testing pathway. Limited access to testing, low rates of test requests, and inconsistent reimbursement all hinder the widespread adoption of biomarker testing, resulting in delays in diagnosis and treatment, ultimately affecting patient outcomes negatively.

Currently, biomarker testing is intrinsically associated with the pathology lab, requiring a sample (tissue and/or blood) to be tested. There are significant efforts to improve reimbursement, access to NGS, and the use of AI for image interpretation for both Immunohistochemistry (IHC) and In Situ Hybridization (ISH) as well as gene signatures from multi-gene analysis that support the work of pathologists and expedite the turnaround time for reporting. Nevertheless, most of these initiatives focus on once the patient sample is already acquired.

In an intriguing new study, Fuster-Matanzo et al. present a novel solution leveraging AI to analyze CT images to identify biomarkers directly from these scans. This innovation, when clinically validated, can significantly enhance accessibility, as CT scans are more commonly performed than specialized biomarker tests and are required by guidelines. It may also speed up the diagnostic process, allowing patients to access treatment more quickly. Importantly, this technology addresses a critical unmet need by providing a testing alternative for patients who cannot provide traditional biopsy samples.

The introduction of CT image analysis for identifying biomarker alterations can not only expedite biomarker test requirements but also open up new possibilities, such as better identifying patients who might not benefit from a treatment despite a positive biomarker result, early recurrence detection, and minimal residual disease (MRD) monitoring, among others. It will also bring radiologists to the center of the precision medicine puzzle.

As we stand on the brink of a new era in precision medicine, AI-driven innovations like image interpretation are paving the way for more accessible, efficient, and effective cancer diagnostics. By integrating these advanced technologies into routine clinical practice, we can ensure that more patients receive the personalized treatments they need, ultimately improving outcomes and transforming lives.