Artificial Intelligence has a huge potential in diagnostic medicine. It can run through vast amounts of data very quickly, noticing details that escape the human eye.
Now scientists from the University of Helsinki have demonstrated the possibilities to use AI to predict the efficacy of a targeted breast cancer therapy. AI can be used to discover hidden patterns in tumour samples.
Histopathology of breast cancer. Image credit: Nephron via Wikimedia (CC BY-SA 3.0)
AI is already used to diagnose cancer and track its progression. Now scientists developed a tool which can detect tumour morphological features typical for breast cancer cases with ERBB2 (oncoprotein that promotes the growth of cancer cells). The usefulness of this technology is quite clear – patients with overexpressed ERBB2 protein can benefit from therapy with monoclonal antibodies against the ERBB2 receptor. Identifying them correctly can lead to more informed treatment choices and better outcomes.
Scientists in Finland showed that the AI-algorithm can learn patterns predictive of the ERBB2 status of a tumour directly from the tumour morphology. Dmitrii Bychkov, one of the authors of the study, explained: “Our results show that morphological features of tumours contain vast information about the biology of the disease that can be extracted with machine learning methods. This valuable data can aid in clinical decision-making”. Researchers tested this with real tissue samples from breast cancer patients and found AI to be very good at determining which patients are more likely to have a more favourable disease outcome. This would be extremely helpful in making choices relating to treatment options.
Furthermore, scientists believe that AI-based technology can not only complement the current molecular diagnostic methods but may one day lead to improved selection of some targeted cancer treatments. Scientists are talking about tailored therapies that take into account all different medical characteristics of a human body. Patients are different and so treatments should be different. AI-based methods can help select the most suitable therapy for every case, while being quick and highly efficient.
AI is going to be more and more common in medicine. Although the importance of doctors and other healthcare professionals is never going to decrease, AI-based solutions can help them notice small details and patterns that are difficult to spot using conventional methods. Hopefully, AI software can become more polished and available worldwide soon.
Source: University of Helsinki