The role of artificial intelligence in the medical and pharmaceutical industries: a patent perspective

In recent years, the integration of artificial intelligence (AI) and machine learning (ML) into the medical and pharmaceutical industries has revolutionised the way we approach healthcare and drug development. From enhancing diagnostic accuracy to accelerating the discovery of new treatments, AI technologies are driving significant advancements. However, at their heart, AI and ML methods are computer implemented inventions, and with this comes complex challenges and opportunities when preparing and prosecuting patent applications in the UK and Europe.

Medical devices

Medical devices can collect, store, and analyse an enormous amount of data from patients, including heart electrocardiogram data, blood pressure, blood chemistry, and medically relevant images. Previously most of these data may have only been analysed with respect to the patient it originated from. However, advances in AI and ML have allowed for latent information to be extracted from these data more efficiently. This has led to medical devices being developed which can employ models trained from these data. For example, a single-lead ECG-enabled Stethoscope has been developed which has correctly identified people with heart failure 9 out of 10 times. The stethoscope uses a deep learning algorithm to assess the ECG to predict the chances of heart failure. The development of such a device can lower the cost of point-of-care screening as it is faster and does not require a full ECG to be taken.

Medical imaging

Computer aided analysis of medical image data has allowed for the automatic assessment of images to aid with the diagnosis, monitoring, and treatment of diseases and conditions. This has increased the speed, accuracy, and efficiency of assessing a wide range of medical images including X-rays, Computer Tomography (CT), ultrasound, magnetic resonance imaging (MRI), and histology images. Recent developments in AI and ML for analysing images have further increased speed and efficiency, but have also allowed for a great leap in the accuracy of diagnoses and the early detection of certain diseases.

By using large databases of images, AI models have been used to identify particular anomalies or features in images that may be difficult for all but the most experienced clinician to spot. Using this technology can therefore allow some of the best image-based assessment and diagnosis to be available to a much wider user base. Additionally, AI analysis of breast screening images has shown that these models are capable of detecting breast cancer four years before it developed into a tumour that would normally be detected by a clinician.

Drug discovery

AI may be used to more rapidly identify potential drug candidates by analysing vast libraries of compounds. For instance, an AI-designed drug for obsessive-compulsive disorder has been developed that reached clinical trials in 12 months, compared to the industry average of five years by developing an automated and adaptive methodology for designing drug ligands. Alternatively, AI may be used to repurpose existing drugs for new purposes and in recent years, this has found applications in rare disorders and conditions. Patent applications for such advances may be directed towards the new uses of known active agents.

The use of AI in drug discovery is not only faster but also more cost-effective. AI can reduce the need for extensive preclinical trials, thereby lowering development costs. Additionally, AI-designed molecules have shown better odds of success in clinical trials, potentially doubling R&D productivity.

Consequently, larger pharmaceutical companies are increasingly forming strategic partnerships with AI firms to leverage this technology. Companies including Bristol-Myers Squibb, Bayer, and Sanofi have collaborated with AI-driven biotech firms to develop new treatments.

Biomarkers

AI-driven identification of biomarkers plays a crucial role in diagnostic medicine. These biomarkers are measurable indicators that help identify diseases, predict outcomes, and guide treatment decisions.

AI has been used to analyse vast omic data sets, such as genomics, proteomics, and metabolomics, to identify disease-related mutations, gene expression patterns, and to analyse protein expression profiles from blood, urine, or tissue samples. Furthermore, AI can be used to predict drug responses based on said genetic variations. These biomarkers aid, for example, in cancer diagnosis, prognosis, and therapeutic monitoring.

There has also been an increase in clinical biomarkers in which AI integrates clinical data (e.g., patient history, lab results and vital signs) to predict disease risk or progression. These are applications of AI’s pattern-recognition capabilities, which can then be used to enhance accuracy, speed, and personalised care.

Protecting AI methods

As AI and ML methods and hardware improve, new assessment and diagnostic methods will be developed, and in many cases the inventor will want to gain patent protection for these. Gaining patent protection for computer-implemented methods such as AI-based assessment and diagnosis at the EPO and in the UK requires careful drafting of the claims to ensure they are not considered to be excluded matter. In the sphere of medical-related inventions as discussed above, it is often possible to tie the method to a real-world technical effect. A number of patents have been granted in Europe which use these and related methods, for example, European patent EP3639191 is for a method that uses AI and deep learning to label histopathological images, improving the scalability and accuracy of pathology diagnostics.

Another consideration is that methods of diagnosis are excluded if they are practiced on the human body. This exclusion does not apply where the device receives data from a patient and analyses it on the device rather than on the body. It is therefore possible to draft a patent for a device like the stethoscope mentioned above which would not fall into this exclusion.

Our patent attorneys have a huge amount of experience and success drafting and prosecuting patent applications for the AI technologies discussed above. Please get in touch to find out more.