MedCity Influencers

Revolutionizing Primary Care: The Role of Pharmacogenomics and AI in Personalized Medicine

This rapidly growing field marries bioinformatics and pharmacology and represents a transformative new era of precision medicine and highly personalized treatments.

Pharmacogenomics (PGx), the study of how genetic profiles impact an individual’s responses to medication, has already begun to help healthcare providers (HCPs) optimize care through its capacity to preemptively enhance drug efficacy, minimize adverse side effects, and improve patient experiences. This rapidly growing field marries bioinformatics and pharmacology and represents a transformative new era of precision medicine and highly personalized treatments, one that serves patients by supporting clinicians to better predict therapeutic responses and more accurately optimize drug dosages.

But, with data-driven solutions come data-driven challenges, not the least of which is the size and complexity of the datasets that pharmacogenomics relies upon. The vastness of genomic data and patient responses to medical treatments requires a herculean human effort to analyze, and because distinguishing meaningful patterns (signal) from irrelevant data (noise) is such a significant challenge in large-scale data analysis, researchers may overlook vital connections between genetic information and patient drug responses. 

AI speeds up PGx insights & expands possibilities

sponsored content

A Deep-dive Into Specialty Pharma

A specialty drug is a class of prescription medications used to treat complex, chronic or rare medical conditions. Although this classification was originally intended to define the treatment of rare, also termed “orphan” diseases, affecting fewer than 200,000 people in the US, more recently, specialty drugs have emerged as the cornerstone of treatment for chronic and complex diseases such as cancer, autoimmune conditions, diabetes, hepatitis C, and HIV/AIDS.

AI has the potential to help PGx manage its data analysis challenges through its capacity to efficiently analyze enormous datasets and identify patterns and correlations that may otherwise remain obscured, aiding researchers and manufacturers in the production of new, more effective medications. Similarly to how AI is used in industries like aerospace for predictive maintenance (e.g., analyzing jet engine data), AI systems in healthcare can excel at cutting through the noise; that is, differentiating normal genetic variations from those that signify disease or predict drug responses, a process for human researchers that is analogous to finding a needle in a haystack. But, AI-driven PGx systems can also help patients directly.  By using their patient’s genetic profile data HCPs can better predict individual responses to specific medications and help make informed treatment decisions that lead to better treatment outcomes.

AI-driven systems can also harness patient data to create digital twins–simulations of a patient’s physiological state–that then can be used to test different treatment strategies and gain new insights from individually-tailored drug interaction data. This technology allows HCPs to swap the traditional trial-and-error approach of many medical treatments with better, more individualized plans that can have better outcomes. For chronic illnesses, like diabetes, the flexibility of digital twin technology also means that providers can monitor, manage, and predict how lifestyle and medication changes can impact things like blood sugar levels, allowing for personalized treatment plans to be more adaptive and responsive to the patient.

Challenges of AI-driven pharmacogenomics

Despite its potential, however, AI in pharmacogenomics faces significant challenges. Because the data sets of genomic information and individual patient responses to medications are so large and so widely distributed across a variety of research platforms, electronic patient record systems, and laboratory information management systems, integrating traditional PGx tools with the data to extract reliable insights becomes difficult. 

HCPs looking to integrate pharmacogenomics systems into their practice also face significant resource challenges themselves. While tool affordability and labor costs for implementation are always top-of-mind, the in-house need that providers face for the genomic expertise necessary to derive clinically relevant, actionable insights from these vast data sets is a significant additional barrier.

Results-driven AI tools

A diversity of emerging AI tools have begun to address such potential challenges and demonstrate tangible results in PGx research and clinical applications while solving these data integration and provider adoption barriers. However, for HCPs choosing which tool to adopt, some differentiators are more important than others. AI-driven extractor tools, for example, that deploy as an interface to other electronic data systems (including Electronic Health Records) would be far-preferred for clinicians because of the resulting enhancement in data integration and improved interoperability, especially if these tools were also more affordable than others on the market. 

The best new tools also leverage AI and advanced deep-learning models to improve the accuracy of variant calling. Variant calling is the process of distinguishing genuine variants from errors, and because pharmacogenes tend to have more complex genetic variations and need to be analyzed differently than typical disease-related genetic variants, the process is complicated for traditional PGx tools. The right AI models, however, that are trained on large, annotated genomic datasets and use established variant-detection algorithms, are reliably better at variant calling and produce much more precise predictions for clinical applications.

Finally, the maintenance plan of a tool – how the data is updated to further train the underlying AI – is also a key differentiator, and some new genomic extractor tools are able to leverage consumer DNA testing and whole-genome sequencing (WGS) by partnering with genetic testing companies and labs, making them attractive candidates for HCPs. These tools can extract PGx data from WGS data, allowing them to expand their genetic services into PGx without collecting additional samples or developing additional tests. The result is the generation of robust clinical insights that can be actioned by the HCP at the point of care without requiring further expert analysis. 

New frontier in pharmacogenomics

Pharmacogenomics as a field is already beginning to revolutionize healthcare, both in the research that providers rely on and the point-of-care, personalized decisions that they make with their patients. With the help of AI, the predictive capabilities of pharmacogenomics are even greater, and with the right tools, HCPs have the potential to create a new standard-of-care from this industry-wide paradigm shift that is as precise and powerful as it is patient-centered. 

Photo: Khanisorn Chaokla, Getty Images

Peter Bannister, DPhil, serves as UGenome's Chief Product Officer for UGenome AI, a precision medicine tools company enabling treatment and dosing to be personalized for every stage of therapeutic development.

Alan Kohler, PhD, serves as UGenome AI's Director of Strategic Communication.

This post appears through the MedCity Influencers program. Anyone can publish their perspective on business and innovation in healthcare on MedCity News through MedCity Influencers. Click here to find out how.