AI-Driven Drug Repurposing: Finding New Uses for Existing Drugs

As Generative AI technology advances at astonishing speed, its effect on the pharma industry is already becoming seismic, going way beyond AI GPs or automated diagnosis.

One of the quicker wins is in the potential for repurposing drugs. Traditionally, developing a new drug from scratch takes over a decade, costs billions of dollars, and has a high failure rate. In contrast, repurposing approved drugs or even failed drug candidates for new indications can significantly reduce timelines, costs, and risks. AI is particularly strong at rapidly sifting through vast amounts of biomedical data to uncover hidden drug-disease connections and predict which existing drugs may be effective for a new therapeutic use.

This is already starting to see success according to Chao Yan and colleagues, who used AI to prioritise repurposing candidates for Alzheimer's disease in their 2024 article published in NPJ Digital Medicine, 'AI-driven predictive modelling techniques analyse molecular, genomic, and proteomic data to predict novel drug-target interactions. By integrating multi-omics data and biological network information, AI algorithms identify disease-associated targets and repurposable drugs that modulate these targets.'

One key advantage of AI-based repurposing is the ability to integrate and analyse diverse types of data, from chemical structures and drug-target networks to patient health records and scientific literature. 'AI can revolutionise the pharmaceutical sector by speeding up medication research and discovery,” states Cecilia Klein, an industry expert from Clarivate, in her 2024 article published on their website. “AI models like drug target identification, De-Novo drug design, virtual screening & molecular docking, drug repurposing, clinical trial optimization, pharmacokinetics & toxicity prediction, and natural product drug discovery may find prospective drug candidates, foretell drug–target interactions, and improve drug design by analysing massive information from clinical trials, academic literature, and biological databases.'

AI approaches are being applied across the drug repurposing pipeline. Machine learning can predict how strongly a drug will interact with potential new targets. Deep learning can generate novel molecular structures optimised for a particular disease. Natural language processing can mine unstructured text data, including social media posts, to find evidence supporting new indications. Mining real-world patient data from electronic health records and social media is increasingly important for validating repurposing hypotheses.

Epidemiological data on factors such as population size, disease incidence and prevalence, comorbidities, mortality, and cancer recurrence can provide valuable insights to help prioritise drug repurposing opportunities. By considering the prevalence and projected trends of different indications, researchers can focus on areas with the greatest potential impact.

There have already been some notable successes with AI-guided repurposing. For example, Yan's 2024 study found that the common diabetes drug metformin was associated with a significantly lower risk of Alzheimer's disease in patient data. Other AI-prioritised candidates like the cholesterol drug simvastatin also showed promise.

Using social media data in drug repurposing

Social media data is also playing an important role in compliant data gathering needed for drug repurposing, offering vital anonymised insights into patients' experiences, opinions, and self-reported outcomes related to their health conditions and treatments, complementing traditional data sources like electronic health records and clinical trials.

We can apply advanced AI techniques like natural language processing and sentiment analysis to social media posts to unlock valuable insights for drug repurposing efforts. For instance, AI algorithms can analyse patient conversations to identify potential new indications for existing drugs based on anecdotal reports of unexpected therapeutic benefits which can then be investigated. If a significant number of patients with a specific condition start mentioning that a particular drug, initially prescribed for another purpose, is alleviating their symptoms, this could signal a potential repurposing opportunity worthy of further investigation. Being able to sample and analyse large amounts of social data means that such anecdotal asides can start to build a genuinely valuable data picture.

Social media data can help researchers compliantly monitor and detect adverse events or side effects that may not have been captured during clinical trials. By tracking patient-reported experiences in real-time, AI tools can quickly surface new safety signals, enabling prompt action to be taken and potentially leading to the discovery of new contraindications or precautions for existing drugs. We wrote about how AI and social data can hugely benefit patient research here.

How Buzz Radar is using AI to extract actionable data from social media

At Buzz Radar, we've developed a game-changing AI assistant called BRIANN capable of doing the astonishing volume of analysis and calculation needed to contribute data to drug repurposing. BRIANN is like having a superhuman analyst on your team, tirelessly scouring social media platforms 24/7 to identify valuable patient insights. With BRIANN, we can monitor millions of social media posts across multiple languages in real-time, a feat that would be impossible with manual methods alone.

What sets BRIANN apart is its ability to understand the nuances of human language, even when people use slang or informal terms to describe their experiences with medications. By leveraging advanced natural language processing techniques, BRIANN can pick up on subtle cues and sentiments that traditional keyword-based approaches might miss. This means we can cast a wider net and capture a more comprehensive range of potential drug responses, and ensure no signals slip through the cracks, tapping into the collective experiences of patients worldwide and using that knowledge to guide research efforts towards the most promising repurposing opportunities.

Of course, with great power comes great responsibility. As we continue to push the boundaries of what's possible with AI and social media data, we must also ensure that we're using these tools ethically and responsibly. At Buzz Radar, we're committed to developing robust methodologies and guidelines that prioritise patient privacy and data integrity. We believe that the future of digital research around pharma lies in the responsible use of AI and real-world data, and we're excited to be at the forefront of this rapidly evolving field.

Ultimately, AI-driven repurposing could help get better treatments to patients faster and more affordably than ever before. As summarised by Rajaram J in a 2024 LinkedIn article, 'With continued innovation, collaboration, and regulatory support, AI-driven drug repurposing holds promise for revolutionising drug discovery and improving healthcare outcomes for patients.' While there is still much work to be done, the future of AI in drug repurposing, especially when combined with novel data streams from social media, looks beyond promising in the acceleration and discovery of life-saving therapies hiding in plain sight within existing drug arsenals.

To talk to us about how our industry-leading tools could open up new avenues and game-changing developments with products already on the market, drop us a line today.

Blog Posts

Category: SOCIAL MEDIA DATA UTILIZATION IN HEALTHCARE

AI-Driven Drug Repurposing: Finding New Uses for Existing Drugs