AI FOR RARE DISEASE: ACCELERATING DRUG REPURPOSING AND DIAGNOSIS

As the complexity of rare diseases continues to challenge traditional treatment and diagnostic methods, artificial intelligence (AI) is emerging as a transformative tool. It enables the repurposing of existing drugs for rare disease treatments and offers faster, more accurate diagnoses for patients.

In this blog, we explore the evolving role of AI in drug repurposing for rare diseases, its potential to facilitate diagnosis, and the profound impact it’s having on the rare disease patient journey.

The potential of drug repurposing

Drug repurposing is about finding new uses for existing drugs where they were not previously indicated, turning challenges into opportunities.¹ This approach has gained traction in response to the notoriously high costs and unsustainability of the conventional drug discovery process. For rare diseases, where resources are often scarce, drug repurposing offers a faster, smarter way to bring treatments to patients in need.

Leveraging AI in identifying repurposing opportunities

AI is revolutionising drug repurposing, making it faster and more effective to uncover treatments for rare diseases. AI analyses large datasets to make predictive models for both repurposing and diagnosis based on known metrics. While data-driven approaches to drug discovery are not new, AI introduces innovative techniques like deep learning, where artificial networks modelled loosely on the human brain generate insights that were previously out of reach.²

One standout example is the novel AI model TxGNN. Developed by scientists at Harvard Medical School, TxGNN is a metric learning graph neural network (GNN) model able to predict drug repurposing opportunities by analysing relationships between drugs, diseases, and molecular interactions:^3,4

• Designed to identify potential drug candidates, ranking drugs as possible indications and contraindications for 17,080 diseases.
• Trained on diverse data sources – including DNA sequences, cell signalling, gene activity levels and clinical notes.
• Identifies patterns in shared disease mechanisms (e.g. genomic aberrations) and makes predictions by extrapolating data from well-understood diseases with established treatments to diseases that are poorly understood.

AI tools like TxGNN don’t just expand the pool of viable drug candidates, they also reduce costs and the risk of clinical trial failures. With nearly 8,000 FDA-approved and experimental drugs analysed, the model accelerates progress for rare diseases, many of which remain neglected.³

Transforming rare disease diagnosis with AI

Machine learning’s predictive capabilities, especially in analysing large and complex data sets, offer vast potential for generating valuable insights into patient diagnoses, particularly where rare disease knowledge is limited, and faster diagnosis is particularly impactful. Its ability to predict survival timeframes, recurrence metastasis risks and potential therapy responses holds exceptional value for both patients and clinicians.⁵

A statement about the TxGNN AI model in transforming rare disease:

“With this tool we aim to identify new therapies across the disease spectrum but when it comes to rare, ultrarare, and neglected conditions, we foresee this model could help close, or at least narrow a gap that creates serious health disparities.”

Algorithms have been designed to compile networks and register patient information to identify new cases of rare diseases. Utilising AI in this way allows for the development of strategies to implement and access therapies for managing these conditions, making these systems a valuable tool for clinicians. Diagnostic decision support systems provide relevant differential diagnoses, enabling faster and potentially more accurate diagnosis.

Ultimately, this technology could facilitate the early detection of rare diseases, leading to improved outcomes and more timely treatment options, potentially lengthening survival timeframes and enhancing patient quality of life.

Though the future success of AI applications in the rare disease space is both exciting and promising, it is important to acknowledge that these models are still in their infancy. One significant drawback is that available data for rare diseases is often small and heterogenous due to limited accessible knowledge surrounding these conditions.⁵ This highlights the need for more research in rare diseases to generate useable data, as well as the implementation of data augmentation techniques, where specific AI models can be trained to identify patterns from smaller data sets.⁵

Despite these challenges, the future remains bright for AI applications within the rare disease field and the importance of advancing research on rare diseases using this technology will remain a trending topic.

The impact on the rare disease patient journey

Rare diseases, by their nature, present significant challenges like limited awareness, small patient populations and diagnostic complexity. AI helps address these issues by streamlining diagnostics, optimising patient identification and personalising outreach. Fundamentally, this is reshaping how rare disease brands connect with healthcare professionals, patients and caregivers.

AI also enables highly personalised marketing strategies. By analysing patient data, marketers can tailor messaging and educational materials based on individual patient profiles. AI can segment audiences more effectively, ensuring that the right message reaches the right healthcare provider or patient group. This increases the relevance and impact of campaigns, fostering stronger connections with a target audience.

For rare disease brands, AI signifies innovation, precision, and hope. It has the potential to speed up diagnosis times, repurpose existing drugs for rare disease treatment, and facilitates marketers in advancing the conversation about these conditions. Ultimately, brands which embrace the potential of AI, position themselves as forward thinking and dedicated to improving the lives and families of those affected by rare conditions.

To find out more about how IGNIFI can work with your rare disease brand click here or contact us.