In the age of artificial intelligence, no trial data should be going to waste. Findacure’s Rick Thompson looks at how these technologies could bring us closer to treatments for underserved rare diseases.
The repurposing of drugs is becoming more common, especially in the field of rare diseases. In the past, repurposing has mostly been driven by academics looking for new possibilities in generics. Now, as part of lifecycle management, pharmaceutical companies are looking more closely at drugs they have on their shelves. These might be licensed drugs that could hold potential for a patent extension, or drugs which failed efficacy trials for an intended indication.
In the quest to repurpose a drug for a rare condition, there is a need to look at any and all available data. The wealth of published scientific literature forms one crucial source of data, with the ever-expanding pool of ‘omic data forming another.
A third pool of clinical evidence is formed by trial data, which will probably only be considered through the published literature. By definition, however, trial master files represent a much richer and more detailed source of data on a drug and how it performs. Published literature tends to catalogue successful clinical trials, but value can also lie in a trial that did not lead to a positive and viable outcome: the data it produced could still provide evidence for repurposing. For instance, provided a drug has not failed a trial on safety, the side effects it caused in one population could constitute on-target effects in another.
With large datasets crucial to gaining an understanding of rare diseases and opening the door to drug development, digital technology is proving transformative. It enables careful collation and organisation of information, but the innovations of artificial intelligence (AI) are now taking things further, facilitating the effective analysis and interrogation of big data to create new treatment hypotheses.“Patient associations are working to develop registries (some using wearable technologies or apps) and natural history studies, which means that ever-greater volumes of data are being produced”
These techniques make the production of and access to high-quality data on rare diseases the gateway to treatment identification, and so are proving more crucial than ever for organisations in pharma.
Digitised, rigorously controlled data lends itself to techniques of processing and analysis which characterise both drug discovery and drug repurposing.
Raw text can be analysed by Natural Language Processing (NLP) techniques which form connections between studies that could otherwise take thousands of hours of human time to identify.
When combined with analyses of ‘omic approaches, and an appropriate level of disease-specific knowledge from patient groups, you can create a powerful resource for the identification of new treatment hypotheses for rare diseases – and an opportunity to address severe unmet needs.
Findacure is a charity that works directly with rare disease patient groups to help them grow and professionalise. Over the last five years we have focused on the power of drug repurposing for rare genetic diseases. It is estimated that, worldwide, just 400 treatments are licensed for 7000 known rare conditions, which tend to be determined by a very specific genetic factor.
As a consequence, most patients are being left with no hope of a treatment in their lifetime. Luckily, patient associations are working to fill the void by uniting patients and driving research forward for their conditions. Many are working to develop registries (some using wearable technologies or apps) and natural history studies, which means that ever-greater volumes of data are being produced.
This drive to generate data on and interest in their condition – along with the collated knowledge of their community’s lived experience of rare disease – can prove transformative to the treatment landscape. We are now seeing patient associations involved in several collaborative efforts that are identifying drugs which, as candidates for repurposing, stand to deliver treatments to rare disease patients more quickly and cheaply.
In 2020, the pharmaceutical industry has not by any means proved immune to the disruption caused by COVID-19. But, as in other industries, the pandemic has accelerated the process of digital transformation that was already underway.
A recent survey of more than 200 life sciences professionals, conducted on behalf of digital archiving specialists Arkivum, found 70% of respondents saying that COVID-19 has triggered a change in the way clinical trials will be conducted. There can be no doubt that digital technology will play a key role in that change. The survey reports that over 90% of sponsors and CROs have already adopted an eClinical application to improve study execution and data collection in live trials.
When a trial is completed, the valuable and extensive data it has produced must be archived – an exercise crucial both to regulatory compliance and to any future efforts at repurposing. 70% of sponsors reported that they use a digital archive rather than the traditional paper-based option, and 45% of respondents cited the role that clinical trial data plays in finding new indications and formulations. Yet at the same time, 38% of sponsor organisations described their ability to access archived clinical data and records as ‘extremely or very inadequate’.
This percentage rose to 65% amongst QA, compliance, legal and regulatory professionals. Moreover, just 31% of life sciences organisations seem to run a digital archive of sufficient sophistication to ensure that data can be managed in accordance with the FAIR data principles.
These were established to further scientific study through keeping data Findable, Accessible, Interoperable and Re-usable – all key attributes when it comes to exploring the new potential of an existing drug.
In the search to repurpose drugs, readier, more reliable access to archived trial data – including trials that produced negative results – can clearly prove highly beneficial.
If data has been well stewarded before and after it reaches the archive, and if its integrity has been maintained through careful curation, it facilitates the application of AI techniques. Natural language processing can be used in conjunction with, say, analysis of ‘omic-level data and patient group insights in order to work through the problems and side effects encountered in the full spectrum of trials. This can open the way to repurposing for different populations, and to new approaches to the design of clinical trials.
The success of these endeavours will also be favoured by the availability of comprehensive rare disease registries which collate patient-level data on disease natural history while also bringing together a pool of patients who could participate in trials.
Meanwhile at the pre-competitive stage of drug development, researchers are adopting a more open, collaborative approach to data. Now is the time to enable further collaboration by increasing access to historical data and releasing its full value. Success in finding treatments for rare disease is above all the product of collaboration, as technological innovation complements and amplifies a compassionate, patient-centred approach.
In all this, it is worth remembering that the people who participate in clinical trials – especially in the field of rare disease, where recruitment of patients is a particular challenge – would appreciate knowing that their participation will have a lasting value, whatever the outcome of the trial.
Trial participants take on a burden by putting in time, effort, hope and commitment. They also put themselves at some degree of risk whenever they take an experimental drug. In the field of rare diseases, trial participants are hoping to help the next generation of patients even more than themselves.It is crucial to maximise the potential value of data they are helping the professionals to collect.
With repurposing on the table, and improved access to all trial data, we can better unlock this potential.