AI for Clinical Trial Control Arms

Problem statement

The challenge, as defined by AION Labs’ partners, led by Teva Pharmaceuticals, is to reduce the number of control patients needed in clinical trials. This will be facilitated by computationally matching control patients from a shared pool of control patient datasets from past clinical trials, leveraging the sharing of data in a federated environment from AION Labs pharma partners and other sources.

It currently takes around a decade and more than a billion dollars in funding to bring a new drug to market, half of which is spent on clinical trials that are growing ever larger and more complex. The traditional approach of double-blinded-randomized control trials is considered the gold standard for validity and regulatory approval of drug candidates.

Previous attempts to produce synthetic patient cohorts have been based on real world evidence (RWE) to reduce the number of patients needed in clinical trials. While promising, RWE differs from clinical trials in data collection, combination of treatments and measurable endpoints.  In contrast, control arms from previous clinical trials can provide the best data source for modeling clinical trial outcomes with the standard of care, assuming the data becomes accessible.

The solution envisioned will be centered around a pool of control patients for various clinical indications, based on historical and current real patient data. The clinical data will be stored with all relevant data commonly collected from clinical trials (inclusion/exclusion data, primary and secondary readouts). When conducting a clinical study, pharma companies will recruit a treatment arm and a smaller control arm

The AI matching solution will expand the selected control arm from the existing pool of shared patients based on the study protocol and the existing arms in the study. We expect the solution to include cohort matching metrics and potential suggested changes to the study protocol that will increase these metrics to enhance confidence in the suitability of the cohort expansion.

What we are looking for

We are seeking an experienced entrepreneur with an outstanding track record and history of successful ventures. Those with a strong interest in the fields of machine learning, language models, clinical studies, computational biology or other relevant fields, are invited to apply for an opportunity to join our team as an EIR.

The successful EIR will build a comprehensive business plan within a six-month period of starting at AION Labs. During this period the AION Labs team and partners will be available to work with the EIR until a final plan is formulated and presented to the investment committee for approval. We expect the plan to address the following points in order of importance:

• Fine tune the problem statement and market research.
• Define the business model that will create incentives for pharma companies to provide access to their control arm data.
• Suggest a pharma compliant data infrastructure.
• Suggest a regulatory and legal framework for reducing control arms in clinical trials.
• Define a business model including competitive landscape, market fit and go-to-market strategy.
• Suggest a novel technological solution for augmenting control arms with data from previous control patients.

Application and selection process

• Applications should be submitted to: eir@aionlabs.com
• As part of the application procedure, please submit: (i) Your curriculum vitae including your publication/track record. (ii) a cover letter describing your envisioned approach of studying this field as an EIR.
• Candidates will be invited for an interview with the AION Labs team and selected strategic partners.
• The selected EIR will have 6 months to prepare for an Investment Committee meeting to secure a pre-seed investment offer.