[Image from Lokavant]
The company, spun out from Roivant Sciences, designed Spectrum v15 to overcome the limitations of traditional feasibility analysis. Instead of static, often inaccurate manual comparisons pulled from disparate sources, the AI-powered platform taps data from some 500,000 trials to enable dynamic scenario planning.
Clinical trials are a significant driver of drug development bottlenecks. Nearly 80% of studies miss enrollment deadlines, and for a blockbuster program a single day of delay can vaporize about millions in lost revenue.
To enable this dynamic planning, Spectrum allows teams to model feasibility at granular levels (site, country, region, study), and configure forecasts for adaptive trial designs. The platform can even determine required screening volumes to meet enrollment goals. Spectrum also integrates real-time study data for continuous re-forecasting, allowing mid-study course corrections. Pfizer’s Head of Predictive Analytics, Jonathan Crowther, noted in the announcement, “This isn’t just operational efficiency, it’s strategic foresight,” enabling development to be “anticipatory and resilient.”
One factor that distinguishes Lokavant’s Spectrum pitch from others is its focus on causal-AI. That is, while it is relatively straightforward to note an enrollment trend at a given cluster of sites, identifying the underlying drivers or enrollment can be more challenging. Causal AI aims to uncover the underlying drivers determining why those trends occur. That could mean linking specific protocol criteria or site characteristics to actual enrollment speed or dropout rates.
This deeper understanding that causal AI offers can help chip away at modern clinical trial complexity. “As an industry experiencing unprecedented volatility, there is a great need to quantify uncertainty while identifying reliable paths to study enrollment success,” said Rohit Nambisan, Lokavant CEO and founder, in the announcement.
Filed Under: clinical trials, machine learning and AI
