Technology
Opyl’s technology revolutionises the landscape of clinical trial exploration, design, optimisation, and prediction, offering unparalleled insights and opportunities for patients, doctors, clinical researchers, pharmaceutical companies, and medical investors.
A testament to the reliability and precision of our technology, Opyl’s newest product, TrialKey, accurately forecasted a significant 74.51% probability of success for the groundbreaking DMX-200 trial, even before the official announcement of positive interim results. In the realm of clinical trials, foresight is paramount for unlocking potential gains, and our predictive capabilities provide a decisive advantage.
How It Works
Data Acquisition: Our technology aggregates data from an extensive database comprising over 65,000 clinical trials spanning more than two decades.
AI Analysis: Employing advanced AI algorithms, the data is analysed through over 700 variables, including crucial drug parameters.
Algorithm Validation: Our methodology is rigorously tested using data from 4,189 clinical trials to ensure precision and recall accuracy.
Simulation: Utilising probability distribution simulations, we evaluate potential trial outcomes, providing insights into success probabilities.
Methodology
Data Collection: We collect publicly available clinical trial data from reputable registries such as ANZCT.
Data Processing: Collected data undergoes meticulous filtering and processing to develop features, incorporating structured data and insights generated by Natural Language Processing models.
Variable Categorisation: Variables derived from the data are segmented into Trial Execution Variables and Technology Variables, streamlining the prediction process.
Target Definition: Our technology aims to predict the likelihood of a trial successfully meeting its endpoint at its respective phase, guiding decision-making processes.
Machine Learning Modeling: We deploy various machine learning models to enhance prediction accuracy, ensuring robust and reliable insights for stakeholders.