Case Study – Ariana Pharma, uniquely positioned to identify strong signal in complex datasets.
At Ariana Pharma, we are at the forefront of revolutionizing drug development through explainable Artificial Intelligence (xAI). Our groundbreaking AI technology and platform enable us to uncover crucial signals hidden within complex datasets.
In the field of drug development and market access strategies, we tackle pivotal questions head-on: For a specific disease, what happens in terms of disease evolutions at the scale of a large population? How to leverage Real-World Evidence in the context of drug development? How do we extract valuable insights from complex data matrices encompassing varying data types and sizes?
Our solution, KEMⓇ, answers these challenges.
Here is a striking example of how we found key information in a very small population and successfully validated it on a much larger scale. By analyzing a few cases in selected small cities in Brazil, we unearthed crucial associations between diagnoses and operations in Brazil as a whole.
Our methodology involved a random selection of five cities in the state of Sao Paulo, among the 5,570 cities in Brazil. These cities, home to a total of 500,000 inhabitants (out of Brazil’s 200 million), witnessed 300,000 hospital admissions from 2008 to 2022 (representing a staggering 174 million admissions nationwide). During this period, 516,115 diagnoses were recorded.
To comprehend the relationship between hospital admissions diagnoses and corresponding treatments or surgeries, we turned to our innovative KEMⓇ xAI platform. Leveraging Formal Concept Analysis, our platform generated an impressive 50,518 association rules, accompanied by related concepts, descriptors, and metrics.
KEMⓇ yielded two concepts that exhibited equally strong associations in both the selected cities and the entire country. Notably, one concept unveiled a remarkable signal: a mere 10 patients, suffering of Astigmatism after corneal transplantation, were treated with a very specific type of keratoplasty surgery, that avoids performing a new corneal transplantation. This Transplant topoplasty procedure was proposed and popularized by an ophthalmologist from Sao Paulo in 2004. Our KEM RWE platform allowed us to extract the time series of this keratoplasty, and visually represented its nationwide spread after 2016 through an illuminating map (refer to the GIF map).
With KEMⓇ’s confirmation across Brazil, we observed 3,000 patients treated with this procedure between 2008 and 2022, contributing to a vast hospital patient record of 174M.
The KEMⓇ Platform stands as a unique tool meticulously designed to unearth strong and consistent signals within intricate datasets. Whether working with clinical data featuring a small patient sample and an extensive array of clinical descriptors, or with Real-World Evidence characterized by a large patient population and limited descriptors, KEM® adapts to any scale. Moreover, our AI-driven hypotheses are easily validated in subsequent stages of development and within new cohorts.
Ariana Pharma’s KEMⓇ Platform provides transformative insights from complex data like never before.