AiCure Announce Open Source Computer Vision Platform
By Clinical Research News Staff
October 29, 2020 | AiCure has announced OpenDBM, an open-source version of its computer vision and AI-powered digital biomarker platform. The platform's transparent framework will give the scientific community access to AiCure's digital biomarker algorithms, according to an announcement from the company, and the ability to apply them to their own datasets to measure patient responses to treatment, including facial, vocal, and motor characteristics.
AiCure’s platform gathers and analyzes visual and auditory cues directly through a patient's smartphone camera, pinpointing critical disease characteristics and behavioral trends. Through accurate and consistent data capture, the platform helps to ensure the integrity of clinical trial data. The AiCure app walks a patient through the process of taking a study medication, recording precise details on when and how a study medication was taken (and if someone tried to fudge the process).
AiCure hopes to encourage scientific scrutiny and collaboration not only to track and understand patient behavior, but to advance the development of digital biomarkers for remote assessment. The company—along with Merck and Kent State University—will discuss the need to democratize access to these algorithms at the 2020 CNS Summit this week.
AiCure's OpenDBM will give the scientific community access to AiCure’s digital biomarker algorithms to measure patient responses to treatment, including facial, vocal and motor characteristics. Scientifically validated digital biomarkers hold great promise to enhance a clinical trial's objectivity, sensitivity and frequency of assessment, the company believes.
Despite the potential, digitally measuring patient behavior is often shrouded in mystery, as proprietary machine learning models are typically not accessible to scientists to independently evaluate. By expanding access to these algorithms, AiCure hopes to empower the pharmaceutical industry and scientific community to improve their understanding of disease symptomology, drug dosing side effects, and stratified disease variations.
"While digital biomarkers help to eliminate the blind spots of infrequent and subjective in-person visits, their exclusivity means many in the scientific community are still flying blind when it comes to measuring the impact and validity of these proprietary algorithms, limiting their use and the weight they carry during regulatory conversations," said Ed Ikeguchi, CEO of AiCure, a press release announcing the open source release. "AiCure's OpenDBM opens this black box. Through an open science framework, we hope to unleash the potential of digital biomarkers to not only safeguard the success of a trial and understanding of a drug's impact, but also better equip sites to offer each patient the support they need."