Companies testing AI analysis tool for early diagnosis of PF
Brainomix e-Lung detects changes on scans to identify deterioration
Written by |
Brainomix, a health software company, and drug developer Boehringer Ingelheim will advance testing of an artificial intelligence (AI)-based tool that may enable earlier diagnosis of progressive pulmonary fibrosis (PF).
The two companies are collaborating to run studies of the Brainomix 360 e-Lung, a U.S. Food and Drug Administration (FDA)-cleared technology for analyzing routine CT scans of the lungs. The e-Lung may help identify subtle changes in lung images that correspond with scarring (fibrosis) and deterioration, hallmarks of progressive PF.
“The evidence generated to date for e-Lung is highly compelling, showing the technology has the potential to accelerate diagnosis by more than two years,” Michalis Papadakis, PhD, CEO and co-founder of Brainomix, said in a company press release.
Brainomix will work with Boehringer on a new study to assess the potential real-world impacts of the e-Lung. Boehringer markets the PF medications Jascayd (nerandomilast) and Ofev (nintedanib).
“This next phase will enable us to evaluate [the e-Lung’s] potential at scale, providing robust real-world validation of what we expect could be a transformative advancement in the patient care pathway,” Papadakis said.
Early diagnosis important, but challenging
Progressive PF is a condition in which permanent scar tissue builds in the lungs, making breathing difficult. Treatment may work best if initiated earlier in disease progression, but diagnosing PF early can be challenging.
“Innovations that enhance our ability to detect disease have the potential to improve patient outcomes,” said Emmanuelle Clerisme-Beaty, MD, Boehringer’s senior vice president of medicine and regulatory affairs.
Brainomix designed the e-Lung to help fill this gap. Using patient CT scans, the company trained the AI to identify signs of fibrosis.
The e-Lung incorporates a metric called the weighted reticulovascular score (WRVS), which assesses tissue and blood vessel features for signs of lung scarring. According to Brainomix, WRVS predicts lung function decline in people with PF.
In the first part of their collaboration, Brainomix and Boehringer sponsored a study, REVISE-PPF, in which investigators retrospectively analyzed CT scans from 253 patients across three lung clinics. The patients weren’t diagnosed with progressive PF, and doctors considered them clinically stable.
Results demonstrated that between 55% and 97% of participants, depending on the center, showed signs of progressive PF. The researchers estimated that the e-Lung could have supported a progressive PF diagnosis up to 28 months earlier for 71% of participants.
Investigators will discuss REVISE-PPF data showing that e-Lung analysis can also help predict risk of progressive PF from early lung scans in a May 17 presentation at the American Thoracic Society’s international conference.
The companies will now conduct a prospective study, PROGRESS-PPF. Researchers will follow participants to determine whether the e-Lung can support earlier diagnosis of progressive PF in real-world clinical practice, leading to earlier treatment initiation and better health outcomes.
