AI-powered Imaging Biomarker May Predict IPF Lung Health, Survival
A novel imaging biomarker identified via an artificial intelligence (AI)-powered analysis could help predict lung function and survival outcomes for people with idiopathic pulmonary fibrosis (IPF).
Brainomix, a U.K.-based software company, is developing a number of AI-based tools to analyze different diseases, including one called e-ILD for IPF and other fibrotic lung diseases, it announced in a press release.
The general idea behind AI-driven analysis is to feed clinical data into a computer for analysis via pre-specified mathematical rules and algorithms. The computer then “learns” from the data, finding patterns based on the specific algorithms given.
For instance, the computer might deduce that patients with a certain constellation of features visible on lung scans are more or less likely to experience a decline in lung function. These deductions then can be used to make prognoses and predictions for other patients.
Brainomix announced this week that the latest version of its e-ILD software has been refined using data from the Open Source Imaging Consortium (OSIC) Data Repository, which was launched last year by OSIC.
OSIC is a global nonprofit collaboration among academics, industry leaders, and advocacy groups that assembles and stores imaging and other data in an effort to better understand IPF and other lung and respiratory diseases.
“We are pleased to see these encouraging data and developments from Brainomix as they highlight the important value of our consortium in enabling the development of novel solutions to combat pulmonary fibrosis,” Elizabeth Estes, OSIC’s executive director, said in the release.
The analysis using e-ILD “demonstrated the value of several AI-powered imaging biomarkers,” notably its own proprietary biomarker, which assesses the amount of lung tissue displaying abnormalities particular to interstitial lung diseases (ILDs), Brainomix stated. And it showed this AI-powered biomarker can predict both lung function decline and survival in IPF patients based on information from an initial CT scan.
In addition, e-ILD can help predict transplant-free survival two times better than forced vital capacity, an established lung function measure that assesses how much a person can exhale in a forced breath, the company reported.
It also expects that this new biomarker could be helpful in selecting patients at highest risk of disease progression for clinical trials and be useful for measuring disease activity over time.