Nijmegen-based AI software company, Thirona has developed an AI algorithm to detect and quantify lung abnormalities related to cystic fibrosis (CF).

PRAGMA-AI aims to process faster and automated analysis of CT scans of patients with the condition, delivering an assessment in several seconds.

WHY IT MATTERS

CF is a rare genetic disorder with approximately 70,000 known cases worldwide, characterised by lung problems and reduced life expectancy.

PRAGMA-CF is a quantitative method for the assessment of CF lung disease on CT scans and lung irregularities related to CF, such as abnormal airways and collapsed lung tissue. These quantifications are critical in determining the status of the patient and treatment paths but are not feasibly performed by humans in clinical care.

The method also takes up to several hours per patient and requires specialised data analysts. Thirona has now automated this method to analyse CT scans of CF patients in several seconds, without human interference.

The new AI-based algorithm PRAGMA-AI aims to enable large-scale usage in clinical trials and clinical care for better diagnostics, patient monitoring, and treatment planning.

The new algorithm is part of Thirona’s lung quantification software package LungQ, which automatically collates measurements supply information on the pattern and extent of CF lung disease.

LungQ PRAGMA-AI has been validated on a large number of scans of CF patients, showing high diagnostic performance comparable to trained human analysts.

THE LARGER CONTEXT

Thirona obtained its AI software patent in October 2020, and has since used the license to integrate the quantitative method it now uses for the analysis of chest CT scans.

Also in the CF space is competitor, NuvoAir which shared three key pulmonology related studies presented at the 2020 North American Cystic Fibrosis (CF) conference showing that patients with the respiratory disease can achieve good quality spirometry at home through the platform.

ON THE RECORD

Professor Harm Tiddens from Erasmus Medical Centre, who co-developed the original PRAGMA-CF method said: “It allows for large scale investigation of CF lung disease in both research studies and clinical trials. This is a crucial component in the evaluation of new expensive treatment options for patients suffering from CF.

 “CF patient registries, collecting information on the health status of CF patients, have already shown interest in the PRAGMA-AI method. They use patient information to create care guidelines, drive quality improvement, and to study CF treatments and outcomes.”