Research Overview

Researchers from National Jewish Health have discovered that minor increases in lung scarring, identified through an AI-based tool on CT scans taken one year apart, correlate with disease progression and survival rates in patients suffering from fibrotic interstitial lung disease. This study, published in the American Journal of Respiratory and Critical Care Medicine, indicates that computer-assisted image analysis could offer a more objective and earlier identification of patients at risk for worsening conditions.

Key Findings

• Even small increases in fibrosis over a year indicate a higher risk of lung function decline and mortality.
• Patients with a fibrosis score increase of 5% or more had over a two-fold increased risk of death or lung transplant compared to those with stable scores.
• Changes were most pronounced in patients with less severe disease at the outset, highlighting the potential for early intervention.

Understanding Fibrotic Interstitial Lung Diseases

Fibrotic interstitial lung diseases, including idiopathic pulmonary fibrosis (IPF), are chronic conditions characterized by lung scarring that progressively hampers breathing. Traditional methods for predicting disease progression often rely on subjective assessments of symptoms, lung function tests, and radiologist interpretations of high-resolution CT scans, which can vary significantly.

Methodology

The study employed a deep learning technique known as data-driven textural analysis (DTA), developed by the Quantitative Imaging Laboratory at National Jewish Health. This method provides precise measurements of lung fibrosis on CT scans. The researchers found a strong association between increases in DTA fibrosis scores over one year and subsequent declines in lung function, as well as heightened risks of death or lung transplant.

Implications for Clinical Practice

According to David Baraghoshi, PhD, a biostatistician at National Jewish Health, this research illustrates how quantitative imaging and statistical modeling can reveal significant patterns in disease progression. By linking changes in fibrosis scores to future outcomes, imaging data can serve as a crucial marker for clinical trajectories.

Future Directions

The findings were validated using data from the Pulmonary Fibrosis Foundation Patient Registry, reinforcing their applicability. These insights could significantly influence clinical trials and patient management, as quantitative CT analysis may become a vital endpoint for trials, a tool for identifying high-risk patients, and a guide for treatment decisions in everyday practice.