Prognostic value of a composite physiologic index developed by adding bronchial and hyperlucent volumes quantified...

Area of Science:

• Pulmonary Medicine
• Radiology
• Medical Imaging Analysis

Background:

• The composite physiologic index (CPI) estimates interstitial lung disease (ILD) extent in idiopathic pulmonary fibrosis (IPF) using pulmonary function tests (PFTs).
• CALIPER-CPI, a revised CPI, estimates ILD volume using automated CT software.
• Artificial intelligence quantitative CT (AIQCT) software can now quantify bronchial and ILD volumes separately in IPF patients.

Purpose of the Study:

• Develop AIQCT-derived CPI formulas to quantify CT abnormalities in IPF.
• Investigate the association of these novel CPI formulas with IPF patient survival.

Main Methods:

• Developed new CPI formulas based on PFTs and AIQCT-derived volumes: ILD-CPI, ILDB-CPI (ILD + bronchial), and ILDBH-CPI (ILD + bronchial + hyperlucent).
• Analyzed associations with survival in two IPF cohorts (n=116 and n=72).
• Compared the predictive performance of original CPI, CALIPER-CPI, and AIQCT-derived CPIs using C-indices.

Main Results:

• All AIQCT-derived CPIs, along with original CPI and CALIPER-CPI, were associated with overall survival.
• The ILDB-CPI demonstrated the highest C-index (0.759) among AIQCT-derived CPIs, comparable to original CPI (0.765) and CALIPER-CPI (0.749).
• Inclusion of hyperlucent volume (ILDBH-CPI) did not improve predictive accuracy (C-index=0.729).

Conclusions:

• The ILDB-CPI, derived from AIQCT, effectively predicts IPF patient outcomes.
• AIQCT-derived ILDB-CPI performance is comparable to established CPI measures.
• Adding hyperlucent lung volume to CPI formulas does not enhance mortality prediction accuracy.