Beyond Visual Interpretation: Quantitative Analysis and Artificial Intelligence in Interstitial Lung Disease Diagnosis "Expanding Horizons in Radiology"

Abstract

Diffuse lung disorders (DLDs) and interstitial lung diseases (ILDs) are pathological conditions affecting the lung parenchyma and interstitial network. There are approximately 200 different entities within this category. Radiologists play an increasingly important role in diagnosing and monitoring ILDs, as they can provide non-invasive, rapid, and repeatable assessments using high-resolution computed tomography (HRCT). HRCT offers a detailed view of the lung parenchyma, resembling a low-magnification anatomical preparation from a histological perspective. The intrinsic contrast provided by air in HRCT enables the identification of even the subtlest morphological changes in the lung tissue. By interpreting the findings observed on HRCT, radiologists can make a differential diagnosis and provide a pattern diagnosis in collaboration with the clinical and functional data. The use of quantitative software and artificial intelligence (AI) further enhances the analysis of ILDs, providing an objective and comprehensive evaluation. The integration of "meta-data" such as demographics, laboratory, genomic, metabolomic, and proteomic data through AI could lead to a more comprehensive clinical and instrumental profiling beyond the human eye's capabilities.

Keywords: AI (artificial intelligence); HRCT (high-resolution computed tomography); ILDs (interstitial lung diseases).

Figure 1

(A,B) (image J, Java open source): In (A), a control normal lung is depicted by a sharply shifted Gaussian curve with a narrow and tall peak. In (B), a representative slice of the right lung from a patient with advanced idiopathic pulmonary fibrosis, obtained through thin-section volumetric CT, is shown and illustrates the density histogram of the fibrotic lung, which exhibits a less pronounced peak and skewness compared to panel A. The segmented area of interest is highlighted in yellow. On the right side, the results of the digital processing analysis for this specific slice are presented. Once the sampling of the entire lung was completed, the software automatically generated averaged data from the analysis of all slices of both lungs.

Figure 2

(Slicer3D): UIP/IPF: Color-coded map differentiated for lung areas. Histogram analysis provides an accurate description, both globally and regionally, of mean lung density (MLD), skewness, and kurtosis parameters, with increased potential to stratify fibrotic damage.

Figure 3

High-resolution computed tomography (HRCT) images and regions classified as fibrosis by data-driven texture analysis of a 53-year-old man former smoker: (A) the baseline of fibrosis score was 11.2%; baseline FVC % pred was 73%, D_LCO % pred was 69.0%. (B) HRCT images in the same subject at a nominal 78-week follow-up. Regions classified as fibrosis by DTA are shown in orange. The DTA fibrosis score increased by 12.6% percentage points at follow-up. FVC declined 10.5% (relative to baseline), D_LCO declined 13.0% (relative to baseline).