• IPF
• HRCT

Introduction
The clinical course of Idiopathic Pulmonary Fibrosis (IPF) is largely unpredictable. For an individual with IPF, the overall prognosis, rate of progression and likelihood of exacerbations are highly variable. The current clinical standard methods of assessing disease severity and change over time, pulmonary function and visual assessment of High-Resolution Computed Tomography (HRCT), are inconsistent, expensive and interpretation can be subjective. Furthermore, even with recently available antifibrotic therapy, and the individual efficacy and impact of treatment is difficult to assess with current tools.
Objectives
Primary aims of the present study were to improve clinical assessment of IPF, the evaluation of disease over time and use these features to better predict outcomes. Specific primary aims include:
1. Evaluate, at baseline, the correlation between High Resolution Computed Tomography (HCRT) abnormalities with accepted measures of disease severity such as functional parameters;
a. Using semi-quantitative (visual score) performed by radiologists
b. Using quantitative image analysis software
2. Identify, at baseline, possible radiological thresholds to predict the survival of IPF patients in 3-year follow-up;
3. Evaluate the role of changes of HRCT abnormalities in IPF patients in relation to treatment
a. Using semi-quantitative (visual score) performed by radiologists
b. Using quantitative image analysis software
Secondary aims were to:
1. Evaluate, at baseline, the role and prognostic value of ultrasound B lines compared with HRCT abnormalities
2. Assess the potential role of procoagulant microparticles derived from endothelial cells in the peripheral blood as a potential screening test, prognostic biomarker or surrogate for disease severity in IPF.
Materials and methods
One-hundred and six consecutive patients with a multidisciplinary diagnosis of IPF, and for whom one supine, non contrast volumetric HRCT (≤2 mm slice thickness) at baseline and pulmonary functions (PFTs) within 3 months were available were included in the study. Images were evaluated by a radiologist using a semi-quantitative visual feature scoring system and also characterized with quantitative HRCT analysis software. The correlations between PFTs and quantitative analysis and semi-quantitative visual scores at baseline were assessed. The best radiological thresholds for survival prediction were calculated based on a 3-year follow-up for all 106 patients. Fifty-nine patients who underwent at least two serial non-contrast, supine, volumetric HRCTs (≤2 mm slice thickness) were classified into two groups based on treatment. The HRCT's at each timepoint were quantitatively analysed and visually scored.
The quantitative HRCT image analysis utilized in this study was Computer-Aided Lung Informatics for Pathology Evaluation and Rating (CALIPER) software, developed at Mayo Clinic in Rochester, MN, USA. CALIPER performs automated anatomic segmentation (lung extraction, tracheobronchial tree segmentation and vascular-related structure identification), volumetric parenchymal histogram pattern matching and morphological feature analysis to characterize lung patterns. CALIPER patterns are determined by algorithmic identification and volumetric quantification of every voxel volume unit into one of eight radiological features: normal lung, three grades of decreased lung attenuation (mild, moderate and severe low attenuation areas; LAA), ground glass opacification (GGO), reticular pattern (R), honeycombing (HC) and pulmonary vascular-related structures (PVRS). Analysis of the total extent of interstitial disease was performed as a percent of interstitial lung abnormalities (ILD%) as a sum of the percent of areas of ground glass, reticular and honeycombing. Analysis of the pulmonary vascular related structure were performed as percent of pulmonary vascular relate structures volume (PVRS%).
Radiologic abnormalities (honeycombing, reticulation, ground-glass and traction bronchiectasis) and fibrotic score (reticulation+ honeycombing) were visually scored at baseline and during follow-up. A semi-quantitative analysis of the percentage of these radiologic abnormalities to the nearest 5% in three zones for each lung, was performed as described by Best et al.
To address the secondary aims of the study, two subgroups of patients were identified:
1) A subset of 24 number of patient had a lung ultrasound evaluation immediately after the HRCT scan for assessing position and number of B-lines
2) A subset of 19 in which circulating endothelial derived MP (EMP) were obtained from peripheral blood within one month of the HRCT for evaluating their role as biomarkers of IPF.
Results
The correlation between PFTs and radiological parameters at baseline was significant for almost all HRCT abnormalities visually and quantitatively assessed with CALIPER. There was a strong correlation for FVC with ILD (r=0.48) and PVRS (r=0.50). There was also a strong correlation for DLCO with PVRS (r=0.52) and also with visual fibrotic score (r=0.47). A baseline ILD score higher than 20 %, PVRS score greater then 5% and a visual fibrotic score higher than 30 % defined a worse prognosis during 3-year follow-up.
A significant progression of radiological abnormalities in treated and in untreated patients was found. In particular, ILD% and PVRS% increases faster in untreated patients. ILD progression was 0.07% per month in treated patients, and 0.37% per month in untreated patients (p=0.002). PVRS increased 0.046% per month in treated and 0.005% per month in untreated patients (p<0.001).
In patients with lung ultrasound evaluation we demonstrated a correlation between B-lines score and both visual and quantitative radiological parameters (PVRS %, r=0.71, p<0.001).
The number of Endothelial-derived MP (EMP) was higher in IPF patients compared to controls, although the difference did not reach the predefined level of significance.
Conclusions
CALIPER quantification of fibrosis (ILD %) and vascular involvement (PVRS) at baseline and during the follow-up could distinguish the progression of disease based on treatment and predict the survivals in IPF patients.
LUS might have a role as a screening test before HRCT.
Based on data available at this point, EMP cannot be justified as a serum biomarker.