Quantification of pulmonary perfusion in idiopathic pulmonary fibrosis with first pass dynamic contrast-enhanced perfusion MRI

Abstract

Introduction: Idiopathic pulmonary fibrosis (IPF) is a fatal disease of lung scarring. Many patients later develop raised pulmonary vascular pressures, sometimes disproportionate to the interstitial disease. Previous therapeutic approaches that have targeted pulmonary vascular changes have not demonstrated clinical efficacy, and quantitative assessment of regional pulmonary vascular involvement using perfusion imaging may provide a biomarker for further therapeutic insights.

Methods: We studied 23 participants with IPF, using dynamic contrast-enhanced MRI (DCE-MRI) and pulmonary function tests, including forced vital capacity (FVC), transfer factor (TL_CO) and coefficient (K_CO) of the lungs for carbon monoxide. DCE-MRI parametric maps were generated including the full width at half maximum (FWHM) of the bolus transit time through the lungs. Key metrics used were mean (FWHM_mean) and heterogeneity (FWHM_IQR). Nineteen participants returned at 6 months for repeat assessment.

Results: Spearman correlation coefficients were identified between TL_CO and FWHM_IQR (r=-0.46; p=0.026), K_CO and FWHM_mean (r=-0.42; p=0.047) and K_CO and FWHM_IQR (r=-0.51; p=0.013) at baseline. No statistically significant correlations were seen between FVC and DCE-MRI metrics. Follow-up at 6 months demonstrated statistically significant decline in FVC (p=0.040) and K_CO (p=0.014), with an increase in FWHM_mean (p=0.040), but no significant changes in TL_CO (p=0.090) nor FWHM_IQR (p=0.821).

Conclusions: DCE-MRI first pass perfusion demonstrates correlations with existing physiological gas exchange metrics, suggesting that capillary perfusion deficit (as well as impaired interstitial diffusion) may contribute to gas exchange limitation in IPF. FWHM_mean showed a significant increase over a 6-month period and has potential as a quantitative biomarker of pulmonary vascular disease progression in IPF.

Keywords: idiopathic pulmonary fibrosis; imaging/CT MRI etc; interstitial fibrosis.

Figure 1

Left panel: Example of signal change through a parenchymal region of interest. The full width at half-maximum (FWHM) signal is often used as a standard threshold for integration. Right panel: CT with best matching coronal slice mean transit time (FWHM) maps derived from dynamic contrast-enhanced MRI from two participants (A+B and C+D). Reticulation and honeycombing on the CT is related to regions of increased transit time in basal and peripheral regions. Colour bar denotes FWHM in seconds.

Figure 2

Reproducibility of regional full width at half maximum (FWHM). An example of one subject undergoing same day repeat imaging at two time points (A and B). Colour bar denotes FWHM in seconds.

Figure 3

Box and whisker diagrams of baseline and 6-month changes in FWHM_mean (A); FWHM_IQR (B); FVC % predicted (C); K_CO SI units (D); TL_CO SI units (E); TL_CO % predicted (). Wilcoxon signed-rank p values quoted for each. FVC, forced vital capacity; FWHM, full width at half maximum; FWHM_IQR, IQRof the FWHM; FWHM_mean, mean FWHM;K_CO, diffusingcoefficient of the lungs for carbon monoxide; TL_CO, diffusingcapacity of the lungs for carbon monoxide.

Figure 4

Representative parametric transit time (FWHM) maps from two participants each with baseline (A and C) and follow up (B and D) scans. Participant 12 (A and B) demonstrates an increase in FWHM in the right lower lobe and left mid-zone peripheral lung tissue. FWHM_mean decreased slightly by 0.13 s, with associated changes in FVC from 60.1% to 63.2%, and TL_CO from 37.1% to 33.0%, respectively. Participant 18 (C and D) shows a gross overall increase in FWHM, again worse in peripheral lung tissue. FWHM_mean increased by 4.35 s, in spite of both FVC and TL_CO remaining relatively stable from 76.0% to 74.5%, and from 87.2% to 88.2%, respectively. Colour bar denotes FWHM in seconds. FVC, forced vital capacity; FWHM, full width at half maximum; FWHM_mean, mean FWHM;TL_CO, diffusingcapacity of the lungs for carbon monoxide.