GOLD grade-specific characterization of COPD in the COSYCONET multi-center trial: comparison of semiquantitative MRI and quantitative CT

Abstract

Objectives: We hypothesized that semiquantitative visual scoring of lung MRI is suitable for GOLD-grade specific characterization of parenchymal and airway disease in COPD and that MRI scores correlate with quantitative CT (QCT) and pulmonary function test (PFT) parameters.

Methods: Five hundred ninety-eight subjects from the COSYCONET study (median age = 67 (60-72)) at risk for COPD or with GOLD1-4 underwent PFT, same-day paired inspiratory/expiratory CT, and structural and contrast-enhanced MRI. QCT assessed total lung volume (TLV), emphysema, and air trapping by parametric response mapping (PRM_Emph, PRM_fSAD) and airway disease by wall percentage (WP). MRI was analyzed using a semiquantitative visual scoring system for parenchymal defects, perfusion defects, and airway abnormalities. Descriptive statistics, Spearman correlations, and ANOVA analyses were performed.

Results: TLV, PRM_Emph, and MRI scores for parenchymal and perfusion defects were all higher with each GOLD grade, reflecting the extension of emphysema (all p < 0.001). Airway analysis showed the same trends with higher WP and higher MRI large airway disease scores in GOLD3 and lower WP and MRI scores in GOLD4 (p = 0.236 and p < 0.001). Regional heterogeneity was less evident on MRI, while PRM_Emph and MRI perfusion defect scores were higher in the upper lobes, and WP and MRI large airway disease scores were higher in the lower lobes. MRI parenchymal and perfusion scores correlated moderately with PRM_Emph (r = 0.61 and r = 0.60) and moderately with FEV1/FVC (r = -0.56).

Conclusion: Multi-center semiquantitative MRI assessments of parenchymal and airway disease in COPD matched GOLD grade-specific imaging features on QCT and detected regional disease heterogeneity. MRI parenchymal disease scores were correlated with QCT and lung function parameters.

Key points: Question Do MRI-based scores correlate with QCT and PFT parameters for GOLD-grade specific disease characterization of COPD? Findings MRI can visualize the parenchymal and airway disease features of COPD. Clinical relevance Lung MRI is suitable for GOLD-grade specific disease characterization of COPD and may serve as a radiation-free imaging modality in scientific and clinical settings, given careful consideration of its potential and limitations.

Keywords: Chronic obstructive pulmonary disease; Computed tomography; Magnetic resonance imaging; Pulmonary emphysema.

Fig. 1

Patient recruitment flowchart

Fig. 2

MIR airway pathologies. A, B Transversal and coronal post-contrast VIBE images showing airway wall thickening of the right lower lobe bronchi (white arrows). C, D Transversal and coronal post-contrast VIBE images showing a tree-in-bud sign in the right upper lobe (white arrows). E, F Transversal and coronal T2 HASTE images showing bronchiectasis and associated wall thickening in both lower lobes (white arrows). G, H Coronal post-contrast VIBE images during inspiration (G) and expiration (H) with the expiratory collapse of the left main bronchus (white arrows)

Fig. 3

Box plots for quantitative CT parameters. Box plots of the parenchymal disease parameters TLV, PRM (PRM_fSAD, PRM_Emph, and PRM_Abnormal), MRI parenchymal and perfusion defect scores, and the airway disease parameters WP (WP_1–2, WP_3–5, and WP_6–10) and MRI central, large and SAD scores for the COPD stages risk COPD–GOLD1–4. Boxes indicate the 25–75th percentile, whiskers indicate the 5th and 95th percentile, and individual outliers are indicated by white-filled circles

Fig. 4

GOLD stage-specific imaging characteristics in QCT and MRI. A1–A5 Quantitative CT images with parametric response imaging (PRM) with normal lung (green = PRM_Normal), functional SAD (yellow = PRM_fSAD), and emphysema (red = PRM_Emph). Note: air trapping and emphysema increase with the GOLD stage. B1–B5 Perfusion MRI images showing homogeneous lung signal in GOLD1 and extensive perfusion defects (white cross) in GOLD4. Note: MRI perfusion defects correspond to CT emphysema. C1–C5 Segmented airway tree on CT (blue = low WP and red = high WP). D1–D5 Axial images of lobe bronchi on CT with segmented airway walls with the highest WP in GOLD3. E1–E5 Transverse T1 VIBE images of lobe bronchi on MRI. Note: assessment of the airway with MRI is much more challenging than with CT

Fig. 5

Spearman rank order correlation coefficients. Spearman rank order correlation coefficients were calculated between semiquantitative MRI scores for parenchymal disease (parenchymal and perfusion defect scores), QCT PRM (PRM_fSAD, PRM_Emph), and the ratio of FEV₁/FVC. In addition, semiquantitative MRI scores for central, large, and SAD were correlated with QCT WP pooled for central (WP_1–2), large (WP_3–5), subsegmental (WP_6–10) airways, and FEV₁/FVC. Spearman’s r correlation coefficients are given in each panel. Solid lines indicate the linear regression