Abstract 5931: Fractal Dimension: A Novel Method of Quantifying Vascular Structural Changes in Pulmonary Hypertension
Introduction - The pulmonary arterial tree is fractal, as it is morphologically self-similar at different scales. Fractal dimension can therefore be used to quantify arterial branching, lower values indicating less branching. In pulmonary arterial hypertension (PAH), occlusion of the arterial lumen leads to a loss of branches and vascular pruning. This is often evident in CT pulmonary angiograms and may reflect disease severity.
Hypothesis - We hypothesised that fractal dimension is inversely correlated with disease severity in PAH.
Methods - Thirty children with PAH underwent CT pulmonary angiography (64 slice, dual source cardiac gated system). The pulmonary arterial tree was segmented (Fig 1a⇓) and skeletonised. The fractal dimension of the skeletonised data was calculated using the box counting method. Fractal dimension was compared to markers of disease severity.
Results - Fractal dimension correlated inversely with resting PVRI (Fig 1b⇓, r=−0.63, p=0.002, n=21) and WHO functional class (Fig 1c⇓, p=0.002, n=30), whilst it correlated positively with percentage of predicted six-minute walk distance (r=0.53, p=0.006, n=26).
Conclusion - We have demonstrated that fractal dimension correlates with disease severity in pulmonary hypertension. This is due to a progressive loss of arterial branches, the hallmark of PAH. We believe fractal dimension provides a window into the pathological processes underlying pulmonary vascular disease. As such, this novel non-invasive method may be useful in monitoring disease progression and response to therapy.