Background Nailfold videocapillaroscopy (NVC) captures high-magnification images of capillaries, allowing non-invasive assessment of systemic sclerosis [SSc] related microvascular disease. We have developed a state-of-the-art NVC system to enable faster and easier acquisition of high quality static images and produce video sequences in which red blood cell velocity can be measured. The system uses novel software to make fully automated quantitative estimates of capillary morphology and blood flow velocity, eliminating the need for (subjective) clinician input to image interpretation.
Objectives To apply the new system in a case-control study to: 1) assess system usability 2) test how well automated measures of capillary morphology differentiate between healthy controls (HC), subjects with primary Raynaud's phenomenon (PRP) and patients with SSc 3) investigate capillary blood flow velocity as a new quantitative measure for NVC.
Methods 112 participants were recruited into the study (50 HC, 12 PRP, 50 SSc). Panoramic video sequences were taken for all 10 digits, each comprising 10,000 – 20,000 frames (size: 640x480 px, spatial resolution: 1μm/px, frame rate: 120 frames/s). The frames were processed and registered to produce a static panorama of the whole nailfold (∼2–5mm wide). The analysis software automatically detected and measured capillary structural properties in this image1, before using in-focus frames at each capillary to create video sequences for blood flow velocity estimation (Fig 1). Capillary-level measurements were averaged across all 10 nailfolds to produce subject-level parameters: capillary density, mean capillary width, maximum capillary width, shape score, derangement and mean flow velocity. For each parameter one-way ANOVA and Tukey's range test were used to check for group differences, and the area under ROC curve (Az) used to measure separation between HC/PRP and SSc.
Results ANOVA tests showed significant group-wise differences for all parameters. Table 1 shows group means and Az for each parameter.
Conclusions The new system successfully generated high-quality video sequences, and allowed fast panoramic imaging of the whole nailfold. Automatically extracted capillary morphology parameters matched both clinical expectations and previous study results1,2 and show considerable promise as outcome measures for SSc. Preliminary flow results suggest reduced blood velocity is associated with abnormal morphology and thus lower in SSc patients, further work is needed to corroborate this.
Berks et al, MICCAI, 8673, 658–665, 2014
Murray et al, Arthritis Care Res, 61(8), 2009
Acknowledgement Funded by the Wellcome Trust
Disclosure of Interest None declared