The influence of measurement location on reliability of quantitative nailfold videocapillaroscopy in patients with SSc

Abstract

Nailfold videocapillaroscopy is being increasingly used as a marker of SSc-related microvascular disease, including in response to treatment. However, it requires further validation. Our aim was to assess the inter-observer, intra-observer and test-retest variability of semi-automated measurement of capillary features as well as of a manual density measurement. All capillary apexes in images from 58 patients with SSc were marked up independently by two trained observers (inter-observer variability). The first observer then re-marked the images (intra-observer variability), and finally, the first observer marked up a second image of the same nailfold (test-retest). Mark-up of capillaries was carried out on cropped mosaic images (cropped independently by the observers to a fixed width, to allow the same length of nail bed to be studied for each patient) and on whole mosaic images (examining the whole nail bed). Reproducibility of independently cropped mosaic images was poor and was due to the variation in the positioning of the cropped area. However, quantification of whole mosaic images was highly reproducible, e.g. for inter-capillary distance, the intra-class correlation coefficient for inter-observer, intra-observer and test-retest reliability was 0.95, 0.98 and 0.90 (compared with 0.88, 0.79 and 0.89 for cropped mosaic images), respectively. Intra-observer limits of agreement for whole mosaic images were better than inter-observer reproducibility. Quantitative assessment of SSc-related change in nailfold capillaries is unreliable if examination of the same set of capillaries cannot be guaranteed. Conversely a wide-field, high-magnification system that allows visualization of the whole nail bed offers a highly reproducible approach for quantitative assessment and therefore has potential as an outcome measure.