Validation of the guideline to identify adherent and rolling leukocytes in skin microvasculature by reflectance confocal videomicroscopy

Summary: Abstract Body: We aimed to validate our previously developed guideline to identify adherent and rolling leukocytes in videos of skin microvasculature by reflectance confocal videomicroscopy. Eight raters with no prior experience in reflectance confocal videomicroscopy reviewed the published guideline for identifying adherent and rolling leukocytes in videos of human skin microvasculature. Subsequently, each rater independently counted adherent and rolling leukocytes in 113 videos in a predetermined random order. The raters were unaware that 13 videos were repeated for intra-observer variability assessment. The average error across eight raters for counting adherent and rolling leukocytes was 0.32 (95% confidence interval: 0.22 – 0.41, n = 113 videos, N = 8 raters) and 0.93 (0.52 – 1.34, n = 113, N = 8), respectively. Inter-observer reproducibility for counting adherent and rolling leukocytes was 0.41 (0.39 – 0.46, n = 100, N = 8) and 1.04 (0.95 – 1.13, n = 100, N = 8), respectively. Intra-observer reproducibility for counting adherent and rolling leukocytes was 0.10 (0.01 – 0.18, n = 13, N = 8) and 0.55 (0.27 – 0.82, n = 13, N = 8), respectively. The average error increased with an increased number of adherent or rolling leukocytes in a video. By analyzing the 113 training videos, the raters’ ability to accurately identify adherent or rolling leukocytes did not change as they gained experience by reviewing more videos. Reviewing the published guideline was sufficient to train all eight raters to identify adherent and rolling leukocytes within an error of less than one leukocyte. The guideline can effectively train raters with no prior experience in analyzing reflectance confocal videomicroscopy data of skin microvasculature. Dhruv Patel¹, Ings Saknite^{1, 2}, Sahil Patel¹, Michael Pogharian¹, Vicky Wang¹, Juliana Fuller¹, X-Zavyer Smith¹, Sage Iwamoto¹, Jessica Wu¹, Eric Tkaczyk^{1, 3} 1. Dermatology, Vanderbilt University Medical Center, Nashville, TN, United States. 2. Biophotonics Laboratory, Latvijas Universitate, Riga, Riga, Latvia. 3. US Department of Veterans Affairs, Nashville, TN, United States. Bioinformatics, Computational Biology, and Imaging