POS1409 AUTOMATED DETECTION OF SCLERODERMIFORM PATTERNS USING CAPILLARY.IO

Background: A nailfold capillaroscopy procedure is a non-invasive, low-cost, and well-established examination that can be used to diagnose several rheumatic autoimmune diseases and support the necessary follow-up of patients. While the clinical implications of the technique are known, a rigorous and in-depth examination of nailfold capillaries remains as one of the major challenges to produce new advances in research and diagnosis, due to practical limitations for analysing the whole nailfold area of each patient. The difference between the different patterns established by Maricq and Cutolo makes it possible to predict the evolution that the patient will present. We introduce Capillary.io, an automatic image reading system able to recognize capillaries in images obtained with any microscope, generate automatic measurements of each capillary and take advantage of this information to report capillary morphology and patterns. Objectives: to determine the ability to detect active and early scerodermiform patterns of Capillary.io. Methods: Forty-nine complete capillaroscopies, reported by expert capillaroscopists according to the different patterns manually (gold standard), were compared with the pattern detection capability of Capillary.io. A scoring system based on the algorithm of the Spanish Capillaroscopy Study Group (GREC) was performed and interpreted by capillary.io for the global interpretation of each of the capillaroscopies analyzed. Results: Overall, 37 of the 49 capillaroscopies reported agreed with the diagnosed pattern (75.51%). Separately, the early pattern presented a concordance of 77.27% and the active pattern of 74.07%. In reference to the findings detected by the Capillary.io system, the mean overall density was 5.01 capillaries/mm in the group with the active pattern compared to 6.46 capillaries/mm in the early pattern. The density of dilations and megacapillaries was 2.81/mm and 1.21/mm in the active pattern group versus 4.69/mm and 0.4/mm in the early pattern group. Global diameters were greater in the active pattern group with an apical mean of 37.3 μm compared to 28.5 μm in the early pattern subgroup. Conclusion: Capillary.io is a simple, easy-to-learn web system for interpreting capillaroscopic images of nail folds. It can be a very useful tool to standardize the interpretation of capillaroscopic images, not only individually for each capillary, but also jointly through the detection of different patterns. References: [1]Chen K, Wang J, Pang J, Cao Y, Xiong Y, Li X, et al. MMDetection: Open MMLab Detection Toolbox and Benchmark. arXiv preprint arXiv:190607155 2019;. [2]Cutolo M, Pizzorni C, Sulli A. Nailfold videocapillaroscopy assessment of microvascular damage in systemic sclerosis - Reply. The Journal of Rheumatology 2000 11;27:2722–2723. [3]Cutolo M, Trombetta AC, Melsens K, Pizzorni C, Sulli A, Ruaro B, et al. Automated assessment of absolute nailfold capillary number on videocapillaroscopic images: Proof of principle and validation in systemic sclerosis. Microcirculation 2018 May;25(4):e12447. [4]Smith V, Vanhaecke A, Herrick AL, Distler O, Guerra MG, Denton CP, et al. Fast track algorithm: How to differentiate a “scleroderma pattern” from a “non-scleroderma pattern”. Autoimmu- nity Reviews 2019 nov;18(11):102394. [5]Tavakol ME, Fatemi A, Karbalaie A, Emrani Z, Erlandsson BE. Nailfold Capillaroscopy in Rheumatic Diseases: Which Parameters Should Be Evaluated? BioMed Research International 2015;2015:1–17. Disclosure of Interests: Eduardo Ramos Shareholder of: Co-founder and shareholder of Capillary.io, Alfredo Guillen-Del-Castillo: None declared, Carmen Pilar Simeon-Aznar: None declared, Borja Gracia Tello Shareholder of: Co-founder and shareholder of Capillary.io, Vicent Fonollosa Pla: None declared, Albert Selva-O’Callaghan: None declared, Luis Saez-Comet: None declared, Elena Martinez Robles: None declared, Juan Jose Rios: None declared, Gerard Espinosa: None declared, Jose Antonio Todoli Parra: None declared, Jose Luis Callejas-Rubio: None declared, Norberto Ortego: None declared, Begona Mari-Alfonso: None declared, Mayka Freire: None declared, Patricia Fanlo: None declared