O15 Using a smartphone app to characterise and quantify skin colour changes in Raynaud's attacks

Abstract

Abstract Background/Aims The lack of objective outcome measures for Raynaud's phenomenon (RP) has been a major limiting factor in development of effective treatments. At present, the Raynaud's Condition Score (RCS) is the only validated outcome measure, and is highly subjective. Mobile phone technology could provide a way forward. We have developed a smartphone app for RP monitoring that guides the patient through the process of capturing images of their hands during RP episodes, as well as capturing other data through post-attack and daily questionnaires. One of the objectives of our research programme (reported here) was to compare digital image (photographic) parameters to the RCS. Methods 40 patients with RP (8 with primary RP, 32 with RP secondary to systemic sclerosis) were recruited (40 female, median age (range): 57 years (25-74), median (range) duration of RP symptoms: 17 (0-53) years). Patients were given a smartphone handset with a pre-installed Raynaud’s Monitoring app and were trained on how to use it/take usable photographs. They were then asked to take photographs of RP attacks over a 14 day period and also to record the RCS for each episode. The app specifically prompts the patient to take a picture of their hand every minute during an attack, until confirmation is given that the attack is complete. At a 2nd visit, the handsets, images, and data were collected for analysis. The mean colour change during each RP attack was quantified (semi-automated method) by the Bhattacharyya distance (BD) in colour space between a region of interest (e.g. a section of a digit) and a control region (dorsal hand). BD was then compared to the RCS using ANOVA, after controlling for patient variability in the range of RCS values used by each patient. Results A total of 3,030 images were collected, describing 229 RP attacks. The median RCS reported was 6 (inter-quartile range [IQR]: 4), while the median for BD was 5.6 (IQR 3.2). ANOVA showed that measured values of the mean image BD were significantly different when different values of RCS were recorded by the patient (p < 0.001), i.e. attacks where patients selected different values of RCS had significantly different values of BD. Across all attacks/patients the F-value from ANOVA for RCS was 76.2, suggesting that the variation in BD for different values of RCS is much greater than the variation in BD for any one value of RCS. Conclusion Patients successfully used a smartphone app to collect photographs and data during episodes of RP. A strong association was found between skin colour change (via BD) and the gold-standard RCS. Mobile phone-documented colour change therefore has potential as an objective measure of RP. Further validation work is now required, as well as studies examining sensitivity to change. Disclosure G. Dinsdale: None. J. Manning: None. A. Herrick: None. M. Dickinson: None. C. Taylor: None.