A predictive model for differential diagnosis between rosacea and sensitive skin: a cross-sectional study

To the Editor: Rosacea is a common chronic inflammatory skin disease of the central facial skin, with typical manifestations such as flushing, erythema, papules, pustules, and phymatous changes.[1] Erythema is the most common manifestation among all types of rosacea. Sensitive skin (SS) is commonly defined as the occurrence of unpleasant sensations (pruritus, burning, pain, and tingling) upon exposure to certain physical, thermal, or chemical stimuli that normally cause no such sensations on healthy skin. It was reported that erythema might also be present in 74% of SS patients.[2] The pathogenesis, medication, and prognosis of rosacea and SS are quite different, so it is necessary to find reliable non-invasive methods to differentiate between the two conditions. The VISIA® complexion analysis system (Canfield Scientific Inc., Parsippany, New Jersey, USA) is commonly used clinically for analyzing facial skin conditions, in which skin images obtained directly using polarized light are further analyzed using the Red/Brown/X (RBX) algorithm.[3] Here, we aimed to explore whether clinical information together with images analyzed by RBX technology could be applied for the differential diagnosis of rosacea from SS in clinical practice. A total of 275 patients (201 rosacea and 74 SS, respectively) with Fitzpatrick skin type III to IV were enrolled from the Dermatology Department of the First Affiliated Hospital of Nanjing Medical University (Nanjing, China) from April 2019 to March 2020. The study was ethically approved by the Institutional Research Committee of the First Affiliated Hospital of Nanjing Medical University (No. 2020-SR-296). All patients provided written informed consent. The clinical diagnosis of rosacea and SS was made by two independent senior dermatologists, according to the criteria in the 2017 update by the National Rosacea Society Expert Committee[1], self-perception sensitive skin questionnaire and a positive lactic acid sting test with the exclusion of other inflammatory skin diseases, respectively.[2] Medical history and clinical characteristics were recorded for all patients. Facial images were taken using the VISIA® 6.0 complexion analysis system, and the red area images were further analyzed. All the data acquired were statistically analyzed as methods described in Supplementary File 1, https://links.lww.com/CM9/A267, using STATA 14.0. (Stata Corp., College Station, TX, USA), including programs of DCA provided by Vickers.[4] There were some differences in medical history and clinical characteristics of patients with rosacea and SS (as shown in Supplementary Table 1, https://links.lww.com/CM9/A268). At the time of the visit, the course of disease among rosacea (64.2%) and SS (60.8%) patients was over and under 20 months, respectively (P < 0.05). More patients with SS (83.8%) reported a history of inappropriate skin care, while the proportion in rosacea patients was only 41.3%. All enrolled patients with rosacea had facial erythema, while 95.9% of SS patients suffered from erythema. The distribution of red area in the VISIA® Red images were classified into four specific patterns and analyzed: peace sign [Figure 1 A1–B3], wing shape [Figure 1 C1–D3], earlobe [Figure 1 E1–F3], and dots and globular [Figure 1 G 1–H3] patterns.Figure 1: Different pattern of red area distribution in rosacea patients. The peace sign pattern: facial images captured using the VISIA® system (A1–A3) and processed by RBX (B1–B3) technology. The illustration of the peace sign is indicated in the right bottom of B2; The wing shape pattern: facial images captured by VISIA® system (C1–C3) and processed by RBX (D1–D3) technology. The illustration of the wing shape is indicated in the right bottom of D2; The wing shape and earlobe pattern: facial images captured by VISIA® system (E1–E3) and processed by RBX (F1–F3) technology (Earlobe erythema is indicated by black arrows); The dots/globular patter: facial images captured using the VISIA® system (G1–G3) and processed by RBX (H1–H3) technology (Dots/globules are indicated by white arrows).The peace sign pattern appeared in 54.7% of rosacea patients, which was significantly higher than that of SS patients (14.9%). No wing shape pattern was observed in SS patients, but it was observed in 19.9% of rosacea patients. The earlobe pattern appeared in nearly half of all rosacea patients (50.7%) and in only 12.2% of SS patients. Univariate regression analysis found that the course of the disease at the time of the visit, inappropriate skin care, and peace sign and earlobe patterns were all related to the diagnosis of rosacea (Supplementary Table 2, https://links.lww.com/CM9/A268). In the multivariate regression analysis, these four clinical factors are independently related to the specific diagnosis of rosacea (Supplementary Table 3, https://links.lww.com/CM9/A268). Among these four clinical factors, the highest positive likelihood of rosacea diagnosis was peace sign pattern (4.547), and then earlobe pattern (4.198), while the lowest was the course of disease at visit (1.579). The negative likelihood ratio of rosacea for improper skin care was 3.143 (Supplementary Table 4, https://links.lww.com/CM9/A268). When modeling to indicate rosacea with two VISIA® red area patterns [peace sign and earlobe patterns] (model 2), we calculated AUC value of 0.788 (95% confidence interval [CI], 0.740–0.835). When modeling with these two VISIA® patterns and two clinical factors (course of the disease at visit and improper skincare) (model 1), we calculated an AUC of 0.861 (95% CI, 0.818–0.904). The decision curve analyses of models 1 and 2 were shown (Supplementary Figure 1, https://links.lww.com/CM9/A268) to predict a correct diagnosis of rosacea. The threshold probabilities of 40% to 100% for the two models were valid; while the threshold probabilities of model 1 (20%–80%) were superior to that of model 2. In our study, we found two clinical variables (the course of the disease at visit and inappropriate skincare) and two VISIA® erythematous patterns (peace sign and earlobe patterns) were found to be related to the specific diagnosis of rosacea. We reported the use of red area distribution patterns detected using the RBX technology of the VISIA® system to distinguish rosacea from SS. More patients enrolled with SS (83.8%) reported a history of inappropriate skin care habits or products, while the proportion in rosacea patients was only 41.3%. In our study, all enrolled patients with rosacea had facial erythema, while the proportion in SS patients was 95.9%. There are some methodological reports on differentiation, including histopathological examination, reflectance confocal microscopy, dermoscopy, and so on, all of which are used to visualize microstructures, while the location of sampling may affect the results. Color imaging and analyzing systems are more widely used and can provide information for facial distribution-enabled mapping (detailing concentration and location of chromophores with 50–100-μm resolution). And quantitative RBX technology is used to separate red skin-color components and visualize vascular features. In this study, we found four specific patterns of red area distribution in rosacea: peace sign, wing shape, earlobe, and dots/globular patterns; in which the peace sign and earlobe patterns were highly predictive for the diagnosis of rosacea. In the peace sign pattern, the central vertical line in the peace sign corresponds to redness on the forehead, nose, and chin, while the diagonal lines on both sides correspond to erythema on the cheeks that form an acute angle with the angle below the central vertical line. In anatomy, these areas receive blood supply from the arteria supratrochlearis, arteria angularis, and arteria infraorbitali.[5] It can be hypothesized that the abnormal dilation of these arteries may be the cause of this pattern. The red area distribution of rosacea was mostly peace sign-shaped (54.7%), which also appeared in 14.9% SS patients. In the Baumann skin typing system, SS skin can be divided into four types (S1–S4), in which S2 is also called rosacea-type and manifests as a proneness to flushing and erythema and can potentially become rosacea. This may explain why peace sign patterns also appear in SS patients in our study. For rosacea, facial redness affects the central face but is not confined to this region, with the lateral cheeks and ears also being affected. In some rosacea patients, only the lateral cheeks were involved and formed the shapes of wings, which we named the wing shape pattern in VISIA®. It is speculated that this pattern may be caused by abnormal vessel dilation of the anastomotic branches of the zygomatic-facial artery and facial artery.[5] The wing shape pattern appeared in 19.9% of rosacea patents, while no SS patient had this pattern. An earlobe pattern was another common pattern in rosacea patients (50.7%), with only 12.2% of SS patients presenting with this pattern, probably because external stimuli rarely affected the earlobe. Erythema on the earlobe may cause inadequate vein drainage due to the anatomic suspension of the earlobe. The dots/globular pattern corresponded to the papule, pustule, and nodule, which was more common in rosacea than SS. While most of rosacea patients were also complicated with erythema, and might be classified into other patterns. No SS patients involved reported a dots/globular pattern. In our study, we detailed the different red area patterns detected using the RBX technology in the VISIA® system and found two clinical variables (the course of the disease at visit and inappropriate skincare) and two erythematous patterns (peace sign and earlobe patterns) that were related to the diagnosis of rosacea. It is the first report about the pattern analysis of facial red areas of rosacea and SS. There are still some limitations to be further improved. The sample size, especially for SS, was not sufficient, and the clinical subtypes of rosacea were not classified to stratify the results. In conclusion, peace sign and earlobe patterns of facial erythema distribution detected using RBX technology together with clinical history are good predictors for distinguishing rosacea from SS. Funding This work was supported by grants from the National Natural Science Foundation of China (No. 81301384), Construction Program of Jiangsu Provincial Clinical Research Center Support System (No. BL2014084). Conflicts of interest None.

Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.