Facial adult female acne in China: An analysis based on artificial intelligence over one million (2024)

Background

To further clarify the acne profile of Chinese adult women, we included 1,156,703 adult women. An artificial intelligence algorithm was used to analyze images taken by high‐resolution mobile phones to further explore acne levels in Chinese adult women.

Method

In this study, we assessed the severity of acne by evaluating patients' selfies through a smartphone application. Furthermore, we gathered basic user information through a questionnaire, including details such as age, gender, skin sensitivity, and dietary habits.

Results

This study showed a gradual decrease in acne severity from the age of 25 years. A trough was reached between the ages of 40 and 44, followed by a gradual increase in acne severity. In terms of skin problems and acne severity, we have found that oily skin, hypersensitive skin, frequent makeup application and unhealthy dietary habits can affect the severity of acne. For environment and acne severity, we observed that developed city levels, cold seasons and high altitude and strong radiation affect acne severity in adult women. For the results of the AI analyses, the severity of blackheads, pores, dark circles and skin roughness were positively associated with acne severity in adult women.

Conclusions

AI analysis of high‐res phone images in Chinese adult women reveals acne severity trends. Severity decreases after 25, hits a low at 40–44, then gradually rises. Skin type, sensitivity, makeup, diet, urbanization, seasons, altitude, and radiation impact acne. Blackheads, pores, dark circles, and skin roughness are linked to acne severity. These findings inform personalized skincare and public health strategies for adult women.

Keywords: adult female acne, artificial intelligence, Chinese, deep learning, skin aging

2.1. Participants and selfie photography

The study was conducted using a smartphone application (You Look Good Today, YLGTD, Hangzhou C2H4 Internet Technology Corporation) for online selfie collection in China. All subjects were required to read the online informed consent and be fully aware that the photos would be used for research purposes and that their private information would be kept strictly confidential before taking the selfie photos. Considering the difference between different smartphone cameras, only smartphones with high‐resolution rear cameras are eligible for photography as previously described.19 Before taking the selfie, the subjects were asked to be in a room with good lighting conditions. Under the voice guidance of the application, subjects used the rear cameras of their smartphones from the frontal angle at a suitable distance with flashlight on, ensuring that the faces were clearly photographed. A total of 1,156,703 identifiable frontal selfie photos of Chinese females were eventually included for subsequent CNN training and recognition. At the meantime, we have designed a questionnaire to acquire the basic information of user. The items of the questionnaire include age, gender, habit of cosmetics, dietary status, city gradient (habitual residence), season and skin sensitivity (Table1). The annual irradiance of total solar radiation was officially divided into four levels in China: level I (<1050kWh/m²), level II (1050–1400kWh/m²), level III (1400–1750kWh/m²), and level IV (≥ 1750kWh/m²) [CHINA WIND AND SOLAR ENERGY RESOURCES BULLETIN, 2022]. The total annual irradiation of the whole country is almost all greater than 1050kWh/m², and there is basically no area of level I.

2.2. Grading of overall facial acne severity based on CNN network

As shown in Figure1, each acne is classified into four levels according to its severity: (1) Grade I, acne only; (2) Grade II, inflammatory papules; (3) Grade III, pustules; (4) Grade IV, nodules, cysts.20 The facial selfies of participants were first transformed for feature extraction by Resnet‐50. A Feature Pyramid Network (FPN) module was then used for feature maps outputs and a Region Proposal Network (RPN) was used for region proposals based on the outputs of FRN. Finally, a region of interest (ROI) head module was used to complete the whole process of target detection by re‐screening the region proposals and accurately classifying and locating the target objects in the candidate regions (Figure2).

The overall severity score (OSS) of facial acne is calculated according to the following formula:

$$\mathrm{OSS}=100\times {\left(\sum \left({{\mathrm{S}}_{\mathrm{i}}}^2\right)\right)}^{\frac{1}{2}}$$

Si represents the score of each acne identified by CNN on a selfie, with 1 being Grade I acne, 2 being Grade II, 3 being Grade III, and 4 being Grade IV.

Datasets were split into training, validation, and test sets at a ratio of 8:1:1. For validation, four physicians individually classified the overall severity of facial acne in a randomly selected group of 100 participants. The physicians then come up with a consensus rating after discussion. OSS is converted to categorical variables for comparison: an OSS between 0 and 300 is classified as mild, between 300 and 700 as moderate, and greater than 700 as severe.

2.3. Privacy protection

All data is collected in compliance with China's data privacy protection regulations and policies, stored and analyzed on a dedicated and protected server, and the images are deleted immediately after the analysis results are obtained.

2.4. Statistics

Consistency of assessment between four physicians and between physicians and AI was assessed by inter‐rater (or intra‐class) correlation coefficient (ICC). One‐way ANOVA with Tukey‐Kramer post‐hoc tests were used for comparison among multiple groups. A value of p<0.05 was considered statistically significant. All statistical tests were conducted in SPSS (Version 25.0. Armonk, NY: IBM Corp.). Data was presented as mean±95% confidence interval (CI) in figures.

3.1. Comparison between artificial intelligence and physicians

A total of 1,156,703 female adult participants (age>25) uploaded their selfies between January 2021 and January 2023 (Table2). The assessments among four physicians, as well as the assessments between physicians and AI both showed good agreements, with an inter‐rater ICC of 0.887 (95% CI: 0.860‐0.935; p<0.001) and an intra‐rater ICC of 0.900 (95% CI: 0.834‐0.938; p<0.001).

3.2. Trends of acne severity with age in adult females

We first analyzed the trend of acne severity with age in adult females. The results showed that the severity gradually declined after age 25 and reached its lowest level between 40 and 44 years of age, before rising gradually again with age (Figure3).

3.3. Intrinsic factors and acne severity

Among the different skin types, we find that dry skin had the lowest acne severity (all p<0.0001), followed by eudermic skin and combination skin with similar acne severity (p>0.05), while oily skin had the highest acne severity (all p<0.0001) (Figure4A). In addition, the acne severity levels increased with skin sensitivity (all p<0.0001) (Figure4B). Acne severity was more severe in females with over 20 cosmetic products on a daily basis than those with 1–5 or 6–20 cosmetic products (all p<0.0001) (Figure4C). Females with very healthy diets had lower acne severity than females with normal diets (p<0.0001), followed by those with unhealthy diets or very unhealthy diets (p<0.0001) (Figure4D).

3.4. Environments and acne severity

External environments may contribute to the aggravation of acne. Across cities with different levels of development in China, Tier 2 cities had the highest level of acne severity (all p<0.0001), followed by Tier 1 cities (all p<0.0001), then Tier 3 and Tier 5 cities with similar levels (p>0.05), and the lowest level of severity was in Tier 4 cities (all p<0.0001) (Figure5A). The ranking of acne severity from worst to least severe throughout four seasons was fall, winter, summer, and spring (all p<0.0001) (Figure5B). China has a very large land area with widely varying sunlight conditions. The results showed that acne severity was highest among females with a long‐term residence in areas of irradiation level IV, followed by those in level III and the lowest in irradiation II (all p<0.0001) (Figures5C and6).

3.5. Acne and other facial skin problems

Acne is sometimes accompanied by other skin problems. With the help of other built‐in AI analysis of the software, we were able to compare the correlation between acne severity and other facial problems. The results showed a positive correlation between the severity of acne and the severity of enlarged pores, blackheads, and skin roughness (all p<0.0001) (Figure7A‐C). In addition, acne severity was found lowest in patients without dark circles, followed by those with mild dark circles, and similarly worst in patients with moderate and severe dark circles (Figure7D).

• 1.Tan JKL, Bhate K. A global perspective on the epidemiology of acne. Br J Dermatol. 2015;172(Suppl 1):3–12. [DOI] [PubMed] [Google Scholar]
• 2.Ali MA, El Taieb MA, Hegazy EM, etal. Evaluation of the effect of oral isotretinoin on the level of serum YKL40 in acne vulgaris patients: a cross‐sectional case‐control. Clin Cosmet Investig Dermatol. 2023;16:3241–3248. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 3.Alshammrie FF, Alshammari R, Alharbi RM, Khan FH, Alshammari SK. Epidemiology of acne vulgaris and its association with lifestyle among adolescents and young adults in hail, kingdom of Saudi Arabia: a community‐based study. Cureus. 2020;12(7):e9277. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 4.Layton AM, Thiboutot D, Tan J. Reviewing the global burden of acne: how could we improve care to reduce the burden?Br J Dermatol. 2021;184(2):219–225. [DOI] [PubMed] [Google Scholar]
• 5.Sachdeva M, Tan J, Lim J, Kim M, Nadeem I, Bismil R. The prevalence, risk factors, and psychosocial impacts of acne vulgaris in medical students: a literature review. Int J Dermatol. 2021;60(7):792–798. [DOI] [PubMed] [Google Scholar]
• 6.Lynn DD, Umari T, Dunnick CA, Dellavalle RP. The epidemiology of acne vulgaris in late adolescence. Adolesc Health Med Ther. 2016;7:13–25. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 7.Adityan B, Kumari R, Thappa DM. Scoring systems in acne vulgaris. Indian J Dermatol Venereol Leprol. 2009;75(3):323–326. [DOI] [PubMed] [Google Scholar]
• 8.Zaenglein AL, Pathy AL, Schlosser BJ, etal. Guidelines of care for the management of acne vulgaris. J Am Acad Dermatol. 2016;74(5):945‐973.e33.e33. [DOI] [PubMed] [Google Scholar]
• 9.Fried RG, Wechsler A. Psychological problems in the acne patient. Dermatol Ther. 2006;19(4):237–240. [DOI] [PubMed] [Google Scholar]
• 10.Schwalbe N, Wahl B. Artificial intelligence and the future of global health. Lancet. 2020;395(10236):1579–1586. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 11.Wang G, Meng X, Zhang F. Past, present, and future of global research on artificial intelligence applications in dermatology: a bibliometric analysis. Medicine (Baltimore). 2023;102(45):e35993. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 12.Esteva A, Kuprel B, Novoa RA, Ko J, Swetter SM, Blau HM, Thrun S. Dermatologist‐level classification of skin cancer with deep neural networks. Nature. 2017;542(7639):115–118. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 13.Tschandl P, Codella N, Akay BN, etal. Comparison of the accuracy of human readers versus machine‐learning algorithms for pigmented skin lesion classification: an open, web‐based, international, diagnostic study. Lancet Oncol. 2019;20(7):938–947. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 14.Tschandl P, Rinner C, Apalla Z, etal. Human‐computer collaboration for skin cancer recognition. Nat Med. 2020;26(8):1229–1234. [DOI] [PubMed] [Google Scholar]
• 15.Du‐Harpur X, Watt FM, Luscombe NM, Lynch MD. What is AI? Applications of artificial intelligence to dermatology. Br J Dermatol. 2020;183(3):423–430. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 16.Bobrov E, Georgievskaya A, Kiselev K, etal. PhotoAgeClock: deep learning algorithms for development of non‐invasive visual biomarkers of aging. Aging (Albany NY). 2018;10(11):3249–3259. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 17.Nouveau‐Richard S, Yang Z, Mac‐Mary S, etal. Skin ageing: a comparison between Chinese and European populations. A pilot study. J Dermatol Sci. 2005;40(3):187–193. [DOI] [PubMed] [Google Scholar]
• 18.Vierkötter A, Hüls A, Yamamoto A, etal. Extrinsic skin ageing in German, Chinese and Japanese women manifests differently in all three groups depending on ethnic background, age and anatomical site. J Dermatol Sci. 2016;83(3):219–225. [DOI] [PubMed] [Google Scholar]
• 19.Li T‐H, Ma X‐D, Li Z‐M, etal. Artificial intelligence analysis of over a million Chinese men and women reveals level of dark circle in the facial skin aging process. Skin Res Technol . 2023;29(11):e13492. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 20.Purdy S, de Berker D. Acne vulgaris. BMJ Clin Evid. 2011;2011. [PMC free article] [PubMed] [Google Scholar]
• 21.Atallah DA, Badran AY, Makhlouf AG, Mekkawy MM. Topical silymarin cream as a novel therapy versus salicylic acid peels in acne vulgaris: a split‐face clinical trial. J Cutan Med Surg. 2023;28(1):22‐28. 12034754231211568. [DOI] [PubMed] [Google Scholar]
• 22.Dal Belo SE, Kanoun Copy L, Lambert C, etal. Efficacy of a multitargeted, salicylic acid‐based dermocosmetic cream compared to benzoyl peroxide 5% in Acne vulgaris: results from a randomized study. J Cosmet Dermatol. 2023;23(3):891‐897. [DOI] [PubMed] [Google Scholar]
• 23.Guler HA, Yildirim SK, Guler D. The relationship between adolescents' perception of family emotional atmosphere and acne severity, self‐esteem, and quality of life in adolescents diagnosed with acne vulgaris. Turk Arch Pediatr. 2023;58(6):646–652. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 24.Li Y, Zeng Y, Chen Z, Nie S, Wu Z. The efficiency and safety of low‐dosage isotretinoin therapy for Chinese acne vulgaris patients. J Cosmet Dermatol. 2023;23(3):926‐930. [DOI] [PubMed] [Google Scholar]
• 25.Mahran A, Ghazally A, Ali AS, Bakr RM. Efficacy and safety of calcipotriol as a potential topical treatment of acne vulgaris: a randomized, controlled, triple blinded, split‐face clinical trial. Clin Exp Dermatol. 2023;49(4):348‐355. [DOI] [PubMed] [Google Scholar]
• 26.Flament F, Lee YW, Lee DH, etal. The continuous development of a complete and objective automatic grading system of facial signs from selfie pictures: Asian validation study and application to women of three ethnic origins, differently aged. Skin Res Technol. 2021;27(2):183–190. [DOI] [PubMed] [Google Scholar]
• 27.Nguyen HTu, Isaacowitz DM, Rubin PAD. Age‐ and fatigue‐related markers of human faces: an eye‐tracking study. Ophthalmology. 2009;116(2):355–360. [DOI] [PubMed] [Google Scholar]
• 28.Flament F, Jiang R, Houghton J, etal. Objective and automatic grading system of facial signs from smartphones' pictures in South African men: validation versus dermatologists and characterization of changes with age. Skin Res Technol. 2023;29(4):e13257. [DOI] [PMC free article] [PubMed] [Google Scholar]
• 29.Flament F, Jacquet L, Ye C, etal. Artificial Intelligence analysis of over half a million European and Chinese women reveals striking differences in the facial skin ageing process. J Eur Acad Dermatol Venereol. 2022;36(7):1136–1142. [DOI] [PubMed] [Google Scholar]
• 30.Law MH, Medland SE, Zhu Gu, etal. Genome‐wide association shows that pigmentation genes play a role in skin aging. J Invest Dermatol. 2017;137(9):1887–1894. [DOI] [PubMed] [Google Scholar]
• 31.Edelkamp J, Lousada MB, Pinto D, etal. Management of the human hair follicle microbiome by a synthetic odorant. J Dermatol Sci. 2023;112(2):99‐108. [DOI] [PubMed] [Google Scholar]
• 32.Kim K, Jang H, Kim E, Kim H, Sung GY. Recent advances in understanding the role of the skin microbiome in the treatment of atopic dermatitis. Exp Dermatol. 2023;32(12):2048‐2061. [DOI] [PubMed] [Google Scholar]
• 33.Di Landro A, Cazzaniga S, Cusano F, etal. Adult female acne and associated risk factors: results of a multicenter case‐control study in Italy. J Am Acad Dermatol. 2016;75(6):1134–1141.e1. [DOI] [PubMed] [Google Scholar]
• 34.Di Landro A, Cazzaniga S, Parazzini F, etal. Family history, body mass index, selected dietary factors, menstrual history, and risk of moderate to severe acne in adolescents and young adults. J Am Acad Dermatol. 2012;67(6):1129–1135. [DOI] [PubMed] [Google Scholar]

Data Availability Statement

The original contributions presented in the study are included in the article/Supplementary Material. Further inquiries can be directed to the corresponding author.