1
|
Turner C, Pardo LM, Gunn DA, Zillmer R, Mekić S, Liu F, Ikram MA, Klaver CCW, Croll PH, Goedegebure A, Trajanoska K, Rivadeneira F, Kavousi M, Brusselle GGO, Kayser M, Nijsten T, Bacardit J. Deep learning predicted perceived age is a reliable approach for analysis of facial ageing: A proof of principle study. J Eur Acad Dermatol Venereol 2024; 38:2295-2302. [PMID: 39360788 PMCID: PMC11587682 DOI: 10.1111/jdv.20365] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [MESH Headings] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 05/28/2024] [Accepted: 09/04/2024] [Indexed: 10/05/2024]
Abstract
BACKGROUND Perceived age (PA) has been associated with mortality, genetic variants linked to ageing and several age-related morbidities. However, estimating PA in large datasets is laborious and costly to generate, limiting its practical applicability. OBJECTIVES To determine if estimating PA using deep learning-based algorithms results in the same associations with morbidities and genetic variants as human-estimated perceived age. METHODS Self-supervised learning (SSL) and deep feature transfer (DFT) deep learning (DL) approaches were trained and tested on human-estimated PAs and their corresponding frontal face images of middle-aged to elderly Dutch participants (n = 2679) from a population-based study in the Netherlands. We compared the DL-estimated PAs with morbidities previously associated with human-estimated PA as well as genetic variants in the gene MC1R; we additionally tested the PA associations with MC1R in a new validation cohort (n = 1158). RESULTS The DL approaches predicted PA in this population with a mean absolute error of 2.84 years (DFT) and 2.39 years (SSL). In the training-test dataset, we found the same significant (p < 0.05) associations for DL PA with osteoporosis, ARHL, cognition, COPD and cataracts and MC1R, as with human PA. We also found a similar but less significant association for SSL and DFT PAs (0.69 and 0.71 years per allele, p = 0.008 and 0.011, respectively) with MC1R variants in the validation dataset as that found with human, SSL and DFT PAs in the training-test dataset (0.79, 0.78 and 0.71 years per allele respectively; all p < 0.0001). CONCLUSIONS Deep learning methods can automatically estimate PA from facial images with enough accuracy to replicate known links between human-estimated perceived age and several age-related morbidities. Furthermore, DL predicted perceived age associated with MC1R gene variants in a validation cohort. Hence, such DL PA techniques may be used instead of human estimations in perceived age studies thereby reducing time and costs.
Collapse
Affiliation(s)
- Conor Turner
- Interdisciplinary Computing and Complex BioSystems (ICOS) Research Group, School of ComputingNewcastle UniversityNewcastle upon TyneUK
| | - Luba M. Pardo
- Department of DermatologyErasmus MC University Medical Center RotterdamRotterdamThe Netherlands
| | - David A. Gunn
- Unilever King's Biosciences Innovation Hub, King's College LondonLondonUK
| | | | - Selma Mekić
- Department of DermatologyUniversity Medical Center UtrechtUtrechtThe Netherlands
| | - Fan Liu
- Department of Genetic Identification, Erasmus MCUniversity Medical Center RotterdamRotterdamThe Netherlands
| | - M. Arfan Ikram
- Department of EpidemiologyErasmus MC University Medical Center RotterdamRotterdamThe Netherlands
| | - Caroline C. W. Klaver
- Department of EpidemiologyErasmus MC University Medical Center RotterdamRotterdamThe Netherlands
- Department of OphthalmologyErasmus MC University Medical Center RotterdamRotterdamThe Netherlands
| | - Pauline H. Croll
- Department of EpidemiologyErasmus MC University Medical Center RotterdamRotterdamThe Netherlands
- Department of Radiology and Nuclear Medicine, Erasmus MCUniversity Medical Center RotterdamRotterdamThe Netherlands
- Department of Otorhinolaryngology, Head and Neck SurgeryErasmus MC University Medical Center RotterdamRotterdamThe Netherlands
| | - André Goedegebure
- Department of Otorhinolaryngology, Head and Neck SurgeryErasmus MC University Medical Center RotterdamRotterdamThe Netherlands
| | - Katerina Trajanoska
- Department of EpidemiologyErasmus MC University Medical Center RotterdamRotterdamThe Netherlands
- Department of Internal MedicineErasmus MC University Medical Center RotterdamRotterdamThe Netherlands
| | - Fernando Rivadeneira
- Department of Internal MedicineErasmus MC University Medical Center RotterdamRotterdamThe Netherlands
| | - Maryam Kavousi
- Department of EpidemiologyErasmus MC University Medical Center RotterdamRotterdamThe Netherlands
| | - Guy G. O. Brusselle
- Department of EpidemiologyErasmus MC University Medical Center RotterdamRotterdamThe Netherlands
- Department of Respiratory MedicineErasmus MC University Medical Center RotterdamRotterdamThe Netherlands
- Department of Respiratory MedicineGhent University HospitalGhentBelgium
| | - Manfred Kayser
- Department of Genetic Identification, Erasmus MCUniversity Medical Center RotterdamRotterdamThe Netherlands
| | - Tamar Nijsten
- Department of DermatologyErasmus MC University Medical Center RotterdamRotterdamThe Netherlands
| | - Jaume Bacardit
- Interdisciplinary Computing and Complex BioSystems (ICOS) Research Group, School of ComputingNewcastle UniversityNewcastle upon TyneUK
| |
Collapse
|
2
|
Xu X, Jigeer G, Gunn DA, Liu Y, Chen X, Guo Y, Li Y, Gu X, Ma Y, Wang J, Wang S, Sun L, Lin X, Gao X. Facial aging, cognitive impairment, and dementia risk. Alzheimers Res Ther 2024; 16:245. [PMID: 39506848 PMCID: PMC11539626 DOI: 10.1186/s13195-024-01611-8] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Key Words] [MESH Headings] [Grants] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 07/11/2024] [Accepted: 10/28/2024] [Indexed: 11/08/2024]
Abstract
BACKGROUND Facial aging, cognitive impairment, and dementia are all age-related conditions. However, the temporal relation between facial age and future risk of dementia was not systematically examined. OBJECTIVES To investigate the relationship between facial age (both subjective/perceived and objective) and cognitive impairment and/or dementia risk. METHODS The study included 195,329 participants (age ≥ 60 y) from the UK Biobank (UKB) with self-perceived facial age and 612 participants from the Nutrition and Health of Aging Population in China Project (NHAPC) study (age ≥ 56 y) with objective assessment of facial age. Cox proportional hazards model was used to prospectively examine the hazard ratios (HRs) and their 95% confidence intervals (CIs) of self-perceived facial age and dementia risk in the UKB, adjusting for age, sex, education, APOE ε4 allele, and other potential confounders. Linear and logistic regressions were performed to examine the cross-sectional association between facial age (perceived and objective) and cognitive impairment in the UKB and NHAPC, with potential confounders adjusted. RESULTS During a median follow-up of 12.3 years, 5659 dementia cases were identified in the UKB. The fully-adjusted HRs comparing high vs. low perceived facial age were 1.61 (95% CI, 1.33 ~ 1.96) for dementia (P-trend ≤ 0.001). Subjective facial age and cognitive impairment was also observed in the UKB. In the NHAPC, facial age, as assessed by three objective wrinkle parameters, was associated with higher odds of cognitive impairment (P-trend < 0.05). Specifically, the fully-adjusted OR for cognitive impairment comparing the highest versus the lowest quartiles of crow's feet wrinkles number was 2.48 (95% CI, 1.06 ~ 5.78). CONCLUSIONS High facial age was associated with cognitive impairment, dementia and its subtypes after adjusting for conventional risk factors for dementia. Facial aging may be an indicator of cognitive decline and dementia risk in older adults, which can aid in the early diagnosis and management of age-related conditions.
Collapse
Affiliation(s)
- Xinming Xu
- Department of Nutrition and Food Hygiene, Ministry of Education Key Laboratory of Public Health Safety, School of Public Health, Institute of Nutrition, Fudan University, 130 Dongan Road, Shanghai, 200030, China
| | - Guliyeerke Jigeer
- Department of Nutrition and Food Hygiene, Ministry of Education Key Laboratory of Public Health Safety, School of Public Health, Institute of Nutrition, Fudan University, 130 Dongan Road, Shanghai, 200030, China
| | - David Andrew Gunn
- Unilever R&D Colworth Science Part, Sharnbrook, Bedfordshire, MK44 1LQ, UK
| | - Yizhou Liu
- State Key Laboratory of Genetic Engineering, School of Life Sciences & Human Phenome Institute, Fudan University, Shanghai, 200438, China
| | - Xinrui Chen
- State Key Laboratory of Genetic Engineering, School of Life Sciences & Human Phenome Institute, Fudan University, Shanghai, 200438, China
| | - Yi Guo
- Department of Biostatistics, School of Public Health, Key Laboratory of Public Health Safety and Collaborative Innovation Center of Social Risks Governance in Health, Fudan University, Shanghai, 200030, China
| | - Yaqi Li
- Department of Nutrition and Food Hygiene, Ministry of Education Key Laboratory of Public Health Safety, School of Public Health, Institute of Nutrition, Fudan University, 130 Dongan Road, Shanghai, 200030, China
| | - Xuelan Gu
- Unilever R&D Shanghai, Shanghai, 200335, China
| | - Yanyun Ma
- Unilever R&D Colworth Science Part, Sharnbrook, Bedfordshire, MK44 1LQ, UK
| | - Jiucun Wang
- State Key Laboratory of Genetic Engineering, School of Life Sciences & Human Phenome Institute, Fudan University, Shanghai, 200438, China
| | - Sijia Wang
- CAS Key Laboratory of Computational Biology, Shanghai Institute of Nutrition and Health, Shanghai Institute for Biological Sciences, CAS-MPG Partner Institute for Computational Biology, University of Chinese Academy of Sciences, Chinese Academy of Sciences, 320 Yue Yang Road, Shanghai, 200031, China
| | - Liang Sun
- Department of Nutrition and Food Hygiene, Ministry of Education Key Laboratory of Public Health Safety, School of Public Health, Institute of Nutrition, Fudan University, 130 Dongan Road, Shanghai, 200030, China.
| | - Xu Lin
- Shanghai Institute of Nutrition and Health, University of Chinese Academy of Sciences, Chinese Academy of Sciences, Shanghai, 200031, China.
- Key Laboratory of Systems Biology, Hangzhou Institute for Advanced Study, University of Chinese Academy of Sciences, Chinese Academy of Sciences, Hangzhou, 310024, China.
| | - Xiang Gao
- Department of Nutrition and Food Hygiene, Ministry of Education Key Laboratory of Public Health Safety, School of Public Health, Institute of Nutrition, Fudan University, 130 Dongan Road, Shanghai, 200030, China.
| |
Collapse
|
3
|
Zhang Z, Shi M, Li J, Chen D, Ren J, Li Z, Tuan H, Zhao Y. The Characteristics and Inheriting Pattern of Skin Aging in Chinese Women: An Intergenerational Study of Mothers and Daughters. Clin Cosmet Investig Dermatol 2024; 17:1773-1782. [PMID: 39132029 PMCID: PMC11315646 DOI: 10.2147/ccid.s468477] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Key Words] [Grants] [Track Full Text] [Download PDF] [Figures] [Journal Information] [Subscribe] [Scholar Register] [Received: 05/13/2024] [Accepted: 07/29/2024] [Indexed: 08/13/2024]
Abstract
Introduction The aging of the skin, which is affected by both external and internal causes, can reflect the external age and the internal health status. While the aging characteristics differ across ethnic groups, the specific changes in skin aging within the Chinese population have been underexplored. Moreover, investigating the similarity of aging skin characteristics between parent-offspring pairs remains uncharted territory. This study aims to fill these gaps by examining the skin aging features of Chinese women and assessing the similarity in aging skin characteristics between mother-daughter pairs. Methods A total of 40 mother-daughter pairs were recruited and analyzed. The perceived ages of the participants were evaluated, and their aging skin traits were systematically graded. Statistical methods were employed to discern the trends of the aging skin characteristics. By introducing a novel similarity parameter, we compared whether various skin aging characteristics have similar patterns between mothers and daughters. Results Our findings indicate that age 50 represents a pivotal point in skin aging. Beyond this age, the increase in rhytides and laxity scores accelerated noticeably, whereas the escalation in dyschromia scores became less marked. By introducing similar parameters between mother-daughter pairs and the radar map, we discovered that the skin aging characteristics are remarkably consistent between mother-daughter pairs. Conclusion Understanding the main aging skin characteristics of different age groups can allow caregivers to devise treatments for preventing skin aging in women of various ages. The mother's skin aging trend is also significant for the daughter's skin aging prevention.
Collapse
Affiliation(s)
- Zhuying Zhang
- Department of Dermatology, Beijing Tsinghua Changgung Hospital, Beijing, People’s Republic of China
- Photomedicine Laboratory, Institute of Precision Medicine, Tsinghua University, Beijing, People’s Republic of China
| | - Mai Shi
- Department of Dermatology, Beijing Tsinghua Changgung Hospital, Beijing, People’s Republic of China
- Photomedicine Laboratory, Institute of Precision Medicine, Tsinghua University, Beijing, People’s Republic of China
| | - Jinghui Li
- School of Medicine, Tsinghua University, Beijing, People’s Republic of China
| | - Dian Chen
- Department of Dermatology, Beijing Tsinghua Changgung Hospital, Beijing, People’s Republic of China
- Photomedicine Laboratory, Institute of Precision Medicine, Tsinghua University, Beijing, People’s Republic of China
| | - Jie Ren
- Department of Dermatology, Beijing Tsinghua Changgung Hospital, Beijing, People’s Republic of China
- Photomedicine Laboratory, Institute of Precision Medicine, Tsinghua University, Beijing, People’s Republic of China
| | - Zhenghui Li
- Department of Dermatology, Beijing Tsinghua Changgung Hospital, Beijing, People’s Republic of China
- Photomedicine Laboratory, Institute of Precision Medicine, Tsinghua University, Beijing, People’s Republic of China
| | - Hsiaohan Tuan
- Department of Dermatology, Beijing Tsinghua Changgung Hospital, Beijing, People’s Republic of China
- Photomedicine Laboratory, Institute of Precision Medicine, Tsinghua University, Beijing, People’s Republic of China
| | - Yi Zhao
- Department of Dermatology, Beijing Tsinghua Changgung Hospital, Beijing, People’s Republic of China
- Photomedicine Laboratory, Institute of Precision Medicine, Tsinghua University, Beijing, People’s Republic of China
| |
Collapse
|
4
|
Ikram MA, Kieboom BCT, Brouwer WP, Brusselle G, Chaker L, Ghanbari M, Goedegebure A, Ikram MK, Kavousi M, de Knegt RJ, Luik AI, van Meurs J, Pardo LM, Rivadeneira F, van Rooij FJA, Vernooij MW, Voortman T, Terzikhan N. The Rotterdam Study. Design update and major findings between 2020 and 2024. Eur J Epidemiol 2024; 39:183-206. [PMID: 38324224 DOI: 10.1007/s10654-023-01094-1] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Key Words] [MESH Headings] [Grants] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 07/21/2023] [Accepted: 12/14/2023] [Indexed: 02/08/2024]
Abstract
The Rotterdam Study is a population-based cohort study, started in 1990 in the district of Ommoord in the city of Rotterdam, the Netherlands, with the aim to describe the prevalence and incidence, unravel the etiology, and identify targets for prediction, prevention or intervention of multifactorial diseases in mid-life and elderly. The study currently includes 17,931 participants (overall response rate 65%), aged 40 years and over, who are examined in-person every 3 to 5 years in a dedicated research facility, and who are followed-up continuously through automated linkage with health care providers, both regionally and nationally. Research within the Rotterdam Study is carried out along two axes. First, research lines are oriented around diseases and clinical conditions, which are reflective of medical specializations. Second, cross-cutting research lines transverse these clinical demarcations allowing for inter- and multidisciplinary research. These research lines generally reflect subdomains within epidemiology. This paper describes recent methodological updates and main findings from each of these research lines. Also, future perspective for coming years highlighted.
Collapse
Affiliation(s)
- M Arfan Ikram
- Department of Epidemiology, Erasmus MC University Medical Center, Rotterdam, Netherlands.
| | - Brenda C T Kieboom
- Department of Epidemiology, Erasmus MC University Medical Center, Rotterdam, Netherlands
| | - Willem Pieter Brouwer
- Department of Hepatology, Erasmus MC University Medical Center, Rotterdam, Netherlands
| | - Guy Brusselle
- Department of Epidemiology, Erasmus MC University Medical Center, Rotterdam, Netherlands
- Department of Pulmonology, University Hospital Ghent, Ghent, Belgium
| | - Layal Chaker
- Department of Epidemiology, and Department of Internal Medicine, Erasmus MC University Medical Center, Rotterdam, Netherlands
| | - Mohsen Ghanbari
- Department of Epidemiology, Erasmus MC University Medical Center, Rotterdam, Netherlands
| | - André Goedegebure
- Department of Otorhinolaryngology and Head & Neck Surgery, Erasmus MC University Medical Center, Rotterdam, Netherlands
| | - M Kamran Ikram
- Department of Epidemiology, and Department of Neurology, Erasmus MC University Medical Center, Rotterdam, Netherlands
| | - Maryam Kavousi
- Department of Epidemiology, Erasmus MC University Medical Center, Rotterdam, Netherlands
| | - Rob J de Knegt
- Department of Hepatology, Erasmus MC University Medical Center, Rotterdam, Netherlands
| | - Annemarie I Luik
- Department of Epidemiology, Erasmus MC University Medical Center, Rotterdam, Netherlands
| | - Joyce van Meurs
- Department of Internal Medicine, Erasmus MC University Medical Center, Rotterdam, Netherlands
| | - Luba M Pardo
- Department of Dermatology, Erasmus MC University Medical Center, Rotterdam, Netherlands
| | - Fernando Rivadeneira
- Department of Medicine, and Department of Oral & Maxillofacial Surgery, Erasmus MC University Medical Center, Rotterdam, Netherlands
| | - Frank J A van Rooij
- Department of Epidemiology, Erasmus MC University Medical Center, Rotterdam, Netherlands
| | - Meike W Vernooij
- Department of Epidemiology, and Department of Radiology & Nuclear Medicine, Erasmus MC University Medical Center, Rotterdam, Netherlands
| | - Trudy Voortman
- Department of Epidemiology, Erasmus MC University Medical Center, Rotterdam, Netherlands
| | - Natalie Terzikhan
- Department of Epidemiology, Erasmus MC University Medical Center, Rotterdam, Netherlands
| |
Collapse
|
5
|
Vladimir K, Perišić MM, Štorga M, Mostashari A, Khanin R. Epigenetics insights from perceived facial aging. Clin Epigenetics 2023; 15:176. [PMID: 37924108 PMCID: PMC10623707 DOI: 10.1186/s13148-023-01590-x] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [MESH Headings] [Grants] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 06/01/2023] [Accepted: 10/23/2023] [Indexed: 11/06/2023] Open
Abstract
Facial aging is the most visible manifestation of aging. People desire to look younger than others of the same chronological age. Hence, perceived age is often used as a visible marker of aging, while biological age, often estimated by methylation markers, is used as an objective measure of age. Multiple epigenetics-based clocks have been developed for accurate estimation of general biological age and the age of specific organs, including the skin. However, it is not clear whether the epigenetic biomarkers (CpGs) used in these clocks are drivers of aging processes or consequences of aging. In this proof-of-concept study, we integrate data from GWAS on perceived facial aging and EWAS on CpGs measured in blood. By running EW Mendelian randomization, we identify hundreds of putative CpGs that are potentially causal to perceived facial aging with similar numbers of damaging markers that causally drive or accelerate facial aging and protective methylation markers that causally slow down or protect from aging. We further demonstrate that while candidate causal CpGs have little overlap with known epigenetics-based clocks, they affect genes or proteins with known functions in skin aging, such as skin pigmentation, elastin, and collagen levels. Overall, our results suggest that blood methylation markers reflect facial aging processes, and thus can be used to quantify skin aging and develop anti-aging solutions that target the root causes of aging.
Collapse
Affiliation(s)
- Klemo Vladimir
- LifeNome Inc., New York, 10018, NY, USA
- Faculty of Electrical Engineering and Computing, University of Zagreb, 10000, Zagreb, Croatia
| | - Marija Majda Perišić
- LifeNome Inc., New York, 10018, NY, USA
- Faculty of Mechanical Engineering and Naval Architecture, University of Zagreb, 10000, Zagreb, Croatia
| | - Mario Štorga
- LifeNome Inc., New York, 10018, NY, USA
- Faculty of Mechanical Engineering and Naval Architecture, University of Zagreb, 10000, Zagreb, Croatia
| | | | - Raya Khanin
- LifeNome Inc., New York, 10018, NY, USA.
- Bioinformatics Core, Memorial Sloan-Kettering Cancer Center, New York, 10065, NY, USA.
| |
Collapse
|