Skin, scalpel and the silicon chip: a systematic review on the accuracy, bias and data governance of artificial intelligence in dermatology, minimally invasive aesthetics, aesthetic, plastic and reconstructive surgery

被引:3
作者
Rahman, Eqram [1 ]
Sadeghi-Esfahlani, Shabnam [2 ]
Rao, Parinitha [3 ]
Garcia, Patricia
Ioannidis, Sotirios [4 ]
Nosta, John [5 ]
Rahman, Zakia [6 ]
Webb, William Richard [1 ]
机构
[1] Innovat Aesthet, Res & Innovat Hub, London, England
[2] Anglia Ruskin Univ, Med Technol Res Ctr MTRC, Chelmsford, England
[3] Skin Address, Aesthet Dermatol Practice, Bengaluru, India
[4] Private Plast Surg Clin, Thessaloniki, Greece
[5] Nosta Lab, Mendham, NJ USA
[6] Stanford Univ, Stanford, CA USA
关键词
Artificial intelligence; Dermatology; Aesthetic surgery; Minimally invasive aesthetics; Plastic surgery; Machine learning; Bias mitigation; Data ethics; CLEFT-LIP; AUTOMATIC DETECTION; NEURAL-NETWORKS; BURN PATIENTS; PREDICTION; DIAGNOSIS; CHILDREN;
D O I
10.1007/s00238-025-02278-6
中图分类号
R61 [外科手术学];
学科分类号
摘要
IntroductionArtificial Intelligence (AI) is becoming increasingly integrated into healthcare, particularly in fields like dermatology, minimally invasive aesthetics, aesthetic surgery, and plastic and reconstructive surgery. AI has the potential to improve diagnostic accuracy, personalised treatment, and patient outcomes. However, issues such as algorithmic bias, ethical concerns, and generalisability of models remain significant barriers to its full adoption in clinical practice.MethodsA systematic review was conducted following PRISMA guidelines. A broad search of databases including PubMed, EMBASE, and Scopus was performed to identify studies on AI applications in dermatology, aesthetic treatments, and surgery. Inclusion criteria focused on studies evaluating AI's impact on clinical outcomes, bias mitigation strategies, and data management. Data extraction and quality assessment were carried out by two independent reviewers.ResultsOut of 103 included studies, AI showed varying accuracy across different fields. In dermatology, AI models, particularly neural networks, achieved an average accuracy of 90%, while in minimally invasive aesthetics and aesthetic surgery, accuracy ranged between 85% and 95%. Bayesian analysis demonstrated a posterior probability of 0.78 that AI outperforms traditional methods. However, challenges in bias, particularly regarding dataset diversity and ethical concerns, were frequently noted, limiting generalisability and applicability across diverse populations.ConclusionsAI offers significant promise in enhancing clinical outcomes, particularly in dermatology and aesthetic surgery. Nonetheless, biases and ethical issues must be systematically addressed. Further research and standardisation are needed to ensure AI's responsible integration into healthcare.Level of evidence: Not gradableConclusionsAI offers significant promise in enhancing clinical outcomes, particularly in dermatology and aesthetic surgery. Nonetheless, biases and ethical issues must be systematically addressed. Further research and standardisation are needed to ensure AI's responsible integration into healthcare.Level of evidence: Not gradable
引用
收藏
页数:27
相关论文
共 155 条
[1]   Multi-Class Diagnosis of Skin Lesions Using the Fourier Spectral Information of Images on Additive Color Model by Artificial Neural Network [J].
Aaron Lopez-Leyva, Josue ;
Guerra-Rosas, Esperanza ;
Alvarez-Borrego, Josue .
IEEE ACCESS, 2021, 9 :35207-35216
[2]   Ethics of Artificial Intelligence in Medicine and Ophthalmology [J].
Abdullah, Yasser Ibraheem ;
Schuman, Joel S. ;
Shabsigh, Ridwan ;
Caplan, Arthur ;
Al-Aswad, Lama A. .
ASIA-PACIFIC JOURNAL OF OPHTHALMOLOGY, 2021, 10 (03) :289-298
[3]   Large Language Models and Artificial Intelligence: A Primer for Plastic Surgeons on the Demonstrated and Potential Applications, Promises, and Limitations of ChatGPT [J].
Abi-Rafeh, Jad ;
Xu, Hong Hao ;
Kazan, Roy ;
Tevlin, Ruth ;
Furnas, Heather .
AESTHETIC SURGERY JOURNAL, 2024, 44 (03) :329-343
[4]  
Aditya S, 2022, IEEE INT C DAT SCI I, DOI [10.1109/ICDSIS55133.2022.9915804, DOI 10.1109/ICDSIS55133.2022.9915804]
[6]   Application of artificial intelligence in the analysis of the facial skin health condition [J].
Alagic, A. . ;
Alihodzic, S. ;
Alispahic, Nejra ;
Becic, E. ;
Smajovic, A. ;
Becic, F. ;
Becirovic, L. Spahic ;
Pokvic, L. Gurbeta ;
Badnjevic, A. .
IFAC PAPERSONLINE, 2022, 55 (04) :31-37
[7]   Artificial Intelligence and Patient Autonomy in Obesity Treatment Decisions: An Empirical Study of the Challenges [J].
Alanzi, Turki ;
Alhajri, Ahlam ;
Almulhim, Sara ;
Alharbi, Sara ;
Alfaifi, Samya ;
Almarhoun, Eslam ;
Mulla, Raghad ;
Alasafra, Zainab O. ;
Alalwan, Zainab ;
Alnasser, Fatima ;
Almukhtar, Fatima ;
Al Ghadeer, Fatemah ;
Amro, Sara ;
Alodhayb, Ibrahim ;
Alanzi, Nouf .
CUREUS JOURNAL OF MEDICAL SCIENCE, 2023, 15 (11)
[8]   Contrasting Explanations for Understanding and Regularizing Model Adaptations [J].
Artelt, Andre ;
Hinder, Fabian ;
Vaquet, Valerie ;
Feldhans, Robert ;
Hammer, Barbara .
NEURAL PROCESSING LETTERS, 2023, 55 (05) :5273-5297
[9]   Development and validation of an algorithm to predict the treatment modality of burn wounds using thermographic scans: Prospective cohort study [J].
Aurelio Martinez-Jimenez, Mario ;
Luis Ramirez-GarciaLuna, Jose ;
Samuel Kolosovas-Machuca, Eleazar ;
Drager, Justin ;
Javier Gonzalez, Francisco .
PLOS ONE, 2018, 13 (11)
[10]   Analysis of surgical outcome after upper eyelid surgery by computer vision algorithm using face and facial landmark detection [J].
Bahceci Simsek, Ilke ;
Sirolu, Can .
GRAEFES ARCHIVE FOR CLINICAL AND EXPERIMENTAL OPHTHALMOLOGY, 2021, 259 (10) :3119-3125