A Narrative Review: Opportunities and Challenges in Artificial Intelligence Skin Image Analyses Using Total Body Photography

被引：10

作者：

Primiero, Clare A. ^{[1
,2
,3
]}

Rezze, Gisele Gargantini ^{[1
,2
]}

Caffery, Liam J. ^{[3
,4
,5
]}

Carrera, Cristina ^{[1
,2
,6
,7
]}

Podlipnik, Sebastian ^{[1
,2
,7
]}

Espinosa, Natalia ^{[1
,2
]}

Puig, Susana ^{[1
,2
,6
,7
]}

Janda, Monika ^{[4
]}

Soyer, H. Peter ^{[3
,8
]}

Malvehy, Josep ^{[1
,2
,6
,7
]}

机构：

[1] Inst Invest Biomed August Pi i Sunyer IDIBAPS, Hosp Clin, Dermatol Dept, Barcelona, Spain

[2] Inst Invest Biomed August Pi i Sunyer IDIBAPS, Fundacio Clin Recerca Biomed, Barcelona, Spain

[3] Univ Queensland, Frazer Inst, Dermatol Res Ctr, Brisbane, Australia

[4] Univ Queensland, Fac Med, Ctr Hlth Serv Res, Brisbane, Australia

[5] Univ Queensland, Ctr Online Hlth, Brisbane, Australia

[6] Univ Barcelona, Med Dept, Barcelona, Spain

[7] Inst Salud Carlos III, CIBER Enfermedades Raras, Barcelona, Spain

[8] Princess Alexandra Hosp, Dermatol Dept, Brisbane, Qld, Australia

来源：

JOURNAL OF INVESTIGATIVE DERMATOLOGY | 2024年 / 144卷 / 06期

基金：

英国医学研究理事会;

关键词：

Artificial intelligence; Dermatology; Melanoma; Total body photography; CONVOLUTIONAL NEURAL-NETWORK; FOLLOW-UP; CANCER; MELANOMA; CLASSIFICATION; DERMOSCOPY; IDENTIFICATION; LIMITATIONS; GUIDELINES; LESIONS;

D O I：

10.1016/j.jid.2023.11.007

中图分类号：

R75 [皮肤病学与性病学];

学科分类号：

100206 ;

摘要：

Artificial intelligence (AI) algorithms for skin lesion classification have reported accuracy at par with and even outperformance of expert dermatologists in experimental settings. However, the majority of algorithms do not represent real-world clinical approach where skin phenotype and clinical background information are considered. We review the current state of AI for skin lesion classification and present opportunities and challenges when applied to total body photography (TBP). AI in TBP analysis presents opportunities for intrapatient assessment of skin phenotype and holistic risk assessment by incorporating patient-level metadata, although challenges exist for protecting patient privacy in algorithm development and improving explainable AI methods.

引用

页码：1200 / 1207

页数：8

共 57 条

[41] Leveraging Artificial Intelligence to Improve the Diversity of Dermatological Skin Color Pathology: Protocol for an Algorithm Development and Validation Study [J].

Rezk, Eman ;

Eltorki, Mohamed ;

El-Dakhakhni, Wael .

JMIR RESEARCH PROTOCOLS, 2022, 11 (03)

[42]

Rotemberg V, 2021, SCI DATA, V8, DOI 10.1038/s41597-021-00815-z

[43] Using deep learning for dermatologist-level detection of suspicious pigmented skin lesions from wide-field images [J].

Soenksen, Luis R. ;

Kassis, Timothy ;

Conover, Susan T. ;

Marti-Fuster, Berta ;

Birkenfeld, Judith S. ;

Tucker-Schwartz, Jason ;

Naseem, Asif ;

Stavert, Robert R. ;

Kim, Caroline C. ;

Senna, Maryanne M. ;

Aviles-Izquierdo, Jose ;

Collins, James J. ;

Barzilay, Regina ;

Gray, Martha L. .

SCIENCE TRANSLATIONAL MEDICINE, 2021, 13 (581)

[44] Independent evaluation of melanoma polygenic risk scores in UK and Australian prospective cohorts [J].

Steinberg, Julia ;

Iles, Mark M. ;

Lee, Jin Yee ;

Wang, Xiaochuan ;

Law, Matthew H. ;

Smit, Amelia K. ;

Nguyen-Dumont, Tu ;

Giles, Graham G. ;

Southey, Melissa C. ;

Milne, Roger L. ;

Mann, Graham J. ;

Bishop, D. Timothy ;

MacInnis, Robert J. ;

Cust, Anne E. .

BRITISH JOURNAL OF DERMATOLOGY, 2022, 186 (05) :823-834

[45]

Stewart BK, 1998, J AM MED INFORM ASSN, P902

[46] Skin Lesion Detection Algorithms in Whole Body Images [J].

Strzelecki, Michal H. ;

Strakowska, Maria ;

Kozlowski, Michal ;

Urbanczyk, Tomasz ;

Wielowieyska-Szybinska, Dorota ;

Kociolek, Marcin .

SENSORS, 2021, 21 (19)

[47] Molecular genetics of human pigmentation diversity [J].

Sturm, Richard A. .

HUMAN MOLECULAR GENETICS, 2009, 18 :R9-R17

[48] Orthorectification of Skin Nevi Images by Means of 3D Model of the Human Body [J].

Szczypinski, Piotr M. ;

Sprawka, Katarzyna .

SENSORS, 2021, 21 (24)

[49]

Tognetti L, 2021, J DERMATOL SCI, V101, P115, DOI 10.1016/j.jdermsci.2020.11.009

[50] The HAM10000 dataset, a large collection of multi-source dermatoscopic images of common pigmented skin lesions [J].

Tschandl, Philipp ;

Rosendahl, Cliff ;

Kittler, Harald .

SCIENTIFIC DATA, 2018, 5

← 1 2 3 4 5 6 →