Skip to main page content
U.S. flag

An official website of the United States government

Dot gov

The .gov means it’s official.
Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

Https

The site is secure.
The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2024 Aug 14;11(1):884.
doi: 10.1038/s41597-024-03743-w.

The SLICE-3D dataset: 400,000 skin lesion image crops extracted from 3D TBP for skin cancer detection

Nicholas R Kurtansky  1 , Brian M D'Alessandro  2 , Maura C Gillis  3 , Brigid Betz-Stablein  4 , Sara E Cerminara  5 , Rafael Garcia  6 , Marcela Alves Girundi  7 , Elisabeth Victoria Goessinger  5 , Philippe Gottfrois  5 , Pascale Guitera  7   8 , Allan C Halpern  3 , Valerie Jakrot  7 , Harald Kittler  9 , Kivanc Kose  3 , Konstantinos Liopyris  10 , Josep Malvehy  11   12 , Victoria J Mar  13   14 , Linda K Martin  7   8   15 , Thomas Mathew  7 , Lara Valeska Maul  16 , Adam Mothershaw  4 , Alina M Mueller  5 , Christoph Mueller  9 , Alexander A Navarini  5 , Tarlia Rajeswaran  17 , Vin Rajeswaran  17 , Anup Saha  6 , Maithili Sashindranath  13 , Laura Serra-García  18 , H Peter Soyer  4 , Georgios Theocharis  10 , Ayesha Vos  14 , Jochen Weber  3 , Veronica Rotemberg  3
Affiliations

The SLICE-3D dataset: 400,000 skin lesion image crops extracted from 3D TBP for skin cancer detection

Nicholas R Kurtansky et al. Sci Data. .

Abstract

AI image classification algorithms have shown promising results when applied to skin cancer detection. Most public skin cancer image datasets are comprised of dermoscopic photos and are limited by selection bias, lack of standardization, and lend themselves to development of algorithms that can only be used by skilled clinicians. The SLICE-3D ("Skin Lesion Image Crops Extracted from 3D TBP") dataset described here addresses those concerns and contains images of over 400,000 distinct skin lesions from seven dermatologic centers from around the world. De-identified images were systematically extracted from sensitive 3D Total Body Photographs and are comparable in optical resolution to smartphone images. Algorithms trained on lower quality images could improve clinical workflows and detect skin cancers earlier if deployed in primary care or non-clinical settings, where photos are captured by non-expert physicians or patients. Such a tool could prompt individuals to visit a specialized dermatologist. This dataset circumvents many inherent limitations of prior datasets and may be used to build upon previous applications of skin imaging for cancer detection.

PubMed Disclaimer

Conflict of interest statement

B. D’Alessandro is an employee of Canfield Scientific, Inc. B. Betz-Stablein is anticipating employment with Canfield Scientific, Inc. H. Halpern receives consultation fees from Canfield Scientific, Inc. H. Kittler has received speaker honoraria from Fotofinder, is an advisor of Fotofinder and AI Medical Technology, and has received license fees from Heine, Casio, MetaOptima, and Barco. H. Kittler has received equipment from Heine, Casio, and DermaMedical. J. Malvehy is co-founder of Athena Tech, scientific advisor of Dermavision, and is chairman of the Task Force of Artificial Intelligence of the EADV. V. Mar has received speaker fees from Novartis, Bristol Myers Squibb, Merck and Janssen, and has participated in Advisory Boards for MSD, L’Oreal, and SkylineDx. L. Martin is funded by the Warwick L Morison Professorship in dermatology. A. Navarini and L.V. Maul received a grant from Canfield Scientific, Inc. for physician salary in a separate study that had no influence on this manuscript. H.P. Soyer is a shareholder of MoleMap N.Z. Limited and e-derm consult GmbH, and undertakes regular teledermatological reporting for both companies. H.P. Soyer is a Medical Consultant for Canfield Scientific, Inc. and a Medical Advisor for First Derm. H.P. Soyer is involved in several committees of the Australiasian College of Dermatology. V. Rotemberg is a consultant for Inhabit Brands, Inc., and receives in kind support from Kaggle and AWS.

Figures

Fig. 1
Fig. 1
Examples of image types. The SLICE-3D dataset is comprised of tile images, which were extracted from 3D TBP images. Tile images display fewer morphologic features than dermoscopic images, which are captured in clinical settings to document and monitor specific lesions.
Fig. 2
Fig. 2
Schematic workflow of dataset curation.
Fig. 3
Fig. 3
Example tiles of varying diagnostic classes.
See this image and copyright information in PMC

Similar articles

See all similar articles

Cited by

See all "Cited by" articles

References

    1. James, O. T. et al. Artificial intelligence and machine learning algorithms for early detection of skin cancer in community and primary care settings: a systematic review. Lancet Digit. Health4, e466–e476 (2022). 10.1016/S2589-7500(22)00023-1 - DOI - PubMed
    1. Beltrami, E. J. et al. Artificial intelligence in the detection of skin cancer. J. Am. Acad. Dermatol.87, 1336–1342 (2022). 10.1016/j.jaad.2022.08.028 - DOI - PubMed
    1. Vaidya, T. et al. Socioeconomic and geographic barriers to dermatology care in urban and rural US populations. J. Am. Acad. Dermatol.78, 406–408 (2017).10.1016/j.jaad.2017.07.050 - DOI - PubMed
    1. Glazer, A. M. & Rigel, D. S. Analysis of trends in geographic distribution of US dermatology workforce density. JAMA Dermatol.153, 472 (2017). 10.1001/jamadermatol.2016.6032 - DOI - PMC - PubMed
    1. Leavitt, E. R. et al. Teledermatology as a tool to improve access to care for medically underserved populations: A retrospective descriptive study. J. Am. Acad. Dermatol.75, 1259–1261 (2016). 10.1016/j.jaad.2016.07.043 - DOI - PMC - PubMed

LinkOut - more resources