Abstract:Background: Previous studies of artificial intelligence (AI) applied to dermatology have shown AI to have higher diagnostic classification accuracy than expert dermatologists; however, these studies did not adequately assess clinically realistic scenarios, such as how AI systems behave when presented with images of disease categories that are not included in the training dataset or images drawn from statistical distributions with significant shifts from training distributions. We aimed to simulate these real-world scenarios and evaluate the effects of image source institution, diagnoses outside of the training set, and other image artifacts on classification accuracy, with the goal of informing clinicians and regulatory agencies about safety and real-world accuracy. Methods: We designed a large dermoscopic image classification challenge to quantify the performance of machine learning algorithms for the task of skin cancer classification from dermoscopic images, and how this performance is affected by shifts in statistical distributions of data, disease categories not represented in training datasets, and imaging or lesion artifacts. Factors that might be beneficial to performance, such as clinical metadata and external training data collected by challenge participants, were also evaluated. 25 331 training images collected from two datasets (in Vienna [HAM10000] and Barcelona [BCN20000]) between Jan 1, 2000, and Dec 31, 2018, across eight skin diseases, were provided to challenge participants to design appropriate algorithms. The trained algorithms were then tested for balanced accuracy against the HAM10000 and BCN20000 test datasets and data from countries not included in the training dataset (Turkey, New Zealand, Sweden, and Argentina). Test datasets contained images of all diagnostic categories available in training plus other diagnoses not included in training data (not trained category). We compared the performance of the algorithms against that of 18 dermatologists in a simulated setting that reflected intended clinical use. Findings: 64 teams submitted 129 state-of-the-art algorithm predictions on a test set of 8238 images. The best performing algorithm achieved 58·8% balanced accuracy on the BCN20000 data, which was designed to better reflect realistic clinical scenarios, compared with 82·0% balanced accuracy on HAM10000, which was used in a previously published benchmark. Shifted statistical distributions and disease categories not included in training data contributed to decreases in accuracy. Image artifacts, including hair, pen markings, ulceration, and imaging source institution, decreased accuracy in a complex manner that varied based on the underlying diagnosis. When comparing algorithms to expert dermatologists (2460 ratings on 1269 images), algorithms performed better than experts in most categories, except for actinic keratoses (similar accuracy on average) and images from categories not included in training data (26% correct for experts vs 6% correct for algorithms, p<0·0001). For the top 25 submitted algorithms, 47·1% of the images from categories not included in training data were misclassified as malignant diagnoses, which would lead to a substantial number of unnecessary biopsies if current state-of-the-art AI technologies were clinically deployed. Interpretation: We have identified specific deficiencies and safety issues in AI diagnostic systems for skin cancer that should be addressed in future diagnostic evaluation protocols to improve safety and reliability in clinical practice. Funding: Melanoma Research Alliance and La Marató de TV3.

International Skin Imaging Collaboration‐Designated Diagnoses (ISIC‐DX): Consensus terminology for lesion diagnostic labeling

A European Consensus on the Consistent Use of the Term “Keratinocyte Cancer”

Melanoma in situ and low risk pT1a melanoma: Need for new diagnostic terminology

Global Harmonization of Morphological Definitions in Hidradenitis Suppurativa for a Proposed Glossary

European consensus-based interdisciplinary guideline for melanoma. Part 1: Diagnostics: Update 2022

Advances in Technology for Melanoma Diagnosis and Prognosis: An Expert Consensus Panel

European consensus-based interdisciplinary guideline for melanoma. Part 1: Diagnostics - Update 2024

A Patient-Centric Dataset of Images and Metadata for Identifying Melanomas Using Clinical Context

Deep Learning-Based Semantic Segmentation of Non-Melanocytic Skin Tumors in Whole-Slide Histopathological Images.

Dermatopathologic features of cutaneous squamous cell carcinoma and actinic keratosis: Consensus criteria and proposed reporting guidelines

Proposed Definitions of Typical Lesions in Hidradenitis Suppurativa

Liver imaging: it is time to adopt standardized terminology

Validation of artificial intelligence prediction models for skin cancer diagnosis using dermoscopy images: the 2019 International Skin Imaging Collaboration Grand Challenge

Teledermatology & Digital Dermatology SymposiumOral presentationsBT01 Developing a national artificial intelligence-assisted skin cancer pathway

Development of a Skin-Directed Scoring System for Stevens-Johnson Syndrome and Epidermal Necrolysis: A Delphi Consensus Exercise

Clinical Utility of a Digital Dermoscopy Image-Based Artificial Intelligence Device in the Diagnosis and Management of Skin Cancer by Dermatologists

From the International Association for the Study of Lung Cancer Early Detection and Screening Committee: Terminology Issues in Screening and Early Detection of Lung Cancer-International Association for the Study of Lung Cancer Early Detection and Screening Committee Expert Group Recommendations

A protocol for annotation of total body photography for machine learning to analyze skin phenotype and lesion classification

P078 Establishing diagnostic criteria for vulval lichen sclerosus (SHELLS): an international electronic-Delphi consensus study

The Impact of Next-generation Sequencing on Interobserver Agreement and Diagnostic Accuracy of Desmoplastic Melanocytic Neoplasms

The 2023 WHO updates on skin tumors: advances since the 2018 edition