Using Multiple Dermoscopic Photographs of One Lesion Improves Melanoma Classification via Deep Learning: A Prognostic Diagnostic Accuracy Study
Abstract
Background: Convolutional neural network (CNN)-based melanoma classifiers face several challenges that limit their usefulness in clinical practice. Objective: To investigate the impact of multiple real-world dermoscopic views of a single lesion of interest on a CNN-based melanoma classifier. Methods: This study evaluated 656 suspected melanoma lesions. Classifier performance was measured using area under the receiver operating characteristic curve (AUROC), expected calibration error (ECE) and maximum confidence change (MCC) for (I) a single-view scenario, (II) a multiview scenario using multiple artificially modified images per lesion and (III) a multiview scenario with multiple real-world images per lesion. Results: The multiview approach with real-world images significantly increased the AUROC from 0.905 (95% CI, 0.879-0.929) in the single-view approach to 0.930 (95% CI, 0.909-0.951). ECE and MCC also improved significantly from 0.131 (95% CI, 0.105-0.159) to 0.072 (95% CI: 0.052-0.093) and from 0.149 (95% CI, 0.125-0.171) to 0.115 (95% CI: 0.099-0.131), respectively. Comparing multiview real-world to artificially modified images showed comparable diagnostic accuracy and uncertainty estimation, but significantly worse robustness for the latter. Conclusion: Using multiple real-world images is an inexpensive method to positively impact the performance of a CNN-based melanoma classifier.
Keywords
Cite
@article{arxiv.2306.02800,
title = {Using Multiple Dermoscopic Photographs of One Lesion Improves Melanoma Classification via Deep Learning: A Prognostic Diagnostic Accuracy Study},
author = {Achim Hekler and Roman C. Maron and Sarah Haggenmüller and Max Schmitt and Christoph Wies and Jochen S. Utikal and Friedegund Meier and Sarah Hobelsberger and Frank F. Gellrich and Mildred Sergon and Axel Hauschild and Lars E. French and Lucie Heinzerling and Justin G. Schlager and Kamran Ghoreschi and Max Schlaak and Franz J. Hilke and Gabriela Poch and Sören Korsing and Carola Berking and Markus V. Heppt and Michael Erdmann and Sebastian Haferkamp and Konstantin Drexler and Dirk Schadendorf and Wiebke Sondermann and Matthias Goebeler and Bastian Schilling and Jakob N. Kather and Eva Krieghoff-Henning and Titus J. Brinker},
journal= {arXiv preprint arXiv:2306.02800},
year = {2026}
}