Can self-training identify suspicious ugly duckling lesions?

TL;DR

This paper presents a self-supervised machine learning approach to automatically detect suspicious 'Ugly Duckling' skin lesions by identifying outliers based on lesion embeddings, aiding melanoma screening.

Contribution

It introduces a novel self-supervised method for outlier detection of skin lesions without requiring expert labels, improving standardization in melanoma screening.

Findings

Achieved 72.1% sensitivity in identifying Ugly Duckling lesions.

Reached 94.2% diagnostic accuracy on test data.

Demonstrated effectiveness compared to dermatologists.

Abstract

One commonly used clinical approach towards detecting melanomas recognises the existence of Ugly Duckling nevi, or skin lesions which look different from the other lesions on the same patient. An automatic method of detecting and analysing these lesions would help to standardize studies, compared with manual screening methods. However, it is difficult to obtain expertly-labelled images for ugly duckling lesions. We therefore propose to use self-supervised machine learning to automatically detect outlier lesions. We first automatically detect and extract all the lesions from a wide-field skin image, and calculate an embedding for each detected lesion in a patient image, based on automatically identified features. These embeddings are then used to calculate the L2 distances as a way to measure dissimilarity. Using this deep learning method, Ugly Ducklings are identified as outliers which should deserve more attention from the examining physician. We evaluate through comparison with dermatologists, and achieve a sensitivity rate of 72.1% and diagnostic accuracy of 94.2% on the held-out test set.