Detecting Skin Reactions in Epicutaneous Patch Testing with Deep Learning: An Evaluation of Pre-Processing and Modality Performance

Abstract

Epicutaneous patch testing is a well-established diagnostic method for identifying substances that may cause Allergic Contact Dermatitis (ACD), a common skin condition caused by exposure to environmental allergens. While the patch test remains the gold standard for identifying allergens, it is prone to observer bias and consumes valuable human resources. Deep learning models can be employed to address this challenge. In this study, we collected a dataset of 1579 multi-modal skin images from 200 patients using the Antera 3D® camera. We then investigated the feasibility of using a deep learning classifier for automating the identification of the allergens causing ACD. We propose a deep learning approach that utilizes a context-retaining pre-processing technique to improve the accuracy of the classifier. In addition, we find promise in the combination of the color image and false-color map of hemoglobin concentration to improve diagnostic accuracy. Our results showed that this approach can potentially achieve more than 86% recall and 94% specificity in identifying skin reactions, and contribute to faster and more accurate diagnosis while reducing clinician workload.

Machine Learning in Allergic Contact Dermatitis: Identifying (Dis)similarities between Polysensitized and Monosensitized Patients

“…In dermatology, image-based screening technologies are developed for the diagnosis and management of skin diseases [14,15,[35][36][37]. Recent studies have been conducted to assess the effectiveness of machine learning algorithms and convolutional neural networks models in precisely detecting and diagnosing ACD from patch test images.…”

“…Recent studies have been conducted to assess the effectiveness of machine learning algorithms and convolutional neural networks models in precisely detecting and diagnosing ACD from patch test images. In both studies, a cohort of 200 ACD patients was examined using a new medical device the Antera ® 3D camera (Miravex Limited, Ireland), while the acquired spectral 3D images were used to map chromophores' concentration (hemoglobin, and melanin) and skin parameters (texture, volume, folds, and fine lines) [15,35]. In the first study, the results indicated that the synergy of convolutional neural networks (CNNs) and machine learning algorithms can achieve a success rate of 85% in ACD detection, indicating a high level of correct diagnostic predictions [15].…”

“…In the first study, the results indicated that the synergy of convolutional neural networks (CNNs) and machine learning algorithms can achieve a success rate of 85% in ACD detection, indicating a high level of correct diagnostic predictions [15]. Convolutional neural networks exhibited high accuracy in ACD diagnosis based on the hemoglobin concentration, while the textural information (texture, volume, folds, and fine lines) was insufficient for classifying a positive allergic reaction [35]. Furthermore, ML algorithms offer the ability to recognize unique patterns in the datasets from extensive repositories of patient information [13,14].…”

Allergic Contact Dermatitis Detection with Machine Learning

“…In the present study, a dataset of 200 patients with ACD collected from Andreas Syggros Hospital in Athens, Greece, was used [14]. The dataset collection involved evaluating skin reactions to patch tests during three consecutive visits and capturing images using an Antera® 3D camera [14].…”

“…In the present study, a dataset of 200 patients with ACD collected from Andreas Syggros Hospital in Athens, Greece, was used [14]. The dataset collection involved evaluating skin reactions to patch tests during three consecutive visits and capturing images using an Antera® 3D camera [14]. The patch testing procedure involved the exposure of the irritated region to a total of 30 allergen substances, involving dyes and chemicals such as Potassium Dichromate, Neomycin Sulphate and Paraphenylenediamine.…”