Exploring convolutional neural networks with transfer learning for diagnosing Lyme disease from skin lesion images
Authors:
Sk Imran Hossain,
Jocelyn de Goër de Herve,
Md Shahriar Hassan,
Delphine Martineau,
Evelina Petrosyan,
Violaine Corbain,
Jean Beytout,
Isabelle Lebert,
Elisabeth Baux,
Céline Cazorla,
Carole Eldin,
Yves Hansmann,
Solene Patrat-Delon,
Thierry Prazuck,
Alice Raffetin,
Pierre Tattevin,
Gwenaël Vourc'H,
Olivier Lesens,
Engelbert Nguifo
Abstract:
Lyme disease which is one of the most common infectious vector-borne diseases manifests itself in most cases with erythema migrans (EM) skin lesions. Recent studies show that convolutional neural networks (CNNs) perform well to identify skin lesions from images. Lightweight CNN based pre-scanner applications for resource-constrained mobile devices can help users with early diagnosis of Lyme diseas…
▽ More
Lyme disease which is one of the most common infectious vector-borne diseases manifests itself in most cases with erythema migrans (EM) skin lesions. Recent studies show that convolutional neural networks (CNNs) perform well to identify skin lesions from images. Lightweight CNN based pre-scanner applications for resource-constrained mobile devices can help users with early diagnosis of Lyme disease and prevent the transition to a severe late form thanks to appropriate antibiotic therapy. Also, resource-intensive CNN based robust computer applications can assist non-expert practitioners with an accurate diagnosis. The main objective of this study is to extensively analyze the effectiveness of CNNs for diagnosing Lyme disease from images and to find out the best CNN architectures considering resource constraints. First, we created an EM dataset with the help of expert dermatologists from Clermont-Ferrand University Hospital Center of France. Second, we benchmarked this dataset for twenty-three CNN architectures customized from VGG, ResNet, DenseNet, MobileNet, Xception, NASNet, and EfficientNet architectures in terms of predictive performance, computational complexity, and statistical significance. Third, to improve the performance of the CNNs, we used custom transfer learning from ImageNet pre-trained models as well as pre-trained the CNNs with the skin lesion dataset HAM10000. Fourth, for model explainability, we utilized Gradient-weighted Class Activation Map** to visualize the regions of input that are significant to the CNNs for making predictions. Fifth, we provided guidelines for model selection based on predictive performance and computational complexity.
△ Less
Submitted 15 February, 2022; v1 submitted 28 June, 2021;
originally announced June 2021.
Glutathione conjugates of the mercapturic acid pathway and guanine adduct as biomarkers of exposure to CEES, a sulfur mustard analog
Authors:
Marie Roser,
David Béal,
Camille Eldin,
Leslie Gudimard,
Fanny Caffin,
Fanny Gros-Désormeaux,
Daniel Léonço,
François Fenaille,
Christophe Junot,
Christophe Piérard,
Thierry Douki
Abstract:
Sulfur mustard (SM), a chemical warfare agent, is a strong alkylating compound that readily reacts with numerous biomolecules. The goal of the present work was to define and validate new biomarkers of exposure to SM that could be easily accessible in urine or plasma. Because investigations using SM are prohibited by the Organization for the Prohibition of Chemical Weapons, we worked with 2-chloroe…
▽ More
Sulfur mustard (SM), a chemical warfare agent, is a strong alkylating compound that readily reacts with numerous biomolecules. The goal of the present work was to define and validate new biomarkers of exposure to SM that could be easily accessible in urine or plasma. Because investigations using SM are prohibited by the Organization for the Prohibition of Chemical Weapons, we worked with 2-chloroethyl ethyl sulfide (CEES), a monofunctional analog of SM. We developed an ultra-high-pressure liquid chromatography - tandem mass spectrometry approach (UHPLC-MS/MS) to the conjugate of CEES to glutathione and two of its metabolites, the cysteine and the N-acetyl-cysteine conjugates. The N7-guanine adduct of CEES (N7Gua-CEES) was also targeted. After synthesizing the specific biomarkers, a solid phase extraction protocol and a UHPLC-MS/MS method with isotopic dilution were optimized. We were able to quantify N7Gua-CEES in the DNA of HaCaT keratinocytes and of explants of human skin exposed to CEES. N7Gua-CEES was also detected in the culture medium of these two models, together with the glutathione and the cysteine conjugates. In contrast, the N-acetyl-cysteine conjugate was not detected. The method was then applied to plasma from mice cutaneously exposed to CEES. All four markers could be detected. Our present results thus validate both the analytical technique and the biological relevance of new, easily quantifiable biomarkers of exposure to CEES. Because CEES behaves very similarly to SM, the results are promising for application to this toxic of interest.
△ Less
Submitted 11 January, 2021;
originally announced January 2021.
