A convolutional neural network architecture for the recognition of cutaneous manifestations of COVID‐19

Abstract

During the COVID-19 pandemic, dermatologists reported an array of different cutaneous manifestations of the disease. It is challenging to discriminate COVID-19-related cutaneous manifestations from other closely resembling skin lesions. The aim of this study was to generate and evaluate a novel CNN (Convolutional Neural Network) ensemble architecture for detection of COVID-19-associated skin lesions from clinical images. An ensemble model of three different CNN-based algorithms was trained with clinical images of skin lesions from confirmed COVID-19 positive patients, healthy controls as well as 18 other common skin conditions, which included close mimics of COVID-19 skin lesions such as urticaria, varicella, pityriasis rosea, herpes zoster, bullous pemphigoid and psoriasis. The multi-class model demonstrated an overall top-1 accuracy of 86.7% for all 20 diseases. The sensitivity and specificity of COVID-19-rash detection were found to be 84.2 ± 5.1% and 99.5 ± 0.2%, respectively. The positive predictive value, NPV and area under curve values for COVID-19-rash were 88.0 ± 5.6%, 99.4 ± 0.2% and 0.97 ± 0.25, respectively. The binary classifier had a mean sensitivity, specificity and accuracy of 76.81 ± 6.25%, 99.77 ± 0.14% and 98.91 ± 0.17%, respectively for COVID-19 rash. The model was robust in detection of all skin lesions on both white and skin of color, although only a few images of COVID-19-associated skin lesions from skin of color were available. To our best knowledge, this is the first machine learning-based study for automated detection of COVID-19 based on skin images and may provide a useful decision support tool for physicians to optimize contact-free COVID-19 triage, differential diagnosis of skin lesions and patient care.

The application of artificial intelligence and data integration in COVID-19 studies: a scoping review

“…All studies analyzed either computed tomography or chest X-ray data, except for 5 studies that analyzed images of lung ultrasound 48–51 or skin lesions. 52 The most common sources of medical images were local hospitals or healthcare systems and image datasets published on public domains, such as GitHub or Kaggle. In these imaging studies, roughly half used the convolutional neural network (CNN)-based models.…”

COVID‐19 and artificial intelligence: Experts and dermatologists perspective

“…Previous works have proposed documenting cutaneous abnormalities using smart phone imagery and AI algorithms integrated into smartphone apps. 20 , 21 Additionally ML techniques for diagnosing urticarial lesions by junior physicians have been studied. 22 Frequent hand washing, use of alcohol‐based sanitizers and gloves have been advocated to combat transmission of COVID‐19.…”

“…Cutaneous features being of short duration, self‐limiting, and non‐emergency unlike systemic symptoms of COVID‐19 have not been prioritized as screening policies till date. Previous works have proposed documenting cutaneous abnormalities using smart phone imagery and AI algorithms integrated into smartphone apps 20,21 . Additionally ML techniques for diagnosing urticarial lesions by junior physicians have been studied 22 .…”

Use of artificial intelligence as a diagnostic support tool for skin lesions in primary care: feasibility study in clinical practice

“…In dermatology, ML using image recognition is especially developed in skin cancer screening [22][23][24][25]. More recently, its use has been extended to a wider range of skin lesions, such as in ammatory and infectious lesions [26][27][28][29], and also in the recognition of cutaneous manifestations of COVID-19 [30]. This suggests that its use in PC as a diagnostic support and screening tool for consultations related to skin problems would standardise and improve the effectiveness and e ciency of the professionals working there.…”