Refine
Has Fulltext
- yes (3)
Is part of the Bibliography
- yes (3)
Document Type
- Journal article (3)
Language
- English (3)
Keywords
- COVID-19 (1)
- SARS-CoV-2 (1)
- artificial intelligence (1)
- deep learning (1)
- dermatology (1)
- imiquimod (1)
- immunology (1)
- machine learning (1)
- melanoma (1)
- neural networks (1)
The COVID‐19 pandemic caused by SARS‐CoV‐2 has far‐reaching direct and indirect medical consequences. These include both the course and treatment of diseases. It is becoming increasingly clear that infections with SARS‐CoV‐2 can cause considerable immunological alterations, which particularly also affect pathogenetically and/or therapeutically relevant factors.
Against this background we summarize here the current state of knowledge on the interaction of SARS‐CoV‐2/COVID‐19 with mediators of the acute phase of inflammation (TNF, IL‐1, IL‐6), type 1 and type 17 immune responses (IL‐12, IL‐23, IL‐17, IL‐36), type 2 immune reactions (IL‐4, IL‐13, IL‐5, IL‐31, IgE), B‐cell immunity, checkpoint regulators (PD‐1, PD‐L1, CTLA4), and orally druggable signaling pathways (JAK, PDE4, calcineurin). In addition, we discuss in this context non‐specific immune modulation by glucocorticosteroids, methotrexate, antimalarial drugs, azathioprine, dapsone, mycophenolate mofetil and fumaric acid esters, as well as neutrophil granulocyte‐mediated innate immune mechanisms.
From these recent findings we derive possible implications for the therapeutic modulation of said immunological mechanisms in connection with SARS‐CoV‐2/COVID‐19. Although, of course, the greatest care should be taken with patients with immunologically mediated diseases or immunomodulating therapies, it appears that many treatments can also be carried out during the COVID‐19 pandemic; some even appear to alleviate COVID‐19.
Psoriasis is a frequent systemic inflammatory autoimmune disease characterized primarily by skin lesions with massive infiltration of leukocytes, but frequently also presents with cardiovascular comorbidities. Especially polymorphonuclear neutrophils (PMNs) abundantly infiltrate psoriatic skin but the cues that prompt PMNs to home to the skin are not well-defined. To identify PMN surface receptors that may explain PMN skin homing in psoriasis patients, we screened 332 surface antigens on primary human blood PMNs from healthy donors and psoriasis patients. We identified platelet surface antigens as a defining feature of psoriasis PMNs, due to a significantly increased aggregation of neutrophils and platelets in the blood of psoriasis patients. Similarly, in the imiquimod-induced experimental in vivo mouse model of psoriasis, disease induction promoted PMN-platelet aggregate formation. In psoriasis patients, disease incidence directly correlated with blood platelet counts and platelets were detected in direct contact with PMNs in psoriatic but not healthy skin. Importantly, depletion of circulating platelets in mice in vivo ameliorated disease severity significantly, indicating that both PMNs and platelets may be relevant for psoriasis pathology and disease severity.
Background
A basic requirement for artificial intelligence (AI)–based image analysis systems, which are to be integrated into clinical practice, is a high robustness. Minor changes in how those images are acquired, for example, during routine skin cancer screening, should not change the diagnosis of such assistance systems.
Objective
To quantify to what extent minor image perturbations affect the convolutional neural network (CNN)–mediated skin lesion classification and to evaluate three possible solutions for this problem (additional data augmentation, test-time augmentation, anti-aliasing).
Methods
We trained three commonly used CNN architectures to differentiate between dermoscopic melanoma and nevus images. Subsequently, their performance and susceptibility to minor changes (‘brittleness’) was tested on two distinct test sets with multiple images per lesion. For the first set, image changes, such as rotations or zooms, were generated artificially. The second set contained natural changes that stemmed from multiple photographs taken of the same lesions.
Results
All architectures exhibited brittleness on the artificial and natural test set. The three reviewed methods were able to decrease brittleness to varying degrees while still maintaining performance. The observed improvement was greater for the artificial than for the natural test set, where enhancements were minor.
Conclusions
Minor image changes, relatively inconspicuous for humans, can have an effect on the robustness of CNNs differentiating skin lesions. By the methods tested here, this effect can be reduced, but not fully eliminated. Thus, further research to sustain the performance of AI classifiers is needed to facilitate the translation of such systems into the clinic.