Coarse-to-Fine Covid-19 Segmentation via Vision-Language Alignment (preprint)

Segmentation of COVID-19 lesions can assist physicians in better diagnosis and treatment of COVID-19. However, there are few relevant studies due to the lack of detailed information and high-quality annotation in the COVID-19 dataset. To solve the above problem, we propose C2FVL, a Coarse-to-Fine segmentation framework via Vision-Language alignment to merge text information containing the number of lesions and specific locations of image information. The introduction of text information allows the network to achieve better prediction results on challenging datasets. We conduct extensive experiments on two COVID-19 datasets including chest X-ray and CT, and the results demonstrate that our proposed method outperforms other state-of-the-art segmentation methods.

A blood atlas of COVID-19 defines hallmarks of disease severity and specificity (preprint)

Treatment of severe COVID-19 is currently limited by clinical heterogeneity and incomplete understanding of potentially druggable immune mediators of disease. To advance this, we present a comprehensive multi-omic blood atlas in patients with varying COVID-19 severity and compare with influenza, sepsis and healthy volunteers. We identify immune signatures and correlates of host response. Hallmarks of disease severity revealed cells, their inflammatory mediators and networks as potential therapeutic targets, including progenitor cells and specific myeloid and lymphocyte subsets, features of the immune repertoire, acute phase response, metabolism and coagulation. Persisting immune activation involving AP-1/p38MAPK was a specific feature of COVID-19. The plasma proteome enabled sub-phenotyping into patient clusters, predictive of severity and outcome. Tensor and matrix decomposition of the overall dataset revealed feature groupings linked with disease severity and specificity. Our systems-based integrative approach and blood atlas will inform future drug development, clinical trial design and personalised medicine approaches for COVID-19.

Dynamic changes of peripheral blood T lymphocytes in COVID-19 patients

Objective: To analyze the dynamic changes of T lymphocytes in patients with COVID-19. Methods: Blood samples were collected from 40 COVID-19 cases and 40 healthy controls in Beihai People's Hospital from January to February, 2020. The counts of CD4 +T and CD8 +T lymphocytes were detected by flow cytometry. Moreover, the T lymphocyte counts in 24 convalescent patients with two consecutive negative nucleic acid test results were also detected. Results: The leukocytes and lymphocytes in the patients with acute COVID-19 were significantly lower than those in the healthy controls [(4.71±1.54)×10 9cell/L vs (6.26±1.44)×10 9cell/L, (1.13±0.41)×10 9cell/L vs (1.51±0.39)×10 9cell/L;both P<0.05]. The counts of CD4 +T and CD8 +T lymphocytes in the patients with acute COVID-19 were significantly lower than those in the healthy controls [(447.15±144.42) cell/μl vs (592.83±146.76) cell/μl, (309.35±173.05) cell/μl vs (397.20±136.94) cell/μl;both P<0.05], while no significant difference was observed in the CD4 +/CD8 +T cell ratio (P>0.05). In the 24 convalescent COVID-19 patients, the counts of CD4 +T and CD8 +T lymphocytes were higher during convalescence than in the acute phase [(598.08±138.71) cell/μl vs (420.67±147.38) cell/μl, (439.08±166.94) cell/μl vs (296.67±151.06) cell/μl;both P<0.05], but there was no significant difference in the T lymphocyte counts between the convalescent patients and the healthy controls (P>0.05). Conclusions: A transient immune deficiency occurred in patients with acute COVID-19, but the impaired immune function could restore to normal level during recovery.