Cross-Site Severity Assessment of COVID-19 From CT Images via Domain Adaptation

Early and accurate severity assessment of Coronavirus disease 2019 (COVID-19) based on computed tomography (CT) images offers a great help to the estimation of intensive care unit event and the clinical decision of treatment planning. To augment the labeled data and improve the generalization ability of the classification model, it is necessary to aggregate data from multiple sites. This task faces several challenges including class imbalance between mild and severe infections, domain distribution discrepancy between sites, and presence of heterogeneous features. In this paper, we propose a novel domain adaptation (DA) method with two components to address these problems. The first component is a stochastic class-balanced boosting sampling strategy that overcomes the imbalanced learning problem and improves the classification performance on poorly-predicted classes. The second component is a representation learning that guarantees three properties: 1) domain-transferability by prototype triplet loss, 2) discriminant by conditional maximum mean discrepancy loss, and 3) completeness by multi-view reconstruction loss. Particularly, we propose a domain translator and align the heterogeneous data to the estimated class prototypes (i.e., class centers) in a hyper-sphere manifold. Experiments on cross-site severity assessment of COVID-19 from CT images show that the proposed method can effectively tackle the imbalanced learning problem and outperform recent DA approaches.

Triglyceride/High-Density Lipoprotein Cholesterol Ratio is Associated with the Mortality of COVID-19: A Retrospective Study in China

Background

Triglyceride to high density lipoprotein cholesterol ratio (TG/HDL-c) is crucial when researching metabolic and vascular diseases, and its involvement in COVID-19 was sparsely elaborated on. The purpose of the study was to explore the inflammatory associations between the TG/HDL-c ratio and COVID-19 prognosis.

Methods

A total of 262 COVID-19 patients consisting of 244 survivors and 18 non-survivors were retrospectively investigated. The clinical features and baseline hematological parameters were recorded and analyzed. The receiver operating characteristic curve (ROC) was used to explore the role of TG/HDL-c in predicting the mortality of COVID-19, the Spearman’s rank correlation coefficients were used to measure the correlation between TG/HDL-c and inflammatory indicators, and the Kaplan–Meier (KM) curve was used to estimate the survival of COVID-19 patients with high and low TG/HDL-c ratio. Logistic regression analyses were performed to investigate the role of TG/HDL-c ratio on mortality of COVID-19 with no underlying diseases.

Results

Compared with the survivors, the non-survivors of COVID-19 had significantly higher levels of white blood cells (4.7 vs 13.0 × 10⁹/L; P < 0.001), neutrophils (3.0 vs 11.6 × 10⁹/L; P < 0.001), C-reactive proteins (15.7 vs 76.7 mg/L; P < 0.001) and TG/HDL-c ratio (1.4 vs 2.5; P = 0.001). The ROC curve [area under the curve (AUC), 0.731; 95% confidence interval (CI), 0.609–0.853; P = 0.001] suggested that the TG/HDL-c ratio could predict the mortality of COVID-19. The TG/HDL-c ratio was positively correlated with white blood cells (r = 0.255, P < 0.001), neutrophils (r = 0.243, P < 0.001) and C-reactive proteins (r = 0.170, P < 0.006). Patients with high TG/HDL-c ratio showed a worse survival compared with those with low TG/HDL-c ratio (Log rank P = 0.003). Moreover, TG/HDL-c ratio was an independent factor in predicting the mortality of COVID-19 patients with no underlying diseases.

Conclusion

Our study demonstrated that TG/HDL-c ratio might potentially be a predictive marker for mortality in COVID-19 patients.

Nonadherence to Treatment and Patient-Reported Outcomes of Psoriasis During the COVID-19 Epidemic: A Web-Based Survey

PURPOSE

The COVID-19 epidemic has caused difficulties in continuous treatment for patients with chronic diseases and resulted in nonadherence to treatment and adverse health outcomes. This study aimed to investigate the associations of nonadherence to treatment with patient-reported outcomes of psoriasis during the COVID-2019 epidemic.

METHODS

A cross-sectional study among Chinese patients with psoriasis was conducted through a web-based questionnaire survey during 25 Feb 2020 and 6 Mar 2020. Demographic and clinical data, nonadherence to treatment, and patient-reported outcomes were collected. The outcomes included deterioration of the disease condition, perceived stress, and symptoms of anxiety and depression. Logistic regression was used to investigate the associations.

RESULTS

A total of 926 questionnaires were collected. A total of 634 (68.5%) reported nonadherence to treatment, and worse adherence was found among patients receiving systemic treatment (adjusted odds ratio [AOR]: 2.67; 95% CI: 1.40-5.10) and topical treatment (AOR: 4.51; 95% CI: 2.66-7.65) compared to biological treatment. Nonadherence to treatment (less than two weeks and more than two weeks) was significantly associated with deterioration of psoriasis (aOR: 2.83 to 5.25), perceived stress (AOR: 1.86 to 1.57), and symptoms of anxiety (AOR: 1.42 to 1.57) and depression (AORs: 1.78). Subgroup analysis by treatment showed consistent results in general.

CONCLUSION

Nonadherence to treatment was associated with the aggravation of psoriasis conditions, perceived stress, and symptoms of anxiety and depression.