Multiclass Segmentation of Breast Tissue and Suspicious Findings: A Simulation-Based Study for the Development of Self-Steering Tomosynthesis.

In breast tomosynthesis, multiple low-dose projections are acquired in a single scanning direction over a limited angular range to produce cross-sectional planes through the breast for three-dimensional imaging interpretation. We built a next-generation tomosynthesis system capable of multidirectional source motion with the intent to customize scanning motions around "suspicious findings". Customized acquisitions can improve the image quality in areas that require increased scrutiny, such as breast cancers, architectural distortions, and dense clusters. In this paper, virtual clinical trial techniques were used to analyze whether a finding or area at high risk of masking cancers can be detected in a single low-dose projection and thus be used for motion planning. This represents a step towards customizing the subsequent low-dose projection acquisitions autonomously, guided by the first low-dose projection; we call this technique "self-steering tomosynthesis." A U-Net was used to classify the low-dose projections into "risk classes" in simulated breasts with soft-tissue lesions; class probabilities were modified using post hoc Dirichlet calibration (DC). DC improved the multiclass segmentation (Dice = 0.43 vs. 0.28 before DC) and significantly reduced false positives (FPs) from the class of the highest risk of masking (sensitivity = 81.3% at 2 FPs per image vs. 76.0%). This simulation-based study demonstrated the feasibility of identifying suspicious areas using a single low-dose projection for self-steering tomosynthesis.

Low muscle mass and high visceral fat mass predict mortality in patients hospitalized with moderate-to-severe COVID-19: a prospective study.

Introduction: The severity of coronavirus disease 2019 (COVID-19) has been positively correlated with several comorbidities. The primary outcome of the study was to assess the relationship between the mortality and severity of COVID-19 and obesity classes according to BMI, visceral adipose tissue (VAT) area, s.c. adipose tissue area, muscle area (MA), and leptin levels. Methods: In this prospective cohort study, 200 patients hospitalized with moderate-to-severe COVID-19 underwent an unenhanced CT of the thorax and laboratory tests, and leptin levels between June and August 2020 were obtained. Results: Our study included 200 patients (male 52%; mean age: 62 (49-74) years; obesity (BMI > 30): 51.5%)). Fifty-eight patients (23.5%) were admitted to the intensive care unit and 29 (14.5%) died. In multivariate logistic regression (corrected for leptin, sex, age, and serum biomarkers) and receiver operating characteristic curve analyses, high VAT > 150 cm2 (odds ratio (OR): 6.15; P < 0.002), MA < 92 cm2 (OR: 7.94; P < 0.005), and VAT/MA ratio > 2 (OR: 13.9; P < 0.0001) were independent risk factors for mortality. Indeed, the Kaplan-Meier curves showed that patients with MA < 92 cm2 and without obesity (BMI < 30) had a lower survival rate (hazard ratio between 3.89 and 9.66; P < 0.0006) than the other groups. Leptin levels were not related to mortality and severity. Conclusion: This prospective study reports data on the largest number of hospitalized severe COVID-19 patients and pinpoints VAT area and MA calculated by CT as predictors of COVID-19 mortality.

Heterozygote Advantage of the Type II Deiodinase Thr92Ala Polymorphism on Intrahospital Mortality of COVID-19.

CONTEXT: The type 2 deiodinase and its Thr92Ala-DIO2 polymorphism have been linked to clinical outcomes in acute lung injury and pulmonary fibrosis. OBJECTIVE: Our objectives were to evaluate were cumulative mortality during admission according to Thr92Ala-DIO2 polymorphism. METHODS: Here we conducted an observational, longitudinal, and prospective cohort study to investigate a possible association between the Thr92Ala-DIO2 polymorphism and intrahospital mortality from COVID-19 in adult patients admitted between June and August 2020. Blood biochemistry, thyroid function tests, length of stay, comorbidities, complications, and severity scores were also studied according to Thr92Ala-DIO2 polymorphism. RESULTS: In total, 220 consecutive patients (median age 62; 48-74 years) were stratified into 3 subgroups: Thr/Thr (n = 79), Thr/Ala (n = 119), and Ala/Ala (n = 23). While the overall mortality was 17.3%, the lethality was lower in Ala/Thr patients (12.6%) than in Thr/Thr patients (21.7%) or Ala/Ala patients (23%). The heterozygous genotype (Thr/Ala) was associated with a 47% reduced risk of intrahospital mortality whereas univariate and multivariate logistic regression adjusted for multiple covariates revealed a reduction that ranged from 51% to 66%. The association of the Thr/Ala genotype with better clinical outcomes was confirmed in a metanalysis of 5 studies, including the present one. CONCLUSION: Here we provide evidence for a protective role played by Thr92Ala-DIO2 heterozygosity in patients with COVID-19. This protective effect follows an inheritance model known as overdominance, in which the phenotype of the heterozygote lies outside the phenotypical range of both homozygous.

Thyroid Hormone Levels During Hospital Admission Inform Disease Severity and Mortality in COVID-19 Patients.

Background: Illness severity in patients infected with COVID-19 is variable. Methods: Here, we conducted an observational, longitudinal, and prospective cohort study to investigate serum thyroid hormone (TH) levels in adult COVID-19 patients, admitted between June and August 2020, and to determine whether they reflect the severity or mortality associated with the disease. Results: Two hundred forty-five patients [median age: 62 (49-75) years] were stratified into non-critical (181) and critically ill (64) groups. Fifty-eight patients (23.6%) were admitted to the intensive care unit, and 41 (16.7%) died. Sixteen (6.5%) exhibited isolated low levels of free triiodothyronine (fT3). fT3 levels were lower in critically ill compared with non-critical patients [fT3: 2.82 (2.46-3.29) pg/mL vs. 3.09 (2.67-3.63) pg/mL, p = 0.007]. Serum reverse triiodothyronine (rT3) was mostly elevated but less so in critically ill compared with non-critical patients [rT3: 0.36 (0.28-0.56) ng/mL vs. 0.51 (0.31-0.67) ng/mL, p = 0.001]. The univariate logistic regression revealed correlation between in-hospital mortality and serum fT3 levels (odds ratio [OR]: 0.47; 95% confidence interval [CI 0.29-0.74]; p = 0.0019), rT3 levels (OR: 0.09; [CI 0.01-0.49]; p = 0.006) and the product fT3 × rT3 (OR: 0.47; [CI 0.28-0.74]; p = 0.0026). Serum thyrotropin, free thyroxine, and fT3/rT3 values were not significantly associated with mortality and severity of the disease. A serum cutoff level of fT3 (≤2.6 pg/mL) and rT3 (≤0.38 ng/mL) was associated with 3.46 and 5.94 OR of mortality, respectively. We found three COVID-19 mortality predictors using the area under the receiver operating characteristic (ROC) curve (AUC score): serum fT3 (AUC = 0.66), rT3 (AUC = 0.64), and the product of serum fT3 × rT3 (AUC = 0.70). Non-thyroidal illness syndrome (fT3 < 2.0 pg/mL) was associated with a 7.05 OR of mortality ([CI 1.78-28.3], p = 0.005) and the product rT3 × fT3 ≤ 1.29 with an 8.08 OR of mortality ([CI 3.14-24.2], p < 0.0001). Conclusions: This prospective study reports data on the largest number of hospitalized moderate-to-severe COVID-19 patients and correlates serum TH levels with illness severity, mortality, and other biomarkers to critical illness. The data revealed the importance of early assessment of thyroid function in hospitalized patients with COVID-19, given the good prognostic value of serum fT3, rT3, and fT3 × rT3 product. Further studies are necessary to confirm these observations.