Evidence of brain metabolism redistribution from neocortex to primitive brain structures in early acute COVID-19 respiratory syndrome.

BACKGROUND: Neuropsychiatric sequelae of COVID-19 have been widely documented in patients with severe neurological symptoms during the chronic or subacute phase of the disease. However, it remains unclear whether subclinical changes in brain metabolism can occur early in the acute phase of the disease. The aim of this study was to identify and quantify changes in brain metabolism in patients hospitalized for acute respiratory syndrome due to COVID-19 with no or mild neurological symptoms. RESULTS: Twenty-three non-intubated patients (13 women; mean age 55.5 ± 12.1 years) hospitalized with positive nasopharyngeal swab test (RT-PCR) for COVID-19, requiring supplemental oxygen and no or mild neurological symptoms were studied. Serum C-reactive protein measured at admission ranged from 6.43 to 189.0 mg/L (mean: 96.9 ± 54.2 mg/L). The mean supplemental oxygen demand was 2.9 ± 1.4 L/min. [18F]FDG PET/CT images were acquired with a median of 12 (4-20) days of symptoms. After visual interpretation of the images, semiquantitative analysis of [18F]FDG uptake in multiple brain regions was evaluated using dedicated software and the standard deviation (SD) of brain uptake in each region was automatically calculated in comparison with reference values of a normal database. Evolutionarily ancient structures showed positive SD mean values of [18F]FDG uptake. Lenticular nuclei were bilaterally hypermetabolic (> 2 SD) in 21/23 (91.3%) patients, and thalamus in 16/23 (69.6%), bilaterally in 11/23 (47.8%). About half of patients showed hypermetabolism in brainstems, 40% in hippocampi, and 30% in cerebellums. In contrast, neocortical regions (frontal, parietal, temporal and occipital lobes) presented negative SD mean values of [18F]FDG uptake and hypometabolism (< 2 SD) was observed in up to a third of patients. Associations were found between hypoxia, inflammation, coagulation markers, and [18F]FDG uptake in various brain structures. CONCLUSIONS: Brain metabolism is clearly affected during the acute phase of COVID-19 respiratory syndrome in neurologically asymptomatic or oligosymptomatic patients. The most frequent finding is marked hypermetabolism in evolutionary ancient structures such as lenticular nucleus and thalami. Neocortical metabolism was reduced in up to one third of patients, suggesting a redistribution of brain metabolism from the neocortex to evolutionary ancient brain structures in these patients.

Improving deep neural network generalization and robustness to background bias via layer-wise relevance propagation optimization.

Features in images' backgrounds can spuriously correlate with the images' classes, representing background bias. They can influence the classifier's decisions, causing shortcut learning (Clever Hans effect). The phenomenon generates deep neural networks (DNNs) that perform well on standard evaluation datasets but generalize poorly to real-world data. Layer-wise Relevance Propagation (LRP) explains DNNs' decisions. Here, we show that the optimization of LRP heatmaps can minimize the background bias influence on deep classifiers, hindering shortcut learning. By not increasing run-time computational cost, the approach is light and fast. Furthermore, it applies to virtually any classification architecture. After injecting synthetic bias in images' backgrounds, we compared our approach (dubbed ISNet) to eight state-of-the-art DNNs, quantitatively demonstrating its superior robustness to background bias. Mixed datasets are common for COVID-19 and tuberculosis classification with chest X-rays, fostering background bias. By focusing on the lungs, the ISNet reduced shortcut learning. Thus, its generalization performance on external (out-of-distribution) test databases significantly surpassed all implemented benchmark models.

Construction of a nomogram for predicting COVID-19 in-hospital mortality: A machine learning analysis.

Background and objectives: We aim to verify the use of ML algorithms to predict patient outcome using a relatively small dataset and to create a nomogram to assess in-hospital mortality of patients with COVID-19. Methods: A database of 200 COVID-19 patients admitted to the Clinical Hospital of State University of Campinas (UNICAMP) was used in this analysis. Patient features were divided into three categories: clinical, chest abnormalities, and body composition characteristics acquired by computerized tomography. These features were evaluated independently and combined to predict patient outcomes. To minimize performance fluctuations due to low sample number, reduce possible bias related to outliers, and evaluate the uncertainties generated by the small dataset, we developed a shuffling technique, a modified version of the Monte Carlo Cross Validation, creating several subgroups for training the algorithm and complementary testing subgroups. The following ML algorithms were tested: random forest, boosted decision trees, logistic regression, support vector machines, and neural networks. Performance was evaluated by analyzing Receiver operating characteristic (ROC) curves. The importance of each feature in the determination of the outcome predictability was also studied and a nomogram was created based on the most important features selected by the exclusion test. Results: Among the different sets of features, clinical variables age, lymphocyte number and weight were the most valuable features for prognosis prediction. However, we observed that skeletal muscle radiodensity and presence of pleural effusion were also important for outcome determination. Integrating these independent predictors was successfully developed to accurately predict mortality in COVID-19 in hospital patients. A nomogram based on these five features was created to predict COVID-19 mortality in hospitalized patients. The area under the ROC curve was 0.86 ± 0.04. Conclusion: ML algorithms can be reliable for the prediction of COVID-19-related in-hospital mortality, even when using a relatively small dataset. The success of ML techniques in smaller datasets broadens the applicability of these methods in several problems in the medical area. In addition, feature importance analysis allowed us to determine the most important variables for the prediction tasks resulting in a nomogram with good accuracy and clinical utility in predicting COVID-19 in-hospital mortality.

Non-Alcoholic Fatty Liver Disease in Long-Term Type 2 Diabetes: Role of rs738409 PNPLA3 and rs499765 FGF21 Polymorphisms and Serum Biomarkers.

Fibroblast growth factor 21 (FGF21) signaling and genetic factors are involved in non-alcoholic fatty liver disease (NAFLD) pathogenesis. However, these factors have rarely been studied in type 2 diabetes mellitus (T2D) patients from admixed populations such as in those of Brazil. Therefore, we aimed to evaluate rs738409 patanin-like phospholipase domain-containing protein (PNPLA3) and rs499765 FGF21 polymorphisms in T2D, and their association with NAFLD, liver fibrosis, and serum biomarkers (FGF21 and cytokeratin 18 levels). A total of 158 patients were included, and the frequency of NAFLD was 88.6%, which was independently associated with elevated body mass index. Significant liver fibrosis (≥F2) was detected by transient elastography (TE) in 26.8% of NAFLD patients, and was independently associated with obesity, low density lipoprotein, and gamma-glutamyl transferase (GGT). PNPLA3 GG genotype and GGT were independently associated with cirrhosis. PNPLA3 GG genotype patients had higher GGT and AST levels; PNPLA3 GG carriers had higher TE values than CG patients, and FGF21 CG genotype patients showed lower gamma-GT values than CC patients. No differences were found in serum values of FGF21 and CK18 in relation to the presence of NAFLD or liver fibrosis. The proportion of NAFLD patients with liver fibrosis was relevant in the present admixed T2D population, and was associated with PNPLA3 polymorphisms.