Search

Background Neutropaenic patients are at a high risk of contracting severe infections. In particular, in these patients, parameters with a high negative predictive value are desirable for excluding infection or bacteraemia. This study evaluated sepsis biomarkers in neutropaenic patients suffering from systemic inflammatory response syndrome ( SIRS). Further, the predictive capacities of evaluated biomarkers in neutropaenic SIRS patients were compared to non-neutropaenic SIRS patients. Material and methods In this prospective observational cohort study, patients with clinically suspected sepsis were screened. The predictive capacities of procalcitonin ( PCT), C-reactive protein and lipopolysaccharide-binding protein ( LBP) in neutropaenic SIRS patients were evaluated in terms of their potential to identify infection or bacteraemia and were compared to results for non-neutropaenic SIRS patients. To select an appropriate control cohort, propensity score matching was applied, balancing confounding factors between neutropaenic and non-neutropaenic SIRS patients. Results Of 3370 prospectively screened patients with suspected infection, 51 patients suffered from neutropaenic SIRS. For the identification of infection, none of the assessed biomarkers presented a clinically relevant discriminatory potency. Lipopolysaccharide-binding protein and PCT demonstrated discriminatory capacity to discriminate between nonbacteraemic and bacteraemic SIRS in patients with neutropaenia [receiver-operating characteristics-area under the curves ( ROC- AUCs): 0·860, 0·818]. In neutropaenic SIRS patients, LBP had a significantly better ROC- AUC than in a comparable non-neutropaenic patient cohort for identifying bacteraemia ( P = 0·01). Conclusion In neutropaenic SIRS patients, none of the evaluated biomarkers was able to adequately identify infection. LBP and PCT presented a good performance in identifying bacteraemia. Therefore, these markers could be used for screening purposes to increase the pretest probability of blood culture analysis. [ABSTRACT FROM AUTHOR]

Aims: We tested the hypothesis that artificial intelligence (AI)-powered algorithms applied to cardiac magnetic resonance (CMR) images could be able to detect the potential patterns of cardiac amyloidosis (CA). Readers in CMR centers with a low volume of referrals for the detection of myocardial storage diseases or a low volume of CMRs, in general, may overlook CA. In light of the growing prevalence of the disease and emerging therapeutic options, there is an urgent need to avoid misdiagnoses. Methods and Results: Using CMR data from 502 patients (CA: n = 82), we trained convolutional neural networks (CNNs) to automatically diagnose patients with CA. We compared the diagnostic accuracy of different state-of-the-art deep learning techniques on common CMR imaging protocols in detecting imaging patterns associated with CA. As a result of a 10-fold cross-validated evaluation, the best-performing fine-tuned CNN achieved an average ROC AUC score of 0.96, resulting in a diagnostic accuracy of 94% sensitivity and 90% specificity. Conclusions: Applying AI to CMR to diagnose CA may set a remarkable milestone in an attempt to establish a fully computational diagnostic path for the diagnosis of CA, in order to support the complex diagnostic work-up requiring a profound knowledge of experts from different disciplines. [ABSTRACT FROM AUTHOR]