Auto-FedRL: Federated Hyperparameter Optimization for Multi-institutional Medical Image Segmentation

Federated learning (FL) is a distributed machine learning technique that enables collaborative model training while avoiding explicit data sharing. The inherent privacy-preserving property of FL algorithms makes them especially attractive to the medical field. However, in case of heterogeneous client data distributions, standard FL methods are unstable and require intensive hyperparameter tuning to achieve optimal performance. Conventional hyperparameter optimization algorithms are impractical in real-world FL applications as they involve numerous training trials, which are often not affordable with limited compute budgets. In this work, we propose an efficient reinforcement learning (RL)-based federated hyperparameter optimization algorithm, termed Auto-FedRL, in which an online RL agent can dynamically adjust hyperparameters of each client based on the current training progress. Extensive experiments are conducted to investigate different search strategies and RL agents. The effectiveness of the proposed method is validated on a heterogeneous data split of the CIFAR-10 dataset as well as two real-world medical image segmentation datasets for COVID-19 lesion segmentation in chest CT and pancreas segmentation in abdominal CT. © 2022, The Author(s), under exclusive license to Springer Nature Switzerland AG.

Acute liver necrosis in a SARS-CoV-2 positive patient with triple positive antiphospholipid syndrome

Background : Several thrombotic manifestations have been reported with SARS-Cov-2 infection including liver vascular involvement. Aims : We present a dramatic case of acute liver necrosis in a 36-year-old SARS-Cov-2 positive Italian woman with no respiratory symptoms and triple positive antiphospholipid syndrome (APS). Methods : The patient was referred to our University Hospital for acute hypertransaminasemia and liver failure (Figure). She had systemic lupus erythematosus (positivity for: ANA, anti-dsDNA, complement activation, Coombs;thrombocytopenia, previous arthritis). Anti-phopspholipid antibodies (aPL) were detected for the first time in 2015 during routine pregnancy screening and chronically treated with aspirin. Apparently, no venous/arterial nor obstetric events were recorded up to the recent hospitalization. FIGURE 1 Results : At arrival, US-Doppler and CT-scan were consistent with signs of chronic liver disease and occlusion of the three hepatic veins defining a Budd-Chiari syndrome. We opted for a stepwise approach considering anticoagulation (clexane 100 UI/Kg b.i.d) the first line of therapy before any invasive intervention. Dexamethasone 6 mg/ day b.i.d., 6 sessions of plasma-exchange, i.v.-immunoglobulin were sequentially planned to revert the liver damage sustained by aPL. After 5-days, two hepatic-veins resulted recanalized in association with amelioration of liver-enzyme/function and aPL quantification. Then we performed hepatic vein catheterization and transjugular liver biopsy. The histology showed multiple areas of necrosis associated with liver cirrhosis. Unexpectedly, no signs of acute Budd-Chiari were observed (e.g. intraparenchymal hemorrhages, centrilobular congestion, sinusoidal dilation). Other etiologies were also excluded and we hypothesized the involvement of small arteries of the liver in a triple positive APS in a patients with SLE. We finally addressed the patient to a liver transplant program and a tight multispecialistic follow-up. Conclusions : Thrombosis of arterial/venous vessels or microcirculation causes liver damage in some patients with aPL. Our report suggests that SARS-Cov-2 can exacerbate this prothrombotic condition determining a life-threatening complication such as acute liver failure.

The interventional radiology service during the covid-19 pandemic: Steps for managing the risk of infection

It is imperative to ensure the safety of health-care workers in the angiographic room during the outbreak of the coronavirus disease 2019 (COVID-19). The selection criteria for interventional radiology (IR) procedures, the prepara-tion of the staff and angiographic suite, ventilation systems, and intra-and post-procedural workflow optimization methods are detailed. The specific measures needed to protect occupational safety and health may result in higher costs, longer procedural times, and greater technical problems. However, these precautions may help to minimize the spread of COVID-19 among IR practitioners. © 2020 CC BY 4.0 – in cooperation with Depts. of Cardiothoracic/ Vascular Surgery, General Surgery and Anesthesia, Örebro University Hospital and Örebro University, Sweden.