Early prediction of ventricular peritoneal shunt dependency in aneurysmal subarachnoid haemorrhage patients by recurrent neural network-based machine learning using routine intensive care unit data.

Aneurysmal subarachnoid haemorrhage (aSAH) can lead to complications such as acute hydrocephalic congestion. Treatment of this acute condition often includes establishing an external ventricular drainage (EVD). However, chronic hydrocephalus develops in some patients, who then require placement of a permanent ventriculoperitoneal (VP) shunt. The aim of this study was to employ recurrent neural network (RNN)-based machine learning techniques to identify patients who require VP shunt placement at an early stage. This retrospective single-centre study included all patients who were diagnosed with aSAH and treated in the intensive care unit (ICU) between November 2010 and May 2020 (n = 602). More than 120 parameters were analysed, including routine neurocritical care data, vital signs and blood gas analyses. Various machine learning techniques, including RNNs and gradient boosting machines, were evaluated for their ability to predict VP shunt dependency. VP-shunt dependency could be predicted using an RNN after just one day of ICU stay, with an AUC-ROC of 0.77 (CI: 0.75-0.79). The accuracy of the prediction improved after four days of observation (Day 4: AUC-ROC 0.81, CI: 0.79-0.84). At that point, the accuracy of the prediction was 76% (CI: 75.98-83.09%), with a sensitivity of 85% (CI: 83-88%) and a specificity of 74% (CI: 71-78%). RNN-based machine learning has the potential to predict VP shunt dependency on Day 4 after ictus in aSAH patients using routine data collected in the ICU. The use of machine learning may allow early identification of patients with specific therapeutic needs and accelerate the execution of required procedures.

Clinical surrogates of dysautonomia predict lethal outcome in COVID-19 on intensive care unit.

BACKGROUND: Unpredictable vegetative deteriorations made the treatment of patients with acute COVID-19 on intensive care unit particularly challenging during the first waves of the pandemic. Clinical correlates of dysautonomia and their impact on the disease course in critically ill COVID-19 patients are unknown. METHODS: We retrospectively analyzed data collected during a single-center observational study (March 2020-November 2021) which was performed at the University Medical Center Hamburg-Eppendorf, a large tertiary medical center in Germany. All patients admitted to ICU due to acute COVID-19 disease during the study period were included (n = 361). Heart rate variability (HRV) and blood pressure variability (BPV) per day were used as clinical surrogates of dysautonomia and compared between survivors and non-survivors at different time points after admission. Intraindividual correlation of vital signs with laboratory parameters were calculated and corrected for age, sex and disease severity. RESULTS: Patients who deceased in ICU had a longer stay (median days ± IQR, survivors 11.0 ± 27.3, non-survivors 14.1 ± 18.7, P = 0.85), in contrast time spent under invasive ventilation was not significantly different (median hours ± IQR, survivors 322 ± 782, non-survivors 286 ± 434, P = 0.29). Reduced HRV and BPV predicted lethal outcome in patients staying on ICU longer than 10 days after adjustment for age, sex, and disease severity. Accordingly, HRV was significantly less correlated with inflammatory markers (e.g. CRP and Procalcitonin) and blood carbon dioxide in non-survivors in comparison to survivors indicating uncoupling between autonomic function and inflammation in non-survivors. CONCLUSIONS: Our study suggests autonomic dysfunction as a contributor to mortality in critically ill COVID-19 patients during the first waves of the pandemic. Serving as a surrogate for disease progression, these findings could contribute to the clinical management of COVID-19 patients admitted to the ICU. Furthermore, the suggested measure of dysautonomia and correlation with other laboratory parameters is non-invasive, simple, and cost-effective and should be evaluated as an additional outcome parameter in septic patients treated in the ICU in the future.

Activity-regulated cytoskeleton-associated protein/activity-regulated gene 3.1 (Arc/Arg3.1) enhances dendritic cell vaccination in experimental melanoma.

Dendritic cell (DC) vaccination has proven to be an effective and safe adjuvant for cancer immunotherapies. As the presence of DCs within the tumor microenvironment promotes adaptive antitumor immunity, enhancement of DC migration toward the tumor microenvironment following DC vaccination might represent one possible approach to increase its therapeutic efficacy. While recent findings suggest the activity-regulated cytoskeleton-associated protein/activity-regulated gene 3.1 (Arc/Arg3.1) as critical regulator of DC migration in the context of autoimmune diseases, we aimed to investigate the impact of Arc/Arg3.1 expression for DC-based cancer vaccines. To this end, DC migration capacity as well as the induction of T cell-mediated antitumor immunity was assessed in an experimental B16 melanoma model with Arc/Arg3.1-/- and Arc/Arg3.1-expressing BMDCs applied as a subcutaneous vaccine. While antigen presentation on DCs was critical for unleashing effective T cell mediated antitumor immune responses, Arc/Arg3.1 expression enhanced DC migration toward the tumor and secondary lymphoid organs. Moreover, Arc/Arg3.1-expressing BMDCs shape the tumor immune microenvironment by facilitating tumor recruitment of antigen-specific effector T cells. Thus, Arc/Arg3.1 may represent a novel therapeutic target in DCs in order to increase the therapeutic efficacy of DC vaccination.