@article{PennigHoyerKrauskopfetal.2021,
  author    = {Pennig, Lenhard and Hoyer, Ulrike Cornelia Isabel and Krauskopf, Alexandra and Shahzad, Rahil and J{\"u}nger, Stephanie T. and Thiele, Frank and Laukamp, Kai Roman and Grunz, Jan-Peter and Perkuhn, Michael and Schlamann, Marc and Kabbasch, Christoph and Borggrefe, Jan and Goertz, Lukas},
  title     = {Deep learning assistance increases the detection sensitivity of radiologists for secondary intracranial aneurysms in subarachnoid hemorrhage},
  series = {Neuroradiology},
  volume    = {63},
  journal   = {Neuroradiology},
  number    = {12},
  issn      = {0028-3940},
  doi       = {10.1007/s00234-021-02697-9},
  url       = {http://nbn-resolving.de/urn:nbn:de:bvb:20-opus-308117},
  pages     = {1985-1994},
  year      = {2021},
  abstract  = {Purpose To evaluate whether a deep learning model (DLM) could increase the detection sensitivity of radiologists for intracranial aneurysms on CT angiography (CTA) in aneurysmal subarachnoid hemorrhage (aSAH). Methods Three different DLMs were trained on CTA datasets of 68 aSAH patients with 79 aneurysms with their outputs being combined applying ensemble learning (DLM-Ens). The DLM-Ens was evaluated on an independent test set of 104 aSAH patients with 126 aneuryms (mean volume 129.2 ± 185.4 mm3, 13.0\% at the posterior circulation), which were determined by two radiologists and one neurosurgeon in consensus using CTA and digital subtraction angiography scans. CTA scans of the test set were then presented to three blinded radiologists (reader 1: 13, reader 2: 4, and reader 3: 3 years of experience in diagnostic neuroradiology), who assessed them individually for aneurysms. Detection sensitivities for aneurysms of the readers with and without the assistance of the DLM were compared. Results In the test set, the detection sensitivity of the DLM-Ens (85.7\%) was comparable to the radiologists (reader 1: 91.2\%, reader 2: 86.5\%, and reader 3: 86.5\%; Fleiss κ of 0.502). DLM-assistance significantly increased the detection sensitivity (reader 1: 97.6\%, reader 2: 97.6\%,and reader 3: 96.0\%; overall P=.024; Fleiss κ of 0.878), especially for secondary aneurysms (88.2\% of the additional aneurysms provided by the DLM). Conclusion Deep learning significantly improved the detection sensitivity of radiologists for aneurysms in aSAH, especially for secondary aneurysms. It therefore represents a valuable adjunct for physicians to establish an accurate diagnosis in order to optimize patient treatment.},
  language  = {en}
}
@article{BohmannKurkaduMesnildeRochemontetal.2019,
  author    = {Bohmann, Ferdinand O. and Kurka, Natalia and du Mesnil de Rochemont, Richard and Gruber, Katharina and Guenther, Joachim and Rostek, Peter and Rai, Heike and Zickler, Philipp and Ertl, Michael and Berlis, Ansgar and Poli, Sven and Mengel, Annerose and Ringleb, Peter and Nagel, Simon and Pfaff, Johannes and Wollenweber, Frank A. and Kellert, Lars and Herzberg, Moriz and Koehler, Luzie and Haeusler, Karl Georg and Alegiani, Anna and Schubert, Charlotte and Brekenfeld, Caspar and Doppler, Christopher E. J. and Onur, Oezguer A. and Kabbasch, Christoph and Manser, Tanja and Pfeilschifter, Waltraud},
  title     = {Simulation-based training of the rapid evaluation and management of acute stroke (STREAM) — a prospective single-arm multicenter trial},
  series = {Frontiers in Neurology},
  volume    = {10},
  journal   = {Frontiers in Neurology},
  issn      = {1664-2295},
  doi       = {10.3389/fneur.2019.00969},
  url       = {http://nbn-resolving.de/urn:nbn:de:bvb:20-opus-369239},
  year      = {2019},
  abstract  = {Introduction: Acute stroke care delivered by interdisciplinary teams is time-sensitive. Simulation-based team training is a promising tool to improve team performance in medical operations. It has the potential to improve process times, team communication, patient safety, and staff satisfaction. We aim to assess whether a multi-level approach consisting of a stringent workflow revision based on peer-to-peer review and 2-3 one-day in situ simulation trainings can improve acute stroke care processing times in high volume neurocenters within a 6 months period. Methods and Analysis: The trial is being carried out in a pre-test-post-test design at 7 tertiary care university hospital neurocenters in Germany. The intervention is directed at the interdisciplinary multiprofessional stroke teams. Before and after the intervention, process times of all direct-to-center stroke patients receiving IV thrombolysis (IVT) and/or endovascular therapy (EVT) will be recorded. The primary outcome measure will be the "door-to-needle" time of all consecutive stroke patients directly admitted to the neurocenters who receive IVT. Secondary outcome measures will be intervention-related process times of the fraction of patients undergoing EVT and effects on team communication, perceived patient safety, and staff satisfaction via a staff questionnaire. Interventions: We are applying a multi-level intervention in cooperation with three "STREAM multipliers" from each center. First step is a central meeting of the multipliers at the sponsor's institution with the purposes of algorithm review in a peer-to-peer process that is recorded in a protocol and an introduction to the principles of simulation training and debriefing as well as crew resource management and team communication. Thereafter, the multipliers cooperate with the stroke team trainers from the sponsor's institution to plan and execute 2-3 one-day simulation courses in situ in the emergency department and CT room of the trial centers whereupon they receive teaching materials to perpetuate the trainings. Clinical Trial Registration: STREAM is a registered trial at https://clinicaltrials.gov/ct2/show/NCT03228251.},
  language  = {en}
}