Browsing by Author "Kara, Batuhan"

Filter results by typing the first few letters
Now showing 1 - 3 of 3
  • Thumbnail Image
    Item
    A joint convolutional-recurrent neural network with an attention mechanism for detecting intracranial hemorrhage on noncontrast head CT
    (NATURE PORTFOLIO, 2022-01-01) Alis, Deniz; Alis, Ceren; Yergin, Mert; Topel, Cagdas; Asmakutlu, Ozan; Bagcilar, Omer; Senli, Yeseren Deniz; Ustundag, Ahmet; Salt, Vefa; Dogan, Sebahat Nacar; Velioglu, Murat; Selcuk, Hakan Hatem; Kara, Batuhan; Ozer, Caner; Oksuz, Ilkay; Kizilkilic, Osman; Karaarslan, Ercan
    To investigate the performance of a joint convolutional neural networks-recurrent neural networks (CNN-RNN) using an attention mechanism in identifying and classifying intracranial hemorrhage (ICH) on a large multi-center dataset
  • Thumbnail Image
    Item
    Four-year Experience of Intravenous and Endovascular Treatment in Acute Ischemic Stroke: A Single Center Study
    (GALENOS PUBL HOUSE, 2022-01-01) Ozdemir, Zeynep; Dindar, Gulsah Zorgor; Aksoy, Sena; Acar, Erkan; Selcuk, Hakan; Kara, Batuhan; Soysal, Aysun
    Objective: Acute ischemic stroke (AIS) is a major cause of mortality and morbidity throughout the world. Intravenous thrombolysis (IVT) and endovasculary treatments (EVT) are recomended currently in eligible patients admitted within the therapeutic window. In this study, the data of AIS patients who were treated with intravenous and EVT methods in Istanbul Bakirkoy Prof. Dr. Mazhar Osman Training and Research Hospital between 2017-2020 were evaluated retrospectively. Materials and Methods: Five hundred and ninety patients who received IVT and/or EVT were included in the study. Demographic, clinical, radiological characteristics, risk factors and post-treatment clinical characteristics of these patients were analyzed. Results: Of the 590 patients, 324 (54.9\%) underwent IVT, 164 (27.8\%) EVT and 102 (17.3\%) combined IVT+EVT. The median National Institutes of Health Stroke Scale (NIHSS) scores were 9 (1-21) in the iv tPA group, 13 (3-27) in the EVT group, 12 (4-23) in the combined treatment group at admission In the IVT group, 220 patients had no artery occlusion (67.9\%), M2 segment of the middle cerebral artery (MCA) was found to be the most frequently occluded artery with 32 patients (9.9\%). In the EVT and combined IVT+EVT groups, the M1 segment of the MCA had the highest occlusion rate {[}76 (44.2\%), 49 (45\%), respectively]. Asymptomatic hemorrhage rate was higher in the EVT group than the other groups. Symptomatic hemorrhage rate was lower in the IVT group compared to the other groups. A total of 182 (56.2\%), 67 (39\%) and 53 (48.6\%) patients in the IVT, EVT and combined IVT+EVT groups had good outcome, respectively. Conclusion: Acute stroke treatment has been proven to significantly reduce the serious burden of stroke on patients, caregivers and society. For this reason, the societal awareness of the importance of urgent admission to emergency departmentsand the number and capacity of centers that provide AIS treatments should be increased.
  • Thumbnail Image
    Item
    Inter-vendor performance of deep learning in segmenting acute ischemic lesions on diffusion-weighted imaging: a multicenter study
    (NATURE PORTFOLIO, 2021-01-01) Alis, Deniz; Yergin, Mert; Alis, Ceren; Topel, Cagdas; Asmakutlu, Ozan; Bagcilar, Omer; Senli, Yeseren Deniz; Ustundag, Ahmet; Salt, Vefa; Dogan, Sebahat Nacar; Velioglu, Murat; Selcuk, Hakan Hatem; Kara, Batuhan; Oksuz, Ilkay; Kizilkilic, Osman; Karaarslan, Ercan
    There is little evidence on the applicability of deep learning (DL) in the segmentation of acute ischemic lesions on diffusion-weighted imaging (DWI) between magnetic resonance imaging (MRI) scanners of different manufacturers. We retrospectively included DWI data of patients with acute ischemic lesions from six centers. Dataset A (n=2986) and B (n=3951) included data from Siemens and GE MRI scanners, respectively. The datasets were split into the training (80\%), validation (10\%), and internal test (10\%) sets, and six neuroradiologists created ground-truth masks. Models A and B were the proposed neural networks trained on datasets A and B. The models subsequently fine-tuned across the datasets using their validation data. Another radiologist performed the segmentation on the test sets for comparisons. The median Dice scores of models A and B were 0.858 and 0.857 for the internal tests, which were non-inferior to the radiologist's performance, but demonstrated lower performance than the radiologist on the external tests. Fine-tuned models A and B achieved median Dice scores of 0.832 and 0.846, which were non-inferior to the radiologist's performance on the external tests. The present work shows that the inter-vendor operability of deep learning for the segmentation of ischemic lesions on DWI might be enhanced via transfer learning