Beste Altınay will give a seminar on “MFDM: MRI Free Decision Model for Diagnosis and Treatment Selection in Patients with Low Back and Neck Pain” on 31 May at 12:45. The abstract of the talk and a short bio is shared below.
Beste Mimaroğlu Altınay is currently a PhD Candidate at the Department of Medical Informatics, Middle East Technical University. She received her B.S. degree in Computer Engineering from Başkent University. She worked as a software specialist in the field of Hospital Information Management Systems. Afterwards, she worked as a research assistant in the Department of Management Information Systems at Cumhuriyet University and has been working as an IT Software Manager at Social Sciences University of Ankara since 2013.
Low back pain (LBP) and neck pain (NP) is a worldwide public health problem that affects life quality. Our goal was to develop a machine learning model that can direct LBP and NP patients for the appropriate treatment without magnetic resonance imaging (MRI) findings, thus reducing the demand for MRI and its burden on the health system. To evaluate the treatment outcomes, demographic information, clinical findings, and preoperative evaluation of pain, movement restriction, and pain data duration are analyzed from patient data in the pain clinic. Support Vector Machine (SVM) models are built by analyzing ten different attributes from 1482 patient data to classify correct treatment: drug, RF/IDET, or surgical intervention. The stepwise model proposed here classifies drug therapy patients with an 84% success ratio and can direct patients to surgery or RF/IDET with a 74.47% success ratio without MRI results. The proposed MRI Free Decision Model (MFDM) can be utilized in primary healthcare facilities to direct the patients to the appropriate treatment options without MRI, reducing the cost and load on the healthcare system while benefiting the patient by reducing the time to initiate the treatment.
Dr. Burçak Otlu will give a seminar on “Evaluating topography of mutational signatures using SigProfilerTopography” on 24 May at 12:45. The abstract of the talk and a short bio is shared below.
Burçak Otlu is a computer scientist with several years of experience in computational biology. She has been working as a postdoctoral scholar in UC San Diego, Cellular and Molecular Medicine Department since March 2018. She is working on mutational signatures and topography of somatic mutations across the genome. She is an academic fellow at the Department of Health Informatics at Middle East Technical University. Formerly, she was a visiting scholar in UC San Diego, Computer Science, and Engineering Department between August 2017 and February 2018. She received her Ph.D. from Middle East Technical University, Computer Engineering Department where she focused on developing tools and techniques for assessing the functional relevance of genomic loci. Prior to coming to UC San Diego, she has worked as a teaching assistant in the same department for several years. Previously, she has also worked in the private sector as a software engineer.
Mutations are found on the genomes of all cancerous and normal somatic cells. These mutations were generated by the activities of endogenous and exogenous mutational processes which were operative throughout the cell lineage. Mutational processes imprint characteristic patterns of somatic mutations, termed, mutational signatures. To exemplify, mutational signature associated with tobacco smoking causes C>A mutations prominently on the transcribed strand, whereas mutational signature due to exposure to ultraviolet light results in C>T mutations on the untranscribed strand during DNA transcription. Previous analyses have demonstrated that somatic mutations in cancer are not uniformly distributed across the landscape of the genome. Importantly, mutational signatures imprinted by different mutational processes exhibit distinct topographical properties including to be located on (i) early or late replicating regions, (ii) genic or intergenic regions, (iii) transcribed strand or untranscribed strand of DNA with respect to the transcription process, (iv) leading or lagging strand of DNA in regard to DNA replication. Remarkably, mutational signatures may accumulate somatic mutations preferably at nucleosome occupied loci, chromatin accessible regions, transcription factor binding sites, and histone modification sites. Here, I present SigProfilerTopography, the most advanced tool for evaluating the topography of mutational signatures. SigProfilerTopography allows examining all types of mutational signatures and reveals topographical dependencies related to chromatin accessibility, nucleosome occupancy, histone modifications, transcription factor binding sites, replication timing, transcription strand bias, replication strand bias, and processivity. The tool also allows performing user-defined custom analysis based on custom assays. Having augmented with realistic simulated mutations, SigProfilerTopography assesses the significance of its findings which finally characterizes the mutational signatures and gives insight about their underlying biological mechanisms.
Kuntay Aktaş ve Osman Tunç, 17 Mayıs 12:45'te "Medikal 3D Printing - Tanı, Tedavi ve Cerrahide Kişiye Özel 3 Boyutlu Çözümler" başlıklı sunumlarını gerçekleştireceklerdir. Sunum özeti ve sunucuların kısa özgeçmişleri aşağıda paylaşılmıştır.
Makina Mühendisliği’nden mezun olduktan sonra Medikal, Havacılık ve Savunma Sanayi alanında bir çok projede yer aldı. Btech Innovation’ın kurucu ortağı, Yönetim Kurulu Başkanı ve CEO’sudur.
Makine Mühendisliğini tamamladıktan sonra bir çok medikal proje içerisinde yer aldı. Anatomik Modelleme ve Medikal 3D Printing alanında birçok eğitim verdi.
Manyetik Rezonans Görüntüleme ve Bilgisayarlı Tomografi cihazlarından alınan 2 boyutlu radyolojik verilerin medikal görüntü işleme yazılımlarında 3 boyutlu olarak modellemeleri gerçekleştirilmektedir. Elde edilen 3 boyutlu anatomik yapıların defekt ya da patolojileri üzerinde ilgili cerrahi ekiple ameliyat öncesi sanal cerrahi planlamaları yapılarak kişiye özel çözümler üretilmektedir. Bu çözümler medikal model, kişiye özel cerrahi kılavuz ya da kişiye özel protez olabilmektedir. Organizasyon kapsamında tasarım ve üretim sürecinin klinik uygulamalarını içeren bir sunum gerçekleştirilecektir.
Ahmet Görkem Er will give a seminar on “Informatics in the Pandemic Era” on 26 April at 12:45. The abstract of the talk and a short bio is shared below.
Ahmet Gorkem Er is a clinical fellow in the Department of Infectious Diseases and Clinical Microbiology at Hacettepe University and a Ph.D. student in Medical Informatics at METU. He holds an M.D. degree from Istanbul University Istanbul Faculty of Medicine and took his internal medicine residency training from Hacettepe University. He believes in interdisciplinary academic research and mainly focuses on implementing health informatics approaches into the medical field.
COVID-19 pandemic continues to impact all aspects of our lives. An interdisciplinary approach where clinicians and informaticians work together is crucial to control the pandemic. From diagnosis to treatment to protective measures, informatics is central to COVID-19 research and for the delivery of healthcare to the patients. Ahmet Görkem Er will highlight the importance of health informatics by giving examples in the pandemic era.
Utku Kaya will give a seminar on “AI Based Medical Device Development ” on 12 April at 12:45. The abstract of the talk and a short bio is shared below.
Utku Kaya (BSc, Electrical & Electronics Engineering, METU), has first been involved in the software industry during his undergraduate years. After 10+ years of software engineering experience, he joined Oracle Turkey where he executed numerous public sector projects for 8 years. In 2019 he founded SmartAlpha, an AI start-up that aims to democratize AI in healthcare globally.
Industry experts agree that AI will revolutionize healthcare in near future. But who will develop these revolutionary tools: Engineers or clinicians? Utku Kaya, the founder of SmartAlpha, believes that the development of such impactful tools requires a transdiciplinary approach where engineers and clinicians understand and serve for each other during the entire production process. This journey to a high impact success is challenging due to strict regulatory needs, massive data dependencies and market complexity.
Utku Kaya outlines the path to develop an AI-in-healthcare solution from idea to product.