Project Title: Small Animal Imaging Application and Innovation Platform for Detecting the Effects of Aging on the Brain and Musculoskeletal System (click here for more information)

Funded by TUBITAK 1004 Grants (2023-2027) 

Principal Investigator: Esin Ozturk Isik

Researchers: Arda Arpak, Said Aldemir

Project Summary: The main goal of this project is to create multimodal imaging protocols and develop innovative software systems at 7T PET/MR to detect and monitor effects of aging that could potentially be translated to human studies in the future. The first aim is to develop a magnetic resonance fingerprinting (MRF) sequence to estimate tissue parameters (T1, T2, R2*, QSM). The second aim is to develop a robust and fast MR spectroscopic imaging sequence. The third aim is to develop a brain atlas using PET/MR in rats. The fourth aim is to develop a deep learning software platform for high-sensitivity small animal imaging at 7T MRI that will have modules for accelerated data acquisition, spatial resolution enhancement, image reconstruction and unsupervised learning techniques to reduce the high data dependency of artificial neural networks.

Keywords: Brain imaging, aging, MRI reconstruction, 7T MRI

Project Title: Assessment of Radiotherapy Planning Efficacy Based on Brain Tumor Shape Analysis (click here for more information)

Funded by Bogazici BAP Grant 19482 (2022-2023) 

Principal Investigator: Esin Öztürk Işık

Researchers:  Esra Sümer

Project Summary: The primary goal of Gamma Knife (GK) dose planning is to cover any shape irregularities of target that strongly affect dose distribution inside and outside the target. However, there is currently no clinically practical tool for estimating the effect of shape irregularity of target on dose plan efficiency. In this study, the main aims are to measure tumor shape irregularity and analyze its effect on GK plan efficiency and treatment outcomes. In the present study, GK treatment plans created by Perfection/ICON platform for vestibular schwannoma. Tumor shape irregularity will be measured using radiomic shape features from segmented magnetic resonance (MR) images. Dose planning efficiency will be measured using the selectivity index (SI), gradient index (GI), Paddick conformality index (PCI), and efficacy index (EI). Correlation and linear regression analyses will be applied between the shape irregularity features and the dose plan indices. Then, machine learning algorithms will be used to identify the best-performing shape feature to predict dose plan efficiency. Finally, treatment outcomes at year 2 will be investigated, including tumor growth control and functional hearing preservation using Cox regression analyses to determine whether tumor shape irregularity has any effect on GK plan efficiency and treatment outcomes. Among the various shape irregularity metrics, the one with the highest predictive performance will be selected. It is expected that the irregularity of tumor shape will provide useful information to the clinician with the initial estimation of treatment efficacy and the comparisons of tumors.

Keywords: Gamma knife radiosurgery, shape irregularity, radiomics, machine learning.

Project Title: The Effect of Super Resolution Deep Learning on Radiomic Shape Features Acquired from Diffusion MRI of Stroke Patients

Funded by Bogazici BAP Grant 19363 (Doctorate Project, 2022-2023) 

Principal Investigator: Esin Öztürk Işık

Researchers:  Asım Samli

Project Summary: Stroke is the 2nd leading cause of death in the World causing life long disabilities and serious decline in patients' quality of life. Diffusion weighted imaging (DWI) and apparent diffusion coefficients (ADC) can detect location and extent of ischemia induced brain damage within minutes of onset with high precision and specificity. Hence, DWI is the standard diagnostic and monitoring tool for stroke. On the other hand, DWI has relatively low spatial resolution compared to other contemporary imaging modalities. Large number of minable and quantitative features, called radiomics, can be extracted from medical images. Previous studies have reported that the resolution of medical images have significant effects on radiomics features. The main purpose of this project is to obtain  high resolution DWI images from low resolution using deep learning, then identifying the effect of the super resolution technique on the radiomics features. In this project, we will assess the robustness of the radiomics shape features obtained from low resolution DWI images, and super resolved images using deep learning that have 2 or 4 fold higher resolution. As a result, robustness of widely used radiomic shape features will be assessed upon application of super resolution and unstable shape features will be determined. 

Keywords: Super resolution, deep learning, radiomics. 

Project Title: Developing BBB-ASL as non-invasive early biomarker of Alzheimer's disease (click here for more information) 

Funded by JPND COFUND TÜBİTAK 1071 Grants (2021-2023) 

Principal Investigator: Alp Dinçer 

Co-Investigators: Esin Öztürk Işık, Matthias Günther, Eric Achten, Henk Mutsaerts, Udunna Anazodo, Tormod Fladby, Catherine Morgan, David Thomas, Jennifer Linn, Saima Hilal

Researchers:  Ayse Irem Cetin, Gulce Turhan

Project Summary: Alzheimer's Disease (AD) is a progressive neurodegenerative disease with multiple pathologies. Biomarkers that reliably detect the onset of the disease before any clinical symptoms appear are crucial for the diagnosis of AD and related dementias. One of the earliest pathological changes in AD is the loss of Blood-Brain-Barrier (BBB) ​​integrity. Non-invasive arterial spin labeling magnetic resonance imaging (ASL-MRI) technique measures blood perfusion. Perfusion change measured quantitatively by the ASL-MRI technique can provide information in the diagnosis and follow-up of BBB integrity. The project '(Developing BBB-ASL as non-invasive early biomarker of Alzheimer's disease) (DEBBIE)' presented to the Joint Program for Neurodegenerative Disease Research (JPND) has proposed to be developed a clinical imaging biomarker called as "BBB-ASL" that maps the loss of BBB integrity in AD. The proposed DEBBIE project was entitled to receive support under the second call of JPND JPCOFUND-2 ERA-NET Cofund-2 2020. The European Union allowed this project consortium to expand, and the consortium Coordinator  Prof. Dr. Matthias Günther  reached Dr. Alp Dinçer at Mehmet Ali Aydınlar Acıbadem University professor and Assoc. Dr. Esin Öztürk Işık at Boğaziçi University in Turkey as partner. As a result of the interviews, the partners in Turkey have agreed to the application of the BBB-ASL technique in brain tumors. In this context, it is aimed to evaluate the BBB differences in different histopathological tumor grades in patients with glioma, a type of brain tumor, using the BBB-ASL technique. Moreover, the parameters of BBB integrity differences in histopathological tumor grades will be classified by machine learning algorithms. Additionally, the disruption in BBB integrity and water exchange differences between AD and glioma patients will be assessed. 

Keywords: Neurodegenerative disorders, dementia, brain imaging, brain tumors, blood-brain-barrier, arterial spin labeling

Project Title: Advanced Magnetic Resonance Imaging and Machine Learning Based Product Development for Noninvasive Detection of Genetic Subgroups of Brain Membrane Tumors (click here for more information)

Funded by TÜBİTAK 1001 Grants (2020-2022) 

Principal Investigator: Esin Öztürk Işık

Co-Investigators:  Alp Dinçer, Koray Özduman, Alpay Özcan, Ayça Ersen Danyeli, Murat Şakir Ekşi, Özge Can

Researchers:  Abdullah Baş, Asım Samlı, Banu Saçlı Bilmez, Buse Buz, Esra Sümer, Gökçe Hale Hatay, Sena Azamat

Project Summary: Brain membrane tumors, or meningiomas, are the most common brain tumors with a 35% incidence rate in adults. Meningiomas are more common in older people, and it is becoming a major health problem in our ever-aging society. While 80% of meningiomas are grade I and benign, grade III meningiomas tend to grow rapidly and spread to the surrounding tissue. In recent years, various genetic biomarkers for better understanding brain tumor biology have been identified, and they have been added to the classification criteria for brain tumors. These genetic biomarkers have become increasingly important in clinical practice today in diagnosing brain tumors, and to more accurately predict tumor biology, treatment responses, recurrence patterns, and survival rates. It is known that more than 60% of meningomas occur as a result of neurofibromatosis type 2 (NF2) gene changes. Molecular studies have shown that this molecular subgroup with NF2 mutations, which constitutes the majority of meningiomas, carry chromosomal instabilities, and they become malign, invade surrounding tissues, and turn into fast growing tumors. The knowledge of NF2 mutations in meningiomas prior to the application of treatment, such as surgery or radiotherapy, has the potential to significantly affect the treatment decision. However, these mutations are usually limited to the active tumor site, and could only be evaluated by pathological and molecular biological tests conducted on the tumor tissue after surgical removal. There is currently no product that might noninvasively provide NF2 mutation information in meningiomas. In recent years, advanced magnetic resonance imaging (MRI) based machine learning classification methods have gained great momentum in brain tumor studies. Diffusion tensor imaging (DTI), perfusion MRI, proton MR spectroscopic imaging (1H-MRSI), and susceptibility weighted imaging (SWI), provide information about white matter tract changes, perfusion and metabolic abnormalities, and magnetic susceptibility variations. Additionally, radiomic features of MRI modalities provide further hidden quantitative information about data characteristics. There has been a great interest in MRI-based classification of diseases using machine learning, and more recently deep learning methods. The aim of this study is to develop a machine learning based product, for detection of a radiological signature specific to the NF2 molecular subgroup of meningiomas, using advanced MR imaging and radiomic features. Preoperative diagnosis of NF2 molecular subtype of meningiomas by machine learning methods based on non-invasive MRI techniques will contribute to appropriate treatment planning and improvement of patient health.

Keywords: Brain membrane tumor, meningioma, magnetic resonance imaging, neurofibromatosis type 2 (NF2), machine learning, deep learning

Project Title: Automatic Assessment of Gait Impairments in Stroke using Artificial Intelligence, Wearable Technology and Neuroimaging

Funded by Aberystwyth University CIDRA Grants (2020-2021) 

Principal Investigator: Otar Akanyeti

Co-Investigators: Esin Öztürk Işık, Hale Saybaşılı, Can Yücesoy, Alp Dinçer, Dilaver Kaya, Nazire Afşar, Federico Villagra Povina

Researchers: Esra Sümer, Meryem Şahin

Project Summary: The interdisciplinary project aims for delivering wearable technology to improve the lives of stroke patients in Turkey. Recently developed by Aberystwyth University (in collaboration with NHS Wales), the wearable tech provides high resolution movement data to quantify walking disability and evaluate the efficacy of stroke treatment. By providing objective measurements, it will promote consistency of service across Turkey and empower patients to be more independent. Collecting data over long periods in more natural settings will reveal new insights into stroke prevalence and outcomes. It will be easier to identify patients with the greatest health needs and offer personalized rehabilitation programs.

Keywords: Stroke, Gait analysis, Magnetic Resonance Imaging, Blood Biomarkers, Brain-Derived Neurotrophic Factor

Project Title: Determination of Metabolic and Perfusion Magnetic Resonance Imaging Based Biomarkers in Parkinson's Disease Dementia and Parkin Gene Mutations

Funded by Bogazici BAP Grant 19XP1 (2019-2020) 

Principal Investigator: Esin Öztürk Işık

Researchers: Sena Azamat

Project Summary: Parkinson's disease (PD) is the second most common neurodegenerative disease. It has been discovered that non-motor (cognitive) findings also accompany the disease from very early on. Cognitive findings are the ones that lead to the most functional loss. Parkin (PARK2) gene mutation is the most common form of autosomal recessive parkinsonism. Early onset, slow clinical course, positive response to low dose levodopa could be cited as specific features. Following the discovery that parkinsonism develops in patients with PARK2 gene mutation without the presence of Lewy Body (LC), it has become important to identify biomarkers that will enable the differentiation of Parkin mutation carriers and Parkinson’s Disease Dementia. The aim of this study is to determine biomarkers indicating metabolic and perfusion changes in the brain ​​of PARK2 mutation carriers and patients with Parkinson's disease dementia by using ASL-MRI and 1H-MRSI techniques at 3T.

Keywords: Parkinson’s Disease Dementia, Parkin Gene Mutation, ASL-MRI, Proton MR Spectroscopy

Project Title: Development of a Diagnostic Tool for Identifying Genetic, Metabolic and Histopathologic Properties of Glial Brain Tumors (click here for more information)

Funded by TÜBİTAK 1003 Grants (2017-2020) 

Principal Investigator: Alp Dinçer

Co-Investigators: Esin Öztürk Işık, Koray Özduman, Alpay Özcan, Özge Can

Researchers: Abdullah Baş, Asım Samlı, Ayhan Gürsan, Banu Saçlı Bilmez, Buse

Project Summary: The aim of this project is to combine multimodal MR information for applying machine learning methodologies with the aim of differentiating glioblastoma types. For this purpose, the MR, genetic and metabolomic data obtained from a prospective cohort will be analyzed and machine learning methods will be developed for predicting survival rate by only using the MR data. The genetic and metabolomic data obtained from biopsy will be included in the ground truth information.

Keywords: Glioma, Magnetic Resonance Imaging, Magnetic Resonance Spectroscopy, Genomics, Metabolomics, Machine Learning.

Project Title: 1H Bilateral, Flexible Breast RF Coil Design for Magnetic Resonance Imaging Systems 

Funded by TÜBİTAK 1001 Grants (2017-2019) 

Principal Investigator: Korkut Yeğin

Co-Investigators: Esin Öztürk Işık

Researchers:  Başak Bayrambaş, Büşra Kahraman

Project Summary: The main aim of this study is to design 1H bilateral, flexible breast RF coil for magnetic resonance imaging systems and to develop a software for analysis of breast MR images obtained with this coil.

Keywords: Breast, RF coil, magnetic resonance imaging.

Project Title: Determination of Multimodality Magnetic Resonance Imaging Based Biomarkers for Mild Cognitive Impairment in Parkinson Disease

(click here for more information in Turkish)

Funded by TÜBİTAK 1001 Grants (2015-2018) 

Principal Investigator: Esin Öztürk Işık

Co-Investigators: Tamer Demiralp, Hakan Gürvit, Başar Bilgiç, Haşmet Hanağası, Aziz Uluğ, Erdem Tüzün

Researchers:  Ani Kıçik, Dilek Betül Arslan, Sevim Cengiz, Gökçe Hale Hatay, Ozan Genç, Muhammed Yıldırım, Kardelen Eryürek

Project Summary: The main aim of this study to develop biomarkers that would indicate the presence of mild cognitive imparment (MCI) in Parkinson disease (PD) and the probability of its’ evolution into dementia by evaluating the findings from multimodality structural, metabolic, and functional MR imaging of patients diagnosed with PD-MCI or cognitively intact Parkinson’s disease, and healthy controls.

Keywords: Parkinson disease, mild cognitive impairment, neuropsychological evaluation, magnetic resonance imaging, genetics, machine learning.

Project Title: Investigation of the Human Brain Metabolism in-vivo in Chronic Liver Failure Using Magnetic Resonance Spectroscopic Editing Techniques

Funded by Boğaziçi University BAP Grants (2015-2018) 

Principal Investigator: Esin Öztürk Işık

Co-Investigators:  Bahattin Hakyemez, Emre Ökeer, Tuba Erürker Öztürk, Aylin Bican Demir

Researcher:  Gökçe Hale Hatay

Project Summary: The main objective of this study is to understand the metabolic changes in the brain that occur due to minimal hepatic encephalopathy, and define metabolic biomarkers that can help in the diagnosis of this disorder.

Keywords: Biocomputing, Signal Processing, MR spectroscopic imaging, Minimal hepatic encephalopathy.

Project Title: Feasibilty Study of Obtaining High-Resolution Spectroscopy Images Using Data Fusion Techniques (click here for more information)

Funded by Royal Society Newton Mobility Grants (2016-2017) 

Investigators: Esin Öztürk Işık / Maria Valdes-Hernandez

Project Summary: The main objective of this study is to develop a pipeline to increase the resolution of magnetic resonance spectroscopy images to provide the international scientific and research community a technique to study tissue microstructure and metabolic changes in neurological diseases, and increase research output by optimising tissue characterisation complementing the information from multimodal and multispectral magnetic resonance imaging with the application of this technique.

Keywords: MR spectroscopic imaging, Cancer, Deep-Learning, Convolutional Neural Network, Signal Processing and Analysis.

Project Title: Accelerated Phosphorus MR Spectroscopic Imaging of Brain Tumors at 3T using Compressed Sensing (click here for more information)

Funded by TÜBİTAK 3501 Career Development Grants  (2012-2014) 

Principal Investigator: Esin Öztürk Işık

Co-Investigator:  Bahattin Hakyemez

Researcher: Gökçe Hale Hatay

Project Summary: Phosphorus magnetic resonance spectroscopic imaging (31P-MRSI) is a non-invasive MR spectroscopic imaging technique that detects the phosphorus containing metabolites of the brain. 31P MRSI provides in-vivo quantitative information about the energy metabolism, oxygen state and pH within a given region of interest. Although, phosphorus magnetic resonance spectroscopic imaging provides vast information, it has not been widely used in the clinical settings yet. One of the major reasons of this problem is the low MR signal of phosphorus, because phosphorus is 15 times less abundant in the body than proton, and its gyromagnetic ratio is less than half of that of proton’s (1H=42.58 MHz/T, 31P=17.2 MHz/T). It is possible to average out multiple phosphorus signal acquisitions to get a higher signal to noise ratio (SNR), but this would result in longer scan times. Faster phosphorus MR spectroscopic imaging techniques should be devised to enable a wider use of 31P-MRSI. In this study, we aimed to implement compressed sensing technique for fast phosphorus magnetic resonance spectroscopic imaging.

Keywords: Phosphorus MR Spectroscopic Imaging, Compressed Sensing, Human Brain, 3T.

Project Title: Phosphorus MR Spectroscopic Imaging of Brain Tumors at 3T (31P_SPECTRA_3T) (click here for more information)

Funded by FP7 Marie Curie International Reintegration Grants  (IRG) (2010-2014) 

Principal Investigator: Esin Öztürk Işık

Project Summary:  The previous studies have shown phosphorus metabolite level differences between the brain tumors and healthy brain tissue. When the brain tissue gets ischemic, ATP production comes from hydrolysis of PCr catalyzed by creatine kinase leading to a reduction of PCr:-ATP ratio in 31P MRSI and the breakdown of glycogen to lactic acid which can be observed as the lactate peak in 1H spectra. Pi also increases at ischemia due to increased ATP hydrolysis that is not matched by ATP synthesis leading to a Pi/PCr increase.

Phosphorus MRSI has some major advantages over proton spectroscopic imaging. First, the phosphorus spectroscopy does not require any water suppression and it does not have any lipid contamination problems. Second, although the detection of lactate peak with 1H spectroscopy requires special spectral editing schemes due to the overlapping lipid resonances, 31P spectra can readily display the PCr, Pi, and ATP peaks that can give an assessment of the ischemic state. Tumor growth is associated with both an increased cell membrane synthesis and a higher degradation of cell membranes. Although both membrane synthesis and degradation are evaluated by observing the changes in a single Cho peak in 1H MR spectroscopy, these pathways can be differentiated using 31P MR spectroscopy.

Despite the advantages of 31P-MRSI, it has not been widely used for clinical applications at lower field strengths. The wider availability of high field scanners and multi-channel radiofrequency (RF) surface coils have increased the sensitivity and accuracy of MR imaging and spectroscopy of brain tumors through higher signal-to-noise ratio (SNR) and improved spectral resolution. Increased spectral dispersion at high field results in less overlap between different peaks and simplifies the appearance of the spectrum. Phosphorus MR spectroscopic imaging highly benefits from high field strength with a resulting increase in SNR, better definition of Pi peak, better separation of PC, PE, GPC and GPE peaks, and better estimation of pH values.

The goal of this project is to apply phosphorus magnetic resonance spectroscopic imaging accurately at high field 3T scanners to add new information regarding the characteristics of brain tumors and produce a new metric that estimates the aggressiveness of a brain tumor using 31P MRSI peak intensities.

Keywords: Phosphorus MR Spectroscopic Imaging, Human Brain, 3T