Gizem Dilara Özdemir | Harvard Medical School (original) (raw)

Papers by Gizem Dilara Özdemir

Machine Learning: Science and Technology

2022 Medical Technologies Congress (TIPTEKNO)

Plasma is the fourth state of matter, and it is species of partially ionized gas generated under ... more Plasma is the fourth state of matter, and it is species of partially ionized gas generated under an electric field that contains photons, free electrons, ions, free radicals, and reactive oxygen/nitrogen species. Plasma can be produced at atmospheric pressure or under a vacuum in two ways; thermal and nonthermal. Furthermore, they can be classified into natural and artificial plasmas. Non-thermal atmospheric plasma, also known as cold atmospheric plasma (CAP), is produced in a cold form under a high electrical field at atmospheric pressure. CAP is primarily produced using two methods which are the dielectric barrier discharge (DBD) and plasma jet. The electrical discharge between two electrodes separated by an insulating dielectric barrier is known as DBD. This method is quite widely used in many studies in the literature and has an important place in the field of plasma medicine. In the DBD method, the electrode configuration, shape, material, and substance from which the dielectric barrier is made are important. There are many studies conducted with different electrode configurations in the literature. Besides the electrode configuration and shape, the barrier and electrode materials can also affect the reactivity of the discharge by changing the discharge electrical power. It is thought that plasma discharge at different times will vary in CAP applications due to the change in conductivity of the conductive material used depending on the capacitive resistance. In this study, deionized water (DIW) activated with CAP treatment using different electrode materials (copper, stainless steel, and aluminum) to compare the physical quantities that can change such as pH and conductivity. The aim of this study is to observe the effect of using different electrode materials (copper, stainless steel, and aluminum) on the biological outcome of CAP treatment and compare the antimicrobial activities of different materials.

2022 Medical Technologies Congress (TIPTEKNO), Oct 31, 2022

Cold atmospheric plasma (CAP) has various therapeutic effects that have been broadly reported in ... more Cold atmospheric plasma (CAP) has various therapeutic effects that have been broadly reported in the literature. Furthermore, the biomedical activity of CAP is primarily attributed to various reactive oxygen species (ROS). Beside the direct treatment, plasma-activated liquids (PALs) may gain similar therapeutic effects as previously reported in the literature. However, plasma parameters that are needed to activate a liquid for desired therapeutic effect cannot be compared due to the diversity in between the plasma sources and electrode geometries from different research groups. Therefore, a standard method has not been achieved. Plasma-generated species are mostly oxidizing agents and the determination of oxidative strength of PALs by different methods and correlation of results from those different methods may assist to fill the gap in between required plasma parameters and desired therapeutic effect of PALs. This study aims to estimate the oxidative strength of plasma-activated water (PAW) with respect to plasma treatment time using paperbased colorimetric sensors and machine learning (ML) methods. Colorimetric detection caused by reactive species was performed using starch+potassium iodide (KI) and tetramethylbenzidine (TMB)+KI. The color change caused by PAW was evaluated with different ML algorithms. Fine Tree Classifier (FTC) yielded 85.9% and 93.7% accuracy for the starch+KI and TMB+KI solutions, respectively. Our results demonstrated the capability of ML algorithms for the prediction of the oxidative strength of PALs. Considering the wide usage of ML, this pilot study may provide a basis to understand the underlying biological outcomes of PALs.

2022 Medical Technologies Congress (TIPTEKNO)

Disruption of the epithelial tissue's integrity leads to wounds. Damaged tissue raises th... more Disruption of the epithelial tissue's integrity leads to wounds. Damaged tissue raises the possibility of infection, which has a negative impact on many aspects of life. Recent interest has focused on several novel approaches to improve quality of life, including photobiomodulation (PBM) research that emphasizes wound modeling. The contribution of the PBM method to the wound healing process is examined by in vitro studies. The size of the area recovered from the microscopic examination images is the primary success criterion in wound healing research, which investigates the effectiveness of various parameters such as the laser wavelength, power, and exposure duration. Therefore, segmentation is a crucial step in analyzing obtained images and has a significant role in conducting accurate analysis. In this study, a U-net structure-based deep learning (DL) approach was presented for accurately segmenting microscopic wound healing images from PBM studies. The success of the developed DL model was evaluated with various performance metrics and compared with ground truth labels, which were manually determined by a blind expert. Most of the performance metrics utilized had success rates of over 90%. The average dice similarity coefficient (DSC) between ground truth labels and the DL model's prediction was obtained as 0.953 and 0.939 for the validation and test image sets, respectively.

Background: Infectious diseases not only cause severe health issues but also burden the healthcar... more Background: Infectious diseases not only cause severe health issues but also burden the healthcare system. Therefore, the effective treatment of those diseases is crucial. A variety of conventional therapies including antibiotics as well as novel treatment methods namely antimicrobial peptides (AMP) are utilized for the treatment of infections. However, because of the drawbacks of current therapies such as the risk of the emergence of drug-resistant microorganisms, low stability of the agent in its environment as well as toxicity problems, new solutions are still being investigated. A recent approach is the usage of these two different antimicrobial agents in combination. Nonetheless, the determination of synergism is time-consuming and depends on several experimental studies. Prediction of different biological outcomes with machine learning (ML) algorithms is a widespread research field recently, and AMP with ML studies is one of the most researched areas further to understand the ...

2019 Medical Technologies Congress (TIPTEKNO)

Squamous cell carcinoma (SCC) is the second most common skin cancer among the white race. Plasma ... more Squamous cell carcinoma (SCC) is the second most common skin cancer among the white race. Plasma is an ionized gaseous state of matter containing chemically active species, such as ions, electrons, photons, reactive oxygen and nitrogen species, and UV light. Cold atmospheric plasma (CAP) has just recently been showing promising anti-cancer activities supported by the ability to induce cell death via apoptosis and cell cycle arrest leading to tumor cell destruction in vitro and in vivo. In this study, two different plasma treatment methods, which are direct plasma treatment and fluid-mediated plasma treatment, apply on SCC and keratinocytes cell lines to determine lethal dose. Also, apoptotic behaviors of two cell types are evaluated with TiterTACS™ apoptosis detection kit. For direct plasma treatment, 60 seconds exposure to CAP found as optimum time and, for fluid-mediated plasma treatment 15 minutes holding of 30 seconds CAP exposure N-Acetyl Cysteine (NAC) solution found as optimum treatment time. Results show that CAP can selectively inactivate SCC cell line through apoptosis while no damage or apoptotic behavior observing in keratinocyte cell line.

2019 Medical Technologies Congress (TIPTEKNO)

Counting of microbial colonies is crucial due to the applications of medical microbiology to sear... more Counting of microbial colonies is crucial due to the applications of medical microbiology to search and detect the causes of diseases. While different tasks performed, the counting process of bacteria colonies is provided either by the searcher manually or by a common software, nowadays. The manual counting of bacteria colonies is tiresome, eye-straining, and time-consuming for the searcher where common softwares require high troublesome with having high error rates. The aim of this study is detecting and counting bacteria colonies without having these limitations in today's non-practical applications. Therefore, an image-processing based bacteria colony counter designed in MATLAB. In the medical plasma laboratory of the Izmir Katip Celebi University three different types of hospital-acquired infection cause bacterias, which are Escherichia coli, Pseudomonas aeruginosa, and Enterococcus faecalis, cultured and examined properly, then, using the Circular Hough Transform (CHT) in MATLAB the detection and counting of bacteria colonies provided. To be able to obtain more practical usage, a Graphical User Interface (GUI) designed.

2019 Medical Technologies Congress (TIPTEKNO), 2019

The aim of the present study is to determine the long term effects of the Cold Atmospheric Plasma... more The aim of the present study is to determine the long term effects of the Cold Atmospheric Plasma (CAP) treatment on the Grade 5 Titanium(Ti) implant surface at different storage conditions. After CAP application, Ti discs were stored in room conditions, in saline and an inert ambient, then contact angle and surface roughness measurements were done. Discs were stored using accelerated aging test for 1 month, 3 months, 6 months, 1 year and 2 years and measurements were made at the end of these time points. Optical Emission Spectra (OES) measurement was performed to determine plasma generate species during CAP application. OES shows that OH and NO generated. As a result of contact angle measurements, showed that the samples stored in saline were capacity of preserving hydrophilicity for an extended period of time compared to the samples stored in room condition, inert ambient. Surface roughness measurements with a profilometer showed no difference in surface roughness compared to the control group (untreated Ti) in plasma-treated groups.

2018 Medical Technologies National Congress (TIPTEKNO), 2018

Plasma is defined as the fourth state of matter and can be produced at atmospheric pressure, unde... more Plasma is defined as the fourth state of matter and can be produced at atmospheric pressure, under high electric field and at room temperature (cold). In the present study, the conversion of the sound data into electrical signals obtained during the generation of plasma discharge and the processing of these signals with different algorithms are aimed to evaluate the sound signal variation on different materials and different frequencies according to the plasma sound characteristics. It is also aimed to determine whether there is a correlation between the sound changes during operation of the plasma device and the electrical conductivity.

BMC Medical Informatics and Decision Making, 2021

Background Coronavirus disease 2019 (COVID-19) has become a pandemic since its first appearance i... more Background Coronavirus disease 2019 (COVID-19) has become a pandemic since its first appearance in late 2019. Deaths caused by COVID-19 are still increasing day by day and early diagnosis has become crucial. Since current diagnostic methods have many disadvantages, new investigations are needed to improve the performance of diagnosis. Methods A novel method is proposed to automatically diagnose COVID-19 by using Electrocardiogram (ECG) data with deep learning for the first time. Moreover, a new and effective method called hexaxial feature mapping is proposed to represent 12-lead ECG to 2D colorful images. Gray-Level Co-Occurrence Matrix (GLCM) method is used to extract features and generate hexaxial mapping images. These generated images are then fed into a new Convolutional Neural Network (CNN) architecture to diagnose COVID-19. Results Two different classification scenarios are conducted on a publicly available paper-based ECG image dataset to reveal the diagnostic capability and ...

Scientific Reports, 2022

Accurate and timely diagnosis of appendicitis in children can be challenging, which leads to dela... more Accurate and timely diagnosis of appendicitis in children can be challenging, which leads to delayed admittance or misdiagnosis that may cause perforation. Surgical management involves the elimination of the focus (appendectomy) and the reduction of the contamination with peritoneal irrigation to prevent sepsis. However, the validity of conventional irrigation methods is being debated, and novel methods are needed. In the present study, the use of cold plasma treated saline solution as an intraperitoneal irrigation solution for the management of acute peritonitis was investigated. Chemical and in vitro microbiological assessments of the plasma-treated solution were performed to determine the appropriate plasma treatment time to be used in in-vivo experiments. To induce acute peritonitis in rats, the cecal ligation and perforation (CLP) model was used. Sixty rats were divided into six groups, namely, sham operation, plasma irrigation, CLP, dry cleaning after CLP, saline irrigation af...

Plasma is defined as the fourth state of matter, and non-thermal plasma can be produced at atmosp... more Plasma is defined as the fourth state of matter, and non-thermal plasma can be produced at atmospheric pressure under a high electrical field. The strong and broad-spectrum antimicrobial effect of plasma-activated liquids (PALs) is now well known. The antimicrobial effects of PALs depend on many different variables, which complicates the comparison of different studies and determining the most dominant parameters for the antimicrobial effect. The proven applicability of machine learning (ML) in the medical field is encouraging for its application in the field of plasma medicine as well. Thus, ML applications on PALs could present a new perspective to better understand the influences of various parameters on their antimicrobial effects. In this paper, comparative supervised ML models are presented by using previously obtained data to predict the in vitro antimicrobial activity of PALs. A comprehensive literature search was performed, and 12 distinct features related to PAL-microorganism interactions were collected from 33 relevant articles to automatically predict the antimicrobial activity of PALs. After the required normalization, feature encoding, and resampling steps, two supervised ML methods, namely classification and regression, are applied to the data to obtain microbial inactivation (MI) predictions. For classification, MI is labeled in four categories, and for regression, MI is used as a continuous variable. Sixteen different classifiers and 14 regressors are implemented to predict the MI value. Two different robust cross-validation strategies are conducted for classification and regression models to evaluate the proposed method: repeated stratified k-fold cross-validation and k-fold cross-validation, respectively. We also investigate the effect of different features on models. The results demonstrated that the hyperparameter-optimized Random Forest Classifier (oRFC) and Random Forest Regressor (oRFR) provided superior performance compared to other models for classification and regression. Finally, the best test accuracy of 82.68% for oRFC and R2 of 0.75 for the oRFR are obtained. Furthermore, the determined most important features of predictive models are in line with the outcomes of PALs reported in the literature. An ML framework can accurately predict the antimicrobial activity of PALs without the need for any experimental studies. To the best of our knowledge, this is the first study that investigates the antimicrobial efficacy of PALs with ML. Furthermore, ML techniques could contribute to a better understanding of plasma parameters that have a dominant role in the desired antimicrobial effect. Moreover, such findings may contribute to the definition of a plasma dose in the future.

Disruption of the epithelial tissue's integrity leads to wounds. Damaged tissue raises the possib... more Disruption of the epithelial tissue's integrity leads to wounds. Damaged tissue raises the possibility of infection, which has a negative impact on many aspects of life. Recent interest has focused on several novel approaches to improve quality of life, including photobiomodulation (PBM) research that emphasizes wound modeling. The contribution of the PBM method to the wound healing process is examined by in vitro studies. The size of the area recovered from the microscopic examination images is the primary success criterion in wound healing research, which investigates the effectiveness of various parameters such as the laser wavelength, power, and exposure duration. Therefore, segmentation is a crucial step in analyzing obtained images and has a significant role in conducting accurate analysis. In this study, a U-net structure-based deep learning (DL) approach was presented for accurately segmenting microscopic wound healing images from PBM studies. The success of the developed DL model was evaluated with various performance metrics and compared with ground truth labels, which were manually determined by a blind expert. Most of the performance metrics utilized had success rates of over 90%. The average dice similarity coefficient (DSC) between ground truth labels and the DL model's prediction was obtained as 0.953 and 0.939 for the validation and test image sets, respectively.

Urinary catheters are used to empty the bladder. However, the use of urinary catheters causes bac... more Urinary catheters are used to empty the bladder. However, the use of urinary catheters causes bacterial colo- nization on the surface and urinary tract infections. Catheter associated urinary tract infections (CAUTI) are one of the most common health problems nowadays. A variety of catheter materials and surface modification methods have been tried to prevent infections, but these methods have failed to achieve the expected success. For this reason, new methods are needed. The use of antimicrobial peptides (AMP) has become popular in recent years for reasons such as having antimicrobial effect on multiple pathogens and not causing antimicrobial resistance. In this study, it is aimed to investigate the efficacy of AMPs on pathogens in the case when they are conjugated to catheter surfaces. Furthermore, it is believed that the cold atmospheric plasma (CAP) treatment to catheter surfaces before peptide conjugation will increase the amount of peptide attached to the surface, and the antimicrobial efficacy will increase further. By surface modification with CAP treatment and conjugation of AMPs to catheter surfaces, it is believed that significant progress will be made in the prevention and treatment of CAUTI. Results revealed that the LfcinB (21−25)P al peptide suppresses bacterial growth for a certain period of time, plasma treatment increases the hydrophilicity of the surface, and thus there is an increase in the amount of peptide conjugated to the surface. Future studies will focus on the evaluation of antibiofilm activity of the peptide.

Machine learning (ML) is an artificial intelligence (AI) technique that makes predictions by obta... more Machine learning (ML) is an artificial intelligence (AI) technique that makes predictions by obtaining inferences from data using mathematical and statistical operations. ML algorithms are used to identify patterns in data. These patterns are also used to create a predictive data model [1]. ML is a great option in scenarios where classical statistical methods are fall short in analyzing an enormous amount of data due to its adaptability. Therefore, ML plays a vital role in a wide variety of research areas. ML models are widely used in the medical and biomedical fields [2]. The proven applicability of ML in the medical field is encouraging for its application in the field of plasma medicine as well. The strong and broad-spectrum antimicrobial effect of plasma-activated liquids (PALs) is now well known [3]. The antimicrobial effects of PALs depend on many different variables, which complicates the comparison of different studies and to determine the most dominant parameters on antimicrobial effect. Thus, ML applications on PALs could present a new perspective to better understand the influences of various parameters on their antimicrobial effects.
This study aims to develop a ML model using pre-obtained data to qualitatively predict in vitro antimicrobial activity of PALs. A literature search was performed on PubMed and Scopus databases and relevant 32 publications were obtained. By the analysis of these publications, 12 different features (plasma type, gas, discharge gap, liquid, treatment volume, treatment time, microbial strain, initial microbial load, PAL/microorganism volume ratio, exposure time, post storage time, and storage temperature of PALs) were gathered to be used in the ML algorithm as input parameters. Beside these features, microbial inactivation was categorized as complete, strong, weak, and none to be used as output labels. The model was developed using the Bagged Classification Tree algorithm [4] based on the ensemble learning method. A multiclass classification strategy was conducted as outcomes have four classes. This method consists of several decision trees for better classification performance and to prevent large bias and variance. The idea is to create several different subsets of data by constantly changing the randomly selected training sample. Each subset is used to train these trees. Hence, the average of all predictions from different trees that are more robust than a single decision tree is used. By training the ML model, a validation accuracy of 72.4% was yielded using a 5-fold cross-validation strategy. The results demonstrated that such predictive models could provide an insight to better understand the plasma parameters to lead a desired antimicrobial effect. Furthermore, such findings may also contribute to define a plasma dose.

Antimicrobial effect of cold atmospheric plasma (CAP) and CAP treated liquids is broad-spectrum a... more Antimicrobial effect of cold atmospheric plasma (CAP) and CAP treated liquids is broad-spectrum and well-known in the literature. The antimicrobial effect is primarily attributed to various reactive oxygen species (ROS) [1]. However, due to diversity in between the plasma sources and electrode geometries from different research groups, plasma parameters that are needed to activate a liquid for antimicrobial effect cannot be compared. Furthermore, the utilization of similar plasma parameters to different plasma generation systems may even lead to differences between plasma activated liquids (PALs) obtained by these devices [2]. Therefore, a standard method for the prediction of required plasma parameters for a desired antimicrobial effect of PALs from different plasma sources has not been achieved. Plasma-generated species are mostly oxidizing agents and the determination of oxidative strength of PALs by different methods and correlation of results from those different methods may assist to fill the gap in between required plasma parameters and desired antimicrobial effect of PALs.
This study aims to measure oxidizing strength of different CAP solutions with respect to plasma treatment time using colorimetric sensors and electrochemical methods. Deionized water (DIW) was treated with an AC microsecond pulsed air dielectric barrier discharge (DBD) plasma for 0, 15, 30, 45, 60, 90, 120, 150, 180, 240 and 300 seconds. pH and conductivity measurements were performed on PALs. Then, a paper-based microfluidic device (μPAD) was fabricated as a colorimetric sensor and 3.3’,5,5’-tetramethylbenzidine (TMB) + potassium iodide (KI) and KI + starch solutions were loaded on paper-based sensor for the detection of plasma generated reactive species. Using μPADs, TMB + KI and KI + starch solutions, color change in the detection zone could be easily determined without complex equipment. KI + starch is a well-known method for ROS detection in CAP [3]. TMB gives blue color after oxidation [4] and is used in the study to make the color change more detectable. Furthermore, different concentrations of ascorbic acid were added to the PALs as a scavenger of reactive species in PALs to avoid the color saturation from PALs that are treated for longer time to increase the resolution in between different plasma treatment time points. Colorimetric measurements have shown a correlation in between plasma treatment time and color intensity. Besides, cyclic voltammetry (CV) analysis was performed to observe oxidizing strength of CAP activated DIW. CV analysis results show that oxidizing strength of PALs increases with plasma treatment time. Furthermore, plasma treatment time dependent antimicrobial strength of PALs was in line and consistent with findings of colorimetric and electrochemical measurements. Further image analysis studies for colorimetric measurements and machine learning studies for detailed correlation of findings are underway.

Background: Coronavirus disease 2019 (COVID-19) has become a pandemic since its first appearance ... more Background: Coronavirus disease 2019 (COVID-19) has become a pandemic since its first appearance in late 2019. Deaths caused by COVID-19 are still increasing day by day and early diagnosis has become crucial. Since current diagnostic methods have many disadvantages, new investigations are needed to improve the performance of diagnosis. Methods: A novel method is proposed to automatically diagnose COVID-19 by using Electrocardiogram (ECG) data with deep learning for the first time. Moreover, a new and effective method called hexaxial feature mapping is proposed to represent 12-lead ECG to 2D colorful images. Gray-Level Co-Occurrence Matrix (GLCM) method is used to extract features and generate hexaxial mapping images. These generated images are then fed into a new Convolutional Neural Network (CNN) architecture to diagnose COVID-19. Results: Two different classification scenarios are conducted on a publicly available paper-based ECG image dataset to reveal the diagnostic capability and performance of the proposed approach. In the first scenario, ECG data labeled as COVID-19 and No-Findings (normal) are classified to evaluate COVID-19 classification ability. According to results, the proposed approach provides encouraging COVID-19 detection performance with an accuracy of 96.20% and F1-Score of 96.30%. In the second scenario, ECG data labeled as Negative (normal, abnormal, and myocardial infarction) and Positive (COVID-19) are classified to evaluate COVID-19 diagnostic ability. The experimental results demonstrated that the proposed approach provides satisfactory COVID-19 prediction performance with an accuracy of 93.00% and F1-Score of 93.20%. Furthermore, different experimental studies are conducted to evaluate the robustness of the proposed approach. Conclusion: Automatic detection of cardiovascular changes caused by COVID-19 can be possible with a deep learning framework through ECG data. This not only proves the presence of cardiovascular changes caused by COVID-19 but also reveals that ECG can potentially be used in the diagnosis of COVID-19. We believe the proposed study may provide a crucial decision-making system for healthcare professionals.

Machine Learning: Science and Technology

2022 Medical Technologies Congress (TIPTEKNO)

2022 Medical Technologies Congress (TIPTEKNO), Oct 31, 2022

2022 Medical Technologies Congress (TIPTEKNO)

2019 Medical Technologies Congress (TIPTEKNO)

2019 Medical Technologies Congress (TIPTEKNO), 2019

2018 Medical Technologies National Congress (TIPTEKNO), 2018

BMC Medical Informatics and Decision Making, 2021

Scientific Reports, 2022