The beginning of this reporting period was mostly devoted to researching background information for my REVAMMAD project. Along with the literature review, my first step in the practical research included the creation of a data structure based on the object oriented programming technique.
Recently I started to create a new algorithm which at the first stage it is able to group the vascular segments present within a certain region, and in the future this algorithm will be able to connect these segments automatically.
The main goal of my project is to define alternative strategies that can prove to be more efficient and/or cost saving in the detection of diabetic retinopathy. The first step of this process was an extensive literature review in order to gain a comprehensive understanding of the nature of the disease, the risk factors that affect its progression and the different frameworks applied for detection. My knowledge was further reinforced by undertaking the level 3 City and Guilds accreditation for screeners and graders and by participating in respective conferences and seminars. Currently, I am trying to define the underlying parameters of the National Diabetic Eye Screening Programme and estimate the relative costs especially in the case of inappropriate decisions. This is the starting point of my practical implementation and the accurate definition of the framework of this model is crucial in the further assessment of alternative approaches.
During this reporting period, I published a paper at the 36th Annual International Conference of the IEEE Engineering in Medicine and Biology Society and gave an oral presentation regarding my work related to retinal image registration and super resolution.
Since then, I started working on a ED model of the eye, and a geometrical approach to the retinal image registration problem, to provide more accurate registrations. So far, the prototype for the method seems to be working fine, although is not complete.
Arranged and started collaborating with Sergio Crespo, ESR from REVAMMAD and the Charite University Hospital in Berlin.
During reporting period, model of the retinal vascular network was improved, network structure was changed fundamentally and made more realistic, mechanical properties of vessel wall were added, as well as autoregulation of vessel diameter. Program for 3D visualization of network structure and dynamics was written. In addition to modeling part of my work, were carried out preparing for animal experiments. I finished course “Laboratory Animal Science – FELASA category C”, that allow me to carry out the experiments on the laboratory animals. To get the rat retinal images we using Laser Speckle Contrast method, that is “full-field” and non invasive technique. Laser Speckle Imaging setup was build. Now, we are ready to obtain retinal images, made some tests, that we will use for improving of the model.
report_network Anastasiia Neganova October 2014
In the first six month working on the REVAMMAD project we developed an automatic algorithm to generate a 3D mesh of the retinal vessels starting from segmented images of retinal fundus from Lincoln University. We also devised an ad hoc fluid structure interaction model for modeling the interaction between blood and vessel wall into the retinal vasculature. We included in the model for the vessel wall the regulation of the blood flow rate. We are currently summarizing the results in view of their publication.
During this reporting period, the first version of the algorithm to segment corneal endothelial cells was developed. The corneal nerves segmentation algorithm was adapted and tested in mosaic images.
NUMBER AND LOCATION OF LESIONS FOR THE OPTIMIZATION OF THE EXAMINATION INTERVAL IN DIABETIC RETINOPATHYA recent study concluded that it is possible to construct a model for optimising the examination interval during screening for diabetic retinopathy in low-‐risk patients. However, the model fails to predict the interval for patients in whom the primary assessment recommended a short screening interval, suggesting that more risk factors should be identified and included. Two sets of fundus photographs, one where the result of the model and recommended interval are concordant and one where they are discordant, were analysed. The number of different lesions and their centre position (see fig. 1) with regards to the areas defined on a previous study (see fig. 2) were stored in an array of features associated with the relative fundus photo (see table 1). Considering the concordance and discordance on the interval as the two possible values of a dependent variable in a classification problem, an extraction of the best isolated features and a 10-‐fold cross validation were performed to assess their significance. The feature selection chose as the two best isolated features the number and the percentage of small haemorrhages in two different areas respectively. The 10-‐fold cross validation performed extracting the first two ranked features only resulted in the correct classification of 64% of the instances. With these two sets the ranking of the feature selection matched the clinicians’ experience while the percentage of correct classifications suggested that the model would benefit from the information held in the location and number of small haemorrhages showing on fundus photography.
Most of the work done during this period has been on researching background information for my REVAMMAD project. Side projects have included work in confocal microscopy related to the detection of features in different layers of the cornea. This work has led to the submission of two abstracts, one of which has already been accepted and will be presented as a poster at the ARVO conference. The other was submitted to the EMBC conference. Corneal nerves are similar to retinal vessels and many of the nerves metrics can also be calculated for vessels, such as tortuosity. This overlap will lead to much of the work being produced now, to be useful when the Retinal vessel work is undertaken.
To approach retinal vascular diseases it is necessary to adopt different strategies to understand the mechanisms of the pathology and transfer that knowledge into practical clinical tools. Vasculature in the retina changes under pathological conditions. However, vessels are not alone and we have to understand the disease as a complex puzzle where every piece has its role. Some other cell types besides the ones that compound vessels might be involved in terms of disease, as it is the case of neurons and microglia. Microglia belongs to the immune system at the retina, but its role in disease it is not clearly understood. Current models of imaging and algorithm analysis are pursuing early diagnosis regarding the dynamics and structure of the vasculature. In our basic medical research, we aim to switch the disease correlation into a new concept that involves vessels inside the neurovascular unit, where microglia might play a role, turning the current analysis model into one more complex and defined status.
My REVAMMAD project concerns developing analytical models for simulation of blood flow and vascular structural adaptation in retinal microcirculation under physiological and pathophysiological conditions based on anatomical data from the segmentation of imaging data. In order to gain comprehensive insight into relevant mechanisms underlying function of retinal microvascular network, some other tissues in addition to retina will also be employed for biological experiments. Those extra tissues are expected to supply an easier and comprehensive way for this research because some experimental methods cannot be applied to retina.Significant research during this reporting period:1. Comprehensive theoretical knowledge gained: Insight into morphology, topology, hemodynamics and structural adaptation of microvascular networks which lays theoretical foundation for practical modelling work;2. Hands-on modelling practice in progress: Learning of Delphi programming and apprehension of in-house modelling software on structural adaptation of microvascular network;3. Biological experiments started: Starting to establish the microvascular network occlusion model on CAM.
During last 6 months I reviewed a number of technologies that will be used in implementing the data warehouse and evaluation framework. Together with my supervisor, Dr Luca Antiga, we came up with a very exciting and revolutionary architecture of the data warehouse that will utilize state-of-the-art technologies. We addressed one of the biggest technological challenges of this project, namely how to efficiently store and execute computations alongside data. To accomplish this task we decided to use Docker containers (https://www.docker.io/) which are considered to be “the next big thing” by many experts in the industry that will revolutionize cloud computing. Moreover we came up with a flexible and reproducible way of extracting information from multiple and diverse datasets.
During this period the main goal was to study the previous and the ongoing research on the early detection of diabetic retinopathy and find all the experiments that have been done until now. At the same time I had to attend a few training events and collaboration meetings at the university of Lincoln with local hospitals as well as internal events. I received some images from the partner in Sunderland which were analyzed for finding some indications of changes in the vasculature before and after diabetic retinopathy. In the meantime, I am trying until now to connect and realize how these geometric changes might be connected with the hemodynamic changes that occur due to diabetes mellitus. Currently I am trying to find more images with specific profile (a few images before and a few after diabetic retinopathy with some information of patients’ history). At the university, we are planning to supply a fundus camera in order to recruit some volunteers for having them screened. This means that we will have images and history record of the volunteers, which will assist on the grouping of the results according to some risk factors like age, gender, duration of diabetes, other diseases etc
During this reporting period, I produced an image registration method with sub-pixel accuracy. I employed such method as the first step for already existing super resolution methods, to produce higher quality images from low resolution datasets and submitted a paper. I also started researching on the topic of 3D reconstruction of the eye model employing fundus images as a base.
A novel method for digital curvature estimation has been proposed during this period. A paper has been written and it will be submitted shortly to a journal. This particular area has been investigated since I am working on corneal nerves tortuosity, in collaboration with Harvard Medical School.The problem is to assess image tortuosity and curvature seems to be a good feature for describing tortuosity, even though they are different things. After reviewing the state-of-the-art on digital curvature estimation, I proposed a new solution to unsolved problems in this field: poor accuracy and window size selection.With our multiple-window approach we are now able to estimate digital curvature regardless of the corneal nerve’s size with high accuracy, according to huge amount of experiments on synthetic (ground truth) data.