End-user: Eugene and Marilyn Glick Eye Institute in Indianapolis (IN, USA), and the Eye Clinic of Lithuanian University of Health Sciences.
Thanks to its special connection to the brain and its accessibility to measurements, the eye provides a unique window on the brain, thereby offering non-invasive access to a large set of potential biomarkers that might help in the early diagnosis and clinical care of Neuro-Degenerative Diseases (NDD). However, characterizing ocular biomarkers as surrogates of cerebral or systemic vascular status is far from trivial. Clinical measurements are influenced by many factors that vary among individuals and cannot be isolated in vivo, thereby posing serious challenges for the interpretation of such measurements.
Motivated by the need of mathematical and computational methods to study the Eye-Brain system (which we refer to as Eye2Brain) and aid the interpretation of ocular measurements as biomarkers for the brain status, we are currently developing a multi-component platform combining detailed descriptions of the eye and the brain with a systemic view of the Eye2Brain connections.
The development of an articulated platform capable of providing physicians with an integrated view of the patient’s status will significantly improve our current ability to monitor health and to prevent, detect, treat and manage disease in a personalized manner. Within this project, we propose to develop such a platform for application in ophthalmology, with the specific goal of developing, testing and delivering a software that can be used in ophthalmology clinics to improve diagnosis and care of ocular diseases (e.g. glaucoma, diabetic retinopathy, age-related macular degeneration) and other pathologies that also manifest in the eye (e.g. diabetes, hypertension, NDD). This application clearly leverages the resources and expertise that we have gathered within the Eye2Brain project, and it extends them to build a tool that can directly impact the clinical practice.
The schematic below describes the types of data that we will integrate in the platform as well as the modeling components that we will use to connect the data within the eye.
Figure 8. The types of data to be integrated, and the modelling components to be used to connect the data within the eye.
Fundus camera images are processed to extract geometrical information on the retinal vasculature; computational techniques developed within the AngioTk (www.cemosis.fr/projects/angiotk/) platform which is adapted to generate computational domains for the blood vessels where we simulate blood flow using Feel++/CFD using the clinically measured values of blood pressure as patient-specific inputs (specifically, we will adapt the mathematical model described in Dziubek et al (2015)). Data from the Heidelberg Retinal Flowmeter will be used to properly tailor the microvascular levels of the model. Retinal Oximetry data will be used to incorporate oxygen dynamics into the vascular model, following a similar procedure as in Causin et al (2015). Color Doppler Imaging data is used to tailor the lumped model describing the blood flow in the central retinal artery and vein to the patient-specific measurements, following a similar procedure to that described in Guidoboni et al (2014). Images obtained via Optical Coherence Tomography is processed to extract geometrical information regarding the structure of the optic nerve head and is integrated within the rest of the ocular platform as a component to be visualized and explored in detail.
This application represents a challenge for integration into MSO4SC due to both the rich and possibly interconnected model components and data flow coming from different sources, which need to be exploited by the different model approaches.
Three approaches are proposed which provide a *Patient-specific virtual simulator of tissue perfusion in the lamina cribrosa* coupling 3D and 0D models but with increasing complexity in the 3D up to a full eye computational model and in parallel a decreasing in complexity of the 0D modeling. The models are fed by the ophthalmological instruments, as described above.
Test Case 1: Improper perfusion of the lamina cribrosa (LC) may lead to severe alterations of the visual function. LC perfusion parameters are difficult to estimate with non-invasive measurements and are affected by many factors that vary among individuals and cannot be easily isolated. Here we utilize a mathematical virtual simulator (MVS) to address these challenges.
The MVS combines i) a three-dimensional porous-media model for LC perfusion with ii) a circuit-based model for blood flow in the retrobulbar and ocular posterior segments. Systems i) and ii) are solved using advanced computational and visualization methods (Feel++). Simulation inputs may include some patient-specific factors that can be measured non-invasively, e.g. systolic blood pressure (SBP), diastolic blood pressure (DBP), intraocular pressure (IOP) and ocular geometry.
Figure 9 shows the MVS visualization of ocular geometry. MVS simulated LC perfusion velocities at time t = 2.4 s (green line) are shown for IOP = 15 mmHg and SDB/DBP = 100/70 mmHg, SDB/DBP = 120/80 mmHg, SDB/DBP = 140/90 mmHg. Increasing SBP and DBP leads to higher LC perfusion velocities, especially near the nasal area. MVS also simulates blood velocities in the central retinal artery and vein (CRA and CRV), which are similar to those obtained via Color Doppler Imaging. Thus MVS can serve as an instrument to visualize and estimate LC perfusion parameters, thereby providing new means to address the increasing demand of information on parts of the eye that are not-easily accessible with standard investigations.
Figure 9. Right: MVS multiscale scheme. Right: MVS perfusion simulations.
Figure 10. Left: 2D view of the eye and its components. Right: 3D view of a section of the computational geometry of the eye used for the pilot simulations.
Test case 2: In Test case 1,we considered a 3D model of the LC and a 0D systemic model of the retinal fluid. In this test case, we complexify the model by coupling the 3D poro-viscoelastic model of the LC with the 3D biomechanical behavior of the Sclera, the Choroid and the Retina, see Figure 10. This enables a more refined modeling of tissue perfusion of the lamina cribrosa. This enables (i) a more refined modeling of tissue perfusion in the lamina cribrosa, choroid and retina, and (ii) the integration of clinical data coming from fundus camera images, heidelberg retinal flowmeter, ocular coherence tomography and retinal oximetry.
Test case 3: This last test case is one step further the last one by adding 3D fluid models for aqueous and vitreous humours, see Figure 10 above. We have a full 3D model of the eye completed by a systemic 0D model ensuring physiological behavior of our model. This enables (i) a more complete modeling of ocular biophysics, and (ii) a virtual laboratory where the efficacy of various therapeutic approaches, including topical medications and surgical interventions, can be tested accounting for patient specific conditions.