Look but don´t touch
Multiphysics simulation and non-contact measurements help keep eyes healthy
Researchers have been exploring for an accurate yet non-invasive method of predicting eye diseases, many of which cannot be remedied because patients are diagnosed when the disease is already at an advanced stage. One potential diagnostic method is to measure the ocular surface temperature (OST) and compare thermal variations in the eyeball to computer models as well as clinical data. An interdisciplinary team of physicians, researchers, and engineers in Singapore have embarked on such a project, and they have found that computer simulations using COMSOL Multiphysics in conjunction with artificial-intelligence systems offer a possible solution to predicting eye abnormalities at an early stage.
A key aspect is determining the interior temperature of the eye. Most recently, efforts for non-contact temperature measurements have focused on infrared (IR) imaging. And while IR gives temperatures only on the cornea surface, with the use of COMSOL Multiphysics researchers can create a model of the human eye and, using the principles of heat transfer and the bioheat equation, predict interior temperatures based on the surface temperature.
Research methodology
In the laboratory, research will follow along these lines: First, a body of knowledge relating to temperatures and diseases is established. The next step is to create computer simulations where the predicted thermal variations of the eyeball can be compared to known data. The simulated finite element results can be correlated and classified according to the measured corneal temperature to obtain their mathematical relationship within a large database of OST measurements.
The researchers hope that when this process is complete, a physician will be able to measure the OST of a patient and predict problems. One idea is then to send simulated temperature data and other physical measurements into an appropriate software-based classification system that can then predict possible diseases at earlier stages.
Clearly a key aspect of this project is creating an accurate mathematical model of the heat flow in the eyeball´s interior. Associate Professor Eddie Y.K. Ng from Singapore's Nanyang Technological University and associates have created such a model to simulate the organ's steady-state conditions based on the properties and parameters reported in the technical literature. They started with the geometry by measuring coordinates from a classic anatomical model of the human eye. They then worked with these coordinates in a CAD package and then imported this CAD geometry into COMSOL Multiphysics.
Left: A cut-away view of the generated mesh.
Middle: Temperature profile, from top-to-bottom, along the outer surface of the eye.
Right: A cut-away view of the temperature distribution on the different part of the eye. The streamlines show the heat flux.
In building the model, the scientists turned to the fact that the governing equation for heat flow in the eye is the Pennes bioheat equation.
The first boundary condition is defined at the sclera (the tough, opaque tissue that serves as the eye´s protective outer coating) with blood outside this boundary acting as a heat source. The second boundary condition is defined at the cornea, where three modes of cooling take place: heat loss through convection, radiation, and tear evaporation.
Finding the dominant parameters
The researchers next ran a sensitivity analysis to study the influence of various factors on the model to determine the dominant parameters. The researchers also looked at the relationship between blood temperature and the ocular temperature. At 39 ºC the corneal surface gives a temperature of 35.17 ºC, whereas at a normal body temperature of 37 ºC the corneal temperature is 33.64 ºC. These differences provide the possibility of detecting fever or sickness based on the ocular temperature.
The researchers compared results of the COMSOL Multiphysics simulation to experimental and computational results from studies done on human as well as animal eyes. Results show a discrepancy of only 0.33% when compared to images from IR screening.
Getting this level of understanding would have taken considerably more time without COMSOL Multiphysics reports Dr. Ng. "With this software, we were able to develop such a model in a short amount of time without worrying much about the details of the finite-element scheme. The package is very easy to use, especially for a few particular cases such as modeling the presence and effect of contact-lens wear and the coupling of three physics-conduction, conduction & convection, plus Navier-Stokes for fluid flow-yet still get fast, accurate results."
The Singapore Eye IR Working Group: (front row, left to right) Dr. Rajendra Achiarya U, Ngee Ann Polytechnic; Assoc Prof Dr. Eddie Y.K. Ng, Nanyang Technological Univ; Dr. Caroline Chee and Dr Manjunath Gupta, National University Hospital; (back row) research assistants E.H. Ool and J.H. Tan, Nanyang Techonogical Univ.