|Main Objective: Cataract is a clouding or opacity that develops in the crystalline lens of the eye. For clinical prognostic and therapeutic purposes it is very important to identify the cataract type and the disease stages especially in its early stages (Paunk 2007). With increasing relevancy, the ultrasounds appear as powerful, noninvasive and less expensive technique for tissue evaluation. Concerning to cataract treatment, phacoemulsification is the most common surgical procedure by using a hand-held ultrasonic device to fragment the clouded lens into small pieces that are then aspirated (Wilson 2005). To emulsify the cataract, the selection of the optimal ultrasound energy level is of fundamental importance to prevent the endothelial cells damage and the posterior capsule rupture. This unrecovered injury can be avoided if the hardness of the cataractous lens is previously and correctly estimated.
The applicability of the ultrasound methods for characterizing both normal and cataractous lenses has been studied for decades. Two important ultrasound parameters, velocity and frequency- dependent attenuation have been used trying to evaluate the hardness of the cataractous lens in porcine and human eye tissue. The most important one is related to the fact the results only provide information about the cataract presence and respective hardness globally, because evaluation is made considering the anterior and posterior reflected lens signals. Thus, important information as the exact regional hardness of cataract is not achieved with this methodology. Also, there is no adequate noninvasive characterization of regional location and hardness of cataract. This is of particular importance for estimating the optimal phacoemulsification energy, which should be adjusted for the different hardness levels that can be found in all extension of cataract. In this project it is intended to use the backscattering signals, instead of the anterior and posterior lens echo signals frequently used in literature, to extract features capable to characterize the mechanical properties of tissues, namely scatter density and mean scatter size for subsequent correlation with the lens hardness. Time and frequency domain features will be evaluated in order to identify the most powerful ones. Several feature classifiers such as Artificial Neural Networks, Support Vector Machines, k-Nearest neighbors, and Naive Bayes will be tested and their potential evaluated for quantification of regional changing of cataract hardness.
To develop a medical device for cataract surgery and clinicians support to access cataract hardness and location for any type of cataract, and determine in real-time the optimal energy for phacoemulsification, pre-clinical studies will be performed in vitro, in porcine lens, and in vivo in rats animal models that simulate different types of cataract. The animal models will allow the translation from pre-clinical to clinical research.
|Start Date: 01-06-2013|
|End Date: 01-06-2015|
|Team: Jaime Batista dos Santos, Marco Alexandre Cravo Gomes, Fernando Manuel Santos Perdigão|
|Groups: Multimedia Signal Processing – Co|
|Partners: University of Coimbra - CEMUC: Santos JB (Inv. Responsável) António Miguel Lino Santos Morgado Jose Silvestre Silva Miguel Caixinha Mário João Simões Ferreira dos Santos Rufino Martins da Silva|
|Local Coordinator: Marco Alexandre Cravo Gomes|