Article Preview
TopIntroduction
According to the World Health Organization (WHO), breast cancer is the most common type of cancer among women, affecting 2.1 million women worldwide, each year (WHO, 2019). Breast cancer is also responsible for the highest number of cancer-related deaths among women. In 2018, it is estimated that 627,000 women died from breast cancer. This number represents 15% of all women who died from cancer (WHO, 2019).
The best strategy to reduce the disease morbidity and mortality rate is early detection (Walker & Kaczor, 2012). Discovering the tumor in its early stages increases the patient's chances of healing.
There are two strategies for early detection of breast cancer: early diagnosis and screening. Early diagnosis consists of identifying breast cancer in symptomatic individuals as early as possible, whereas screening is the identification of breast cancer in asymptomatic individuals (WHO, 2019). As a strategy, screening adopts the execution of relatively simple tests in healthy people, aiming at identifying cancer in its preclinical (asymptomatic) phase.
The best-defined screening test for accuracy, cost, access, and risk is mammography (Walker & Kaczor, 2012). However, this technique has low sensitivity (especially for dense breast patients), a high rate of false positives and the risks of exposing the patient to ionizing radiation, especially in young patients (who have essentially glandular tissue which is sensitive to radiation). Moreover, the mammography is performed by compressing the breast, causing discomfort to the patient (Walker & Kaczor, 2012).
Besides the mammography, other techniques such as ultrasound, nuclear magnetic resonance, scintigraphy, thermography, and electrical impedance tomography may be used in screening as complementary tools in the diagnosis of breast cancer (Walker & Kaczor, 2012).
One technique that has been gaining prominence in recent decades is Electrical Impedance Tomography (EIT), which is an imaging tool that uses the electrical properties of a medium to form its tomographic images. The examination is performed by positioning the electrodes circumferentially around the region of interest (such as the chest or breasts). Then, an excitation current pattern is applied through the electrodes and the distribution of electrical potentials is measured. These measurements are used to reconstruct an EIT image, which represents a map of the conductivity distribution of the medium (Tehrani, Jin, McEwan, Schaik, 2010).
EIT is free of ionizing radiation (Cheney, Isaacson & Newell, 1999) which makes it a promising technique. It also has a cheaper and smaller equipment when compared to other imaging techniques, such as magnetic resonance imaging and computed tomography. However, it is still a recent procedure and is not strongly established as a tool for medical diagnosis. Among EIT disadvantages are low resolution images and high reconstruction time when compared to other tomography techniques that are commonly used (Tehrani et al., 2010; Kumar, Sriraam, Benakop, Jinaga, 2010; Barbosa et al., 2017). Therefore, the above mentioned factors, strongly affects the technique's reliability for medical analysis. In addition, its application in breast imaging and breast cancer detection is still poorly investigated (Cherepenin et al., 2001; Zou & GUO, 2003; Holder, 2004; Choi, Kao, Isaacson, Saulnier, & Newell, 2007; Halter, Hartov, Paulsen, 2008).
In order to solve this problem, new reconstruction methods have been studied. One of the approaches is through an optimization problem, wherein the objective function is an error function between the measured distribution and a candidate image for solution (Barbosa et al., 2017).
In this paper we propose a reconstruction method for EIT images based in Fish School Search (Filho, Lima, Lins, Nascimento, & Lima, 2008; Madeiro, Lima- Neto, Bastos-Filho, Figueiredo, 2011) and Non-Blind Search (Saha & Bandyopadhyay, 2008). We also compare our finds with reconstructed images obtained by using Genetic Algorithm and Particle Swarm Optimization. As objective function we used the relative quadratic error between the measured potentials and the potential from an artificial image candidate.