1.2. Bioimagen markers in breast cancer detection

1. Introduction

1.1. Breast cancer and early detection

162 Advanced Applications for Artificial Neural Networks

cases of this disease are expected to occur until 2030 [2].

through the lymphatic systems or blood streams [3].

its potential harm, is divided into benign and malignant types.

for patient will increase to more than 97% [10].

classifying the malignant and benign cases.

Nowadays, cancer is a massive public health problem around the world. According to the International Agency for Research on Cancer (IARC) [1], part of the World Health Organization (WHO), there were 8.2 million deaths caused by cancer in 2012 and 27 million of new

Cancer, medically defined as a malignant neoplasm, is a board group of disease involving unregulated cell growth [2]. In cancer, cell divides and grows uncontrollably forming malignant tumors, invading nearby parts of the body. The cancer can spread to all parts of the body

Cancer can be diagnosed by classifying tumors in two different types such as malignant and benign. Benign tumors represent an unnatural outgrowth but rarely lead to a patient's death; yet, some types of benign tumors, too, can increase the possibility of developing cancer [4]. On the other hand, malignant tumors are more serious and their timely diagnosis contributes to a successful treatment. As a result, predication and diagnosis of cancer can boost the chances of treatment, decreasing the usually high costs of medical procedures for such patients [5].

Breast cancer (BC) is the most commonly diagnosed cancer and the most common cause of death in women all over the world. Among the cancer types, BC is the second most common cancer for women, excluding skin cancer [6]. Besides, the mortality of BC is very high when compared to other types of cancer [7]. BC, similar to other cancers, starts with a rapid and uncontrolled outgrowth and multiplication of a part of the breast tissue, which depending on

Generally, there are two types of BC that are in situ and invasive. In situ starts in the milk duct and does not spread to other organs even if it grows [8]. Invasive breast cancer on the contrary, is very aggressive and spreads to other nearby organs, and destroys them as well [9]. It is very important to detect the cancerous cell before it spreads to other organs; thus, the survival rate

A major class of problems in medical science involves the diagnosis of disease, based upon various tests performed upon the patient. The evaluation of data taken from patients and decisions of experts are the most important factors in diagnosis. The correct diagnosis of BC is one of the major problems in the medical field. As BC can be very aggressive, only early detection can prevent mortality. Clinical diagnosis of BC helps in predicting the malignant cases and timely diagnosis can increase the chances of a patient's life expectancy from 56 to 86% [11]. BC has four early signs: microcalcification, mass, architectural distortion, and breast asymmetries [12]. The various common methods used for breast cancer diagnosis (BCD) are positron emission tomography (PET), magnetic resonance imaging (MRI), CT scan, X-ray, ultrasound, photoacoustic imaging, tomography, diffuse optical tomography, elastography, electrical impedance tomography, opto-acoustic imaging, ophthalmology, mammogram, etc. [13]. The results obtained from these methods are used to recognize the patterns, which are aiming to help the doctors for Bioimagen markers allow to characterize and to study different diseases using some kind of information, such as genetic, histological, clinical imaging, etc. These biomarkers can be used to detect abnormalities in the data as genetic mutations that cause some diseases and can also be used in the clinic for the detection of patients with some types of disease [17].

The application of quantification of bioimaging markers to aid in the diagnosis, treatment, and follow-up of pathologies provides added value throughout the clinical practice process by providing additional information to conventional diagnostic tests [18]. From imaging tests processed in the right way, abnormalities in a tissue are evidenced before they are perceptible in the reading of the radiologist, fundamental objective of this type of biomarkers [19]. In addition, they allow the monitoring of the treatment effect from a quantitative point of view.

As mentioned before, BC is one of the leading causes of death in women around the world, accounting for nearly one-third of cancer-related deaths. Currently, the clinical screening by mammography is the most effective way for the early detection of this disease. Using analysis of mammograms obtained through X-rays allows radiologists to visualize early signs of cancer, such as calcifications, masses, and architectural distortions among other early signs of cancer. However, this analysis is a routine, monotonous, and exhausting task and it is estimated that only 0.3–0.4% of the cases are actually carcinogenic [20].

It has been shown that because of these problems and other factors intrinsic to cancer such as obscuration of abnormalities by fatty tissue, a radiologist can omit up to 30% of cancers. Moreover, because this type of analysis produces many false positives, the number of unnecessary biopsies is increased up to 35%, causing a high level of stress in the patient, and in turn saturating the health systems [21].

Due to all the problems presented by mammography screening, great efforts have been made to support the radiologist in the search for these lesions through computer-aided detection (CAD) or biomarker systems, which try to help the radiologist by taking advantage of the latest advances in computer vision and their manipulation in digital form [22].

At present, computer-aided detection/diagnosis (CAD/CADx) systems are one of the numerous major research topics in diagnostic radiology and medical imaging. CAD systems allow the radiologist to manipulate mammography to highlight certain features that would otherwise be difficult to visualize. One of the most used techniques is the improvement of contrast, which allows to highlight objects in areas of low intensity. To date, CAD is a more suitable method for primary diagnosis of cancer in computed tomography, X-ray, MRI, or mammogram images. CAD system is an effective intermediate between input images and the radiologist. The output from CAD is not considered as an end result; nevertheless, the result is used as reference with regard to additional testing in the related field [23].

The CAD approach helps medical doctors to diagnose diseases with a higher degree of efficiency, while minimizing examination time and cost, as well as avoiding unnecessary biopsy procedures. However, CAD systems not only allow a better visualization of mammograms, but also using different digital image processing (DIP), knowledge discovery from data (KDD), artificial intelligence (AI) techniques such artificial neural networks (ANN) allow to preselect certain regions of interests (ROIs) for later analysis by the radiologist [24].

Classification of histopathology images into distinct histopathology patterns, corresponding to the noncancerous or cancerous condition of the analyzed tissue, is often the primordial goal in image analysis systems for cancer automatic-aided diagnosis applications. Recent advances in DIP, KDD, and AI techniques allow to build CAD/CADx that can assist pathologists to be more productive, objective, and consistent in diagnosis. The main challenge of such systems is dealing with the inherent complexity of histopathological images.

The aim of this research is to use advanced DIP to investigate and develop specific imaging biomarkers for Mexican patients through the quantitative mammography analysis and with this information to develop technology based on advanced KDD and AI techniques, aiming to detect breast cancer in the early stages in order to support the diagnosis and prioritization of high-risk patients.
