Page 107 - AIH-2-2

Artificial Intelligence in Health AI in early breast cancer diagnosis: A review

The incidence and mortality rates of breast cancer As highlighted by tumor metastasis charts, early
exhibit variations correlated with the human development detection of breast cancer is critical for improving
index (HDI). In countries with high HDI, the lifetime treatment efficacy and patient prognosis. However,
risk of breast cancer diagnosis is approximately 1 in 12 traditional screening methods, such as mammography
women, with a mortality risk of 1 in 71. Conversely, in and ultrasound, have limitations, including false-negative
countries with low HDI, typically developing nations, the results, particularly in young patients with dense breast
lifetime risk of diagnosis is 1 in 27, with a mortality risk tissue. A comprehensive discussion on imaging techniques
4
of 1 in 48. 2 for cancer detection is presented in the following section
Breast cancer originates in the mammary tissue and summarized in Table 1. It is important to note that
and is characterized by abnormal cellular proliferation misdiagnosis and medical errors are significant issues in
forming a neoplasm. The disease typically originates in healthcare, with medical mistakes causing more deaths
the cells surrounding the milk ducts. The mammary gland annually than breast cancer, human immunodeficiency
5
comprises adipose tissue, glandular tissue organized into virus, and car accidents combined. This underscores
lobes (each containing multiple lobules), an intricate the need to develop computer-aided diagnosis (CAD)
network of milk ducts, and connective tissue. Lobules, systems to assist radiologists and healthcare professionals
which are responsible for milk production, connect to the in making life-altering decisions to improve patient care.
ducts, serving as the primary site of carcinogenesis. With advancements in machine learning (ML) and artificial
intelligence (AI) in healthcare, smart CAD systems can be
Tumors exist in two forms: benign (non-cancerous) developed to assist pathologists and radiologists in their
or malignant (cancerous). Malignant tumors can spread decision-making process.
locally to neighboring tissues or distantly to other organs
through metastasis. The disease is categorized into stages Another method to detect breast cancer before it
ranging from stage 0 to stage 4 (early to late), with treatment progresses beyond the benign stage is through testing
6
options and prognosis being highly dependent on the for certain biomarkers obtained from liquid biopsies.
cancer stage of the patient. Several factors are assessed to Cancer biomarkers are essential for the rapid detection
determine the stage of breast cancer, which helps to decide and monitoring of breast cancer, providing a non-invasive
treatment options and prognosis. One important factor and more convenient alternative to conventional screening
is the tumor size, typically measured in millimeters or techniques. Key biomarkers include circulating tumor cells
centimeters, and is assigned a T value based on this size. (CTCs), cell-free tumor nucleic acids, and microRNAs
A larger tumor generally indicates a later stage of cancer. (miRNAs), which will be discussed in more detail in the
next section.
Another factor is whether the cancer has spread to the
lymph nodes, which are a network of lymph glands near This systematic literature review, motivated by
the breast. Lymph nodes are examined under a microscope advancements in molecular biology and biomedical
and assigned an N value based on the findings, as shown in sciences, covers novel methods in both biological and
Table 1. The greater the number of involved lymph nodes, computational aspects of early detection of breast cancer.
the higher the N value. If lymph nodes are cancer-free, it is To our knowledge, this is the first effort of its kind.
less likely that the cancer has spread. Previous reviews on breast cancer are based either solely on
biomarkers or on deep learning (DL) methods applied to
Metastasis, which refers to cancer spreading to distant imaging modalities. This systematic review aims to provide
organs, is also a key variable in determining the cancer a comprehensive exploration of breast cancer detection
stage. An M value is assigned based on this assessment, as and diagnosis through screening, highlighting challenges
outlined in Table 1. In addition to these factors, healthcare in early detection using CAD, innovative detection
professionals also consider hormone receptor status and methods to isolate cancer biomarkers, the importance of
human epidermal growth factor receptor 2 status through these biomarkers, and the need for standardized research
specific tests that help determine if the cancer cells are approaches.
likely to respond to certain therapies. Finally, examining
cancer cells under a microscope allows for an assessment 2. Background
of their appearance relative to normal body cells, providing Breast cancer continues to be one of the most widespread
insights into how aggressive the cancer may be. cancers globally, with diagnostic challenges and high
Doctors can assign a stage to breast cancer by incidence rates. Contributing factors include aging
considering all of these factors, with earlier stages generally populations, increased risk factors, genetic predisposition,
associated with a better prognosis than later stages. 3 and widespread use of mammography screening. Early
7

Volume 2 Issue 2 (2025) 101 doi: 10.36922/aih.4197