Abstract
Breast cancer is a real public health problem in Morocco. It is the cause of a significant number of deaths caused by late diagnosis. Mammography plays an essential role in the detection of breast cancer and in the early management of its treatment. Despite the existence of screening programs, there are still high rates of false positives and false negatives. Indeed, women were called back for additional diagnoses based on suspicious results that eventually led to cancer. Artificial intelligence (AI) algorithms represent a promising solution to improve the accuracy of digital mammography offering, on the one hand, the possibility of better cancer detection, and, on the other hand, improved efficiency for radiologists for good decision-making. In this work, through a review of the literature on the tools used to evaluate the performance of AI systems dedicated to early detection and diagnosis of breast cancer. We set out to answer the following questions: Is the ethics relating to patient data during the development phase of this software is respected? Do these tools take into consideration the specificities of the field? What about the specification, accuracy and limitations of these applications? At the end, we show through this work recommendations to adapt these evaluation tools of AI applications for breast cancer screening for an optimized and rational consideration of the principle of health vigilance and compliance with the regulatory standards in force governing this field.
Highlights
Breast cancer is the most frequently diagnosed tumor worldwide [1] and represents the first cancer affecting women in Morocco [2] (31.9 per 100,000)
We show through this work recommendations to adapt these evaluation tools of Artificial intelligence (AI) applications for breast cancer screening for an optimized and rational consideration of the principle of health vigilance and compliance with the regulatory standards in force governing this field
In the Hybrid convolutional neural networks (CNN) +k-nearest neighbor method (KNN) (K-Nearest Neighbors) classifier proposed, we exploited the features extracted by our CNN model and used these features as inputs for a KNN classifier
Summary
Breast cancer is the most frequently diagnosed tumor worldwide [1] and represents the first cancer affecting women in Morocco [2] (31.9 per 100,000). The cancerous (malignant) tumor is a group of cancerous cells that can invade and destroy nearby tissue This tumor can spread to other parts of the body (metastasis). Breast cells sometimes undergo changes that make their growth pattern or behavior abnormal These changes can lead to non-cancerous (benign) breast conditions such as atypical hyperplasia and cysts. They can lead to intraductal papillomas that form in the breast ducts and are usually detectable near the nipple. This type of (benign) breast tumor is a mass that does not spread to the rest of the body (no metastasis) and is usually not lifethreatening [3]
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
