Artificial intelligence detecting breast cancer risk?

Artificial Intelligence (AI) is indeed a thriving field that has influenced various industries, sectors and the way of life of humans. In fact, AI is paving the way for more breakthroughs in the medical field. One example of these breakthroughs is the study published in the journal Radiology, which suggested that AI can be a tool in detecting high-risk breast lesions that can be cancerous.

Considered the most common form of cancer among women, Breast Cancer has caused an estimated 522,000 deaths worldwide, Globocan noted. In the United States, 40,000 women die from it annually. However, if detected early, cancers can usually be cured.

Even though there are some tests available for detecting breast cancer like mammograms, MedicalXpress noted that many are “imperfect,” and these often lead to false positive results, unnecessary surgeries, and biopsies. Fortunately, medical experts found a way to address such imperfections through artificial intelligence.

The study

Mammography plays a significant role in breast detection.

However, a team of researchers at Massachusetts Institute of Technology’s (MIT) Computer Science and Artificial Intelligence Laboratory (CSAIL), Harvard Medical School and Massachusetts General Hospital (MGH) believe that AI can help reduce or eliminate the unnecessary surgeries and improve the detection and diagnosis of breast cancer.

The collaboration allows the experts to develop an AI-driven tool that uses machine learning to detect a high-risk lesion that has the potential to be cancerous after its identification through a needle biopsy following a mammogram procedure. The tool, which was trained to detect the lesions, looks into the patterns among a variety of data points including pathology reports, biopsies, family history and demographics.

The researchers considered their findings “useful” after the system tested 335 high-risk lesions and accurately diagnosed 97 percent of the breast cancers as malignant. Furthermore, the AI system also reduced the number of benign or unnecessary surgeries by more than 30 percent, BBC News noted.

Overtreatment and over screening

According to Regina Barzilay, a breast cancer survivor and MIT’s Delta Electronics Professor of Electrical Engineering and Computer Science, doctors have the tendency to “over-screen” patients for breast cancer due to the inaccuracies of the diagnostic tools. Since there are uncertainties in the data, Barzilay said that machine learning is the exact tool needed for the improvement of cancer detection and prevention of overtreatment.

Harvard Medical School professor and MGH’s Department of Radiology’s Breast Imaging Chief Constance Lehman, on the other hand, explained that their study is the first to apply machine learning technique to determine high-risk lesions. Lehman added that they believed their findings could help women in making “more informed decisions about their treatment,” as well as offer “more targeted” healthcare approaches.

Improve patient care

Apart from the detection of cancer, MGH’s Breast Imaging Fellowship Program director and one of the study authors Dr. Manisha Bahl said machine learning can also be used for patient care improvement. The tool does not only aim to reduce unnecessary surgeries but also provide more valuable information to the patients, HealthDay reported.

Due to the potential of their study, researchers are optimistic that they can incorporate mammography and pathology slides into their machine learning tool. They also hope that this system will be included in the clinical practice.

Breast Cancer Awareness Month

Meanwhile, October is considered the Breast Cancer Awareness Month worldwide. That is why the World Cancer Research Fund International and The Cancer Atlas (via U.S. News and World Report) released the top 10 nations with the highest breast cancer diagnosis rates. These include Ireland, United States, Barbados, United Kingdom, Iceland, Bahamas, Netherlands, France, Denmark, and Belgium.