- On the occasion of Pink October, breast cancer screening awareness month, 20 minutes looks at the latest technological innovations aimed at improving the detection and treatment of the disease.
- Among the avenues explored: artificial intelligence, which could make it possible to assess the risks of developing breast cancer before it appears.
- But also cell-by-cell analysis tools, to better understand the origin of breast cancer.
The figures are chilling: during their lifetime, one in eight women will be affected by breast cancer. A disease which, although it is increasingly treatable, still remains the leading cause of cancer mortality in women. Each year, more than 60,000 cases of breast cancer are detected, with, for each woman who receives this diagnosis, the same questions running through her head: “Why me? Why now ? Where does this breast cancer come from? »
Understanding the triggering factors of the disease are therefore at the heart of researchers’ concerns. In this race against time, how can artificial intelligence (AI) and technology improve breast cancer screening and treatment? On the occasion ofPink Octoberbreast cancer screening awareness month, 20 minutes takes stock of the latest innovations.
Detecting cancer before it appears using AI
Health Insurance reminds us that “detected at an early stage, this cancer with a good prognosis can be cured in 9 out of 10 cases”, with a survival rate of 88% five years after diagnosis. Faced with this observation, what if the ultimate precocity consisted of detecting cancer before it even appears? This is the feat that Danish and Dutch researchers are trying to accomplish using artificial intelligence. How ? By combining two diagnostic AI tools, they detail in a study recently published in the journal Radiology.
A first AI model identifies suspicious lesions likely to evolve in the short term on a mammogram showing no signs of cancer [dans les deux ans] in breast cancer. And a second analyzes breast density and helps assess long-term risks – dense breast tissue being associated with a higher risk of developing breast cancer. To combine the two, they used the deep learning, or deep learning, by compiling the results of 39,000 mammograms to train this combined AI model. A model tested on a cohort of 119,000 women who participated in a Danish breast cancer screening campaign.
And the combination of the two AIs looks promising. “Compared to diagnostic and texture models alone, the combined AI model showed improved overall risk assessment for short- and long-term cancer detection,” said Dr. Andreas D. Lauritzen, co-author of the study. In practice, all women undergoing a screening mammogram could instantly know their five-year risk of developing breast cancer. And by prescribing additional MRI for women at high risk of breast cancer, up to 36% of breast cancer cases could be detected earlier.”
Improving disease prevention
Another step that science is already working on: improving the prevention of breast cancer by analyzing the factors favoring its appearance. This is the mission given to itself the Curie Institute, the leading European center for the fight against breast cancer, which treats more than 7,000 patients each year. “Understanding the very first events at the origin of cancer is essential, because the challenges and prospects are considerable in terms of prevention, very early diagnosis, even before the appearance of clinical signs, but also therapeutic management for our patients, explains Professor Alain Puisieux, director of the Institut Curie Research Center. There is a very wide variety of breast cancers which present different characteristics, from a genetic, epigenetic and microenvironment point of view. Characteristics influencing their aggressiveness and their response to treatments.”
Among these breast cancers, the triple negative. An aggressive cancer which mainly affects young women, and against which chemotherapy is ineffective in half of cases and hormone therapy does not work. And on which the work of Dr Céline Vallot focuses. Research director at the CNRS, she studies the role of epigenetics at the Institut Curie [l’incidence de facteurs environnementaux tels que stress, tabagisme, pollution ou encore alimentation] in the appearance of triple negative breast cancer and the adaptation of cancer cells to treatments. “We are seeking to establish the epigenetic identity map of each cell, individually, in a tumor,” she explains.
Thanks to this methodology of Single cell, or single cell analysis, “with extreme precision, we identify at the earliest stages the epigenetic profiles responsible for the transformation of healthy cells into cancerous cells,” she continues. “This cutting-edge technology makes it possible to analyze tumors at unprecedented levels and to study the capacity of cells to adapt and diversify during the evolution of the tumor,” indicates the Institut Curie.
Ultimately, thanks to the data collected cell by cell using high-throughput DNA sequencing, then modeled using “powerful and innovative calculation tools”, “certain epigenetic modifications could allow us to develop new diagnostic tools or to identify new therapeutic targets relevant for the treatment of triple negative breast cancer,” says Dr. Vallot. A new bearer of hope for the “triplets”, the women affected by this cancer, who plead for improved access to care.