Swiss researchers use AI for breast cancer characterization

No time? blue News summarizes for you

• Swiss researchers from PSI and MIT have developed an AI-based image analysis to estimate the stage of breast cancer more precisely and cost-effectively.
• The aim of the study is to improve prognoses, especially for tumors that grow very slowly or rarely progress to a life-threatening form.
• Many more studies are needed before the new method can be used in practice, although initial results are promising.

Summary created with

The PSI in Villigen AG announced on Monday that the aim is to make the predictions reliable. It was pointed out that some tumors grow very slowly or hardly ever change stage from a rather harmless precursor to a life-threatening form.

AI opens up new possibilities for a more precise and cost-effective diagnosis of breast cancer. Researchers at the Paul Scherrer Institute (PSI) and the Massachusetts Institute of Technology (MIT) have developed an image analysis based on AI technology that enables a precise assessment of the stage of the disease. The aim is to make prognoses more reliable. In some tumors, growth is very slow or the stage hardly ever changes from a rather harmless preliminary stage to a life-threatening form.

G.V. Shivashankar is currently developing various methods for the diagnosis and prognosis of cancer at PSI. The current study gives hope that one form of breast cancer can be better characterized.

© Paul Scherrer Institut PSI/Markus Fischer

Patients live with uncertainty

In women, there is the possibility of a preliminary stage of breast cancer in the milk ducts, known as "ductal carcinoma in situ". This form, known as DCIS for short, develops into an invasive "breast carcinoma" in 30 to 50 percent of cases.

Due to the good curability of DCIS, doctors generally recommend treatment. However, according to PSI, there is still a lack of reliable evidence to determine with certainty which tumor remains harmless and which will develop into a life-threatening carcinoma.

AI Health

How an AI app could extend our lives

AI enables better DCIS staging

According to the PSI media release, the study showed that the use of AI can improve the assessment of the extent of a tumor disease (staging) with the help of data that is easy and inexpensive to collect. According to the researchers, they provided a learning algorithm with 560 tissue samples from 122 patients.

A dye was added to the samples to make the chromatin in the cell nucleus glow fluorescently. Chromatin consists of the genetic material DNA and proteins, among other things. The appearance allows conclusions to be drawn about the organization and thus the activity of the DNA contained in the cell nucleus.

The combination of inexpensive and easy-to-obtain "chromatin images" with powerful AI algorithms could provide sufficient information to investigate changes in cell status and tissue organization. However, numerous further studies are required before practical application.