JERUSALEM, Aug. 19 -- Israeli researchers have developed a new, deep learning (DL) technology that is expected to significantly improve personalized cancer treatments - the northern Israel Institute of Technology (Technion) reported on Monday.

This is a method for mapping critical receptors on cancer cells, based on biopsy images of breast cancer patients.

The new method, published in the journal JAMA, extracts molecular information from hematoxylin and eosin stain (H&E) biopsy images - a common staining used to test tissues taken in a biopsy test.

This staining allows the pathologist to identify in the tissue, under the microscope, the type of cancer and its severity.

However, staining alone does not allow to identify critical characteristics that are crucial in determining the appropriate treatment, such as the tumor's molecular profile, its biological pathways, the genetic code of the cancerous cells, and the common receptors on the cell membrane.

Mapping of these receptors is particularly relevant to personalized medicine, allowing a treatment which will block the receptors and inhibit the development of the cancerous tumor.

The Technion researchers' conceptual innovation is in extracting molecular information from the cell shape and environment (the morphology of the tissue) as reflected in H&E scans.

According to the researchers, pathologists can't deduce the tumor properties from its shape because of the huge number of variables, but on the other hand, artificial intelligence (AI), especially DL, are able to do so and characterize the cancer with a complex analysis of its morphology.

Thus, with the help of image processing and AI tools, the researchers showed, for the first time, the possibility of predicting the molecular profile of cells from tumor morphology, only from looking at the tissue as it appears in H&E scans.

DL systems require a tremendous amount of information, so the researchers have written software code to scan network sources and automatically download thousands of biopsy samples and the relevant medical information approved for research.

Although the study focused on breast cancer, the researchers say this is proof of feasibility relevant to all cancer types.