Terahertz waves can image early-stage breast cancer without staining

A team of researchers from the University of Osaka, in collaboration with the University of Bordeaux and the Bergonié Institute in France, has managed to obtain terahertz images of early-stage breast cancer of less than 0.5 mm without staining, which is difficult to identify even by pathological diagnosis. Their work provides a breakthrough toward rapid and accurate in situ diagnosis of various types of cancer and accelerates the development of innovative terahertz diagnostic devices.

Breast cancer is roughly divided into two types: invasive and non-invasive. The first, invasive ductal carcinoma (ICD), begins in the cells of a breast duct, grows through the walls of the duct and reaches the surrounding breast tissue and potentially spreads to other parts of the body. The latter, ductal carcinoma in situ (DCIS), is an early-stage small breast cancer confined to the mammary duct, but which can lead to invasive cancer. Therefore, early detection of DCIS is crucial.

For the pathological diagnosis of cancer, the tissue sample is chemically stained and a diagnosis is made by a pathologist using an image of the stained tissue. However, the staining process takes time and it is difficult to distinguish DCIS from malignant IDC as they appear almost identical.

Terahertz images can distinguish cancerous tissue from normal tissue without staining or exposure to radiation. However, it was still difficult to identify a single DCIS lesion (typically ranging from 50 to 500 μm) using terahertz imaging due to its diffraction-limited spatial resolution of only several millimeters.

“To overcome this drawback, we developed a unique imaging technique in which terahertz light sources that are generated locally at laser beam irradiation spots on a non-linear optical crystal interact directly with a breast cancer tissue sample. Consequently, we were able to clearly visualize a DCIS lesion of less than 0.5 mm, “explains lead author Kosuke Okada. The precision of this technique is approximately 1000 times greater than that of conventional techniques using terahertz waves.

The researchers also found that the terahertz intensity distributions were different between DCIS and IDC, suggesting the possibility of a quantitative determination of cancer malignancy.

The breast cancer tissue sample was provided and histologically evaluated by collaborators from the University of Bordeaux and the Bergonié Institute. “One of the challenges in this research is preparing a high-quality breast cancer tissue sample made from a non-linear optical glass. It is one of the great achievements of international joint research,” says corresponding author Masayoshi Tonouchi.

“Combining our technique with machine learning will aid in early cancer detection and cancer malignancy, as well as the development of innovative terahertz diagnostic devices using microelectromechanical systems.”