Hydrogel fibers offer new route to imaging early-stage breast cancer

Researchers at China's Harbin Engineering University have developed a form of optical fiber that could ultimately enable easier imaging operations within the human body.

Described in Optics Express, the project's biocompatible hydrogel multimode optical fibers (HMMFs) are initially intended specifically to help in the imaging and intelligent recognition of breast tumors, the leading cause of cancer-related mortality among women worldwide.

Key to this goal is the new fiber's small cross-section, measuring just hundreds of microns in diameter. This should allow detection of early malignancies hidden inside the breast's small mammory ducts.

Optical approaches to imaging of breast malignancies have included platforms based on photoacoustic imaging and elastography, but the Harbin project set out to develop optical fiber probes flexible and durable enough to reach the locations of interest within the breast and collect image data directly.

Although soft and biocompatible, hydrogel materials tend to easily lose water and become stiff or crack when exposed to air, which affects long-term use. In addition, most existing hydrogel fibers are millimeters in diameter, noted the project, which is too large to enter narrow spaces such as breast ducts only a few hundred micrometers wide.  

"While traditional, relatively rigid fiber probes may cause mechanical damage when entering narrow, curved or soft tissue spaces, our fibers are very soft with mechanical properties more similar to those of human soft tissues," said research team leader Yu Zhang. "We made these fibers using a draw-spinning method that was inspired by spider-silk spinning."

The project's polyacrylamide hydrogel fibers were draw-spun down to diameters of approximately 120 to 200 microns, a procedure that preserved tightly bound water within the polymer network and promoted long-term stability over timescales measured in days, according to the project's paper.  The fibers enabled "stable optical guiding from the visible to the near-infrared region, with an optical loss as low as 1.05dB/cm at 980 nanometers."

Imaging integrated with intelligent diagnosis

In trials, the HMMFs were incorporated into a multimode fiber imaging platform and used to assess samples of breast tissue, delivering light to the area and recording a speckle pattern from the illumination. The complex speckle data were then assessed by deep learning algorithms designed by the project, and turned into visual images of the samples.

Using H&E-stained breast tumor pathological sections for a proof-of-concept study, the researchers showed that their two-stage deep learning framework could classify tissue as normal, invasive carcinoma or ductal carcinoma with an accuracy of 93.97 percent.

The next steps will include moving the technology closer to clinical use, through improved image reconstruction accuracy for complex tissue structures and training the algorithms on larger, more diverse medical datasets. Further aims are to optimize the hydrogel fiber for in vivo use; conduct additional safety and animal studies; and integrate the fiber, imaging system and AI algorithms into a compact platform that can provide useful diagnostic information in practical settings.

"These results show that the hydrogel multimode fiber can not only transmit image information but also has the potential to be integrated with intelligent diagnosis," said Yu Zhang. "In the future, similar technologies may be used not only for breast cancer, but also for minimally invasive detection of other diseases, surgical navigation, in vivo sensing and intelligent medical imaging."