Fudan University images tumor boundaries within 30 minutes

Accurately identifying the borders of malignant tissues is one of the critical challenges in cancer therapy, and multiple optical approaches have been applied to the task as alternatives to time-consuming tissue excision and biopsy.

These include photoacoustic imaging using ultraviolet wavelengths rather than infrared, to highlight cell nucleii more strongly than surrounding cell components; dual-mode fluorescence and visible light imaging to map the contours of the malignancy; and elastography combined with optical imaging to monitor the physical properties of a tumor.

A project led by Fudan University Institute of Science and Technology has now demonstrated a new approach based on stimulated Raman scattering microscopy, intended to compress deep 3D histological analysis into a workflow of about 30 minutes, according to the project team. The findings were published in Cell.

The Fudan platform - named ULTRA for ultra-rapid cleared stimulated Raman with AI - combines rapid tissue clearing, stimulated Raman scattering microscopy and AI-based virtual staining, to perform deep 3D histological imaging of surgical tissue without conventional staining or sectioning.

ULTRA operates in three stages. First a rapid tissue-clearing method compatible with stimulated Raman scattering microscopy allows fresh or fixed brain tissue to become transparent enough for millimeter-scale volumetric imaging. The tissue is then imaged by stimulated Raman scattering microscopy, which captures intrinsic chemical-bond vibrational signals without external labels.

Finally, AI algorithms reconstruct and convert the imaging data into H&E-like 3D virtual histology, presenting tissue morphology in a format closer to ones familiar to pathologists.

The platform's data processing in fact employs three sequential AI modules, first to restore image quality as signal declines with imaging depth; then predict the desired virtual-staining of the protein channel from the lipid channel to simplify data acquisition; and finally convert stimulated Raman imaging data into simulated H&E staining.

Bringing 3D histology into the operating room

"Conventional intraoperative pathology often relies on a limited number of two-dimensional sections, while tumors themselves are highly heterogeneous three-dimensional entities," commented Lixue Shi of Fudan University. "The goal of ULTRA is to obtain a more complete 3D histological view within a timeframe close to intraoperative pathology, without destroying the tissue."

In trials using human surgical glioma specimens, ULTRA revealed 3D pathological features that conventional two-dimensional sections may not fully capture, according to the project. The platform visualized key histological features across different glioma subtypes observed continuously across three-dimensional tissue volumes, rather than appearing only in isolated sections.

Most importantly, ULTRA was applied to identifying the margins of glioma infiltration, something MRI imaging cannot reliably do. In a glioblastoma margin specimen, the researchers processed and analyzed approximately 1 mm³ of tissue using ULTRA, and then used cellularity analysis and a 3D convolutional neural network to generate an "ULTRAscore," representing the probability that a given 3D tissue block contained tumor.

Further large-scale prospective clinical studies will be needed to evaluate the reliability, efficiency and clinical value of ULTRA in real intraoperative workflows, but the project believes it presents a new route for bringing 3D histology into the operating room.

"We hope this technology will help doctors better understand glioma infiltration margins in the future and provide more tissue-level information for intraoperative decision-making,” said Lixue Shi.