Introduction
Diffuse gliomas are among the most challenging brain tumors to treat surgically. Their infiltrative nature means that tumor cells can spread into surrounding healthy brain tissue, making it difficult for surgeons to distinguish between cancerous and normal areas during an operation. The goal of glioma surgery is to remove as much tumor as possible while preserving critical brain functions, but this balance is hard to achieve when the boundaries are unclear. Traditional imaging methods, such as MRI, provide a preoperative view but cannot fully resolve the microscopic infiltration that occurs at the edges of the tumor. Intraoperative pathology, typically using frozen sections, offers rapid feedback but is limited to small, two-dimensional samples that may not represent the entire tumor margin.
Now, a team of researchers from Fudan University, Huashan Hospital, Zhongshan Hospital, Beijing Neurosurgical Institute, and Capital Medical University have developed a new platform called ULTRA (ultrarapid cleared stimulated Raman with AI) that can generate three-dimensional images of glioma margins in about 30 minutes. This technology, reported in the journal Cell, combines rapid tissue clearing, stimulated Raman scattering (SRS) microscopy, and artificial intelligence to provide a comprehensive view of tumor infiltration without the need for conventional staining or sectioning.
The Challenge of Glioma Surgery
Gliomas are primary brain tumors that arise from glial cells, which support and protect neurons. They are notorious for their diffuse growth pattern, with tumor cells migrating along white matter tracts and blood vessels, often far beyond the main tumor mass. This makes complete surgical resection nearly impossible, and residual tumor cells can lead to recurrence. Neurosurgeons must therefore rely on intraoperative tools to identify tumor margins in real time.
Currently, the standard method is frozen section analysis, where a small piece of tissue is removed, frozen, sliced, and stained for microscopic examination. While this provides rapid pathological information, it has several limitations. The samples are typically tiny, representing only a fraction of the tumor boundary, and the two-dimensional sections may miss infiltrating cells in other planes. Additionally, freezing can introduce artifacts, and staining quality can vary, potentially leading to misinterpretation.
Three-dimensional histology offers a more complete picture by preserving the spatial architecture of the tissue, allowing pathologists to see how tumor cells are distributed throughout a volume. However, existing 3D methods are slow, often requiring hours or days for tissue clearing, labeling, and imaging, making them impractical for use during surgery.
ULTRA: A 30-Minute Workflow
ULTRA addresses these limitations by integrating three key components: a rapid tissue-clearing technique, stimulated Raman scattering microscopy, and AI-based virtual staining. The workflow begins with a tissue sample, either fresh or fixed, that is treated with a clearing solution to make it transparent. This allows light to penetrate deeper into the tissue, enabling imaging at millimeter-scale depths.
Next, the cleared tissue is imaged using stimulated Raman scattering microscopy, a label-free technique that detects molecular vibrations. SRS provides chemical contrast without the need for external dyes, and it can image tissue rapidly. The researchers optimized the SRS system to capture high-resolution images of the cleared tissue, generating data that reflects the cellular composition and structure.
Finally, an AI algorithm performs virtual staining, converting the label-free SRS images into images that resemble traditional histology stains. This step is crucial because it allows pathologists to interpret the images using familiar staining patterns, such as H&E, without the time-consuming staining process. The AI was trained on paired SRS and stained images to accurately predict the virtual stains.
The entire process—from tissue receipt to final 3D reconstruction—takes approximately 30 minutes, making it feasible for intraoperative use. This is a significant improvement over existing 3D histology techniques, which can take hours or even days.
Validation and Potential Impact
The researchers validated ULTRA on surgical tissue samples from glioma patients. They demonstrated that the platform could accurately identify tumor margins, showing infiltrating tumor cells in areas that might have been missed by conventional frozen sections. The 3D images allowed them to visualize the tumor's irregular shape and its invasion into surrounding brain tissue, providing a more complete picture for surgical decision-making.
One of the key advantages of ULTRA is its ability to provide volumetric information in real time. Surgeons could potentially use this technology during surgery to assess whether all tumor tissue has been removed, reducing the risk of leaving behind residual disease. This could lead to more complete resections and improved patient outcomes.
Moreover, ULTRA is not limited to gliomas. The platform could be adapted for other types of cancer where surgical margins are critical, such as breast, lung, or prostate cancer. The combination of rapid clearing, SRS, and AI has broad applications in pathology and surgical oncology.
Future Directions
While ULTRA shows great promise, there are still challenges to overcome. The current system requires specialized equipment and expertise, which may limit its availability to major medical centers. The researchers are working on making the technology more accessible and user-friendly. They also plan to expand the AI training to include a wider range of tissue types and staining protocols.
Another area of development is the integration of ULTRA with other imaging modalities, such as MRI or fluorescence-guided surgery, to provide a comprehensive view of the tumor. Combining these techniques could enhance the accuracy of tumor margin assessment even further.
In conclusion, ULTRA represents a significant advancement in intraoperative pathology. By providing rapid, 3D, AI-enhanced histology, it has the potential to transform how surgeons approach glioma resection, ultimately improving outcomes for patients with these challenging tumors.
This article is based on reporting by Medical Xpress. Read the original article.
Originally published on medicalxpress.com
