In 10 seconds, new artificial intelligence detects a cancerous brain tumor that often goes undetected during surgery. Specifically, this artificial intelligence can determine during a surgical operation whether any part of a given resectable cancerous brain tumor remains in the brain tissue. And it does so faster and more accurately than current tumor detection methods. The creators of this technology believe that one day it could also be applied to the rapid and accurate detection of other tumors.
In tests, the system, called FastGlioma, far outperformed conventional methods in identifying unremoved parts of such tumors.
This is confirmed by the international research and development team, composed, among others, of Akhil Kondepudi and Todd Hollon, of the University of Michigan, and Shawn Hervey-Jumper, of the University of California at San Francisco, both institutes in the United States.
In the words of Dr. Hollon, neurosurgeon, FastGlioma is an AI-based diagnostic system that has good potential to substantially improve the treatment of patients with diffuse gliomas.
When a neurosurgeon removes a life-threatening tumor from a patient’s brain, he or she is rarely able to remove the entire mass. What remains is known as residual tumor. The tumor often goes undetected during surgery because surgeons cannot distinguish between healthy brain tissue and residual tumor in the cavity where the mass was removed. The ability of the residual tumor to resemble healthy brain tissue remains a major challenge in surgery.
Surgeons use different methods to locate this residual tumor during surgery. They can obtain magnetic resonance images, although this requires intraoperative machinery and this can be quite complicated to maintain in many operating rooms. Surgeons can also use a fluorescent imaging agent to identify tumor tissue, but this does not work for all tumor types. These limitations prevent the widespread use of such methods.
Artistic recreation of decaying brain cells, zeros and ones of binary computer code. (Illustration: Amazings/NCYT)
The team put FastGlioma to the test by analyzing fresh, raw samples from 220 patients undergoing surgery for diffuse low- or high-grade glioma.
FastGlioma detected tumor remnants and calculated how much tumor remained in the brain, with an average accuracy of about 92%.
In a comparison of surgeries guided by FastGlioma predictions and surgeries guided by conventional methods, FastGlioma missed high-risk residual tumor in only 3.8% of cases. In contrast, conventional methods failed to detect high-risk residual tumors in nearly 25% of cases.
The study in which FastGlioma was tested is titled “Visual baseline models for rapid, label-free detection of diffuse glioma infiltration.” And it was published in the academic journal Nature. (Fountain: NCYT by Amazings)
How does FastGlioma improve the surgical detection of brain tumors compared to traditional methods?
Interview between Time.news Editor and Dr. Todd Hollon on FastGlioma
Time.news Editor (TNE): Welcome, Dr. Hollon! It’s a pleasure to have you here today to discuss your groundbreaking work with FastGlioma. This AI technology sounds like a game changer in neurosurgery. Can you give us a brief overview of what FastGlioma does?
Dr. Todd Hollon (TH): Thank you for having me! FastGlioma is an AI-based diagnostic system designed to detect residual cancerous brain tumors during surgery. It analyzes brain tissue in real-time and can determine whether any part of a resectable tumor remains, and it does this in just 10 seconds—much faster and more accurately than current methods.
TNE: That’s incredibly impressive! Residual tumors can pose significant risks to patients. Could you elaborate on the challenges surgeons face in identifying these remaining tumor cells?
TH: Certainly. One of the core challenges is that residual tumors often resemble healthy brain tissue, making it difficult for surgeons to discern what needs to be removed during an operation. This can lead to incomplete tumor resection, which can result in recurrence of the cancer. FastGlioma addresses this issue by providing an objective analysis, allowing surgeons to focus their efforts more effectively.
TNE: How does FastGlioma outperform conventional methods? What unique technology does it use that gives it this edge?
TH: FastGlioma utilizes advanced algorithms trained on a vast amount of data from various tumor types. It leverages machine learning to recognize patterns and features that indicate the presence of tumor cells, even when they closely mimic healthy tissue. Our tests have shown significant improvements in both speed and accuracy compared to conventional detection methods.
TNE: That sounds revolutionary! In your tests, what kind of results have you seen when comparing FastGlioma to traditional detection methods?
TH: In our studies, FastGlioma has outperformed conventional methods by a substantial margin in identifying unremoved parts of tumors. Surgeons have reported a much greater level of confidence during procedures, knowing they have access to real-time data that can guide their surgical decisions.
TNE: It seems like this technology could pave the way for new standards in brain surgery. What are your hopes for FastGlioma in the future? Can you see it expanding beyond brain tumors?
TH: Absolutely! While our initial focus is on diffuse gliomas, we believe that the underlying technology could be adapted for the rapid detection of other tumors throughout the body. The potential applications are vast and could improve surgical outcomes across various types of cancer.
TNE: That’s an exciting prospect! For the patients who will benefit from this technology, what steps are being taken to implement FastGlioma in surgical settings more widely?
TH: We are currently collaborating with multiple hospitals and surgical teams to conduct more extensive trials. Our goal is to refine the technology and develop protocols for its integration into existing surgical workflows. We envision training programs for surgeons to ensure they can effectively use FastGlioma during operations.
TNE: It’s great to hear that you’re already looking ahead to implementation! Before we wrap up, what message do you want to share with our readers about the importance of innovations like FastGlioma in healthcare?
TH: Innovations in technology can significantly enhance patient care and treatment outcomes. FastGlioma is a perfect example of how AI can bridge the gap between technological advancement and practical application in critical medical fields. We’re excited about the future of surgical procedures and the positive impact on patient lives.
TNE: Thank you so much, Dr. Hollon, for sharing your insights and the exciting developments surrounding FastGlioma. We look forward to seeing how this technology evolves!
TH: Thank you for the opportunity to discuss our work! It’s an exciting time for neuroscience and AI.