AI Technique Automatically Traces Tumors in Large MRI Datasets to Guide Real-time Glioblastoma Treatment
Posted on 16 Nov 2023
Treating glioblastoma, a prevalent and aggressive brain cancer, involves the use of radiation therapy guided by CT imaging. While this method is effective in targeting radiation, it doesn't provide real-time information about the tumor's response to the treatment. This gap means that clinicians are unable to determine whether a patient's cancer is responding to the treatment or progressing until follow-up images are taken, sometimes months later. Given the rapid progression of glioblastoma, such delays can have critical consequences.
To address this challenge, a team of researchers at Sylvester Comprehensive Cancer Center, part of the University of Miami Miller School of Medicine, (Coral Gables, FL, USA) is utilizing a technique called MRI-guided radiation therapy. This method integrates daily MRI scans with radiation treatments. MRI technology illuminates the brain tumor, assisting in guiding the radiation beams. Importantly, the detailed images produced by MRIs also offer the possibility of near real-time monitoring of the tumor's response or progression. However, this advanced approach generates a substantial volume of data. For the 36 patients with glioblastoma in their study, each of the 31 time points included between four and six distinct images. To efficiently analyze this wealth of information, the team has employed artificial intelligence.
![Image: The researchers are using MRI-guided radiation therapy that pairs daily MRIs with radiation treatment (Photo courtesy of Sylvester) Image: The researchers are using MRI-guided radiation therapy that pairs daily MRIs with radiation treatment (Photo courtesy of Sylvester)](https://globetechcdn.com/mobile_medicalimaging/images/stories/articles/article_images/2023-11-16/SDD-294799325.jpg)
The AI-based solution developed by the researchers automatically delineates glioblastoma tumors and the resection cavities—spaces remaining after surgical removal of tumors—within these extensive MRI datasets. This automated tracing of tumors and cavities allows for tracking of tumor growth or reduction throughout the treatment. The algorithm, an adaptation from previous work in cervical cancer, can swiftly calculate the precise volume of the tumor and track changes over time. This AI method also offers a significant reduction in time compared to manual analysis, which can take over 20 hours per patient. The AI can process the same data in approximately 90 minutes.
Looking forward, the team plans to enhance the machine learning approach to include additional data from the MRI images. A key focus is identifying pseudo-progression, a condition where the tumor appears to grow due to treatment-induced swelling but ultimately recedes. This distinction between actual tumor growth and pseudo-progression is a crucial but challenging aspect of the research. The researchers are designing a study to evaluate tumor progression in glioblastoma patients undergoing MRI-guided radiation therapy on a weekly basis. They aim to adjust treatments in real-time based on the response of the tumors or changes in their size, utilizing the new machine learning method to facilitate swift treatment modifications.
“You can monitor so many different qualities of the tumor with MRI. That’s an untapped frontier,” said Adrian Breto, a doctoral student and programmer. “We haven’t gone yet to the center of the earth as far as what MRI can tell us about the patient’s disease and quality of life. That’s what we’re trying to do, squeeze as much information as we can out of these images for the benefit of the patient.”
Related Links:
Miller School of Medicine