Rapid DNA sequencing and in situ digital image processing could change the future of intraoperative anatomic pathology procedures
Researchers at the Center for Molecular Medicine (CMM) at UMC Utrecht, a leading international university medical center in the Netherlands, are combining artificial intelligence (AI) and machine learning with DNA sequencing to help cancer surgeons diagnose cancer during surgery. We have developed a diagnostic tool that can be used to diagnose. Verification that all cancerous tissue has been completely removed can be determined within minutes while the patient is still on the operating table.
The method “involves a computer scanning segments of a tumor's DNA for specific chemical modifications that may lead to a detailed diagnosis of brain tumor type or subtype.” new york timesfurthermore, “The diagnosis generated during the early stages of a surgery lasting several hours helps the surgeon decide how aggressively to operate. … In the future, this method could be used to identify tumors It may also help guide doctors to tailor treatment to different subtypes.”
With further development and research, this technology has the potential to reduce the need for frozen sectioning if it is confirmed to accurately and reliably indicate to surgeons that all cancer cells have been completely removed. Many anatomic pathologists would welcome such a development given the time pressure and stress associated with this procedure. The pathologist knows that the patient is still in surgery and the surgeon is waiting for frozen section results. Most pathologists would believe that fewer frozen sections would improve patient outcomes and improve patient care.
Scientists from UMC Utrecht publish their research results in a magazine Nature The title is “Ultra-fast deep learning CNS tumor classification during surgery.”
“It is essential to know the subtype of the tumor at the time of surgery,” Dr. Jeroen de Ridder (above), associate professor at the UMC Utrecht Center for Molecular Medicine and one of the study leaders, told The New York Times. “What we are uniquely able to do is to be able to perform this very fine-grained, robust and detailed diagnosis already during surgery. “We can do a lot of classification,” he added. It remains to be seen how this discovery will impact the role of anatomic pathologists and pathology laboratories during cancer surgery. (Photo credit: UMC Utrecht)
Rapid DNA sequencing impacts brain tumor surgery
Scientists at UMC Utrecht will develop Oxford Nanopore's “addressing the challenges posed by central nervous system (CNS) tumors, one of the most lethal tumors, particularly in children,” according to an Oxford Nanopore news release. For this purpose, we employed Oxford Nanopore's real-time DNA sequencing technology.”
Researchers have named the new machine learning AI application “Sturgeon.”
according to new york times“The new method uses faster gene sequencing technology and applies it to only a small slice of a cell's genome, allowing results to be returned before surgeons start operating on the edge of the tumor.”
Dr. Jeroen de Ridder, Associate Professor at UMC Utrecht Center for Molecular Medicine, said: new york times “It's powerful enough to make diagnoses using sparse genetic data, similar to someone recognizing an image from an unknown part of the image based on just 1% of the pixels,” Sturgeon said. ” Ridder is also a principal investigator at the Oncode Institute, an independent research center in the Netherlands.
The researchers examined the sturgeon during 25 live brain surgeries and compared the results to an anatomical pathologist's standard microscopic tissue examination method. “While the new approach yielded 18 correct diagnoses, seven others did not reach the required confidence threshold. The study reports that the diagnosis can be reversed in less than 90 minutes and surgery It was short enough to inform internal decisions.” new york times report.
But there were also problems. When tiny samples contain healthy brain tissue, identifying the right number of tumor markers can be a problem. Under these circumstances, the surgeon can ask the anatomic pathologist to “raise the flag.” [tissue samples] “This tumor is the most commonly sequenced tumor,” said Marc Pagès Gallego, a doctoral candidate and bioinformatician at UMC Utrecht and co-author of the study. new york times I got it.
“Implementation itself is not as simple as it is often made out to be,” said Sebastian Brandner, MD, professor of neuropathology at University College London. times. “Sequencing and classifying tumor cells often still requires advanced bioinformatics expertise and workers who can implement, troubleshoot, and repair the technology,” he added.
“Brain tumors are also best classified by chemical modifications analyzed using new methods. Not all cancers can be diagnosed that way.” times It pointed out.
In this way, research continues. This new method has also been applied to other surgical samples. The study authors said other institutions have also used the method on their own surgical tissue samples, “suggesting that it can be applied to other people's hands.” But more work is needed, so times report.
UMC Utrecht researchers receive Hanars grant
To expand research into the capabilities of sturgeon, the UMC Utrecht research team recently established a research team founded in 2018 to “promote and enhance the use of artificial intelligence and machine learning to improve diagnosis, treatment, and outcomes in sturgeon.” received funding from the Hanars Foundation. cancer patients,” the organization's website states.
A UMC Utrecht news release states that researchers will study how to use the Sturgeon AI algorithm to identify tumors in the central nervous system during surgery. Researchers say this type of tumor is difficult to test without surgery.
“This poses a challenge for neurosurgeons. They have to operate on tumors without knowing what type of tumor they are. As a result, patients are at risk of needing another surgery. “There is a strong sense of gender,'' De Ridder said in a news release.
The Sturgeon application solves this problem. It identifies the “exact type of tumor” during surgery. This allows us to immediately apply the appropriate surgical strategy,” the news release states.
Hanars' funding will allow Jeroen and his team to develop a sturgeon variant that “uses cerebrospinal fluid instead of (part of) the tumor.” This allows the type of tumor to be determined already before surgery. The main challenge is that cerebrospinal fluid contains a mixture of tumor and normal DNA. AI models will be trained with this in mind. ”
The breakthrough by UMC Utrecht scientists is another example of how organizations and research groups are working to reduce time to answer compared to standard anatomical pathology methods. Here are two examples. They are combining the technologies they are developing in ways that achieve these goals.
—Christine Althea O'Connor
Related information:
CNS tumor classification using ultra-fast deep learning during surgery
New AI tool diagnoses brain tumors on the operating table
Types of pediatric brain tumors discovered during surgery using nanopore sequencing and AI
AI speeds up identification of brain tumor types
Four new cancer research projects at UMC Utrecht receive Hanars grants
Rapid nanopore sequencing and machine learning enable intraoperative tumor classification
