Surgeons faced with the daunting task of removing brain tumors encounter a challenging dilemma during their procedures: whether to sacrifice healthy brain tissue to ensure complete tumor removal or take a conservative approach, potentially leaving behind cancerous cells. However, researchers in the Netherlands have recently made a significant breakthrough by leveraging artificial intelligence to provide surgeons with crucial information that may assist in making this critical decision.
The Breakthrough
This groundbreaking method, as outlined in a study published in the prestigious journal Nature, involves a computer scanning segments of a tumor’s DNA and identifying specific chemical modifications. These modifications offer a comprehensive diagnosis of the brain tumor’s type and even its subtype, all in real-time during the early stages of surgery.
The Impact On Surgery
The real-time diagnosis provided by this innovative approach empowers surgeons to determine the appropriate level of aggressiveness required during surgery. Not only does this facilitate more precise tumor removal, but it also paves the way for tailored treatments for specific tumor subtypes in the future.
Why It Matters
According to Jeroen de Ridder, an associate professor at the Center for Molecular Medicine at UMC Utrecht in the Netherlands, knowing the tumor subtype during surgery is of utmost importance. The new method enables fine-grained, detailed diagnoses to be conducted during the surgery itself, providing a significant advantage.
The Deep Learning System
The artificial intelligence system at the core of this breakthrough is aptly named Sturgeon. Initially tested on frozen tumor samples from prior brain cancer surgeries, Sturgeon managed to accurately diagnose 45 out of 50 cases within just 40 minutes of genetic sequencing initiation. In the remaining five cases, the system refrained from offering a diagnosis due to insufficient information.
Subsequently, Sturgeon was put to the test during 25 live brain surgeries, most of which involved pediatric patients. These surgeries were conducted alongside the traditional method of examining tumor samples under a microscope. Impressively, the AI system delivered 18 correct diagnoses and met the required confidence threshold within 90 minutes. This quick turnaround allows it to inform surgical decisions while the operation is ongoing.
Challenges Addressed
Currently, doctors often rely on microscopic examination of brain tumor samples and, in some cases, send them for more extensive genetic sequencing. However, not every medical facility has access to this advanced technology, and even for those that do, the results may take several weeks to arrive. Dr. Alan Cohen, the director of the Johns Hopkins Division of Pediatric Neurosurgery, emphasized the significance of this innovation in reducing the time and resources required for tumor diagnosis and decision-making during surgery.
In conclusion, the integration of artificial intelligence in brain tumor surgery represents a significant advancement in the field of neurosurgery. By providing real-time, detailed diagnoses of brain tumors, surgeons can make more informed decisions during surgery, potentially leading to improved patient outcomes and a more efficient use of medical resources.