A groundbreaking tool that combines rapid DNA sequencing with artificial intelligence (AI) is revolutionizing the classification of brain tumours during surgery. Developed by a team of researchers from UMC Utrecht, Amsterdam UMC, and the Princess Máxima Center for Pediatric Oncology in the Netherlands, the tool, named Sturgeon, enables neurosurgeons to make more informed decisions about tumour resection while the surgery is ongoing.
Traditionally, neurosurgeons have had limited knowledge of the tumour type prior to surgery. During the operation, sections of tumour tissue are removed for histological assessment, which typically takes a week to provide a definitive diagnosis. However, Sturgeon can accurately diagnose most central nervous system (CNS) tumours within just 90 minutes, empowering surgeons to modify their surgical strategies in real-time based on the tumour classification.
The tool utilizes rapid nanopore sequencing technology to obtain a sparse methylation profile during surgery. Methylation patterns are unique DNA modifications that can differentiate between different tumour types. Sturgeon employs a neural network classifier that is patient-agnostic and runs on a laptop computer in a matter of seconds.
In a Nature publication, the researchers described how they created, trained, and tested the Sturgeon tool. During 25 surgeries, Sturgeon accurately classified 72% of tumours in less than 45 minutes. The team simulated thousands of unique nanopore sequencing experiments from each tumour methylation profile to develop and fine-tune the Sturgeon models, which were ultimately trained on 36.8 million simulated nanopore runs.
While Sturgeon has shown promising results, it does have some limitations. It performs best on samples that are well-represented in the training data and is less effective when analyzing samples with less than 50% abnormal cells. Additionally, it requires relatively large tissue samples to provide sufficient DNA concentration.
In the future, the researchers plan to expand the application of Sturgeon to other tumour types, such as sarcoma or leukemia, and validate its effectiveness in larger patient populations. This groundbreaking tool has the potential to significantly improve surgical outcomes for brain tumour patients by enabling more informed decision-making during surgery.
How does Sturgeon work?
Sturgeon utilizes rapid nanopore sequencing technology to obtain a sparse methylation profile during brain cancer surgery. Methylation patterns, which are unique DNA modifications, can differentiate between different types of brain tumours. Sturgeon employs a neural network classifier that can accurately diagnose most central nervous system (CNS) tumours within 90 minutes.
What are the benefits of using Sturgeon?
Sturgeon allows neurosurgeons to make more informed decisions about tumour resection during brain cancer surgery. By providing rapid tumour classification, surgeons can adjust their surgical strategies in real-time, potentially improving patient outcomes and minimizing the need for additional surgeries.
Are there any limitations to Sturgeon?
Sturgeon performs best on samples that are well-represented in the training data and may not perform as well on rare types of CNS tumours. It is also less effective when analyzing samples containing less than 50% abnormal cells. Additionally, relatively large tissue samples are required to provide sufficient DNA concentration for accurate classification.