A critical challenge in glioma treatment is detecting tumour infiltration during surgery to achieve safe maximal resection. Unfortunately, safely resectable residual tumour is found in the majority of patients with glioma after surgery, causing early recurrence and decreased survival.
The article present FastGlioma, a visual foundation model for fast (<10 s) and accurate detection of glioma infiltration in fresh, unprocessed surgical tissue. In a head-to-head, prospective study (n = 129), the FastGlioma study arm had a 3.8% false-negative/tumour miss rate compared with 24.0% in the standard-of-care surgical adjuncts arm. The results indicate a potential 6.3× decrease in the relative risk of residual tumour within resection cavities by using FastGlioma to guide tumour resections.
The performance of FastGlioma remained high across diverse patient demographics, medical centres and diffuse glioma molecular subtypes as defined by the World Health Organization.
All code and scripts to reproduce the main experiments of this paper are available at GitHub (https://github.com/MLNeurosurg/fastglioma) under an MIT license.
【MORE】