Search for a command to run...
Maximal safe surgical resection is a foundational principle in brain tumor surgery. To date, many intraoperative modalities have been developed to help facilitate the identification of brain tumor versus normal brain tissue so that surgical resection is maximized but limited to the boundaries of the tumor for preservation of neurological function. Of note, Raman spectroscopy has been adapted into one of these modalities because of its ability to provide rapid, non-destructive, label-free intraoperative evaluation of tumor borders and molecular classifications and help guide surgical decision-making in real time. In this review, we performed a literature review of the landmark studies incorporating Raman spectroscopy into neurosurgical care to highlight its current applications and limitations. In this modern day, Raman spectroscopy is able to detect tumor cells intraoperatively for primary glial neoplasms, meningiomas, and brain metastases with greater than 90% accuracy. For glioma surgery, a major recent advancement is the ability to detect different mutations intraoperatively, specifically IDH, 1p19q co-deletion, and ATRX, given their implications on survival and how much extent of resection should be ideally achieved. With recent advancements in artificial intelligence and their integration into stimulated Raman histology, many of these tasks can be completed in as fast as ~10 s and on average 2-3 min. Despite the incorporation of artificial intelligence, spectral data can still be heavily influenced by background noise, and its preprocessing has significant variability across platforms, which can impact the accuracy of results. Overall, Raman spectroscopy has significantly changed the intraoperative workflow of brain tumor surgery, and this review highlights the capabilities that neurosurgeons can currently take advantage of in their practice, the existing data to support it, and the areas that researchers can further optimize to improve accuracy and patient outcomes.