Shaping the future of molecular neurosurgery: Toward epigenetic precision in surgical neuro-oncology?

Surgery represents the first clinical step of a glioblastoma patient’s way through therapy.¹ The resected tissue not only allows to establish a neuropathological diagnosis, but more extensive resection also translates into clinically meaningful survival benefits.² Driven by a series of prominent articles published within the last 5 years,^3–5 surgical approaches emerged that move the resection target from the contrast-enhancing tumor core toward the noncontrast-enhancing tumor portion. The oncological value of such ever-increasing extents of resection in patients with glioblastoma reaches a limit when surgery results in diminished functional outcomes, particularly as resection can never cure glioblastoma given the infiltrative nature of the disease. Neurosurgeons are, therefore, regularly facing a delicate dilemma: Is it oncologically worth hunting down each small tumor remnant during an operation, even when exposing the patient to the unavoidable risk inherent to operations in functionally relevant areas? Drexler and Khatri with colleagues now published in Nature Medicine their translational findings regarding an association of different epigenetic tumor signatures with varying oncological values concerning the extent of resection.⁶ The authors conclude that their contribution might be suitable to identify patients for whom it appears even more justified to pursue a more aggressive resection.

Based on the presence of a neural epigenetic fingerprint,⁷ Drexler and Khatri retrospectively classified patients with newly diagnosed glioblastoma according to their epigenetic tumor signature into high- or low-neural tumors. Not only did high-neural glioblastomas exhibit upregulation of synapse-related genes and enrichment of intratumoral malignant stem-cell-like cells, but they were also characterized by a peculiar spatial microenvironment which was accompanied by increased neuron-to-glioma synapse formation in mice and higher functional peritumoral connectivity in patients. In line with a more profound neural–tumor interaction in tumors designated as high-neural glioblastomas, affected patients had less favorable outcomes than patients with low-neural tumors. Using plasma samples from affected patients, DNA analytes and brain-derived neurotrophic factors pointed toward the presence of a high-neural tumor phenotype. Highly emphasized was their finding that only complete resection of the contrast-enhancing tumor (in their definition equivalent to ≤1 cm³ residual contrast-enhancing tumor volume, ie, RANO class 2) was associated with prolonged survival in high-neural glioblastoma, while less extensive resection of 90%–99% of the contrast enhancement (or ≤5 cm³ residual contrast-enhancing tumor volume, ie, RANO class 3A) was sufficient to be associated with improved survival in low-neural glioblastoma. Once the respective threshold for extent of resection has been met in patients with high- or low-neural glioblastoma, a stepwise increase in median overall survival was seen for more extensive surgery; and patients undergoing a “supramaximal” resection beyond the contrast-enhancing tumor borders had most favorable outcome.

It is tempting to envision a world in which such (epi-)genetic details on biological glioblastoma properties are readily available to the neurosurgeon as early as during the operation as a tool to guide judicious decision-making on whether the patient is more likely to benefit from more extensive resection. In turn, such information might not only be used to confirm the rationale for aggressive resection but also to highlight cases in which a more cautious surgical approach (or even a biopsy) should be excelled when a complete resection cannot be reasonably achieved due to critical white matter tracts or highly functional cortical areas constraining the resection cavity. The introduction of intraoperative sequencing methods allowing near real-time intraoperative subclassification of brain tumors represents the first step into a clinical reality where neurosurgical decision-making is assisted by tissue sampled at early stages during the resection.⁸ Once an accurate intraoperative diagnosis is reliably established, the next step might be to leverage the tumor diagnosis through the intraoperative availability of off-the-shelf molecular diagnostics including the detection of IDH mutations or the prediction of the MGMT promoter methylation status. This is of particular relevance given that the expected benefits of more aggressive resection substantially vary between tumors depending on the presence of a methylated MGMT promoter, with survival benefits of lower residual tumor remnants being rather uniformly shown in unmethylated glioblastoma while conflicting evidence has been produced for methylated tumors.⁹ Confirmation of the interactive effects between the oncological value of resection and molecular tumor features in large prospective cohorts will be required to convince neurosurgeons to adapt their surgical strategy based on intraoperative diagnostics. A widespread implementation of developments enabling intraoperative standard diagnostics across academic institutions will probably precede the implementation of the current findings from Drexler and Khatri into clinical practice.

A first glimpse of how intraoperative diagnostics may influence surgical decision-making during glioblastoma resection is illustrated by a recent study from Massaad and colleagues published in Neuro-Oncology.¹⁰ Based on samples obtained from the infiltration zone of glioblastomas undergoing complete resection of the contrast enhancement, tumor cell burden was quantified by detection of a TERT promoter mutation, and patients with clean “molecular margins” had improved local tumor control. At present, however, such an approach as well as the findings from Drexler and Khatri must certainly be viewed as experimental and application should be limited to prospective trials given both studies rest upon rather small sample sizes. While it is, therefore, to be assumed that the observations from Drexler and Khatri are not yet practice changing, the idea behind their study indeed is; and the message is clear: We will be witnessing a future of precision neuro-oncology also from a surgical perspective, and this future is coming soon.

Funding

Nothing to report.

Acknowledgments

The text is the sole product of the authors and no third party had input or gave support to its writing.

Conflict of interest statement

P.K.—none. J.-C.T.—Consultant: Novartis, Servier.

Authorship statement

Study concept and design: P.K. Data collection: P.K., J.C.T. Data analysis and interpretation: P.K., J.C.T. Manuscript drafting: P.K., J.C.T. Manuscript revising: P.K., J.C.T.

References