Introduction: Glioblastoma multiforme (GBM) poses a substantial challenge within the field of neuro-oncology as the most aggressive, malignant primary brain tumor, with overall patient survival around 9 months¹. Although treatments, which include combined surgical resection and chemoradiotherapy, are still largely palliative,^2,3 timely tumor resection and frequent monitoring is critical to improving patient quality of life. Even with advances in intraoperative neuronavigation and optical tumor paints, there are significant limitations to achieving precise delineation of tumor margins from normal parenchyma^4–7. It is also imperative to vigilantly monitor the patient for tumor recurrence. Magnetic resonance imaging (MRI) is the gold standard for diagnosis and postoperative surveillance, but reduced frequency of imaging poses a significant challenge to effective longitudinal tracking of tumor recurrence for patients who live in resource-constrained settings or are too clinically unstable to undergo an MRI. Ultrasound is a versatile point-of-care modality that provides fast, deep-tissue imaging with superb spatiotemporal resolution along with extensively documented patient safety data. Ultrasound can provide high resolution imaging intraoperatively, and offers lower-resolution detection across intact skull^8–10. We can utilize gas vesicles (GVs) – air-filled protein nanostructures – as an effective ultrasound contrast agent that generates robust nonlinear signal^11,12. Roughly an order of magnitude smaller than FDA-approved microbubbles, which are too large to exit vasculature,¹³ GVs are ideal candidates for extravasating out of leaky tumor vasculature akin to the dynamic contrast enhancement seen in small molecule MRI agents. In this study, we use gas vesicles to enhance ultrasound imaging for precise tumor delineation, intraoperative guidance, and longitudinal monitoring of glioblastoma multiforme (Fig.1).

Objective: This study introduces an ultrasound-based imaging technique that can be used for non-invasive brain tumor diagnosis, intraoperative resection guidance, and longitudinal postoperative recurrence monitoring. 

Methods and Results: We implanted mice intracranially with U87 or GL261 GBM tumor cells to establish orthotopic and syngeneic tumors, respectively. GVs were shown to selectively localize to the tumors with robust ultrasound signal following tail vein injection (Fig. 2A). We initially observed diffuse ultrasound signal across the entirety of the coronal plane during infusion, followed by a rapidly increased signal preferentially within the tumor with loss in the surrounding healthy parenchyma over the following 20 minutes (Fig. 2B). Histology demonstrated that GVs extravasate out of tumor neovasculature and are then degraded by tumor-associated macrophages. For our resection and recurrence model, we resected tumors using vacuum aspiration and sharp dissection, and the skull flap was replaced with a sonotransparent TPX window (Fig. 2C). Pre-resection and post-resection ultrasound and MRI were taken to show that GV-enhanced ultrasound can be effective at determining the extent of tumor resection (Fig. 2D). Over the next 10 – 14  days, GVs were used to visualize tumor recurrence and invasion into surrounding tissue (Fig. 2E). With sham negative controls (saline injection in place of tumor implant), we saw no GV accumulation in the resection cavity, demonstrating that GVs preferentially localize to neoplastic tissue.

Discussion: With a tool that can span initial diagnosis to resection guidance and recurrence monitoring, gas vesicle-based acoustic tumor paint presents a promising avenue to complement existing modalities, in both diagnostic and therapeutic interventions, for clinical management of GBM. This imaging paradigm has the potential to transform patient care and clinical outcomes by providing a rapid, non-invasive, and highly-accessible point-of-care tool for clinicians to use. We are now exploring its application in larger preclinical models.

Author