AI-powered 3D visualisation that helps surgeons plan complex procedures faster, safer, and with greater precision.
OneBonsai developed a proof-of-concept 3D visualization tool to support neurosurgeons in planning brain tumor surgeries. Using AI-powered segmentation of CT scans, the system generates an interactive 3D model of the tumor and surrounding anatomy, which can be accessed securely on a tablet. This intuitive approach simplifies preoperative planning, reduces preparation time, and enhances spatial understanding of tumor morphology, giving surgeons clearer insights to make safer and more precise decisions.
Simplifying 3D Tumor Visualization for Surgeons
Intuitive 3D Tumor Visualization Tool
Collaboration with Medical Experts and Advanced Tech
A Promising Proof of Concept
Neurosurgeons face some of the most complex challenges in modern medicine: planning precise and safe interventions for brain tumor patients. Traditionally, preoperative planning relied heavily on CT scans and 2D imaging tools. While effective to an extent, these methods required surgeons to mentally reconstruct three-dimensional structures from flat slices, which was time-consuming and left room for interpretation errors. Existing 3D segmentation software was often overly complex, slow, and not designed for the specific workflows of surgeons, making it impractical for day-to-day use in clinical settings. This created a bottleneck: surgeons needed a tool that was both accurate and intuitive, allowing them to visualize tumors in their full context without adding unnecessary complexity. The challenge was clear: how to give neurosurgeons fast, reliable, and accessible 3D visualizations that could improve planning, reduce preparation time, and ultimately contribute to safer surgical outcomes.
“Relying only on 2D CT scans made it difficult for surgeons to clearly understand tumor size and position before operating.”
To address this, OneBonsai designed a proof-of-concept system that transforms standard CT scans into interactive 3D models of the skull, brain, and tumor. The core of the solution is an AI-driven segmentation algorithm that isolates tumor tissue with high accuracy, creating a clear distinction between healthy and affected regions. The resulting model is rendered into a user-friendly 3D visualization that can be manipulated on a tablet or workstation. Surgeons are able to rotate, zoom, and interact with the model in real time, giving them a precise view of tumor morphology and its relation to surrounding anatomy. With this tool, what once required lengthy preparation and mental visualization can now be achieved in minutes. The solution offers surgeons not only speed but also clarity, giving them greater confidence as they plan critical interventions.
The development approach focused on usability and clinical relevance. OneBonsai collaborated with neurosurgeons to deeply understand their workflows, challenges, and decision-making processes. Instead of building another complex research tool, the team created a streamlined interface designed for quick and intuitive use in a busy clinical environment. The AI segmentation was locally deployed on secure GPU servers to ensure both data privacy and high performance, with visualization streamed to the surgeon’s device over a reliable 5G connection. This eliminated the need for heavy local installations or time-consuming data transfers, making the system lightweight and accessible. Furthermore, we trained the model on patient data to improve the accuracy of the segmentation. The design philosophy was to remove friction: give surgeons only the tools they need, cut out unnecessary complexity, and make the process as fast and natural as possible. The system thus combined cutting-edge AI technology with human-centered design, tailored specifically for real-world surgical planning.
“Reviewing a tumor in 3D can now be done quickly and easily, further improving patient care.”
The proof-of-concept has already shown promising results. Surgeons using the tool can achieve accurate tumor visualization and determine optimal entry angles in under three minutes, compared to much longer preparation times with traditional methods. The simplified interface, paired with high-quality AI segmentation, can reduce cognitive load and allow surgeons to focus on strategy rather than software navigation. While still in controlled testing, the tool has demonstrated its potential to significantly improve preoperative planning, leading to safer surgeries and better patient outcomes. For hospitals, the solution also offers efficiency gains by reducing planning time and standardizing workflows. Looking forward, the system provides a foundation for broader deployment and integration into neurosurgical practices, marking an important step toward AI-augmented precision medicine.
3D Pre-Operative Visualisation Tool
3D Pre-Operative Visualisation Tool
“The tool saves time in preparation and gives our surgeons a new level of precision.”