About the project

The initial research phase focused on understanding the challenges associated with prostate biopsy procedures and identifying potential technological solutions. This involved a comprehensive review of existing biopsy techniques, imaging modalities, and robotic systems used in medical applications.

In order to integrate our system with existing clinical workflows, we analyzed current practices and identified key areas where robotic assistance could enhance precision and reduce patient discomfort. This phase also included consultations with medical professionals to gather insights and validate our approach.

Prostate phantoms used for testing and calibration

Additionally, we explored state-of-the-art AI techniques for medical imaging analysis, focusing on their potential to improve the detection and localization of suspicious prostate regions. This research laid the groundwork for the subsequent development phases of the project.

Overall, the initial research phase was crucial in shaping the direction of our project and establishing a solid foundation for the development of our robotic prostate biopsy system.

Most importantly, we assembled a multidisciplinary team with expertise in robotics, medical imaging, and clinical practice to drive the project forward. We planned for seamless clinical integration by engaging with healthcare professionals early in the process to ensure our solutions align with real-world needs.

Developing precise and reliable control algorithms for the robotic system to ensure accurate needle placement during the biopsy procedure. This includes kinematic modeling, motion planning, and control strategies to achieve the desired performance and safety standards.

A crucial aspect to consider is movement compensation. Since the prostate can shift due to patient movement or physiological factors, the robotic system must be capable of adapting in real-time to these changes. This involves integrating sensors and feedback mechanisms to monitor prostate position and adjust the robot's actions accordingly.

Robot control setup used during test biopsy procedures

Clinicians contribute valuable insights in this area to ensure that the robotic system can be seamlessly integrated into existing clinical workflows. This includes considerations for ease of use, safety, and compatibility with other medical equipment.

Currently, we use a force sensor and visual servoing to minimize the error caused by prostate motion during the biopsy procedure. This allows the robot to adjust its position in real-time, ensuring accurate needle placement despite any movement of the prostate.

Graph showing vertical motion compensation during test biopsy procedure

Our next step was to optimize robot position to meet the clinician's demands. As they traditionally work with a lowered setup, we lowered the prostate phantom to make our prototype easier to test and more realistic. This also improves the usability of the system in a clinical setting.

Improved robot setup

Integrating advanced imaging techniques such as micro-TRUS and AI algorithms to enhance the detection and localization of suspicious prostate regions. This includes developing machine learning models to analyze imaging data and assist in biopsy planning.

Most importantly, we use AI to segment the prostate and suspicious regions from micro-TRUS images. This allows us to create a detailed 3D model of the prostate, which is essential for accurate biopsy planning and needle guidance. This approach reduces the cognitive load on clinicians and improves the overall accuracy of the biopsy procedure.

Based on the MicroSegNet architecture, we developed a custom neural network tailored for the segmentation of prostate and suspicious regions from micro-TRUS images. This model was trained on a diverse dataset of annotated images to ensure robust performance across various clinical scenarios.

Neural network architecture used for prostate and suspicious region segmentation from micro TRUS images