This repository contains a prototype app for breast biopsy planning with precise needle navigation and augmented visualization based guidance for the biopsy needle. This prototype was developed in 3 day Healthcare Hackathon Bayern.
Innovating Mammography – Enhancing Precision and Comfort in Needle Insertion for Biopsy by Siemens Healthineers
In the realm of mammography, precise needle insertion is crucial for both diagnostic and therapeutic procedures. However, the current process can be challenging, leading to discomfort and anxiety for patients. We challenge you to transform the mammography journey into a more comfortable and reassuring experience for women. How can it be less daunting and what innovative solutions can be employed to ensure a smoother, more supportive experience for patients?
- Targeting lessions accurately is difficult from 2D scans as they lack the 3D perception.
- Most of the time is spent over referencing a scan while putting the needle inside, resulting into more time needed for the biopsy procedure which is painful for the patient.
- Patient movement might cause issues in needle guidance through prior scans.
- Needle might pass through critical structures like blood vessels.
- There aren't much navigation system available for breast biopsy, as it is a soft tissue organ which is makes it difficult for navigation systems to work due to the ability of the breast to get compressed, resulting into change in position of the lesion.
We proposed a solution to the above pain points through an application for accurate biopsy needle guidance. The application works in two stages-
- A needle navigation and planning module
- A augmented reality viewer to visualize a digital phantom over physical anatomy, to see the needle movement without continous image guidance.
The needle paths are calculated automatically and a feedback mechanism is established via visual and auditory feedback to let the user know when to stop the needle to avoid excessive penetration and further trauma.
- Streamline 2D to 3D Mesh Model
- Xray/Mammogram to 3D using generative models.
- Statistical shape modelling.
- 3D reconstruction using a reference mesh of 3D breast and scaling with respect to measured anterio-posterior, latero-medial and superior-inferior dimensions of the breast through a digital image.
- Fast and continuous 3D registration for the AR App using detected landmarks.
- Simulating breast deformation on the mesh if the patient moves, by tracking relative position change within each landmark.
- Integrate the app in Microsoft HoloLens or Apple Vision Pro for more seamless experience with the navigation system.