Tissue characterization based on the analysis on i3DUS data for diagnosis support in neurosurgery

TL;DR

This paper presents a novel computer-aided diagnosis system using a mixed-attention Res-U-net architecture with an asymmetric loss function for tissue characterization in intraoperative ultrasound images, improving accuracy over existing methods.

Contribution

The study introduces a new deep learning model that outperforms current pixel-level classification methods for ultrasound tissue analysis during neurosurgery.

Findings

Achieved state-of-the-art classification accuracy on ultrasound data

Outperformed U-net and FCN in all evaluation metrics

Provides a robust second opinion for intraoperative diagnosis

Abstract

Brain shift makes the pre-operative MRI navigation highly inaccurate hence the intraoperative modalities are adopted in surgical theatre. Due to the excellent economic and portability merits, the Ultrasound imaging is used at our collaborating hospital, Charing Cross Hospital, Imperial College London, UK. However, it is found that intraoperative diagnosis on Ultrasound images is not straightforward and consistent, even for very experienced clinical experts. Hence, there is a demand to design a Computer-aided-diagnosis system to provide a robust second opinion to help the surgeons. The proposed CAD system based on "Mixed-Attention Res-U-net with asymmetric loss function" achieves the state-of-the-art results comparing to the ground truth by classification at pixel-level directly, it also outperforms all the current main stream pixel-level classification methods (e.g. U-net, FCN) in all the evaluation metrices.