Clinical Report: Detecting Tissue Structures Using Intraoperative Ultrasound
Overview
This report summarizes a novel method for detecting tissue structures and contact-related acoustic shadowing (AS) in intraoperative ultrasound (IOUS) images. The approach uses iterative Gaussian filtering and topological data analysis to generate confidence maps that quantify probe-tissue contact quality during surgery.
Background
Intraoperative ultrasound is a valuable imaging tool for surgeons, offering real-time feedback at low cost and ease of integration. However, its adoption is limited by operator variability and challenges in interpreting images due to artefacts such as acoustic shadowing caused by poor probe-tissue contact. Accurate detection of these artefacts is critical, especially in delicate surgeries like brain operations, to avoid tissue damage and improve surgical outcomes. Recent algorithmic advances have focused on shadow detection but often do not specifically address probe-tissue coupling, which this work aims to resolve.
Data Highlights
A dataset of 51 intraoperative ultrasound images from 11 patients undergoing brain surgery was used to evaluate the proposed method. Images were acquired with a Canon i900 ultrasound machine using linear and convex probes. A consultant neuroradiologist annotated the images to classify acoustic shadowing related to probe-tissue contact.
Key Findings
A novel method using iterative Gaussian filters with vertical bias effectively identifies visible tissue areas in ultrasound images by detecting high spatial intensity variation.
Topological representation of tissue features via Vietoris–Rips complexes enables robust detection of contact-related acoustic shadowing.
The method generates a confidence map quantifying perceptual salience, reflecting tissue contrast and resolution quality on a pixel-wise basis.
Contact-related acoustic shadowing is classified by analyzing the density of 2D simplices overlapping scan lines, providing a threshold-based detection approach.
The approach addresses limitations of previous methods that assumed hyperechoic lines correlate with signal attenuation, offering a more generalized and reliable assessment of probe-tissue contact.
Clinical Implications
This method provides surgeons with an objective, real-time assessment of probe-tissue contact quality during intraoperative ultrasound imaging, potentially reducing interpretation ambiguity and improving surgical decision-making. By detecting contact-related acoustic shadowing accurately, it may help prevent inadvertent tissue damage and optimize image reliability, especially in complex procedures like brain surgery.
Conclusion
The proposed topological and filtering framework advances intraoperative ultrasound imaging by enabling reliable detection of tissue structures and contact-related artefacts. This enhances confidence in ultrasound interpretation and supports safer, more effective surgical interventions.
References
Author/Source/Year -- Detecting Tissue Structures Using Intraoperative Ultrasound: Techniques and Their Utilization
