Interactive Shape Sonification for Tumor Localization in Breast Cancer Surgery

Title of the Paper

Interactive Shape Sonification for Tumor Localization in Breast Cancer Surgery

Paper Information

• Subject Area: Human-Computer Interaction, Medical Assistive Technology, Breast Cancer Surgery Guidance
• Keywords: Sonification, Breast Cancer Localization, Surgical Sonification, Shape Sonification, Augmented Reality, Auditory Display, Breast Cancer Surgery, Surgical Navigation System, Human-Computer Interaction, Tumor Boundary

Research Background and Problem

• Problem or Challenge:

  • Currently, 20% of patients undergoing breast-conserving surgery for breast cancer require reoperation due to residual cancerous tissue in the breast.
  • Existing breast cancer localization systems use auditory feedback to indicate the position of seeds (markers implanted preoperatively in the tumor) but fail to provide information about the tumor boundary's location and shape.
  • During surgery, the variability of breast tissue can lead to discrepancies in tumor position and size inferred from preoperative imaging.

• Significance: Accurate localization of tumor boundaries can reduce postoperative recurrence rates and avoid unnecessary removal of healthy tissue, thereby improving patient outcomes and cosmetic results.

• Related Work:

  • Current tumor localization technologies (e.g., Savi Scout®) only provide auditory feedback for seed positions.
  • Augmented reality surgical navigation systems have been explored for breast tumor localization but have not yet achieved the required accuracy.
  • Research on auditory substitution and related shape sonification has demonstrated that auditory information can be used for shape recognition and position guidance.

Solution

• Method and Innovation:

  • The authors propose an interactive auditory display based on shape sonification to accurately localize tumor boundaries and seed positions during breast cancer surgery.
  • A multi-parameter auditory mapping is provided, encoding tumor boundary (shape information) and seed position (point location) into intuitive auditory feedback.
  • Tumor contour information is integrated into the surgical sonification navigation system to enhance localization precision.

• Implementation Steps and Techniques:

  • An electromagnetic tracking system and sound synthesis software are used to generate real-time auditory feedback.
  • The sonification design includes the following models:
    • Beep: Represents the current clinical auditory feedback, encoding only seed position.
    • Rhythm: Uses varying rhythmic sounds to represent tumor boundaries and seed positions.
    • Synth and Sine: Provide continuous and distinguishable sounds for identifying tumor boundaries and seed positions.
  • Real-time mapping of tumor shapes and auditory parameters is implemented in Unity software.
  • Three sets of experiments are designed to evaluate the effectiveness of sonification in simulated breast surgery scenarios.

Research Outcomes

• Specific Outcomes:

  • Experiments show that shape sonification significantly improves the accuracy of tumor boundary and seed position localization.
  • The sonification scheme achieved better user-perceived usability, particularly receiving high evaluations from breast surgeons.
  • After using shape sonification, the overlap (Dice coefficient) of tumor boundaries and seed positions improved from 0.6 to 0.74, significantly outperforming traditional auditory feedback methods.

• Advantages and Comparisons:

  • Compared to traditional auditory feedback, shape sonification provides additional tumor boundary information, reducing errors caused by reliance on preoperative medical imaging.
  • The sonification model achieves higher localization accuracy than augmented reality guidance technology (average Dice coefficient 0.74 vs. 0.2 for AR guidance).
  • Breast surgeons reported that shape sonification significantly enhanced their confidence and satisfaction during operations.

• Experimental or Evaluation Results:

  • Experimental Design:
    • Two-dimensional evaluations using cardboard models.
    • Simulated real surgical scenarios using silicone breast models and agar breast models.
    • Three user studies involving non-clinical participants, medical professionals, and breast surgeons.
  • Result Data:
    • Dice coefficient improved from 0.57 to 0.74.
    • System usability scores increased from 65.63 to 85.63.

• Limitations and Future Directions:

  • Current experiments cannot fully simulate the elasticity and complex deformation characteristics of breast tissue.
  • Tumor depth and directional information are not integrated into the sonification design.
  • Future work should focus on reducing system latency and designing convenient marking and tracking methods to improve precision.
  • Further research is needed on the impact of sonification design for visually impaired users and its application in multi-sensory interaction environments.

This study demonstrates the potential of shape sonification in breast cancer localization tasks, providing precise guidance that conventional visual feedback cannot replace. This work not only offers valuable insights for the field of surgical navigation but also opens new avenues for innovation in multimodal interaction technologies.