Challenges with segmenting intraoperative ultrasound for brain tumours - Scorecard - MDSpire

Challenges with segmenting intraoperative ultrasound for brain tumours

  • By

  • Alistair Weld

  • Luke Dixon

  • Giulio Anichini

  • Neekhil Patel

  • Amr Nimer

  • Michael Dyck

  • Kevin O’Neill

  • Adrian Lim

  • Stamatia Giannarou

  • Sophie Camp

  • August 1, 2024

  • 0 min

Share

Clinical Scorecard: Obstacles in the Segmentation of Intraoperative Ultrasound Images for Brain Tumor Surgery

At a Glance

CategoryDetail
ConditionBrain tumors requiring maximal-safe resection
Key MechanismsIntraoperative ultrasound (iUS) imaging for real-time tumor detection and delineation
Target PopulationPatients undergoing brain tumor surgery with intraoperative ultrasound guidance
Care SettingNeurosurgical operating room with intraoperative imaging

Key Highlights

  • iUS offers real-time tumor visualization integrated into surgical workflow and is more affordable than intraoperative MRI.
  • Challenges in iUS include limited field of view, artefacts, steep learning curve, and variability in tumor appearance and intraoperative changes.
  • Current evidence shows iUS has moderate sensitivity (72.2%) and high specificity (93.5%) for glioma resection assessment but requires accuracy improvements.

Guideline-Based Recommendations

Diagnosis

  • Use iUS co-registered with preoperative MRI/CT for intraoperative tumor boundary delineation.
  • Cross-reference iUS images with preoperative MRI to ensure accurate tumor boundary identification.

Management

  • Employ iUS to guide maximal-safe tumor resection to improve symptoms, quality of life, and survival.
  • Incorporate standardized training and new supporting techniques to reduce segmentation errors and improve iUS utility.

Monitoring & Follow-up

  • Assess tumor resection completeness intraoperatively using iUS with reference to postoperative MRI.
  • Monitor interobserver variability in tumor boundary segmentation to identify areas needing training or tool improvement.

Risks

  • Potential for incomplete tumor resection or inadvertent damage due to inaccurate tumor boundary detection on iUS.
  • Steep learning curve and variability in image interpretation may impair surgical outcomes.

Patient & Prescribing Data

Patients with brain tumors undergoing iUS-guided resection

iUS-guided resection achieves approximately 77% gross total resection rate, comparable to other navigation methods, but accuracy improvements are needed for standard care adoption.

Clinical Best Practices

  • Utilize experienced operators and standardized protocols for iUS image acquisition and interpretation.
  • Combine iUS with preoperative MRI for improved tumor boundary delineation.
  • Implement training programs to overcome the steep learning curve and improve segmentation consistency.
  • Explore simplified annotation methods such as bounding boxes to complement detailed segmentation and reduce variability.

References

Original Source(s)

Related Content