Lung tumor tracking in fluoroscopic video based on optical flow

Area of Science:

• Medical Physics
• Radiotherapy Technology
• Image-Guided Interventions

Background:

• Accurate real-time lung tumor localization is crucial for dynamic radiotherapy delivery.
• External surrogates for tumor tracking have significant uncertainties.
• Implanted fiducial markers offer accuracy but pose practical challenges.

Purpose of the Study:

• To develop and validate a non-invasive method for real-time lung tumor tracking using optical flow algorithms.
• To enable precise tumor position determination without implanted fiducial markers.

Main Methods:

• An optical flow algorithm was employed to track lung tumor projections in fluoroscopic images.
• Tumor centroid displacement was estimated using motion vectors.
• Template matching refined contour positions for accurate localization.

Main Results:

• The optical flow method achieved high accuracy in tumor tracking.
• Position differences were less than 1.4 mm in 95% of frames in the best case.
• The method demonstrated feasibility in a small patient cohort.

Conclusions:

• This non-invasive optical flow technique provides accurate real-time lung tumor tracking for radiotherapy.
• It offers a practical alternative to fiducial marker implantation.
• Further validation in larger studies is warranted.