A Bayesian approach for three-dimensional markerless tumor tracking using kV imaging during lung radiotherapy

Area of Science:

• Medical Physics
• Radiotherapy Technology
• Image-Guided Therapy

Background:

• Accurate lung tumor motion monitoring is crucial for precise radiotherapy.
• Current kilovoltage (kV) imaging methods struggle with continuous tumor tracking due to poor visibility and 2D to 3D estimation challenges.
• Markerless tumor tracking offers potential for broader access to advanced radiotherapy techniques.

Purpose of the Study:

• To develop and validate a Bayesian approach for markerless tumor tracking in lung radiotherapy.
• To address the limitations of current kV imaging in continuous tumor monitoring and 3D position estimation.
• To demonstrate the clinical feasibility of a novel markerless tumor tracking system.

Main Methods:

• Proposed a Bayesian approach using an extended Kalman filter for markerless tumor tracking.
• Combined a respiratory-correlated model for tumor position prediction with template matching for 2D measurement.
• Validated the method retrospectively on 13 lung cancer patient datasets, comparing against implanted marker ground truths.

Main Results:

• The markerless tumor tracking method demonstrated reduced 3D tracking errors compared to the standard of care.
• Mean, standard deviation, and 95th percentile errors ranged from 1.6-2.9 mm, 0.6-1.5 mm, and 2.6-5.8 mm, respectively.
• Achieved 96.5% of errors with 3D magnitude <5 mm, significantly outperforming the standard of care (84.1%).

Conclusions:

• The study successfully demonstrated the clinical feasibility of 3D markerless tumor tracking in realistic scenarios.
• This novel approach enables more accurate and precise lung radiotherapy using existing infrastructure.
• Future work will focus on clinical implementation and real-time application of the markerless tracking system.