Encrypted login | home

Program Information

Theoretical Estimation of Margin Necessary for Markerless Motion Tracking


R Patel

R Patel*, A Block , M Harkenrider , J Roeske , Loyola Univ Medical Center, Maywood, IL

Presentations

SU-E-J-188 (Sunday, July 12, 2015) 3:00 PM - 6:00 PM Room: Exhibit Hall


Purpose:To estimate the margin necessary to adequately cover the target using markerless motion tracking (MMT) of lung lesions given the uncertainty in tracking and the size of the target.

Methods:Simulations were developed in Matlab to determine the effect of tumor size and tracking uncertainty on the margin necessary to achieve adequate coverage of the target. For simplicity, the lung tumor was approximated by a circle on a 2D radiograph. The tumor was varied in size from a diameter of 0.1 – 30 mm in increments of 0.1 mm. From our previous studies using dual energy markerless motion tracking, we estimated tracking uncertainties in x and y to have a standard deviation of 2 mm. A Gaussian was used to simulate the deviation between the tracked location and true target location. For each size tumor, 100,000 deviations were randomly generated, the margin necessary to achieve at least 95% coverage 95% of the time was recorded. Additional simulations were run for varying uncertainties to demonstrate the effect of the tracking accuracy on the margin size.

Results:The simulations showed an inverse relationship between tumor size and margin necessary to achieve 95% coverage 95% of the time using the MMT technique. The margin decreased exponentially with target size. An increase in tracking accuracy expectedly showed a decrease in margin size as well.

Conclusion:In our clinic a 5 mm expansion of the internal target volume (ITV) is used to define the planning target volume (PTV). These simulations show that for tracking accuracies in x and y better than 2 mm, the margin required is less than 5 mm. This simple simulation can provide physicians with a guideline estimation for the margin necessary for use of MMT clinically based on the accuracy of their tracking and the size of the tumor.


Contact Email: