Poster Poster Program Therapy Physics

Tracking Lung Tumor Regression during Radiotherapy Using 5DCT-Driven Motion-Compensated CBCT

Abstract

Purpose

Standard Cone-Beam CT (CBCT) has limited utility in longitudinal lung tumor response assessment due to the presence of motion artifacts. Our proposed method of motion-compensated Simultaneous Algebraic Reconstruction Technique (mcSART) addresses these artifacts using a 5DCT-derived motion model. Here, we demonstrate that mcSART-reconstructed CBCT enables reliable daily tumor tracking over a full course of lung radiotherapy.

Methods

Daily CBCT projection data were acquired for a patient being treated for lung cancer with 30 fractions. CBCT Images were reconstructed using the mcSART workflow, which incorporates a patient-specific deformation vector field (DVF) derived from a planning 5DCT and a bellows breathing surrogate to track real-time breathing amplitude and rate. To evaluate motion model changes and determine if such rescans were necessary over a 30-fraction treatment course, 5DCT scans were repeated at fractions 12 and 22. Tumor response was quantified by measuring the longest axial diameter (RECIST 1.1) of the target volume on all 30 mcSART datasets using MIM software. These measurements were analyzed to track regression trends relative to the baseline established by the three planning 5DCTs.

Results

mcSART reconstruction successfully recovered tumor edges obscured by motion blur in standard reconstructions. The tumor diameter decreased from 69.6±0.75 mm at Fraction 1 to 56.0±0.75 mm at Fraction 30, representing a total reduction of 19.5% which was consistent with the regression observed in the three 5DCTs for the corresponding fractions. No significant deviations from the regression trend were observed among the three motion models.

Conclusion

The proposed mcSART process enables quantitative daily tumor tracking not feasible with standard clinical CBCT. Our method allows for the early detection of tumor regression, potentially informing adaptive replanning intervention. In future work, we will apply this process to multiple patients and use multiple observers and a more robust tumor measurement protocol to account for intra-observer variability.

People

Peter Boyle, MSAuthors · Department of Radiation Oncology, University of California, Los Angeles Ryan AndoscaPresenting Author · Department of Radiation Oncology, University of California, Los Angeles Claudia R. Miller, MS, BSAuthors · Department of Radiation Oncology, University of California, Los Angeles Michael Vincent Lauria, PhDAuthors · Department of Radiation Oncology, University of California, Los Angeles Drew Moghanaki, MD, MPHAuthors · Department of Radiation Oncology, University of California, Los Angeles Jack Neylon, PhDAuthors · Department of Radiation Oncology, University of California, Los Angeles Dylan P. O'ConnellAuthors · Department of Radiation Oncology, University of California, Los Angeles Daniel A. Low, PhDAuthors · Department of Radiation Oncology, University of California, Los Angeles