Author: B. Gino Fallone, Gawon Han, Keith D. Wachowicz, Mark G. Wright, Eugene Yip, Jihyun Yun 👨🔬
Affiliation: Medical Physics Division, Department of Oncology, University of Alberta, Dept. of Medical Physics, Cross Cancer Institute and Dept. of Oncology, University of Alberta, Medical Physics Division, Department of Oncology, University of Alberta and Department of Medical Physics, Cross Cancer Institute, Dept. of Medical Physics, Cross Cancer Institute and Dept. of Oncology, University of Alberta; MagnetTx Oncology Solutions, www.magnetTX.com 🌍
Purpose: MRI-radiotherapy hybrid systems can guide the therapeutic beam, dynamically adjusting to a moving tumor in real-time. However, there is a time delay from imaging and beam control, requiring prediction of the tumor’s future position and shape for accurate targeting. We aim to predict intrafractional MR images 1–2 frames beyond the current image.
Methods: Sagittal dynamic images from 10 lung and 10 liver patients were acquired using a 3T MRI with a 2D bSSFP protocol at 4 frames/s (FOV of 40x40cm, resolution of 128x128). A window of past 60 frames was selected to characterize recent variations using principal component analysis (PCA). PCA was applied in k-space, and the 16 most significant principal components (PCs) were used for prediction. The temporal trajectory formed by the PC scores was extrapolated independently to 1 and 2 frames beyond the window, using auto-regression. Using the 16 extrapolated scores as weights, the PCs were combined to generate predictive images. Image-based metrics including structural similarity (SSIM) were used to assess accuracy. The same auto-contouring technique was applied to both predicted and acquired images, and the consistency of tumor position and shape was measured using Dice coefficient (DC).
Results: The predicted images compared well against acquired images corresponding to the same time point, with an average SSIM of 0.93 for 1-frame, and 0.92 for 2-frame predictions for both liver and lung. The contour analysis yielded agreement with corresponding acquired images for 1- and 2-frame predictions, with average DC of 0.94 and 0.90 for liver, and 0.91 and 0.88 for lung. Average prediction time was 49 ms/frame.
Conclusion: The presented methodology offers a computationally efficient and accurate means of predicting images in real-time for radiotherapy guidance. The predicted contours have a sufficient level of consistency to consider for use in directing MLC motions in advance of anatomic motion.