Comparison of Data-Driven Respiratory Signal Extraction Methods From Cone-Beam CT (CBCT) — a Preliminary Clinical Study

Abstract

The difficulty of defining a data driven gold standard ground truth for internal motion has posed a challenge to clinically validate developed methods to extract respiratory motion especially during a 60-second cone-beam CT (CBCT) scan in Image-Guided Radiotherapy Treatment (IGRT). A methodology to manually track respiratory motion on clinically acquired lung cancer patient CBCT projection data over a 360° view angle is presented in this paper that serves as a ground truth respiratory signal for our work. The tracked signal is used as a reference to assess the performance of four data-driven methods in respiratory motion extraction, namely: Amsterdam Shroud (AS), Intensity Analysis (IA), Local Principal Component Analysis (LPCA), and Fourier Transform (FT)-based methods. The clinical assessment using this reference signal includes both quantitative and qualitative analysis. It is found out quantitatively that all four methods managed to extract respiratory signals which are highly correlated with the ground truth, with the LPCA method displaying the highest correlation coefficient value at 0.9071. This result is further supported by qualitative analysis and discussion via visual inspection of each extracted signal plotted with the reference signal on the same axes.