Abstract:PURPOSE: Breathing management can reduce breath-to-breath (intrafraction) and day-by-day (interfraction) variability in breathing motion while utilizing the respiratory motion of internal and external surrogates for respiratory guidance. Audiovisual (AV) biofeedback, an interactive personalized breathing motion management system, has been developed to improve reproducibility of intra- and interfraction breathing motion. However, the assumption of the correlation of respiratory motion between surrogates and tumors is not always verified during medical imaging and radiation treatment. Therefore, the aim of the study was to test the hypothesis that the correlation of respiratory motion between surrogates and tumors is the same under free breathing without guidance (FB) and with AV biofeedback guidance for voluntary motion management.METHODS: For 13 lung cancer patients receiving radiotherapy, 2D coronal and sagittal cine-MR images were acquired across two MRI sessions (pre- and mid-treatment) with two breathing conditions: (a) FB and (b) AV biofeedback, totaling 88 patient measurements. Simultaneously, the external respiratory motion of the abdomen was measured. The internal respiratory motion of the diaphragm and lung tumor was retrospectively measured from 2D coronal and sagittal cine-MR images. The correlation of respiratory motion between surrogates and tumors was calculated using Pearson's correlation coefficient for: (a) abdomen to tumor (abdomen-tumor) and (b) diaphragm to tumor (diaphragm-tumor). The correlations were compared between FB and AV biofeedback using several metrics: abdomen-tumor and diaphragm-tumor correlations with/without ≥5 mm tumor motion range and with/without adjusting for phase shifts between the signals.RESULTS: Compared to FB, AV biofeedback improved abdomen-tumor correlation by 11% (p = 0.12) from 0.53 to 0.59 and diaphragm-tumor correlation by 13% (p = 0.02) from 0.55 to 0.62. Compared to FB, AV biofeedback improved abdomen-tumor correlation by 17% (p = 0.01) and diaphragm-tumor correlation by 15% (p < 0.01) while correcting 0.3 s (p = 0.54) and 0.2 s (p = 0.19) phase shifts, respectively. In addition, AV biofeedback with ≥5 mm tumor motion range, compared to FB improved abdomen-tumor correlation by 14% (p = 0.18) and diaphragm-tumor correlation by 17% (p = 0.01). The highest abdomen-tumor and diaphragm-tumor correlations were found using ≥5 mm tumor motion range and phase shifts, resulting in a 12% improvement in AV biofeedback.CONCLUSIONS: Our results demonstrated that AV biofeedback improves the correlation of respiratory motion between surrogates and the tumor. This suggests a need for AV biofeedback for respiratory guidance utilizing respiratory surrogates during image-guided and MRI-guided radiotherapy in thoracic regions.

Comparison between GA-BP algorithm and SVM in respiratory motion estimation

A Comparative Study on the Generalization Ability of Back Propagation Neural Network and Support Vector Machine for Tracking Tumor Motion in Radiotherapy

SVM Based Human Respiratory Pattern Classification Method for Stereo Radiotherapy Robot

A markerless beam's eye view motion monitoring algorithm based on structure conversion and demons registration in medical image with partial data

Audiovisual biofeedback improves the correlation between internal/external surrogate motion and lung tumor motion

Comparison of Initial Learning Algorithms for Long Short-Term Memory Method on Real-Time Respiratory Signal Prediction

Variations of the Respiration Signals for Respiratory-Gated Radiotherapy Using the Video Coached Respiration Guiding System

Parameters Selection in Gene Selection Using Gaussian Kernel Support Vector Machines by Genetic Algorithm

A back propagation neural network based respiratory motion modelling method

Lung Respiratory Motion Estimation Based on Fast Kalman Filtering and 4D CT Image Registration

Investigation of a Breathing Surrogate Prediction Algorithm for Prospective Pulmonary Gating

Respiratory Motion Estimation of Tumor Using Point Clouds of Skin Surface

Research on the geometric difference between different ITV generation methods in non-small cell lung cancer treated with stereotactic radiotherapy

Diffeomorphic respiratory motion estimation of thoracoabdominal organs for image‐guided interventions

Lung respiration motion modeling: a sparse motion field presentation method using biplane x-ray images

Towards real-time respiratory motion prediction based on long short-term memory neural networks

Technical evaluation of a laser-based optical surface scanning system for prospective and retrospective breathing adapted computed tomography

Multi-task autoencoder based classification-regression model for patient-specific VMAT QA.

A novel external/internal tumor tracking approach to compensate for respiratory motion baseline drifts

Predicting respiratory motion for real-time tumour tracking in radiotherapy

3D tumor localization through real-time volumetric x-ray imaging for lung cancer radiotherapy