TH-C-17A-10: A Novel Volumetric Respiratory Surrogate Using Optical Surface Imaging




Respiration induces motion of the entire torso because the semifluid internal organs are movable driven primarily by the diaphragm within the connected body cavities: thoracic, abdominal and pelvis. Therefore, surface motion of the entire torso represents all respiration-induced external motion and could serve as a more informative respiratory surrogate.


Four-dimensional optical surface (4DOS) imaging was used to monitor the entire torso as a volumetric respiratory surrogate. High speed image capture yields high spatial-resolution images at 5 frames per second using 3 ceiling-mounted stereoscopic cameras, capturing all moving surface of the torso. 4DOS images were retrospectively reconstructed. An in-house MATLAB program was designed to automatically process the surface images with 1×1mm2 grid and calculate torso volume variation as a function of time with a common volume of interest defined by the torso surface, a posterior cut-plane (PCP), and 2 vertical cut-planes at superior-inferior borders. Torso volume variation during quiet breathing (<±3mmHg) represents time-resolved tidal volume (TV) since the tissues and reserved body air within the torso conserve. The spatial distribution of TV in the torso is quantified as breathing pattern (BP=ΔVthorax/TV) using the rib cage as the thorax. Two volunteers were examined performing 4 different breathing patterns, including fake breathing during breath hold.


Using different PCPs at or below sagittal midline of the torso, 4DOS produces same TV (<±2%) in free-/chest-/belly-breathing, suggesting that posterior body motion is negligible. Fake breathing during breath hold shows large BP variations (e.g., −ΔVthorax=ΔVabdomen≤750cm3) but small TV changes (−6±39cm3), depicting volume conservation. Surface motion is location-dependent, suggesting intrinsic uncertainties and limited usefulness of point-fiducial surrogates.


The 4DOS-based respiratory surrogate produces dynamic TV and BP values, which are potentially useful in developing a more reliable tumor motion surrogate. Further study of this tidal-volume surrogate with more volunteers is on-going under an IRB-approved protocol.

This research is in part supported by NIH (U54CA137788/132378).