Characterization of optical‐surface‐imaging‐based spirometry for respiratory surrogating in radiotherapy
Guang Li,Jie Wei,Hailiang Huang,Qing Chen,Carl P. Gaebler,Tiffany Lin,Amy Yuan,Andreas Rimner,James Mechalakos
DOI: https://doi.org/10.1118/1.4941951
IF: 4.506
2016-01-01
Medical Physics
Abstract:Purpose: To provide a comprehensive characterization of a novel respiratory surrogate that uses optical surface imaging (OSI) for accurate tidal volume (TV) measurement, dynamic airflow (TV') calculation, and quantitative breathing pattern (BP) estimation during free breathing (FB), belly breathing (BB), chest breathing (CB), and breath hold (BH). Methods: Optical surface imaging, which captures all respiration-induced torso surface motion, was applied to measure respiratory TV, TV', and BP in three common breathing patterns. Eleven healthy volunteers participated in breathing experiments with concurrent OSI-based and conventional spirometric measurements under an institutional review board approved protocol. This OSI-based technique measures dynamic TV from torso volume change (Delta V-torso = TV) in reference to full exhalation and airflow (TV'= dTV/dt). Volume conservation, excluding exchanging air, was applied for OSI-based measurements under negligible pleural pressure variation in FB, BB, and CB. To demonstrate volume conservation, a constant TV was measured during BH while the chest and belly are moving ("pretended" respiration). To assess the accuracy of OSI-based spirometry, a conventional spirometer was used as the standard for both TV and TV'. Using OSI, BP was measured as BPOSI = Delta V-chest/Delta V-torso and BP can be visualized using BPSHI = SHIchest/(SHIchest+ SHIbelly), where surface height index (SHI) is defined as the mean vertical distance within a region of interest on the torso surface. A software tool was developed for OSI image processing, volume calculation, and BP visualization, and another tool was implemented for data acquisition using a Bernoulli-type spirometer. Results: The accuracy of the OSI-based spirometry is -21 +/- 33 cm(3) or -3.5%+/- 6.3% averaged from 11 volunteers with 76 +/- 28 breathing cycles on average in FB. Breathing variations between two separate acquisitions with approximate 30-min intervals are substantial: -1%+/- 34% (ranging from -64% to 40%) in TV, 4%+/- 20% (ranging from -50% to 26%) in breathing period (T), and -1%+/- 34% (ranging from -49% to44%) in BP. The airflow accuracy and variation (between two exercises) are -1 +/- 54 cm(3)/s and -5%+/- 30%, respectively. The slope of linear regression between OSI-TV and spirometric TV is 0.93 (R-2 = 0.95) for FB, 0.96 (R-2 = 0.98) for BB, and 0.95 (R-2 = 0.95) for CB. The correlation between the two spirometric measurements is 0.98 +/- 0.01. BP increases from BB, FB to CB, while TV increases from FB, BB, to CB. Under BH, 4% volume variation (range) on average was observed. Conclusions: The OSI-based technique provides an accurate measurement of tidal volume, airflow rate, and breathing pattern; all affect internal organ motion. This technique can be applied to various breathing patterns, including FB, BB, and CB. Substantial breathing irregularities and irreproducibility were observed and quantified with the OSI-based technique. These breathing parameters are useful to quantify breathing conditions, which could be used for effective tumor motion predictions. (C) 2016 American Association of Physicists in Medicine.