Development and clinical evaluation of a simple optical method to detect and measure patient external motion

Abstract

A simple and independent system to detect and measure the position of a number of points in space was devised and implemented. Its application aimed to detect patient motion during radiotherapy treatments, alert of out-of-tolerances motion, and record the trajectories for subsequent studies. The system obtains the 3D position of points in space, through its projections in 2D images recorded by two cameras. It tracks black dots on a white sticker placed on the surface of the moving object. The system was tested with linear displacements of a phantom, circular trajectories of a rotating disk, oscillations of an in-house phantom, and oscillations of a 4D phantom. It was also used to track 461 trajectories of points on the surface of patients during their radiotherapy treatments. Trajectories of several points were reproduced with accuracy better than 0.3 mm in the three spatial directions. The system was able to follow periodic motion with amplitudes lower than 0.5 mm, to follow trajectories of rotating points at speeds up to 11.5 cm/s, and to track accurately the motion of a respiratory phantom. The technique has been used to track the motion of patients during radiotherapy and to analyze that motion. The method is flexible. Its installation and calibration are simple and quick. It is easy to use and can be implemented at a very affordable price. Data collection does not involve any discomfort to the patient and does not delay the treatment, so the system can be used routinely in all treatments. It has an accuracy similar to that of other, more sophisticated, commercially available systems. It is suitable to implement a gating system or any other application requiring motion detection, such as 4D CT, MRI or PET.

Figure 1

In‐house calibration phantom showing the six accurately located calibration points.

Figure 2

Sketch of (a) the in‐house phantom made to perform circular motion and (b) the in‐house phantom made to perform oscillations of low amplitude. As the eccentric disk rotates, the points of the bar move up and down.

Figure 3

Sketch of the measurement setup, and an example of the paired images obtained. The circles mark the points that have to be tracked.

Figure 4

A view of the user interface of the program. The algorithm is tracking the motion of only one point placed on the patient skin (black arrow). Motion magnitudes in mm are shown besides their respective tolerances; they are depicted over green color if tolerances are being met, or over yellow or red colors otherwise.

Figure 5

Examples of the position of the labels on patients surface: (a) bladder with electrons; (b) breast with electron beam applicator; (c) lung; (d) rectum.

Figure 6

Amplitude of the motion of Anzai phantom in z‐ (vertical) axis, for frequencies 10 rpm (○) and 15 rpm (□), and for quasi‐respiratory (up) and sinusoidal (down) modes. Only one of each three points is represented. Lines are the fitting of the data.

Figure 7

Relative displacements (a), with respect to the initial position, of a point on the thorax of a breast cancer patient, showing respiratory motion in 3D (X, Y, and Z directions); percentiles ((b) and (c)) of respiratory amplitude in antero–posterior (Y) direction in each treatment session, for 21 breast tumor volunteer patients. For patients in (b), points to track were placed on the sternum; for patients in (c), points were on the epigastrium. Vertical lines divide trajectories of different patients.

Figure 8

Smoothed trajectories of all points (399) (a) tracked for breast cancer patients, for lateral motion; variance of the 399 motion measurements (b) at each time, for the three axis; all the smoothed trajectories (c), in lateral direction, of a particular patient; a particular smoothed trajectory (d), in left to right direction.

Figure 9

Graph of the trajectories of the 15 central points (a) (marked in the photo) detected by the system; projections ((b),(c),(e)) of the points detected in three planes; photograph (d) of a volunteer wearing the T‐shirt with the marks.