Investigation of the feasibility of a simple method for verifying the motion of a binary multileaf collimator synchronized with the rotation of the gantry for helical tomotherapy

Abstract

In this paper, we suggest a new method for verifying the motion of a binary multileaf collimator (MLC) in helical tomotherapy. For this we used a combination of a cylindrical scintillator and a general-purpose camcorder. The camcorder records the light from the scintillator following photon irradiation, which we use to track the motion of the binary MLC. The purpose of this study is to demonstrate the feasibility of this method as a binary MLC quality assurance (QA) tool. First, the verification was performed using a simple binary MLC pattern with a constant leaf open time; secondly, verification using the binary MLC pattern used in a clinical setting was also performed. Sinograms of simple binary MLC patterns, in which leaves that were open were detected as "open" from the measured light, define the sensitivity which, in this case, was 1.000. On the other hand, the specificity, which gives the fraction of closed leaves detected as "closed", was 0.919. The leaf open error identified by our method was -1.3 ± 7.5%. The 68.6% of observed leaves were performed within ± 3% relative error. The leaf open error was expressed by the relative errors calculated on the sinogram. In the clinical binary MLC pattern, the sensitivity and specificity were 0.994 and 0.997, respectively. The measurement could be performed with -3.4 ± 8.0% leaf open error. The 77.5% of observed leaves were performed within ± 3% relative error. With this method, we can easily verify the motion of the binary MLC, and the measurement unit developed was found to be an effective QA tool.

Figure 1

Schematic (a) and photograph (b) of the measurement setup. The center of the plastic scintillator is irradiated with X‐rays. The coordinate system follows the Left‐hand coordinate system. SAD in this figure stands for the source–axis distance.

Figure 2

Scintillation light emitted from a $2.5\hspace{0.17em}\times \hspace{0.17em}10\hspace{0.17em}{\text{cm}}^2$ field (a), and scintillation light emitted in the case in which the even number leaf positions are open (b). The white boxes are ROIs used to measure the detected light. The ROIs are set at the center of the leaves.

Figure 3

Relationship between the leaf open time and detected light. ${\mathrm{Q}}_{\text{center}}$, on the vertical axis, is calculated from Eq. (3).

Figure 4

Measured light on each ROI in the case in which each leaf was open individually. These are normalized by the light on the 32nd ROI.

Figure 5

Field dependency of the measured light: the light measured at each ROI and its corresponding leaf position (a); the number of open leaves and the output factor at the central axis and in the field edge region (b). The vertical axis is the detected light per frame.

Figure 6

Stability of the data measured with an ion chamber and the plastic scintillator. The measured data from the scintillator are based on ${\mathrm{q}}_{\text{center}\hspace{0.17em}\mathrm{i}}$. The data from the ion chamber are from measurements made at a depth of 10 cm. Each plot is averaged over a time of 10 sec and normalized by the value at 100 sec. The readings of both the plastic scintillator and ion chamber decreased with exposure time.

Figure 7

Stability of the gantry rotational speed. The solid axis represents the theoretical value.

Figure 8

ROC curves for a simple binary MLC pattern model (a) and a clinical binary MLC pattern model (b). The Youden index is at its maximum value at $\text{Th}=44$ pixel value in the simple model and at $\text{Th}=28$ pixel value in the clinical model.

Figure 9

Sinogram for planned simple binary MLC pattern model (a), measured sinogram (b), difference between these sinograms (c), and the resulting histogram (normalized) (d).

Figure 10

Sinogram on planned clinical binary MLC pattern model (a), measured sinogram (b), difference between these sinograms (c), and the resulting histogram (normalized) (d).