Quality assurance device for four-dimensional IMRT or SBRT and respiratory gating using patient-specific intrafraction motion kernels

Abstract

Emerging technologies such as four-dimensional computed tomography (4D CT) and implanted beacons are expected to allow clinicians to accurately model intrafraction motion and to quantitatively estimate internal target volumes (ITVs) for radiation therapy involving moving targets. In the case of intensity-modulated (IMRT) and stereotactic body radiation therapy (SBRT) delivery, clinicians must consider the interplay between the temporal nature of the modulation and the target motion within the ITV. A need exists for a 4D IMRT/SBRT quality assurance (QA) device that can incorporate and analyze customized intrafraction motion as it relates to dose delivery and respiratory gating. We built a 4D IMRT/SBRT prototype device and entered (X, Y, Z)(T) coordinates representing a motion kernel into a software application that 1. transformed the kernel into beam-specific two-dimensional (2D) motion "projections," 2. previewed the motion in real time, and 3. drove a recision X-Y motorized device that had, atop it, a mounted planar IMRT QA measurement device. The detectors that intersected the target in the beam's-eye-view of any single phase of the breathing cycle (a small subset of all the detectors) were defined as "target detectors" to be analyzed for dose uniformity between multiple fractions. Data regarding the use of this device to quantify dose variation fraction-to-fraction resulting from target motion (for several delivery modalities and with and without gating) have been recently published. A combined software and hardware solution for patient-customized 4D IMRT/SBRT QA is an effective tool for assessing IMRT delivery under conditions of intrafraction motion. The 4D IMRT QA device accurately reproduced the projected motion kernels for all beam's-eye-view motion kernels. This device has been proved to, effectively quantify the degradation in dose uniformity resulting from a moving target within a static planning target volume, and, integrate with a commercial respiratory gating system to ensure that the system is working effectively. Such a device is discussed as a potential tool to optimize the gating duty cycle to maximize delivery efficiency while minimizing dose variability.

Figure 1

The two‐dimensional diode array mounted atop high‐precision X–Y steppers with independent motion. The X–Y steppers receive power and motion instructions from a controller device.

Figure 2

The prototype software interface, showing a three‐dimensional (X,Y,Z) versus time (T) motion kernel entry, together with axial, sagittal, and coronal projections at the plane of isocenter. Beam‐specific projections to the detector plane are also calculated and displayed.

Figure 3

Schematic of integration with the gating system: linking the superior–inferior movement of the projected motion kernel with the gating mechanism that detects anterior–posterior movement of surrogate markers. (a) Maximum exhalation creates the most posterior marker position, and (b) maximum inhalation creates the most anterior marker position (analogous to movement of structures in the lung vs. position of markers on a breathing belt). $\text{IR}=\text{infrared}$.

Figure 4

Row A): Target centroid motion kernel plots of patient SBRT1, projected to axial, sagittal, and coronal isocenter planes (panels 1 – 3). Panels 4 and 5 show two sample beam's‐eye‐view (BEV) projections for plan beams with gantry angles of 180 and 270 degrees. Row B): Target centroid motion kernel plots of patient SBRT2, projected to axial, sagittal, and coronal isocenter planes (panels 1 – 3). Panels 4 and 5 show two sample BEV projections for plan beams with gantry angles of 48 and 213 degrees.

Figure 5

Photographs of the positions of the moving detectors inside the internal target volume (top row). The target detectors are those overlapping with the smaller clinical target volume projection. These photographs were captured at these approximate time points: (a) 0 s, (b) 0.7 s, (c) 2.3 s, (d) 3 s, and (e) 4.6 s. The bottom row shows photographs of the external gating markers and how they move in correlation with the projected motion kernels.

Figure 6

Real‐time Position Manager (RPM: Varian Medical Systems, Palo Alto, CA) respiratory gating console. The upper window shows a real‐time camera image of the two RPM markers, caught in time during their up–down cycle. The lower window plots the vertical rise and fall of the marker position versus time and the duty cycle indicating when the delivery is turned on and off. Here, the pulses are turned on when the markers are lowest, representing the phase centered on maximum exhalation.

Figure 7

A sample of three fractions delivered to the patient SBRT1 moving target for the 180‐degree gantry angle beam. The left column (non‐gated delivery) shows not only blurring of the dose over the target diodes, but also fraction‐to‐fraction variation (that is, between the rows). The right column (gated delivery) shows that dose blurring and fraction‐to‐fraction variation are reduced.

Figure 8

Histograms of the coefficient of variation for the target diodes of the stereotactic body radiation therapy (SBRT) 180 field measured over 10 separate fractions. Without gating, the open field delivers a much more consistent dose (that is, each target diode receives a dose that is very similar fraction‐to‐fraction) than in dynamic multileaf collimation (DMLC). When gating is employed, the coefficients of variation across the target diodes were much reduced for DMLC delivery of this beam.

Figure 9

A possible decision flowchart for four‐dimensional (4D) intensity‐modulated radiation therapy (IMRT) / stereotactic body radiation therapy (SBRT) patient plans. This chart covers clinics both with and without respiratory gating technology for treatment delivery. $\text{CT}=$ computed tomography; $\text{QA}=\text{quality assurance}$; $\text{CV}=\text{coefficient of variation}$.