MRI-based polymer gel dosimetry for validating plans with multiple matrices in Gamma Knife stereotactic radiosurgery

Abstract

One of treatment planning techniques with Leksell GammaPlan (LGP) for Gamma Knife stereotactic radiosurgery (GKSRS) uses multiple matrices with multiple dose prescriptions. Computational complexity increases when shots are placed in multiple matrices with different grid sizes. Hence, the experimental validation of LGP calculated dose distributions is needed for those cases. For the current study, we used BANG3 polymer gel contained in a head-sized glass bottle to simulate the entire treatment process of GKSRS. A treatment plan with three 18 mm shots and one 8 mm shot in separate matrices was created with LGP. The prescribed maximum dose was 8 Gy to three shots and 16 Gy to one of the 18 mm shots. The 3D dose distribution recorded in the gel dosimeter was read using a Siemens 3T MRI scanner. The scanning parameters of a CPMG pulse sequence with 32 equidistant echoes were as follows: TR = 7 s, echo step = 13.6 ms, field-of-view = 256 mm × 256 mm, and pixel size = 1 mm × 1 mm. Interleaved acquisition mode was used to obtain 15 to 45 2-mm-thick slices. Using a calibration relationship between absorbed dose and the spin-spin relaxation rate (R2), we converted R2 images to dose images. MATLAB-based in-house programs were used for R2 estimation and dose comparison. Gamma-index analysis for the 3D data showed gamma values less than unity for 86% of the voxels. Through this study we accomplished the first application of polymer gel dosimetry for a true comparison between measured 3D dose distributions and LGP calculations for plans using multiple matrices for multiple targets.

Figure 1

Phantoms fixed in Leksell frame with indicator box for MRI scanning.

Figure 2

T2‐weighted image of phantom A with five 18 mm collimator shots delivering five different doses. The maximum dose to the shot at the center was 16 Gy. The maximum doses to the peripheral shots were 3, 5, 8, and 12 Gy.

Figure 3

Dose distributions plotted on three orthogonal planes for the calculated dose data stored in the global matrix of phantom B.

Figure 4

Calibration curve used to convert R2 into dose.

Figure 5

Comparison of BANG3 measured (solid line) and LGP calculated (thin line) dose distributions plotted on transverse planes: (a) $\mathrm{Z}=95\hspace{0.17em}\text{mm}$, (b) $\mathrm{Z}=99\hspace{0.17em}\text{mm}$, (c) $\mathrm{Z}=100\hspace{0.17em}\text{mm}$, (d) $\mathrm{Z}=101\hspace{0.17em}\text{mm}$, (e) $\mathrm{Z}=105\hspace{0.17em}\text{mm}$, (f) $\mathrm{Z}=64\hspace{0.17em}\text{mm}$. The labels of isodose curves indicate the percentage dose normalized to 7.2 Gy; or 100 indicates 7.2 Gy.

Figure 6

Comparison of BANG3 measured (solid line) and LGP calculated (thin line) dose distributions: (a) coronal plane at $\mathrm{Y}=110\hspace{0.17em}\text{mm}$, showing shot #1; (b) sagittal plane at $\mathrm{X}=80\hspace{0.17em}\text{mm}$, showing shot # 1 (right) and #3 (left).

Figure 7

Dose‐volume comparisons of measurements (solid line) and LGP calculations (thin line): (a) DDVH; (b) DVH. The value of 227.5 in the horizontal axis corresponds to 16 Gy.

Figure 8

Histogram showing the Gamma index distribution. The criteria for the gamma‐index calculations were 3% dose difference and 3 mm distance‐to‐agreement. The calculation volume was limited to a subspace: $\mathrm{x}=70$ to 110, $\mathrm{y}=70$ to 120, and $\mathrm{z}=90$ to 110.

Figure 9

Dose‐dependent‐dose‐difference‐diagram (D4 diagram) (a), with the diamond symbols indicating the mean dose difference and the error bars indicating one standard deviation; dose‐difference histogram (DDH) (b), a frequency histogram of dose difference representing the mean, median, standard deviation, minimum and maximum dose difference between the measured and calculated dose.

Figure 10

An spin‐echo image of phantom B presenting the artifacts. The image was acquired at echo time of 54.4 ms for slice number 23 (out of 45 slices). The two dark areas at the top edge of the phantom indicate the susceptibility artifacts due to the metallic fixation pins. We can also observe the stripes running in the top‐to‐bottom direction.