Simulational study of a dosimetric comparison between a Gamma Knife treatment plan and an intensity-modulated radiotherapy plan for skull base tumors

摘要

We treated 14 patients with a skull base (n = 5) or cavernous sinus (n = 4) meningioma, craniopharyngioma (n = 4), or pituitary adenoma (n = 1) by intensity-modulated stereotactic radiotherapy (IM-SRT) using Novalis from April 2011 through June 2012. In this study, a GK PFX treatment plan was additionally made as a simulation plan for each patient. Actual treatment plans using Novalis and simulation plans of GK PFX were compared. The treatment-planning images were acquired with magnetic resonance imaging (MRI) using a 1.5-Tesla or 3.0-Tesla scanner (Signa Echo Speed Plus 1.5 T, Signa HDxt 3.0 T; GE Healthcare, Tokyo) and 4-detector computed tomography (CT) (Light Speed Plus; GE Healthcare, Tokyo). The references for dose calculation in treatment planning were the CT images. A CT image resolution of 512 × 512 pixels in the axial plane and slice thickness of 1.25 mm was adopted to reduce partial volume effects. To determine gross tumor volume (GTV), contrast-enhanced CT and MRI were acquired. Conditions for non-contrast and contrast-enhanced CT were the same except for the size of the field of view. The slice thickness of MRI was specified from 1–2 mm depending on the tumor size by 3D-SPGR (3D fast-spoiled gradient-recalled acquisition in the steady state) sequence with gadolinium enhancement and 3D fast spin echo. All 14 patients underwent treatment-planning CT and MRI. Planning CT and MRI were fused on the iPlan RT Image version 4.1.2 (BrainLAB). Delineation of target and risk organs was performed with an autosegmentation function and manually by a radiation oncologist and a neurosurgeon on iPlan RT image treatment-planning software. The PTV margin was determined to be 2 mm, considering the spatial uncertainty, including the patient setup error, isocenter mechanical deviation, and ExacTrac image-guidance error [6]. When the CTV was in close contact with critical organs, the CTV-PTV margin was adjusted manually to avoid overlapping the PTV and these critical organs. The Novalis equipped with a micro-multileaf collimator (mMLC) with 3-mm thick leaves (m3; BrainLAB) was used. SRS/SRT with the Novalis system has been described previously [2, 3, 5]. The targets were covered with a ≥ 95% isodose level. The PTV ranged from 1.1–102.2 ml (median, 19.5 ml). The algorithm of dose calculation in iPlan RT dose version 4.1.2 software was the pencil beam convolution (PBC) method. Dosing for all patients was planned with a single isocenter IMRT by radiation oncologists and neurosurgeons. Parameter evaluation of the dose–volume histogram (DVH) was performed considering target coverage and the dose limitation for OARs. The dose constraints for OARs used in the IMRT optimization process were determined according to tolerance dose; e.g. 55 Gy for brain stem (0.1 ml-volume), 50 Gy for optic nerve (0.1 ml), 10 Gy for lens (max), 50 Gy for eye (1 ml), and 50 Gy for acoustic nerve (0.1 ml) with a 2 Gy per fraction regime. The patients underwent IM-SRT calculated by PBC with 6-MV photon beams in 14–18 fractions to a total dose of 40–48 Gy (median 42.5 Gy) (at 100% isodose = at normalization point) over 3–4 weeks. The IM-SRT treatment times were estimated for a dose rate of 320 monitor units/min, calculated in log files of patient management software for treatment. The same CT image data and structure sets including PTV and OARs of patients that were used in Novalis treatment were transferred to LGP version 10.1.1 treatment-planning software from iPlan treatment-planning software via a DICOM-RT protocol. We determined a 2-mm PTV margin around the GTV (= CTV, clinical target volume) in multisession GK-SRT using the Extend system. The dose algorithm, available in LGP software is a simple tissue maximum ratio (TMR) 10 method employing the measurement-based dose calculation by replacing all anatomical structures with water-equivalent material [7]. Multi-isocenter beam delivery, different from the Novalis single isocenter technique, was used to cover the target volume. The OARs were spared as much as possible (e.g. optic nerves less than approximately half of the prescribed isodose) even when it was adjacent to the target. The prescribed dose-fractionation schedule for the target was defined as the same as that used in Novalis IM-SRT. The treatment times with GK PFX-SRT were also estimated using a dose rate of 2.722 Gy/min for 60Cobalt, calculated by LGP treatment software. In this study, we evaluated each dosimetric characteristic using dosimetric metrics of both Novalis IM-SRT as LINAC-based SRT and multisession GK PFX-SRT using the Extend system with a relocatable frame system. We used the dose-calculation algorithm of PBC for the Novalis plan and TMR 10 for the GK plan. Collected dosimetry data were analyzed using SPSS version 18.0 (IBM, Japan). The paired t test was used to examine differences between indices of Novalis IM-SRT and those of PFX-SRT treatment plans. Differences with P 0.05 were regarded as significant.