Clinical evaluation of photon optimizer (PO) MLC algorithm for stereotactic, single-dose of VMAT lung SBRT

Abstract

Recently implemented photon optimizer (PO) MLC optimization algorithm is mandatory for RapidPlan modeling in Eclipse. This report quantifies and compares the dosimetry and treatment delivery parameters of PO vs its predecessor progressive resolution optimizer (PRO) algorithm for a single-dose of volumetric modulated arc therapy (VMAT) lung stereotactic body radiation therapy (SBRT). Clinical SBRT treatment plans for 12 early-stage non–small-cell lung cancer patients receiving 30 Gy in 1 fraction using PRO-VMAT were re-optimized using the PO-VMAT MLC algorithm with identical planning parameters and objectives. Average planning target volume derived from the 4D CT scans was 13.6 ± 12.0 cc (range: 4.3 to 41.1 cc) Patients were treated with 6 MV flattening filter free beam using Acuros-based calculations and 2.5 mm calculation grid-size (CGS). Both treatment plans were normalized to receive same target coverage and identical CGS to isolate effects of MLC positioning optimizers. Original PRO and re-optimized PO plans were compared via RTOG–0915 protocol compliance criteria for target conformity, gradient indices, dose to organs at risks and delivery efficiency. Additionally, PO-VMAT plans with a 1.25 mm CGS were evaluated. Both plans met RTOG protocol requirements. Conformity indices showed no statistical difference between PO 2.5 mm CGS and PRO 2.5 mm CGS plans. Gradient index (p = 0.03), maximum dose to 2 cm away from planning target volume in any direction (D_2cm) (p < 0.05), and gradient distance (p < 0.05) presented statistically significant differences for both plans with 2.5 mm CGS. Some organs at risks showed statistically significant differences for both plans calculated with 2.5 mm CGS; however, no clinically significant dose differences were observed between the plans. Beam modulation factor was statistically significant for both PO 1.25 mm CGS (p = 0.001) and PO 2.5 mm CGS (p < 0.001) compared to clinical PRO 2.5 mm CGS plans. PO-VMAT plans provided decreased beam-on time by an average of 0.2 ± 0.1 minutes (up to 1.0 minutes) with PO 2.5 mm and 1.2 ± 0.39 minutes (maximum up to 3.22 minutes) with PO 1.25 mm plans compared to PRO 2.5 mm plans. PO-VMAT single-dose of VMAT lung SBRT plans showed slightly increased intermediate-dose spillage but boasted overall similar plan quality with less beam modulation and hence shorter beam-on time. However, PO 1.25 mm CGS had less intermediate-dose spillage and analogous plan quality compared to clinical PRO-VMAT plans with no additional cost of plan optimization. Further investigation into peripheral targets with PO-MLC algorithm is warranted. This study indicates that PO 1.25 mm CGS plans can be used for RapidPlan modeling for a single dose of lung SBRT patients. PO-MLC 1.25 mm algorithm is recommended for future clinical single-dose lung SBRT plan optimization.

Introduction

Due to the recent technological advances in radiotherapy, stereotactic body radiotherapy (SBRT) treatment to solitary primary or metastatic lung lesions for medically inoperable non–small-cell lung cancer (NSCLC) patients is safe, effective and has a high cure rate comparable to surgery1., 2., 3., 4. including SBRT for elderly patients.⁵ Moreover, SBRT is better tolerated by patients with respect to surgery due to its minimal adverse effects.⁶ RTOG-0915 protocol (Arm 1) allowed a single dose of 34 Gy SBRT treatment for early-stage I peripheral NSCLC patients when dosimetric compliance criteria were met.⁷ Videtic et al⁸ reported long-term follow-up data which revealed no excess late toxicity in either arm (34 Gy in 1 fraction and 48 Gy in 4 fractions) and demonstrated consistent high rates of local control. They reported a median overall survival of 4 years for each arm suggesting similar efficacy. Their study concluded that single-fraction SBRT of 34 Gy remains a suitable treatment option for patients with early stage inoperable lung cancer. In another study, Videtic et al.⁹ compared 2 single-fraction SBRT dose schemes of 30 Gy and 34 Gy for 80 medically inoperable early-stage I NSCLC patients. Both treatment schedules provided equivalent tumor local-control and overall survival rates with minimal toxicity. Thus, a single dose of 30 Gy is an equally effective treatment as 34 Gy for the selected NSCLC patients and is gaining popularity in the clinics.

It has been demonstrated that with respect to traditional SBRT planning methods, 3D conformal radiation therapy, VMAT provides equal or improved dosimetric delivery.¹⁰ Utilizing volumetric modulated arc therapy (VMAT) and flattening filter free (FFF) beams have reduced SBRT treatment time significantly for a single high dose of radiation and improved patient compliance.^11,12 Removal of the flattening filter from the gantry provides benefits by reducing head scatter, out-of-field dose, residual electron contamination, and delivers treatments with higher dose rates up to factors of 2.33 for 6X-FFF and 4 for 10X-FFF beams compared to the traditional flattened beams.^10,13 Because of the reduced treatment times, VMAT with FFF beams is particularly appealing for delivering a single high dose of SBRT treatment to lung lesions, potentially minimizing intrafraction motion errors as well.

Recently, Varian Eclipse treatment planning system (TPS, Varian Medical Systems, Palo Alto, CA, Version 13.5) has implemented a new multileaf collimator (MLC) optimization algorithm called photon optimizer (PO).¹⁴ PO-MLC algorithm was created to be more efficient for IMRT/VMAT optimization over its predecessor, progressive resolution optimizer (PRO) algorithm. The main difference between PO and PRO algorithms is that PO uses a new model for defining structures. For the PO algorithm, the structures, dose-volume histogram calculations and dose sampling are defined spatially using a single matrix over the image instead of a point-cloud model defining structures that was used in the PRO algorithm. In this setting, the PO-MLC algorithm under-samples voxels at the periphery of the target. However, the PO-MLC setting in Eclipse uses multiresolution dose calculation approach to increase the dose calculation accuracy. Fixed voxel resolutions of 1.25 mm, 2.5 mm, or 5 mm can be used during multiresolution optimization.

A few investigators have reported the dosimetric differences of PO-MLC algorithm for IMRT/VMAT planning in a digital phantom,¹⁵ conventional prostate, head and neck, and brain treatments,¹⁶ knowledge-based planning to rectal cancer patients¹⁷ as well as fractionated lung SBRT patients and stereotactic brain treatments.¹⁸ For instance, the advantages and limitations of PO algorithm compared to its predecessor PRO for IMRT plans were evaluated by Binny et al.¹⁶ Eleven plans including prostate, brain, and head and neck treatments were optimized using both PO and PRO-MLC algorithms in their study. For similar target coverage and dose to critical structures, they reported that the PO algorithm gave higher MLC variability and more monitor units. Liu et al.¹⁸ compared PO with PRO algorithms for VMAT planning of fractionated lung SBRT and brain stereotactic treatments. Their retrospective study included 20 lung SBRT patients (10 received total dose of 54 Gy in 3 fractions and 10 patients received total dose of 50 Gy in 5 fractions) and 10 brain stereotactic patients received total dose of 25 Gy in 5 fractions. They reported for identical target coverage, PO algorithm provided comparable plan quality to PRO, with less MLC complexity, thus improving the treatment delivery efficiency, but contradicting the results presented by Binny et al.¹⁶ Although dosimetric differences with PO algorithm for lung SBRT plans have been studied previously by Liu et al.,¹⁸ the dosimetric impact and treatment delivery complexity of this algorithm with a FFF-beam in the treatment of single high dose of 30 Gy in 1 fraction using noncoplanar VMAT lung SBRT planning with fine resolution dose calculation grid size (CGS) of 1.25 mm has not yet been reported.

Single-fraction lung SBRT (30 Gy in 1 fraction) is an extreme form of hypofractionation used in our clinic for extracranial lesions where the dose calculation accuracy could potentially suffer by tumor size, tumor location and the presence of inhomogeneities in the lung. Moreover, due to under sampling of the voxels at the periphery of tumor volume by the PO-MLC algorithm, there is a potential for higher nontarget normal tissue dose to the organs-at-risk (OAR) adjacent to the tumor periphery. This consequence will be amplified when delivering a single high dose of radiation. This prompted us to quantify the effect of PO-MLC algorithm for our clinical implementation of a single high dose of 30 Gy in one fraction protocol using our noncoplanar VMAT lung SBRT approach. Dose to radiosensitive nontarget OAR is a major worry in VMAT lung SBRT treatments,^19,20 specifically while delivering a single large fraction dose as described here. Therefore, herein, we have retrospectively evaluated 12 consecutive early-stage NSCLC patient's plans who underwent VMAT-SBRT treatment in our clinic using the PRO algorithm. For comparison, the clinical PRO-VMAT plans were re-optimized with the PO algorithm with identical beam geometry and planning objectives. Additionally, PO-VMAT plans were re-optimized with a fine resolution of 1.25 mm CGS for evaluation. The original PRO-VMAT and re-optimized PO-VMAT plans were compared by lung SBRT protocol compliance criteria for the target conformity, gradient indices and dose to OAR per RTOG requirement.⁷