SBRT treatment of abdominal and pelvic oligometastatic lymph nodes using ring‐mounted Halcyon Linac

Abstract Purpose/objectives This work seeks to evaluate the plan quality, treatment delivery efficiency, and accuracy of single‐isocenter volumetric modulated arc therapy (VMAT) of abdominal/pelvic oligometastatic lymph nodes (LNs) stereotactic body radiation therapy (SBRT) on Halcyon Linac. Materials and Methods After completing the in‐house multitarget end‐to‐end phantom testing and independent dose verification using MD Anderson’s single‐isocenter/multi‐target (lung and spine target inserts) thorax phantom, eight patients with two to three abdominal/pelvic oligometastatic LNs underwent highly conformal single‐isocenter VMAT‐SBRT treatment using the Halcyon Linac 6MV flattening filter free (FFF) beam. Targets were identified using an Axumin PET/CT scan co‐registered with planning CT images and a single‐isocenter was placed between/among the targets. Doses between 25 and 36.25 Gy in 5 fractions were delivered. Patients were treated every other day. Plans were calculated in Eclipse with advanced AcurosXB algorithm for heterogeneity corrections. For comparison, Halcyon VMAT‐SBRT plans were retrospectively generated for SBRT‐dedicated TrueBeam with a 6MV‐FFF beam using identical planning geometry and objectives. Target coverage, conformity index (CI), dose to 2 cm away from each target (D2cm) and dose to adjacent organs‐at‐risk (OAR) were evaluated. Additionally, various treatment delivery parameters including beam‐on time were recorded. Results Phantom measurements showed acceptable spatial accuracy of conebeam CT‐guided Halcyon SBRT treatments including compliance with MD Anderson’s single‐isocenter/multi‐targets phantom credentialing results. For patients, the mean isocenter to tumor center distance was 3.4 ± 1.2 cm (range, 1.5–4.8 cm). The mean combined PTV was 18.9 ± 10.9 cc (range, 5.6–39.5 cc). There was no clinically significant difference in dose to LNs, CI, D2cm and maximal doses to OAR between single‐isocenter Halcyon and Truebeam VMAT‐SBRT plans, although, Halcyon plans provided preferably lower maximal dose to adjacent OAR. Additionally, total monitor units, beam‐on time and overall treatment time was lower with Halcyon plans. Halcyon’s portal dosimetry demonstrated a high pass rate of 98.1 ± 1.6% for clinical gamma passing criteria of 2%/2 mm. Conclusion SBRT treatment of abdominal/pelvic oligometastatic LNs with single‐isocenter VMAT on Halcyon was dosimetrically equivalent to TrueBeam. Faster treatment delivery to oligometastatic LNs via single‐isocenter Halcyon VMAT can improve clinic workflow and patient compliance, potentially reducing intrafraction motion errors for well‐suited patients. Clinical follow‐up of these patients is ongoing.


| INTRODUCTION
Stereotactic body radiation therapy (SBRT) treatment of single or multiple abdominal and pelvic lymph nodes (LNs) is a fast, safe, and effective treatment option with one-and five-year tumor local control rates of up to 100% and 70% and a low risk of treatmentrelated toxicity. [1][2][3][4][5][6] In addition to providing higher therapeutic dose to multiple LNs, SBRT can reduce the number of patient hospital visits and will help to improve patient compliance and clinic workflow.
Traditionally, LNs SBRT have been delivered using SBRT dedicated C-arm Linac via conebeam CT image guidance or a robotic Cyber-Knife unit 3,6 and recently, utilizing volumetric modulated arc therapy (VMAT) for fast and effective treatment delivery. 7,8 Due to the advancement of MRI-Linac technology, adaptive SBRT treatments to single or multiple LNs can be delivered in real time, significantly improving targeting accuracy. [9][10][11][12][13][14] However, MRI-Linac treatments are very time consuming and not readily available to every patient's cohort. Longer treatment times are very inconvenient for the patient (who is lying down in the treatment position), and can hinder clinic workflow.
For fast patient throughput, Varian recently introduced a fastrotating ring-mounted Linac called the Halcyon V2.0 (Varian Medical Systems, Palo Alto, CA) platform for conventionally fractionated image-guided radiation therapy (IGRT) treatments. 15 This novel, but coplanar Linac was designed under tight performance specifications in order to improve patient safety and treatment delivery accuracy.
In brief, the Halcyon V2.0 is equipped with a single-energy 6MV-FFF beam with a rapid gantry rotation speed of 4 revolutions per minute with a mean energy and nominal depth of maximal dose at 1.3 MeV and 1.3 cm respectively. [15][16][17][18][19] In contrast with the SBRT-dedicated Carm TrueBeam Linac, Halcyon Linac is equipped with newly designed double-stacked and staggered 1 cm width MLC layers. The proximal and distal layers are offset by 5 mm allowing for a projected 5 mm effective MLC width at isocenter, like that of a standard SBRTdedicated TrueBeam Linac with Millennium 120 MLC. The jawless Halcyon Linac has a maximal field size of 28 × 28 cm 2 with full MLC travel of 28 cm and is twice as fast as the standard Millennium 120 MLC. The stacked/staggered design allows for ultra-low MLC leakage and transmission dose of <0.5%. [15][16][17][18][19] Per machine specifications, Halcyon provides an improved penumbra with a smaller dosimetric leaf gap of approx. 0.1 mm. In addition to the MV-conebeam CT imaging system, the Halcyon Linac is equipped with fast kilovoltage conebeam CT (kV-CBCT) imaging that includes a high-quality iterative CBCT reconstruction (iCBCT) algorithm for better online image quality. 20,21 This novel Linac is designed for a 'one-step patient set up and verification' that automatically applies couch shifts for patient set-up followed by the image-guidance procedure for each treatment. 15 Therefore, therapists do not need to manually apply the isocenter shifts inside the treatment room, significantly decreasing patient set-up time and consequently reducing overall treatment time.
We have recently installed the Halcyon V2.0 in our center. Initial acceptance testing and commissioning results of the Halcyon Linac showed all acceptance and commissioning criteria met with the Varian's specifications as described above. 15 with multiple high and low contrast inserts in it that were used to quantify the spatial resolution directly. A conebeam CT scan of the QUART phantom was obtained at the treatment position and the geometric accuracy was measured by 3D-to-3D matching with the planning CT images with respect to the multiple inserts at different locations. The second was the machine performance check (MPC) phantom on Halcyon that comes with 16 balls bearings (BBs) at the different locations up to 15 cm away from the center of the phantom. Similar to the QUART phantom, to further quantify the spatial localization accuracy as a function of distance to isocenter, a kV 3D/ 3D registration of the MPC phantom was performed by matching the BBs with the planning CT images. Moreover, the independent dose validation of the single-isocenter/multitarget treatment on Halcyon was performed using the MD Anderson's SBRT credentialing thorax phantom with two targets (spine and lung tumors with dosimetry system). This phantom was imaged, planned using a single-isocenter VMAT, and irradiated with a SBRT prescription dose of 6.0 Gy in 1 fraction to both targets simultaneously for credentialing the NRG-BR001 protocol. 23 The tumor distance between the spine and lung targets was about 9 cm. Advanced Acuros-based dose calculation was used. The phantom irradiation results were in compliance with the MD Anderson's standards. Thus, after completing all the phantom validation testing, we are currently treating both conventionally fractionated and selected hypofractionated SBRT patients including abdominal/pelvic oligometastic LNs on our Halcyon Linac.

2.B | Patient cohort
After obtaining Institutional Review Board approval for our institute, eight consecutive oligometastatic abdominal and pelvic LNs patients who underwent SBRT treatments on ring-mounted Halcyon Linac were selected for this study. These patients received 25-36.25 Gy in 5 fractions. All patients involved with this study had been diagnosed with nonmetastatic prostate cancer. In all patients, the primary prostate cancer was treated with local therapy including prostatectomy, radiation therapy and one patient treated with cryotherapy followed by salvage radiation therapy. Pathology of the original tumors ranged from Gleason score 3 + 4 (grade group 2) to Gleason score 4 + 5 (grade group 5). Follow-up PSA evaluations revealed a rising trend.
Patients were evaluated by Axumin PET/CT scans [31][32][33] and were found to have two or more PET positive LNs. For all LNs patients, the evidence of two or more enlarged PET positive LNs metastases were found on Axumin PET/CT scans with no other sites of diseases.

2.C | CT simulation and target delineation
All patients were immobilized using the Body Pro-Lok TM device (CIVCO system, Orange City, IA) in the supine position with their arms above the head. A knee cushion was used to immobilize the knees and legs. Patients were instructed to present for CT simulation and treatment with a comfortably full bladder and empty rectum.
Patients were instructed not to urinate for one and a half to two hours before simulation and each fraction of SBRT treatment to help with daily patient set up reproducibility. To ensure a relatively empty rectum, patients were instructed to use miralax beginning 3 days prior to simulation and daily throughout the course of SBRT treatment. A free-breathing 3D planning CT was acquired on a GE Light- Tumor characteristics for this patient cohort are summarized in Table 1. Each patient had 2-3 enlarged LNs on abdominal and pelvic F I G . 1. Co-registered Axumin PET/CT scan with the planning CT images for locating the enlarged pelvic lymph nodes for example patient #6. Two-bilateral iliac chain LNs were highlighted on Axumin scan.
sites. The combined PTV ranged from 5.57 to 39.5 cc (Table 1).
Bladder and small bowel positioning are most important when targeting LNs in the pelvis. The OAR contours included duodenum, rectum, bladder, and small bowel for dose reporting.      Table 2).

2.F | Plan comparison
Similar relative dose to the GTVs was obtained for both plans. Table 3 shows the number of LNs per patient basis, their prescription dose and the average 3D Euclidian distance between the targets. The isocenter to tumor center distance for each patient ranged from 1.5 cm to 4.8 cm, all less than 5 cm. Table 3 also shows immediately adjacent OAR and the maximal dose to the OAR achieved for the both plans. In this cohort, for each patient, the Halcyon VMAT provided similar but relatively lower maximal dose to OAR compared to TrueBeam VMAT plans (Table 3), systematically

| 167
As described above, treatment delivery accuracy of the singleisocenter VMAT-SBRT plans for multiple LNs was evaluated by delivering each plan in QA measurement mode to both Linacs using an on-board EPID imager and analyzing the gamma pass rates via the PD QA procedure. [36][37][38] We have compiled SBRT QA pass rates following the TG-218 criteria recommended on pre-treatment QA tolerance level (overall pass rate ≥95% at a γcriteria of 2%/2mm with low threshold of 5%). 38 With PD QA, all plans satisfied this criterion and were similar on both machines. The dose delivery accuracy of the single-isocenter Halcyon VMAT and the corresponding TrueBeam VMAT plans were 98.1 ± 1.6% (range 96.9-100.0%) and 97.8 ± 1.8% (range 96.6-100%) respectively. F I G . 4. Axial and coronal views of Halcyon kV-CBCT images (see inset in the coronal view) co-registered with planning CT images used for image-guided single-isocenter (see, cross-hair) VMAT-SBRT treatment to both lesions on Halcyon is shown. In addition to anatomical landmarks, the planned dose colorwash (50%-110% isodose cloud) is overlaid for this treatment to demonstrate the delivery accuracy. Halcyon kV-CBCT images were acquired in the treatment position in free breathing and rigid-registration was performed via automatic imageregistration mode on Halcyon followed by manually fine-tuning the registration for better alignment of both nodes via soft-tissue alignment.
parameters (125 kV, 1080 mAs) were used with 512 × 512 pixels and 2.0 mm slice thickness for iCBCT reconstruction. An in-house SBRT/IGRT protocol was applied registering the fast pretreatment kV-iCBCT with the planning CT scans (see Fig. 4). For each treatment, rigid-registration was performed automatically based on region of interest and bony landmarks followed by a manual refinement of the soft-tissues matching and confirmed by the treating physician and physicist for the alignment of both LNs. The patient was then repositioned by applying the 3DoF couch corrections from the original isocenter and the treatment was delivered. Our image-guided SBRT protocol limits translational 3DoF couch corrections to less than ±3.0 mm, on average in each direction for all LNs SBRT treatments.

ACKNOWLEDGMENT
This work is partially supported by the Varian Medical Systems.

CONF LICT OF I NTEREST
The authors have no relevant conflict of interest to disclose.

AUTHOR CONTRI BUTIONS
DP conceived the project. DP, AW, and JS collected and analyzed the data. WSC provided clinical expertise and definitive supervision of the paper. DP drafted the preliminary manuscript and all coauthors revised and approved the final manuscript for submission.

D A T A A V A I L A B I L I T Y S T A T E M E N T
Research data are not shared.