Dosimetric comparison of multiple vs single isocenter technique for linear accelerator‐based stereotactic radiosurgery: The Importance of the six degree couch

Abstract Purpose Single isocenter technique (SIT) for linear accelerator‐based stereotactic radiosurgery (SRS) is feasible. However, SIT introduces the potential for rotational error which can lead to geographical miss. Additional planning treatment volume (PTV) margin is required when using SIT. With the six degrees of freedom (6DoF) couch, rotational error can be minimized. We sought to evaluate the effect of the 6DoF couch on the dosimetry of patients with multiple brain metastases treated with SIT. Materials and Methods Ten consecutive patients treated with SRS to ≥3 metastases were identified. Original treatments had MIT plans (MITP). The lesions were replanned using SIT. Lesions 5‐10 cm from isocenter had an additional 1mm of margin. Patients were replanned with these additional margins to account for inability to correct rotational error (SITPM). Multiple dosimetric variables and time metrics were evaluated. Dosimetry planning time (DPT) and patient treatment time (PTT) were evaluated. Statistics were calculated using the Wilcoxon signed‐rank test. Results A total of 73 brain metastases receiving SRS, to a median of 6 lesions per patient, were identified. MITPs treated 73 lesions with 63 isocenters. On average, MITPs had a 19.2% higher brain V12 than SITPs (P = 0.017). For creation of SITPM, 30 lesions required 1 mm of additional margin, while none required 2 mm of margin. This increased V12 by 47.8% on average per patient (P = 0.008) from SITP to SITPM. DPT was 5.5 hours for SITP, while median for MITP was 12.5 hours (P = 0.005) PTT was 30 minutes for SITP, while median for MITP was 144 minutes (P = 0.005). Conclusions SITPs are comparable to MITPs if rotational error can be corrected with the use of a 6DoF couch. Increasing margin to account for rotational error leads to a nearly 50% increase in V12, which could result in higher rates of radiation necrosis. Time savings are significant using SIT.


| INTRODUCTION
Stereotactic radiosurgery (SRS) is a commonly used method of high dose radiation, pinpointed to areas of radiographically visible disease within the brain. There is increasing interest in the maximum number of brain metastases (BMs) that can be safely treated with SRS rather than whole brain radiation therapy (WBRT), the historical standard.
However, the traditional technique of one isocenter per lesion may result in significantly elongated treatment times for an individual patient. Recent studies have demonstrated the feasibility of treating multiple intracranial lesions with a single isocenter. [1][2][3][4] However, this method has brought up concerns of rotational error, with potential for diminished PTV coverage if not well accounted for 5-7.
The six degrees of freedom (6DoF) couch is a relatively recent advance that has allowed for improvements in patient reproducibility. 8 Traditional radiation therapy couches allowed for only longitudinal movements in the x, y, and z axes while rotational errors of yaw, pitch, and roll, were not able to be corrected. The necessity of the 6DoF couch in controlling for rotational error and its effect on dosimetric variables has not been previously studied. We hypothesized that having the 6DoF couch to correct for rotational error would allow for minimal changes to normal organ dosimetry when converting multilesion, multi-isocenter plans to a multilesion, single isocenter plan. We also hypothesized that adding additional PTV margin to account for that error in the multilesion single isocenter plans (SITP+1) would lead to worsening of normal organ dosimetry. Lastly, we hypothesized that treated with a single isocenter would be more efficient from both a dosimetrist and physicist perspective, while simultaneously reducing how long the patient was on the treatment table.

2.A | Planning and dosimetric factors
We retrospectively identified ten consecutive patients treated with SRS at our institution to ≥3 metastases. The patients were originally planned using MIT, with the exact number of isocenters at the discretion of the treating dosimetrist, physicist, and physician. Contouring and plan formation was done on the Eclipse treatment planning system (Varian, Palo Alto, CA). All plans were created using volumetric arc therapy (VMAT), with goals of covering 95% of all PTVs to 100% of the prescription dose. All patients were treated on a True-Beam STX linear accelerator with 2.5 mm Micro multileaf collimators SITPs were evaluated to identify PTVs ≥5 cm from the isocenter.
The threshold for cut-off to increase margin, in our model not having the 6DoF couch, was chosen based on a previous publication evaluating distance uncertainty in SRS, as well as a more recent publication from our institution showing concordance. 5,9 Assuming the max rotational error of 1.4 degrees, a distance from isocenter to target of 5 cm leads to distance uncertainty of 1 mm, while a distance of 10 cm leads to distance uncertainty of 2 mm (Fig. 1). These are listed as SITP with margin (SITPM). The rotational tolerance of a TrueBeam treatment couch, as per manufacturer specifications is ≤0.3 degrees, which leads to a distance uncertainty of ≤0.5 mm at 10 cm from isocenter as per Figure 1. These potential distance uncertainties were not accounted for within our data analyses. Prescription dose and normalization were maintained on all three plans.
Prescription dose was between 18 and 21 Gy in 1-3 fractions. All PTVs had max heterogeneity between 105% and 135%. We evaluated mean brain dose (MBD), volume of brain receiving 4 Gy and 12 Gy in cubic centimeters (V4 and V12, respectively), brainstem max dose, lens max dose, and optic chiasm and nerves max dose as our dosimetric factors of interest across all three sets of plans.

2.B | Time metrics
The cumulative time required for dosimetry to adequately plan the lesions was calculated as the dosimetry planning time (DPT) for each patient. This included an estimate of 0.5 hours for creation of normal structures. For MITPs, 2 hours per isocenter were estimated; for SITPs, due to increased complexity of the plan and treating multiple lesions simultaneously, 5 hours was estimated as average planning F I G . 1. Determination of additional margin necessary based on distance from isocenter to target, based on variable degrees of rotational error. Red stars show that at maximum rotational error, distance from isocenter to target of 5 cm necessitates 1 mm of additional margin, while distance from isocenter to target of 10 cm necessitates 2 mm of additional margin. time. The time required for physics verification of the plan to ensure phantom agreement on dosimetry was calculated as the quality assurance time (QAT) for each patient. QAT was estimated as requiring 20 minutes for the first isocenter, and 10 minutes for each additional isocenter. The patient time on table (PTT) for the duration of their treatment was estimated at 20 minutes to allow for cone beam CT verification per isocenter, with a beam on time of 2 minutes per arc that the patient was treated with.

2.C | Statistical analyses
Statistics were calculated using the paired samples Wilcoxon signedrank test. 10

| RESULTS
A total of 10 patients with 73 brain metastases receiving SRS to a median of 6 lesions (range 3-16) were analyzed. MITPs treated 73 lesions with 63 isocenters while SITPs treated the same 73 lesions with a total of 10 isocenters (one per plan). The prescription dose to the PTV, and normalization, was kept the same for both the MITPs and the SITPs with attempts at keeping hot spots similar as well.

| DISCUSSION
This study evaluates the dosimetric feasibility of performing single isocentric plans when treating multiple lesions with SRS. Previous studies have also shown the feasibility of this approach. [1][2][3][4] However, recent publications have demonstrated concerns regarding potential for rotational error and, if unaccounted for, potential for compromised coverage in SRS plans. [5][6][7] In this study, we accounted for rotational error by either using a six-degree couch or adding an additional mm or margin for lesions greater than 5cm away. Our study shows that treating with SIT while controlling for rotational error with the 6DoF couch decreases the primary dosimetric parameter critical for SRS plan evaluation, brain V12. However, this benefit is lost when an extra mm of margin is the method of accounting for rotational error, as is required by theoretical modeling as well as verification by our institution's physics staff. 5,9 Additionally, not being able to control for rotational error, and thus requiring an extra mm of margin, increases brain V12 by an average of 48% in this series. Brain V12 has been shown to be a consistent predictor of radiation necrosis in multiple series. [11][12][13] Other dosimetric parameters, such as brain mean and brain V4, may be increased slightly due to transitioning from MIT to SIT. An increase in MBD in SITP may have been driven by a higher V4, potentially as a result of more noncoplanar arcs. While these param- The step-wise increase in feasibility of SRS alone for increasing number of intracranial lesions has been well documented, initially starting with one to three metastases, eventually increasing to ≤10. [16][17][18] Case reports and retrospective series have also described the feasibility of SRS for patients with greater than ten metastases. 19,20 SRS in the future will be primarily constrained not by the number of lesions requiring treatment, but rather by the volume of metastatic disease and by the length of time the patient can be on the necessary for dosimetrists/physicists, as well as for radiation therapists at the treatment machine.
In conclusion, we report that accounting for rotational error with a six degrees of freedom couch, when treating multiple (≥3) lesions with a single isocenter technique, results in comparable dosimetry, with significant time savings for the dosimetrist, the physicist, and the patient. However, without a six degrees of freedom couch, the additional margin necessary to account for rotational error results in large changes in critical dosimetric parameters, such as V12, potentially putting patients at higher risk for developing radiation necrosis.
We encourage radiation oncologists to consider implementing a single isocenter technique when treating patients with a large number of brain metastases to increase departmental efficiency but to exercise caution in situations where rotational error cannot be appropriately accounted for.

CONF LICT OF I NTEREST
The authors report no conflict of interest.

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
The data that support the findings of this study are available from the corresponding author upon reasonable request. Preliminary data for this research was presented as a poster at the ASTRO annual meeting in 2019.