Field‐in‐field breast planning for a jawless, double‐stack MLC LINAC using flattening‐filter‐free beams

Abstract Background This study intends to develop an efficient field‐in‐field (FiF) planning technique with the Eclipse treatment planning system (TPS) to determine the feasibility of using the Halcyon treatment delivery system for 3D treatment of breast cancer. Methods Ten treatment plans were prepared on the Halcyon treatment planning system and compared to the same patients’ clinically delivered TrueBeam plans which used flattened 6 MV and 10 MV beams. Patients selected for this study were treated via simple, tangential breast irradiation and did not receive radiotherapy of the supraclavicular or internal mammary lymph nodes. Planning target volumes (PTV) volumes ranged from 519 cc to 1211 cc with a mean target volume of 877 cc. Several planning techniques involving collimator, gantry rotation, and number of FiF segments were investigated as well as the use of the dynamically flattened beam (DFB) — a predefined MLC pattern that is designed to provide a flattened beam profile at 10 cm depth on a standard water phantom. For comparison, the clinically delivered TrueBeam plans remained unaltered except for normalization of the target coverage to more readily compare the two treatment delivery techniques. Results Using the physician defined PTV, normalized such that 98% of the volume was covered by 95% of the prescribed dose, the Halcyon plans were deemed clinically acceptable and comparable to the TrueBeam plans by the radiation oncologist. Resulting average global maximum doses in the test patients were identical between the TrueBeam and Halcyon plans (108% of Rx) and a mean PTV dose of 102.5% vs 101.6%, respectively. Conclusions From this study a practical and efficient planning method for delivering 3D conformal breast radiotherapy using the Halcyon linear accelerator has been developed. When normalized to the clinically desired coverage, hot spots were maintained to acceptable levels and overall plan quality was comparable to plans delivered on conventional C‐arm LINACs.


| INTRODUCTION
Breast carcinoma is the most commonly diagnosed cancer in females.
About one in eight women in the United States (12.4%) will develop invasive breast cancer over the course of their lifetime. 1 Radiotherapy, in conjunction with breast conservation surgery and chemotherapy, plays a major role in the treatment of these cancers. 2 Breast radiotherapy is typically delivered by conventional 3D beams or intensity modulated radiation therapy (IMRT). While IMRT is often better suited to treat complex targets with involved lymph nodes, 3 conventional 3D tangent beams continue to make up a large percentage of radiotherapy treatments with acceptable clinical outcomes and widespread availability.
Recently, a compact ring-shape medical linear accelerator (LINAC) system, Halcyon (Varian Medical Systems, Inc.), was released. The Halcyon system provides a single 6FFF beam with double-stack multi-leaf collimator (MLC) beam shaping system and faster gantry rotation when compared with conventional C-arm LINACs. TrueBeam linear accelerators can deliver treatments at a maximum gantry rotation of one rotation per minute vs four rotations per minute with Halcyon. Maximum dose rate for a flattened 6 MV or 10 MV is 600 MU/min at isocenter for TrueBeam and 800 MU/min at isocenter for the 6FFF beam provided on the Halcyon LINAC. Halcyon was designed to address the global need for access to radiation therapy with integrated imaging guidance, improved clinical efficiency, and shorter installation and commissioning time as well as operational demands. [4][5][6] However, the jawless design and unflattened beam provide some challenges when creating breast 3D plans with the FiF techniques. The jawless design might suggest that there is increased out-of-field leakage dose to the patient compared to traditional C-arm accelerators where jaws are needed to minimize this out-of-field leakage. For example, Varian TrueBeam jaws collimate the radiation that is not defined by the MLC aperture and minimize interleaf leakage (radiation that permeates two adjacent leaves of the same bank). There are two sets, one for each x-y collimation direction. They are 78 mm thick along the beam axis and are comprised of a tungsten alloy. Halcyon uses a double-stacked MLC arrangement with thicker leaves to account for a lack of jaws. The Millenium-120 MLCs fitted to most TrueBeams have 67-mm-thick tungsten leaves, whereas the Halcyon has 77 mm thick leaves in each bank, creating a combined thickness of 154 mm which results in an average transmission factor of 0.0047 vs 0.015 for a Millenium-120 system. Because the MLCs are offset in the direction perpendicular to leaf travel, the interleaf leakage is attenuated by the distal leaf bank. This design with thicker, offset leaves obviates the need for jaws.
Typically, the use of 6 or 10 MV flattened beam and a set of static MLC apertures can provide dose uniformity that is clinically acceptable. To better control the dose homogeneity, the field-in-field (FiF) technique is also often used in breast 3D planning. 7

2.A | Radiotherapy delivery equipment
As shown in Fig. 1(a), Halcyon is a ring gantry LINAC with 100 cm diameter bore size and is capable of providing a 6FFF beam utilizing a compact gantry head design. This compact head allows a 1 m diameter patient bore LINAC to be installed in the same room as required for a conventional C-arm LINAC. A typical C-arm LINAC has an effective "bore diameter" of approximately 80 cm, since the physical measurement from the treatment head enclosure to mechanical isocenter is 40 cm. Halcyon employs a staggered, double-stack MLC system, as illustrated in Fig. 1(b), which is the only custom beam shaping device between the x-ray source and the patient. The MLC leaf width is 1.0 cm when projected at isocenter with 100 cm source-axis distance (SAD). The proximal and distal layers are offset, or staggered, by 50% of the leaf width such that the effective resolution of the leaf stack pair becomes 0.5 cm at isocenter. The maximum leaf speed is 5.0 cm/s at isocenter with 100% interdigitation capability. There are 29 leaf-pairs leafs in the proximal layer and 28 leaf-pairs in the distal layer with the maximum field size of 28 × 28 cm 2 at isocenter. One of the advantages of the doublestack design is that the combined transmission is less than 0.01%, negating the need for physical jaws to reduce peripheral dose to nontargeted organs.

2.B | Patient selection
Ten breast patients clinically treated with 3D FiF technique on Carm LINACs (TrueBeam, Varian Medical System) from our institution were randomly selected and anonymized. To confine the study to feasibility of utilizing Halcyon for tangential breast radiotherapy, patients with involved supraclavicular, axillary, or internal mammary lymph node involvement were excluded. Across the 10 patients selected, two physicians and three dosimetrists were involved in the planning and approval of the treatment plans. This particular selection was performed to minimize potential bias due to a single planner. The prescribed dose to the PTV was either 4256cGy MORRIS ET AL.  Typically in clinical plan, the physician will define the medial block edge by placing a block margin around the medial aspect of the PTV with consideration for minimizing heart dose. The block was designed based on individual patient's geometry. In this study, we attempt to develop a simplified and standardized procedure to test the feasibility of FiF technique on Halcyon. Therefore, the heart block was not individually drawn for each patient. Also, at our clinic, this physician defined block edge is not modified by the planner, only MLCs from the other bank are moved to accomplish the FiF modulation. In this study, we attempt to develop a simplified and standardized procedure to test the feasibility of FiF technique on Halcyon. This is not the case in this study and MLCs from either bank were used to modulate dose as needed to achieve target coverage and minimize OAR doses. Then approximately 2 cm of flash is added on the nipple side of the target to account for potential respiratory motion and setup uncertainty (Fig. 2). Dose is then computed without normalization. Initial hotspots may exceed 140% or be as low as 95% depending on the patient anatomy. The subsequent planning steps focus on minimizing dose heterogeneity within the beams eve-view.

2.C.2 | Dynamically flattened beam sequence
The initial tangent fields incorporate the use of the TPS dynamically flattened beam (DFB) option. 10 The DFB is a predefined MLC sequence unique to Halcyon which sweeps the MLCs, whereas the beam is delivering dose to provide a flattened beam profile for a fixed gantry beam delivery. This profile, however, is only flattened at depth when delivered on a flat, homogeneous phantom. The DFB sequence does not take into account patient anatomical variations.
The DFB tangents alone did not produce clinically acceptable dose homogeneity for any of our tested patients (global Dmax < 115% of prescription dose when the PTV was normalized such that V 98 = 95%). However, we found that the use of DFB tangents as opposed to non-DFB tangents markedly reduces the planning time initiate their closing sequence until the beam is again fully blocked.
The DFB sequence is a binary option in the treatment planning system and has no change in control points as a function of patient anatomy, and only varies with field size. The number of monitor units for a DFB field is significantly increased compared to an open field that would deliver the same dose at treatment depth. This is because the beam is continuously being delivered while the MLCs sweep across so the high-intensity portion of the beam is blocked to make a flat beam. Table 3

2.C.3 | Collimator rotation optimization
Optimal collimator rotation is needed to aid in subsequent modulation of hot spots with the highest resolution achievable with the MLC shaping system. A similar concept is used with TrueBeam, but is usually slightly different due to the changes in hot spot extent and location as a result of the FFF beam. The optimal rotation is deter-  limitations of the MLC is achieved by setting the collimator rotation angle such that the direction of leaf motion is parallel to the isodose falloff at the widest section of the breast in the medial-lateral direction, typically at the level of the nipple (Fig. 4). Observed factors that influence optimal collimator rotation are breast volume, anatomic tissue distribution, and simulation setup conditions including slant board angle.
After optimal collimator angle is determined, the MLCs are re-fitted to the PTV with the 0.3 cm isotropic margin and approximately 2 cm of flash added by manually selecting and dragging the leaves of the anterior banks for the two DFB tangent beams. Once dose is recomputed, the field weight of the DFB tangents is adjusted such that the global Dmax is minimized and/or evenly distributed in the craniocaudal axis.

2.C.4 | Manual FiF segments
The initial FiF segment typically utilized only the anterior leaf bank for hot spot reduction. For ease of MLC placement, an isodose is selected that yields a conventional FiF segment size in the BEV. In our clinic, the planning strategy is to select an isodose that is approximately 10%-15% lower than the hot spot but this technique does not work well for the 6FFF beam quality and typically results in a plan that has more FiF segments and required much more time to plan. An "acceptable" FiF segment with this planning strategy yields a 1-2 cm collective leaf motion from the previous segment ( Fig. 5). This method attempts to limit the number of segments while keeping a dosimetrically meaningful number of monitor units (MU) per segment. Institutionally, we have a minimum restriction of 5 MU per segment for conventional LINAC FiF planning as well as Stepand-Shoot IMRT delivery. This reduces the uncertainty due to nonlinearities in the delivery of monitor units. The Halcyon FiF planning method has adopted this policy.
The process of selecting an isodose line as an MLC placement guide is repeated approximately 2-4 times for each tangent field at the initial gantry and collimator rotation angles resulting in 4-8 total FiF segments for the plan. These FiF segments do not utilize the DFB feature and are delivered in a step-and-shoot fashion with the LINAC in a beam-hold state during MLC movements. Generally speaking, larger breast volumes will tend to require a higher number of FiF segments to control hot spots and improve low dose coverage. In our testing, optimal field weight tends to straighten the DVH curve as much as possible while minimizing the high dose tail.

2.C.6 | Plan normalization
The final planning step involves normalization. In this planning study all patients were normalized such that 98% of the volume was covered by 95% of the prescribed dose. If the global D max was in excess of 115% the normalization was reduced until the global D max was below 115% of prescription. This situation was not encountered with any of the 10 test plans. In clinical practice the target coverage and normalization objectives are defined by physician preference. At our institution, it is standard practice to achieve a PTV coverage of V 95% = 98% (V 95 = 95% minimum acceptable) with a global D max less than 115% (Fig. 7)

2.D | Deliverability and plan measurement
In order to test the deliverability of the ten plans, PSQA was performed on each plan. First, ion chamber measurements were taken with a PTW 31010, in a solid water phantom (15 cm 3 ). Passing criteria were percent dose deviation of less than ±3% from the expected dose. Second, portal dosimetry was performed for each field. A composite image was created, and gamma criteria was set according to the recommendations of AAPM task group 218 (3% dose difference, 2 mm distance-to-agreement, global normalization, 10% threshold with a tolerance level of 95% of pixels passing). 11 Third, the plans were delivered to a Delta4 Phantom+ (ScandiDos, Sweden).
The same criteria were set as for portal dosimetry on the composite dose delivered to the phantom.

3.A | Overall plan summary
All Halcyon plans were successfully generated within 30 min following the standardized process developed and were evaluated by a radiation oncologist that specializes in breast radiotherapy. Overall, each Halcyon plan tested was deemed clinically acceptable. The physician evaluated the organs at risk as well as target coverage via isodose and DVH analysis. OAR doses were analyzed for the heart, ipsilateral lung, and con-  Table 2.
The mean number of monitor units used for the Halcyon delivered plans was 732, whereas the TrueBeam delivered plans used 279. In Table 3, it is evident that for larger PTVs (particularly those that were clinically treated with 10 MV) more FiF segments are necessary with the Halcyon LINAC due to the shallower PDD 10 .   Figure 6. For this segment, 4% field weight was optimal but optimal field weights will vary from patient to patient as well as segment to segment. DVH, dose volume histogram; PTV, Planning target volumes. F I G . 7. DVH comparison of PTV coverage using 6FFF tangent beams alone, the DFB tangents alone, and the DFB tangents paired with the manual FiF segments. All three plans were normalized such that 98% of the PTV volume was covered by 95% of prescription. DFB, dynamically flattened beam; DVH, dose volume histogram; PTV, Planning target volumes. reduction. The DVH in Figure 12 shows comparable PTV coverage and OAR sparing between the two plans. Overall, the Halcyon plan only incorporated 4 more FiF segments compared to the TrueBeam plan (4 vs 8) (13).

3.C | Deliverability and Plan QA
All FiF plans were successfully delivered on Halcyon. Table 4 shows the ion chamber measurements and gamma passing rates for all ten plans. The average IC measurement had a percent dose deviation of

| DISCUSSION
We demonstrated a FiF breast planning strategy utilizing a 6FFF beam on the Halcyon platform. All ten plans were determined clinical acceptable. Conventionally, FFF beams have not typically been used in breast 3D planning due to the insufficient ability to deliver homogeneous dose to targets at depth. However, the percent-depth-dose   of the jaws at the center of the field which will abut subsequent treatment fields. The fact that Halcyon is jawless does not prevent an equivalent approach using the onboard MLC. Leakage is significantly reduced but not eliminated by the Halcyon MLCs compared to the Millenium-120 which is standard on most TrueBeams. The biggest challenge of Halcyon compared to TrueBeam is the reduced field size at isocenter paired with the increased limitations in isocenter placement due to the physical bore. Future work will include investigation of using Halcyon to treat these complex breast cases.

| CONCLUSION
In this study, a practical and efficient planning method for delivering 3D conformal breast radiotherapy using the Halcyon linear accelerator has been developed. When normalized to the clinically desired coverage, hot spots were maintained to acceptable levels and overall plan quality was comparable to plans delivered on conventional Carm LINACs. Using the original isocenter and gantry angles that were used for TrueBeam delivery, the Halcyon patients would not have experienced clearance issues as indicated by the TPS. Intrinsically, the 6FFF beam's slightly shallower PDD did not prohibit its use for achieving deep target coverage while keeping superficial hot spots below 115% of prescription.

ACKNOWLEDG MENTS
This work was partially supported by a research grant from Varian Medical Systems.
T A B L E 4 Patient-specific QA results (%