Using flattening filter free beams in electronic tissue compensation whole breast irradiation with deep inspiration breath hold

Abstract Purpose In order to reduce heart dose, DIBH has become a common practice in left‐sided whole breast irradiation. This technique involves a significant strain on patients due to the breath‐hold requirements. We hereby investigate the dosimetric and delivery feasibility of using flattening filter free (FFF) energies with electronic tissue compensation (ECOMP) planning technique to reduce the required breath‐hold lengths and increase patient compatibility. Methods Fifteen left‐sided, postlumpectomy patients previously receiving DIBH whole‐breast radiotherapy (266cGy x 16fx) were retrospectively planned using ECOMP for both 6X and 6X‐FFF. A dosimetric comparison was made between the two plans for each patient using various dosimetric constraints. Delivery feasibility was analyzed by recalculating the 6X ECOMP plan with 6X‐FFF without replanning (6X‐FFF QA) and delivering both plans for a one‐to‐one comparison using Gamma analysis. Beam‐on times for the 6X and 6X‐FFF plans were measured. For all tests, Wilcoxon signed‐rank test was used with P < 0.05 as significant. Results No statistical difference was observed between 6X and 6X‐FFF plans for most dosimetric endpoints except contralateral breast Dmax (P = 0.0008) and skin Dmax (p = 0.03) and Dmin (P = 0.01) for which 6X‐FFF showed favorable results when compared with 6X. 6X‐FFF significantly reduced beam‐on times for all patients by 22%–42% (average 32%). All plan QAs passed departmental gamma criteria (10% low‐dose threshold, 3%/3mm, >95% passing). Conclusion ECOMP planning with FFF was found feasible for left‐sided breast patients with DIBH. Plan quality is comparable, if not better, than plans using flattened beams. FFF ECOMP could significantly reduce beam‐on time and required breath‐hold lengths thereby increasing patient compatibility for this treatment while offering satisfactory plan quality and delivery accuracy.


| INTRODUCTION
Breast cancer is the most diagnosed form of noncutaneous cancer in American women and results in the second highest number of cancer-related deaths behind only lung cancer. It was estimated that in the United States, one in eight women will be diagnosed with breast cancer at some point in their lifetimes. 1 Radiotherapy plays an essential role in breast cancer treatment and is used for over half of all breast cancer patients. 2 It has been shown to both decrease loco-regional recurrence rates and improve overall survival for breast cancer patients. 3 Due to the effective tumor control for these patients and the long-term survival of most of them, one central theme of radiotherapy advances in the recent years has been to reduce the treatment toxicity and long-term side effects.
In 2013, a seminal paper by Darby et al. showed that a non-negligible risk of heart disease and coronary events is associated with breast cancer radiotherapy and the risk is estimated to increase 4-7% for each 1 Gy in mean heart dose. 4 Enlightened by the study, clinicians have grown increasingly cautious about the cardiac risk.
Special techniques have been more commonly employed in modern breast radiotherapy to minimize the radiation dose to the heart, especially for left breast cancer where heart dose tends to be higher due to the close proximity to the treatment targets. Above all, Deep Inspiration Breath Hold (DIBH) is the most popular technique used to reduce the heart dose in breast radiotherapy, especially with the increasing availability of the surface guidance technology. It increases the distance from the breast to the heart by having patients hold their breath for the duration of each treatment beam, which typically takes 20-50 s. DIBH has shown a reduction of dose to the heart between 31% and 80%. 5,6 While the benefits are undeniable, there are many patients who are not well suited for DIBH due to the physical demand it requires. Patients with chronic obstructive pulmonary disease (COPD) or many other respiratory diseases often cannot perform the breath hold of the required length and are hence unable to use the DIBH technique. For patients that can, radiotherapy delivery uncertainty could also increase with prolonged treatment time due to possible drifts at the end of a long breath hold and other patient motions. In contrast, external beam breast radiotherapy has a very large target size and a strict target dose uniformity requirement (commonly only up to 7% hot). Conventional 3D conformal forward planning with tangential fields is therefore challenging to achieve with FFF beams. However, a recently available planning technique, electronic tissue compensation (ECOMP), could easily employ FFF beams. In ECOMP, the planner creates and manually optimizes fluences (or beam intensity) instead of beam apertures, and the plan is then delivered via a sliding window technique. The ECOMP planning technique has been gaining popularity as it is efficient and has been shown to increase target dose uniformity. [7][8][9] Since in ECOMP planning the planner operates on fluences instead of beam apertures, it makes a perfectly feasible planning technique to test FFF beams for DIBH breast radiotherapy. In this work, we report a designed experiment to test the dosimetric and delivery feasibility of FFF ECOMP plans, in order to reduce required breath-hold lengths and increase patient compatibility of DIBH breast radiotherapy while not sacrificing plan quality. Target contours were drawn by the attending radiation oncologist according to our departmental protocol, including breast PTV (determined by wire placed at time of simulation), evaluation PTV or PTVe (breast PTV minus 5 mm skin), Lumpectomy GTV (lumpectomy cavity), Lumpectomy PTV (1.7 cm margin around GTV), and Lumpectomy evaluation PTVe (Lumpectomy PTV minus 5 mm skin). OARs include heart, ipsilateral lung, contralateral breast, and skin (defined as the 5 mm inner wall from the external body contour that overlaps with the breast PTV contour).

2.B | Treatment planning
The plans for this study were retrospectively created using the ECOMP planning technique in Eclipse v.15 (Varian Medical Systems, Palo Alto, CA, USA). A process map of the study is shown in Fig. 1.
For the dosimetric assessment, two parallel plans were created for each patient, one using 6X and the other using 6X-FFF of a True-Beam linear accelerator (Varian Medical Systems, Palo Alto, CA, USA). On each case, the two plans used identical beam angles (two tangential fields), field sizes (including 2cm flash), isocenter, and aperture shapes determined by the attending physician. For ECOMP planning, a block was drawn copying the aperture shape so that when the new irregular surface compensator was added and the fluence map was created, the block can be used to erase fluence outside the block area. Patients were planned to a prescribed dose of 2.66 Gy × 16 (42.56 Gy) to the PTVe. The plans were normalized to maximize the dose to lumpectomy GTV while still keeping a hotspot of < 107%. The penetration depth was defined as 50% depth. This identifies the percent depth at which a uniform dose will be delivered for the two tangent beams. The fluence map generated was then modified by the planner to reduce hotspots and boost cold spots.

2.C | Plan analysis and dosimetric assessment
To evaluate the plan quality using 6X vs 6X-FFF, dosimetric comparison was performed between the two parallel plans on each case using our institutional criteria for postlumpectomy whole breast irradiation as the dosimetric endpoints. The dose endpoints and ideal planning objectives are shown in Table 1 for both the targets and OARs. Note that we added skin D max , D min , and D mean for the purpose of this study although our institution does not currently consider skin constraints in the planning process.

2.D | Plan delivery assessment
To evaluate the delivery time, both 6X and 6X-FFF plans were delivered and the beam-on time was recorded. A dose rate of 600 MU/ min was used for 6X and that of 1200 MU/min was used for 6X-FFF plans.
In addition, to assess plan delivery accuracy a 6X-FFF QA plan was created and QA gamma passing rates were compared between the 6X and the corresponding 6X-FFF QA plans. To create the 6X-FFF QA plan, the 6X plan was copied and the energy changed to 6X-FFF with the corresponding changed dose rate while keeping all other plan parameters (MUs, field weights, fluence, etc.) constant.
Despite keeping all the parameters the same, the leaf sequence must be recalculated when energy and dose rate are changed which results in a slightly different leaf sequence between the 6X and 6X-FFF QA plans. The decision to create the 6X-FFF QA plan for evaluation, rather than using the 6X-FFF plans created for dosimetric assessment, is in an effort to make QA comparison between the two energies more objective. The QA analysis was done using the gamma passing rate (3%/3 mm, 10% low-dose threshold) with portal dosimetry (Varian Medical Systems, Palo Alto, CA, USA).

2.E | Statistical analysis
The dosimetric endpoints, QA gamma passing rates, and delivery beam-on time were compared between the two energy groups using a Wilcoxon signed-rank test, with P < 0.05 considered as significant.  Table 1. For target coverage, no statistically significant difference was noted between the two plans. The mean differences are also very small and likely insignificant clinically. A Pvalue for the lumpectomy PTVe was unable to be obtained since there was no difference in the V95% coverage between the 6X and 6X-FFF plans. For OARs, statistically significant differences were only observed in the contralateral breast D max and skin D max and D min . For the contralateral breast, the D max for 6X-FFF showed a mean reduction of 68.7 cGy from 6X. Figure 2 shows the contralateral breast D max for both 6X and 6X-FFF for all patients. Note that while a couple of plans did not meet the ideal planning objective for contralateral breast, this is due to the use of deep tangents to cover internal mammary nodes for these patients in the original clinical plans which were deemed acceptable. The skin D max for 6X-FFF was slightly lower than 6X (by an average of 28.3 cGy) and the D min for 6X-FFF was slightly higher than 6X (by an average of 28 cGy), indicating a slightly more uniform skin dose when using 6X-FFF. No significant differences were noted for the heart or ipsilateral lung. While not statistically significant, the heart had a reduction in D max between 6X-FFF and 6X (1264.4 cGy vs 1388.8 cGy, respectively).

3.A | Dosimetric comparison
The ECOMP plans of both energies for a representative patient are shown with isodose lines in axial, sagittal, and coronal views (Figure 3) and in the comparative dose-volume histogram ( Figure 4). As can be seen, the two plans are similar with nearly identical target coverage and OAR doses except slightly lower contralateral breast and skin dose on the 6X-FFF plan.
MUs for each plan as well as the difference between the 6X and 6X-FFF plans are shown in Table 2. Higher MUs were needed in the 6X-FFF plans due the peaked beam profile and softer beam energy.

3.B | Delivery comparison
The QA gamma results are shown in Table 3. A p-value for the difference in gamma could not be obtained since many patients had identical passing gamma results. All plans met our institutional criteria for passing QA (10% threshold, 3%/3 mm gamma < 1 for> 95% of points). In a few cases, the gamma passing rates for the 6X-FFF plans were slightly lower than the 6X plans with the greatest difference being 2.9% on one case.
As expected, with a higher dose rate, the 6X-FFF plans recorded shorter beam-on time than the corresponding 6X plans. The percentage decrease in beam-on time for the 6X-FFF plan when compared to the 6X plan is shown in Table 4. The difference was statistically significant (P = 0.0006) with a mean decrease of 32%.

| DISCUSSION
Traditionally, whole breast or chest wall irradiation has been treated with two tangential fields, which has superior delivery robustness compared with IMRT plans with alternative beam arrangements yet often shows an equivalent dosimetry. With the tangential beam  another study comparing different techniques for general breast treatment planning, ECOMP plans were shown to have within 3% mean target volume metrics and comparable OAR doses compared with wedged tangential, IMRT, and hybrid IMRT plans. 10 Additionally, the increase in dose rate could cause concern about possible increased toxicity. While there are few reports discussing standard C-arm linac FFF toxicities for breast radiotherapy, studies of O-ring linacs have shown acceptable toxicities. 15 Also, while not widespread, breast SBRT is used with acceptable toxicities. 16

CONF LICT OF I NTEREST
No conflict of interest.