To assess dose errors caused by the interplay effects of free-breathing (FB) motion and to assess the value of breath-hold (BH) in terms of cardiac dose reduction for scanning beam proton therapy (SBPT).
Three patients with left-sided breast cancer previously treated with photon therapy were included in this dosimetric study: 2 following breast-conserving surgery with 2 hypothetical target volumes (whole breast alone and whole breast plus regional nodes, including supraclavicular, axillary, and internal mammary lymph nodes); and 1 postmastectomy, with the target volume including the chest wall plus regional nodes. SBPT plans were generated with various beam angles that ranged between 2 tangential directions. For treatment with FB, nominal dose and dose with interplay effects considered were calculated based on FB 4-dimensional computed tomography scans. SBPT plans on the BH computed tomography were also calculated for one of the patients, who was selected to be treated with photon therapy with BH.
Dosimetric differences between nominal and interplay dose were small (average target mean dose, −0.06 Gy; range, −0.23 to 0.06 Gy; average heart mean dose, 0.001 Gy; range, −0.12 to 0.05 Gy). The largest dose deviations occurred in plans calculated with tangential beam arrangements; the smallest was noted with the en face beam. The average value of the mean heart dose with FB was <1 Gy. For the selected patient, the mean heart doses were 0.5 and 0.2 Gy for FB and BH, respectively.
Dose deviations caused by the interplay effects of respiratory motion during FB do not have a significant impact in SBPT with en face beam arrangement. BH does not significantly reduce cardiac dose. SBPT delivery is feasible with FB and can provide optimal target coverage and maximal sparing of the cardiopulmonary system, which can translate into improved clinical outcomes and a decrease in treatment-related morbidity in left-sided breast cancer patients or those who require internal mammary node coverage.
A large meta-analysis of more than 10 000 patients in 17 randomized trials has demonstrated that adjuvant radiotherapy following breast conservation surgery in node-positive postmastectomy patients results in a substantial reduction in local and regional failures, which translates into a significant survival benefit . However, this breast cancer–specific survival benefit can be partially diluted by radiation-induced cardiac mortality, especially for left-sided cancer patients when the internal mammary lymph nodes are irradiated [2, 3], with the effect linearly correlated with mean heart dose and no threshold dose below which further reduction in cardiac morbidity is not observed . For every 1 Gy increase in the mean heart dose, the rate of major cardiac events increased by 7.4%, and an even greater effect was observed for women categorized as “higher risk” based on underlying clinical risk factors . Other studies utilizing single-photon emission computed tomography (CT) have reported that more left-sided than right-sided breast cancer patients are found to have myocardial deficits after radiation therapy [3, 5]. Therefore, to maximize the therapeutic ratio of radiation therapy for breast cancer, maximal possible reduction of cardiac dose should be considered, using all feasible techniques.
In external-beam photon therapy, inspiration breath-hold (BH) techniques have been commonly used to reduce the heart dose [6, 7]. Based on a review and meta-analysis of 19 studies reported in the literature, the average heart mean doses for free breathing (FB) and BH were 4.2 and 2.1 Gy, respectively, a 50% relative reduction; the range of heart mean dose reduction with BH was 0.5 to 4 Gy [6, 7]. However, the reproducibility of interfraction and intrafraction BH treatment positioning can be a concern.
As proton therapy becomes more readily available for breast cancer treatment, a similar question naturally arises: will the BH technique still be effective for sparing the cardiac dose in proton therapy? The reported heart mean dose for proton treatments planned with FB is generally within 1 Gy (RBE) [8–11]. The already low heart dose with FB implies that room for dose reduction with BH would be limited.
On the other hand, if proton therapy were given during FB, when breast volumes move with respiration, would the motion affect the planned dose for coverage and organs at risk such as the heart? For pencil-beam scanning beam proton therapy (SBPT), dose is delivered spot by spot. The temporal sequence of beam spot delivery and the target intrafraction motion can cause significant interplay effects, which have been investigated in lung SBPT treatments [12–17]. For breast treatments, to the best of our knowledge, the interplay effects with FB treatments have not been reported. In a study of postmastectomy radiation using SBPT , the averaged dosimetric effect with respiratory motion after multiple fractions was evaluated. In another study of partial-breast irradiation with intensity-modulated proton therapy , the motion effect was estimated by the differences between planned dose on the FB CT and the recalculated dose on the BH CT. Both studies indicated that overall dose deviations resulting from FB motion were small. However, neither study reported the interplay effect in a single fraction, which is an important aspect in evaluating motion effects.
The aim of our study was to determine whether the FB SBPT treatment is motion robust and whether the addition of a BH technique to SBPT would significantly lead to reduction in cardiac dose when compared with FB SBPT. To answer these questions, motion-induced dose deviations in the FB SBPT were evaluated based on the interplay dose. We simulated the interplay dose of respiratory motion based on both 4-dimensional (4-D) CT data and the timing structures of the proton beam delivery system. The resulting interplay dose was compared with the nominal dose for target volume coverage and the dose to cardiopulmonary structures. In addition, cardiac doses in the FB and BH SBPT plans were compared.
Materials and Methods
The study included 3 patients with left-sided breast cancer who had been previously treated with photon therapy. Two were treated with FB. One was treated with BH, because the tangential fields could irradiate a large volume of heart consistent with the patient's anatomy. For each patient, a 4-D CT scan was acquired, and for the patient treated with BH, a BH CT was also acquired. Clinical target volume (CTV) and heart contours were defined per the Radiation Therapy Oncology Group and a validated heart atlas, respectively . Two patients with postlumpectomy intact breasts were planned with 2 hypothetical target volumes: whole breast alone and comprehensive breast, which included whole breast and supraclavicular, axillary, and internal mammary nodes. One postmastectomy patient was planned with the target volume composed of the chest wall (CW) and the comprehensive regional nodal regions.
We estimated the maximum range of breast or CW motion during FB based on rigid regional registration between end of inhale and exhale CT image sets from the 4-D CT dataset. Table 1 lists the displacement along 3 orthogonal directions as well as the volumes of treated breast or CW.
SBPT plans were calculated using the Eclipse treatment planning system (Varian Medical Systems, Palo Alto, California). The nominal spot size in σ, defined to be at the isocenter at 10 cm depth in water (∼110 MeV) is 4 mm. The spot spacing of 5 mm was used for planning. Nominal doses for FB treatment were calculated on the average phase of the 4-D CT. Plans were generated with 1 or 2 fields with various beam angles, which ranged between 2 tangential directions (eg, an en face beam plus a lateral beam). Plans were calculated with single field optimization. Optimization target volumes were generated by expanding the CTVs by 5 mm, excluding anterior expansion into the skin and posterior expansion into the ribs and intercostal muscles, to account for the setup errors of 5 mm and the range uncertainties of 3.5%. A total of 26 SBPT plans were created for the 3 patients. For the patient treated with BH, SBPT plans were calculated on the BH CT as well.
The prescription was 50 Gy at 2 Gy (RBE) per fraction. The optimization objectives were as follows: 100% of the prescribed dose to cover 100% of the CTV and 95% of the optimization target volume; no more than 110% of the prescribed dose within the optimization target volume; and no more than the prescribed dose outside of the target volume. For all SBPT plans, 95% of the CTV received 100% of the prescription dose (D95% = 100%), and the maximal dose (Dmax) was <110%. Doses to organs at risk (heart, lungs, contralateral breast, and esophagus) were optimized to be as low as reasonable. Table 2 lists the dosimetric parameters for the CTV Dmax, the heart mean dose, the mean dose and V20 Gy for the ipsilateral lung, and the skin Dmax. The listed data are the average and the range values of the 26 plans.
Abbreviations: CTV, clinical target volume; D95%, minimum dose received by 95% of the CTV; Gy, Gray; RBE, relative biological effectiveness; Dmax, maximal dose.
Values are mean (range).
Interplay Dose Calculation
For a given SBPT plan, the plan was binned into 10 subplans by a custom-built code. The temporal binning assigned beam spots to a temporal phase based on the timing of the proton beam delivery system and the patient's respiratory cycle. Timing characteristics of a commercial synchrotron system for proton therapy were simulated. The timing structure consists of 4 machine characteristics: energy layer switching, proton extraction, scanning, and spot verification. The energy layer switching time is 2.3 seconds, during which the unused protons from the previous extraction are dumped, and the protons for next extraction are accelerated to the programmed energy. The proton extraction time has small dependence on the energy and is approximately constant (2.2 s/nC). Within an energy layer, the proton spots are delivered in a consecutive order: spot by spot, and line by line. The scanning time is 10 m/s in line and 6 m/s between lines. The verification time is 1 millisecond per spot. The patients' FB cycles were extracted from the 4-D CT datasets, and regular periodicity during intrafraction and interfraction treatment delivery was assumed. The 10 subplans generated 10 subdoses on each of 10 respiratory phases. The 10 subdoses were deformed back to the average phase of the 4-D CT and summed together to yield the final dose, accounting for the interplay effects of periodic free-breathing motion (abbreviated as interplay dose).
To assess the dose errors caused by the interplay effects, the average and range values of the differences between the interplay and the nominal dose (interplay dose–nominal dose) are reported for CTV (mean dose and V50 Gy), heart (mean dose, V25 Gy, and V2 Gy), and ipsilateral lung (mean dose and V20 Gy).
To investigate the dependence of FB motion on beam angles, plans were divided into 2 groups: the plans with 1 or 2 beams with 1 of the beams en face; and the plans with 1 or 2 beams with neither beam en face and at least 1 of the beams tangential. The en face beam angle is typically ∼30° and approximately perpendicular to the breast curvature. The absolute differences of the integrated dose–volume histogram (DVH), ΔDVH, were compared between the 2 groups for CTV, ipsilateral lung, and heart. A Wilcoxon signed rank test was conducted for statistical comparisons of the 2 groups. P values < .05 were considered statistically significant.
To evaluate the dosimetric effect of BH, dose–volume parameters were compared between the FB and BH SBPT plans for CTV (D95%, mean dose, maximum dose), heart (mean dose, V25 Gy, and V2 Gy), and ipsilateral lung (mean dose and V20 Gy).
The dose deviation caused by the interplay effect was small. Table 3 lists the average and range values for differences between the interplay and the nominal dose for CTV (mean dose and V50 Gy), ipsilateral lung (mean dose and V20 Gy), and heart (mean dose, V25 Gy and V2 Gy).
Abbreviations: CTV, clinical target volume; Gy, Gray; RBE, relative biological effectiveness.
Values are mean (range).
Although dosimetric perturbations from respiratory motion were small, CTV coverage was in fact significantly dependent on beam geometry. Figure 1 displays box plots of ΔDVH for CTV, heart, and ipsilateral lung for the 2 groups of plans. The vertical line within the box represents the median value. The ends of the box represent the 75th and 25th quantiles, respectively. The whiskers extending from the ends of the box show the maximum and minimum values. Beam angle dependence was statistically significant for CTV coverage (Figure 1A; P < .001) but not for the dose to heart (Figure 1B; P = .36) and ipsilateral lung (Figure 1C, P = =.23).
As an example, Figure 2 illustrates the dose color wash for a whole breast treatment with BH (Figure 2A) and a treatment with FB (Figure 2B). Two en face–like beams were used in the shown dose distribution. The absence of exit dose for the proton beam avoids heart irradiation, regardless of BH or FB. Table 4 summarizes the dose–volume parameters for the whole breast treatment with BH and FB and for the comprehensive breast treatment with BH and FB.
We have investigated interplay dose associated with FB respiratory motion for 2 common target volumes treated in women with breast cancer (breast alone and breast/CW with regional nodes). The 3 patients were chosen to represent cases in which target volumes are large (1048 cm3), medium (847 cm3), and small (442 cm3). Motion-induced dose deviation was found to be small for the en face beam arrangement for all patients. Fractionation would further mitigate dose deviation such that overall dose errors after multiple fractions would be even less. This was confirmed in a study of postmastectomy radiation therapy using scanned proton beams , where accumulated motion dose was evaluated. In addition, a study on accelerated partial-breast irradiation  reported small respiratory motion effects, although dose errors were represented by differences between the planned dose on FB CT and the recalculated dose on BH CT.
The dosimetric perturbation caused by respiratory motion was small, in part because the range of breast motion was small for the studied patients (Table 1). However, this is also in the range reported by other investigators , thereby making the routine use of 4-D CT simulation for breast cancer uncommon across academic radiation therapy practices and certainly not clinically relevant for SBPT. In other studies, the reported typical magnitude was within 5 mm  and even smaller for regional nodes . In addition, the fact that the breast or CW treatment volumes are large and homogeneous makes the relative magnitude of motion and the change of radiobiological beam path with motion even less.
We found that the en face beam was more robust to the motion induced by FB, whereas the tangential beam was relatively sensitive to this motion. For the en face beam, beam direction is approximately parallel with the breast motion trajectory, whereas for the tangential beam, beam direction is almost perpendicular to the motion trajectory of the breast, which can result in a large change in radiological beam path length. The same beam angle dependence was also observed in a partial-breast study .
In SBPT, the freedom of adapting the range of the proton beam per spot gives rise to superior distal conformality, and the absence of exit dose minimizes cardiac irradiation. The change of anatomy from FB to BH results in negligible differences to dose conformality, and therefore has a minute impact on cardiac dose. This contrasts with photon therapy with tangential beams, where a considerable portion of the heart can be irradiated, and the increased dorsal separation of heart from breast caused by BH can effectively shift the heart away from the radiation fields, resulting in dosimetric cardiac sparing. In fact, for the selected patient in this study who received photon therapy with BH, the received heart mean dose was 1 Gy. If the patient had been treated with photon therapy with FB, the calculated heart mean dose would have been 4 Gy. However, for the same patient, the SBPT with FB already yielded low heart dose (the average value of heart mean dose is 0.5 Gy), and change to the SBPT with BH gave rise to a similar heart dose (the average value of heart mean dose is 0.2 Gy) (Table 4). In studies on larger patient populations [8, 9], the reported average values of heart mean dose with FB were within 1 Gy (RBE). Hence, it is likely that, in general, additional heart dose reduction with BH will be insignificant.
BH techniques require synchronization of the breathing signal with the delivery system. For a scanning proton beam generated by a synchrotron, such as the system simulated in this study, it takes about 2 seconds to switch energy layers, which means that the beam delivery only occurs for a portion of time during one period of BH. Hence, the BH treatment could considerably prolong treatment time. The prolonged treatment time can challenge a patient's cooperation and tolerance, which could affect the reproducibility of BH positioning. To the best of our knowledge, the dosimetric impact of BH reproducibility for scanning proton beam has not been reported, but it can be anticipated that the effect would depend on the beam angle, as it is a similar issue to the FB respiratory motion versus the beam direction. Considering that the heart dose is already low with FB and the challenges in the BH implementation mentioned above, we believe the value of implementing BH technique in SBPT for breast treatments is limited in terms of cardiac dose reduction, and treatment with en face beams is adequate for a robust dose delivery in general. Therefore, the combination of en face beams and free-breathing planning CT acquisition, as well as treatment during FB, should suffice for utilization of SBPT for common breast cancer radiotherapy indications, including whole breast and CW, with or without accompanying nodal coverage.
Future studies will focus on a subset of patients with high cardiac risk factors or preexisting cardiac diseases, for whom any further dose reduction with BH to heart and the cardiac substructures (such as the left anterior descending artery) may be clinically significant, as well as patients with more atypical CW anatomy (such as an extremely thin CW) where even a slight degree of cardiac displacement might be valuable.
Breast SBPT irradiation using an en face beam was robust to the effects of respiratory motion and is feasible. Superior treatment volume coverage as well as cardiopulmonary sparing was maintained in the presence of FB motion with SBPT. Further heart dose reduction with the BH technique was limited and significantly smaller than with photon treatments.
ADDITIONAL INFORMATION AND DECLARATIONS
Conflicts of Interest: The authors have no conflicts of interest to disclose.
Acknowledgments: The authors would like to thank Dr. Nancy Knight for assistance with manuscript preparation.