To investigate the feasibility of using scanned proton beams as adjuvant radiation therapy for breast cancer. Long-term cardiopulmonary complications may worsen the quality of life and reduce the positive contribution of radiation therapy, which has been known to improve long-term control of locoregional disease as well as the long-term survival for these patients.
Ten patients with stage I-III cancer (either after mastectomy or lumpectomy, left- or right-sided) were included in the study. The patients were identified from a larger group where dose heterogeneity in the target and/or hotspots in the normal tissues qualified them for irregular surface compensator planning with photons. The patients underwent planning with 2 scanned proton beam planning techniques, single-field uniform dose and intensity-modulated proton therapy, and the results were compared with those from irregular surface compensator. All volumes of interest were delineated and reviewed by experienced radio-oncologists. The patients were prescribed 50 GyRBE in 25 fractions. Dosimetric parameters of interest were compared with a paired, 2-tailed Student t test.
The proton plans showed comparable or better target coverage than the original photon plans. There were also large reductions with protons in mean doses to the heart (0.2 versus 1.3 GyRBE), left anterior descending artery (1.4 versus 6.4 GyRBE), and the ipsilateral lung (6.3 versus 7.7 GyRBE). This reduction is important from the point of view of the quality of life of the patients after radiation therapy. No significant differences were found between single-field uniform dose and intensity-modulated proton therapy plans.
Spot scanning technique with protons may improve target dose homogeneity and further reduce doses to the organs at risk compared with advanced photon techniques. The results from this study indicate a potential for protons as adjuvant radiation therapy in breast cancer and a further step toward the individualization of treatment based on anatomic and comorbidity characteristics.
According to the latest statistics, breast cancer is the most common cancer in women worldwide, both in the more developed and the less developed regions, with the highest incidence rates reported in the developed countries [1, 2]. Radiation therapy is successfully used nowadays to increase long-term control of local and regional disease as well as for long-term survival in patients with breast cancer . Given the high rates of survival after radiation therapy for these patients, great efforts are made to reduce the risk for iatrogenic side effects in the heart and the lung, which may worsen the quality of life or even threaten the life of the patients. Consequently, a broad array of techniques has been developed for breast cancer radiation therapy. The most commonly used technique today is 3-dimensional conformal photon radiation therapy with multileaf collimators, dynamic wedges, and compensating fields, but advanced techniques that use intensity modulation of photon fields have also been explored for their potential to improve the dose homogeneity in the target and to reduce the cardiopulmonary doses [4–6]. More recently, proton radiation therapy has also been proposed as an option in the arsenal of anticancer therapies to further improve the outcome for patients with breast cancer [7, 8]. Indeed, the finite range of the protons and the steep dose falloff at the distal edge of their range could lead to favorable dose distributions , which may be of interest from the point of view of the irregular geometry of patients with breast cancer, with their curved surfaces and heterogeneous anatomy. Consequently, several dose planning studies and prospective clinical trials have been initiated to explore the potential of protons for breast radiation therapy [10–20]. These studies reflect the technologic evolution in proton beam technology, with most of the clinical studies using passively scattered proton fields; however, the potential of active scanning beams has been explored in planning studies, as the latter have recently become clinically available. Scanned proton beams have the advantage of allowing further modulation of the beam  and reduce the amount of neutrons from in-beam activation in comparison to passively scattered beams . The present study aims to add to the existing knowledge by comparing the potential of protons with the irregular surface compensator technique with photons, for increasing target homogeneity and reducing the doses to organs at risk (OARs) in breast cancer cases not covered by previous studies. Thus, the purpose of this study is to further explore the potential of pencil beam scanning proton technique for breast cancer radiation therapy.
Methods and Materials
Ten patients with breast cancer (5 left-sided and 5 right-sided) who received postoperative radiation treatment to the whole breast only (WBO)—4 patients—or to the breast and the supraclavicular lymph nodes (BSC)—6 patients—were included in the analysis. The patients represented difficult cases for which routine conformal radiation therapy with photons led to high heterogeneity in the planning target volume (PTV) or hotspots outside the PTV—which were unacceptable according to the criteria used as standards of care—thus making them candidates for planning with the irregular surface compensator (ISC) technique . This technique is a forward planning technique implemented in the Eclipse treatment planning system (Varian Medical Systems, Palo Alto, California), using dynamic multileaf collimators to modulate beamlets across the photon fields to improve dose distributions for rounded body contours and target volumes.
All patients underwent computed tomography (CT) scanning after surgery, in supine position, with 2-mm slice thickness, for treatment planning. The relevant structures of interest, the clinical target volume of the original tumour, the PTV, and the OARs were delineated or approved by experienced radiation oncologists according to local practice. The OARs included the heart, the left anterior descending artery (LAD), the lungs, and the contralateral breast. For 4 patients, bolus was used around the mastectomy scar for photon treatment according to local standards of care. Photon plans were created in the Eclipse treatment planning system, with tangential fields for WBO patients and tangential fields and anteroposterior fields for BSC patients, with a monoisocentric technique with the isocenter placed at the junction between the breast and the supraclavicular region. The tangential fields were planned with the ISC technique by using a transmission penetration depth of 50% .
The patients subsequently underwent planning with the Eclipse treatment planning system with scanned proton pencil beams, using both single-field uniform dose (SFUD) and full intensity-modulated proton therapy (IMPT) techniques. In the SFUD technique each proton field attempts to deliver a uniform dose to the PTV, while in the IMPT technique the fields are simultaneously optimized to deliver a uniform dose over the PTV. A 3-field technique has been used in all cases, as it has been suggested that a multifield technique has the flexibility and robustness required for breast cancer radiation therapy with protons . The beam angles used were 20°, 60°, and 340° for patients with left-sided breast cancer and 20°, 300°, and 340° for patients with right-sided breast cancer to provide adequate coverage of the PTV and also limit the dose spilling to the contralateral side. A generic value of 1.1 was used for the relative biological effectiveness (RBE) of protons as recommended by the International Commission on Radiation Units and Measurements (ICRU) . All photon and proton plans were normalized so that the mean dose to the PTV was equal to 50 GyRBE in 25 fractions.
The resulting plans were dosimetrically evaluated for target coverage and radiation burden to the OARs. For the PTV, the following variables were included in the analysis: the volume receiving at least 93% of the prescribed dose (V93%), the ICRU-recommended  near minimum dose, D98% (the dose to 98% of the volume), the near maximum dose, D2% (the dose to 2% of the volume), and the heterogeneity index (HI), defined as (1)
where D50% is the median dose to the volume. The comparison for the OARs was performed for dosimetric parameters used either as constraints in international and national recommendations or as parameters for radiobiological models for normal complications [24, 25] and included the mean dose (Dmean) and near maximum dose (D2%) (D2%) to the OARs.
Integral doses in Gy·kg for photons and in GyRBE·kg for protons were also calculated from the average doses and volumes of the structures of interest delineated on the planning CT by using generic values for the volumetric mass densities: 260 kg/m3 for lungs and 1060 kg/m3 for all other tissues.
A paired, 2-tailed Student t test was used to evaluate the statistical significance of the differences in the dosimetric parameters between the proton and the photon plans.
The dosimetric parameters for target coverage and normal tissue irradiation for all the patients are summarized in Tables 1 and 2. These results show the potential of both SFUD and IMPT proton plans to reduce the size of the hotspots (V105% = 4.7% for photons versus 0.4% for SFUD and 0.0% for IMPT) and to improve the dose homogeneity in the target (HI = 12.6% for photons versus 9.0% for SFUD and 5.9% for IMPT). Furthermore, IMPT has potential for improving the coverage of the PTV with the 93% isodose (V93% = 98.3% with photons and 99.5% with protons, P = .01). These results with the proton technique are quite promising for improving target coverage for difficult cases beyond what could be achieved with advanced photon techniques such as ISC planning. Indeed, a previous study by the same group  has shown that ISC is capable of increasing the homogeneity index in the target, as well as reducing the dose hotspots, in comparison to routine 3-dimensional conformal photon radiation therapy.
With respect to the irradiation of the normal tissues, Table 2 shows a trend toward reducing the radiation burden to the heart (Dmean = 1.3 Gy for photons versus 0.2 GyRBE for protons and D2% = 9.8 Gy for photons versus 3.2 GyRBE for protons) and the LAD (Dmean = 6.4 Gy for photons versus 1.4 GyRBE for protons, and D2% = 13.5 Gy for photons versus 4.7 GyRBE for protons). The trend neared statistical significance for patients with left-sided breast cancer (Figure 1, Table 3). For the ipsilateral lung, the protons generally appear to reduce the radiation burden in terms of mean and near maximum dose, but they might increase the low-dose component of the dose volume histogram (DVH) (Figure 2). However, the trend does not seem to be valid for all patients, as illustrated by the DVH curves in Figure 3. Proton plans also showed a tendency toward increasing the radiation burden of the contralateral breast, probably because of the chosen beam arrangement, but the doses are quite low and would not pose a problem in the case of a previous or a later treatment to the other side (Table 2).
Large differences were also found in terms of the integral dose to the normal tissues outside the PTV, with average values of 67 Gy·kg for photons versus 48 GyRBE·kg for both SFUD and IMPT when all patients were included in the analysis. This reduction is mainly caused by the limited range of the protons, as illustrated in Figure 4, and therefore the difference is even higher for the subgroup of BSC patients included in the analysis (79 Gy·kg for photons versus 53 GyRBE·kg for protons).
The results of this study have shown that proton therapy with scanned pencil beams has the potential to improve target coverage and also to reduce the irradiation of the normal tissues in patients with breast cancer beyond what could be achieved with advanced photon techniques. Few similar studies have been found in the recent literature. Thus, Ares et al  studied the potential of scanned proton therapy in 20 patients with left-sided breast cancer and more recently, Jimenez et al  reported on the use of scanned protons for 5 patients with bilateral breast implants. The present study adds to the existing knowledge by including both WBO and BSC patients to explore a broader range of possible indications. Furthermore, the focus in the present study has been on patients for whom routine planning with photons led to high target heterogeneity and/or unacceptable hotspots. This is an important aspect because, for a considerable fraction of patients, 3-dimensional conformal photon radiation therapy routinely used nowadays for breast cancer may lead to good target dose homogeneity and reduced radiation burden to the OARs. Nevertheless, the normal tissue radiation burden should still be decreased for some patients, and therefore the use of more advanced and costly techniques with photons and protons might have to be explored. With this group of patients in mind, the photon references used for comparison in the present study have been plans created with the ISC technique. Irregular surface compensator is an electronic compensation planning technique that improves dose homogeneity in the target for patients with breast cancer who receive photon radiation therapy, leading to similar dose distributions compared with full volume-based intensity modulated radiation therapy (IMRT) technique [5, 6]. From this perspective, the choice of the ISC technique for the photon plans ensures an advanced clinically used reference against which the real potential of protons to improve dose distributions could be tested.
The significant reduction of the mean cardiac doses predicted with protons, from 2.2 Gy with photons to 0.4 GyRBE with P = .05 with SFUD and P = .04 with IMPT, has deep implications for the risk of ischemic heart disease for survivors of radiation therapy. Thus, a recent study  showed that the rates of major coronary events increased linearly with the mean dose to the heart by 7.4% per gray, with no apparent threshold. Consequently, the potential 5-fold reduction identified in this study with protons, in comparison to the ISC technique, makes proton therapy a potential candidate for patients at high risk for coronary events due to unfavourable anatomy or coexisting cardiac morbidities. Similarly, proton therapy has potential for reducing the doses to the LAD, which is a substructure of interest for the development of cardiac complications . The decreasing trend did not reach statistical significance, possibly because of the uncertainties in delineating such a small structure as the LAD on images with standard radiation therapy quality . However, the use of contrast agents for improving the delineation of this structure will further increase the uncertainty of the conversion of CT images into stopping powers for protons needed for dose calculations, and therefore this approach was not used in the present study.
The reduction of the dose to the lungs with the use of protons has been identified as a trend only in the present study, while previous studies [15, 18] have reported a significant decrease of the radiation burden for this organ. The difference could result from the present study's use of mainly difficult cases, which required much finer balancing of target coverage and irradiation of the lung to produce clinically acceptable plans. A potentially concerning aspect is the trend toward increasing the low-dose component of the ipsilateral lung DVH with protons, in light of a recent study  reporting a strong correlation between radiologic changes in lung morphology and lung volumes receiving low doses. However, patients with radiologic pneumonitis after photon irradiation in the cited study had considerably higher irradiated lung volumes than those reported in the present study, either with photons or with protons; therefore, the clinical impact of the increase is probably quite limited. Nevertheless, further investigations of this aspect are warranted.
The present study also found a reduction in the integral body dose with protons versus photons. This was partly expected, since it is generally believed that the finite range of the protons lead to near zero dose behind the target volume [9, 29]. Nevertheless, it is worth mentioning that the largest difference was found for BSC patients, as the photon irradiation technique uses opposing fields through a large volume of normal tissue. For WBO patients the difference in integral doses was much smaller, 48 Gy·kg for photons versus 40 GyRBE·kg for protons, as the photon technique uses tangential fields over the breast to limit the amount of normal tissue irradiated. However, the relative biological effectiveness of the protons is a source of uncertainty in the estimation of integral body dose for protons. Thus, the generic value of 1.1 for the proton RBE has been determined in the mid-position of a spread-out Bragg peak , while several studies [30–32] have shown that proton RBE varies along the particle track and with the irradiated tissue. Largest variations are expected for late-reacting normal tissues situated in the distal part of the proton's track and beyond, from where most of the contribution to integral dose in proton breast radiation therapy comes. More accurate determinations of integral doses, taking into account the distributions of dose and linear energy transfer in the tissues , are beyond the purpose of the present study. It should also be pointed out that the preferred unit for integral body dose from protons should be GyRBE·kg, rather than joule, as the joule is a derived unit of energy in the International System of Units (SI). Whereas for photons Gy·kg is equivalent to joules, the unit used for protons, GyRBE·kg, would be equivalent to “J (RBE),” which does not correspond to the SI definition of joule.
Besides the uncertainties in relative biological effectiveness for protons, several other uncertainties may appear for proton dose planning. Thus, CT calibration and range uncertainties are known to influence the calculated proton dose distributions; however, an analysis of this aspect, performed by Ares and colleagues  for a similar beam arrangement, found quite a small impact of range uncertainties for breast cancer planning with protons. This is because the beam arrangement used in the present study follows some general principles that provide robustness in proton beam planning, such as the avoidance of a single beam that may go through many anatomic heterogeneities or be directed toward an OAR, and that also allows the sparing of the skin. Furthermore, a small impact was also reported by the same group for respiratory motion, even for the relatively complex spot-scanned IMPT plans. From the perspective of these results, the expected impact of these uncertainties on the resulting dose distributions is rather small.
The findings of the present study suggest that proton therapy with scanned beams has the potential to improve target coverage and to reduce the radiation burden to OARs in breast radiation therapy beyond what could be achieved with advanced photon techniques. Improving target coverage is particularly important for patients with multifocal or lobular disease for which cold spots in the PTV should be avoided. Similarly, pencil beam scanning techniques with protons could also reduce the doses to OARs, which in turn may translate into reduced iatrogenic toxicity and improved quality of life for the patients after radiation therapy. Depending on anatomic and comorbidity characteristics, this could be a step toward further individualization of treatment for patients with breast cancer.
ADDITIONAL INFORMATION AND DECLARATIONS
Conflicts of Interest: The authors have no conflicts to disclose.