Radiation therapy has an essential role in the management of breast cancer that includes either postlumpectomy radiation for breast conservation in early stages, or postmastectomy radiation for the chest wall in multiple node-positive or locally advanced stages. Large meta-analyses of prospective randomized trials have confirmed that radiation reduces locoregional recurrences and reduces breast cancer mortality. However, the risk for late cardiac effects caused by the proximity of the heart and coronary vessels to the chest wall or regional nodes has historically mitigated some of these benefits of adjuvant radiation. The Early Breast Cancer Trialists’ Collaborative Group reported a meta-analysis of prospective randomized trials of postmastectomy radiation that noted improved survival in node-positive women.1 The 20-year improvement in breast cancer mortality comparing radiation to no radiation was 8.1% (p = 0.001), but the gain reducing death was only 5.0% (p = 0.01). For women with 1-3 positive nodes, a group in whom the controversy about routine radiation has been particularly intractable, the difference between breast cancer mortality and survival was 7.9% (p = 0.01) and 3% (p = ns).
This 3% to 5% difference between breast cancer mortality and overall survival may, in large part, be due to an excess of cardiac disease caused by the radiation of that era of studies in the meta-analysis from 1964 to 1986.1 In the Surveillance Epidemiology and End Results (SEER) database from 1973-1992, there was an excess rate of fatal myocardial infarction of 1% to 2% over the course of 8 to 18 years from treatment for patients receiving left-sided vs. right-sided adjuvant radiation.2 A loss of 1% was also seen between the improvement in breast cancer mortality and overall survival in the postlumpectomy radiation setting from randomized prospective trials conducted between 1976-1999.3 This greater difference of death from nonbreast cancer causes between the postmastectomy and postlumpectomy trials may be decreasing over decades due in part to technical improvements, but the difference may also be due to the greater use of regional node — specifically internal mammary node — radiation in the earlier postmastectomy trials. Radiation of the chest wall and internal mammary nodes (IMNs) has been specifically linked to coronary stenosis in distributions consistent with the radiation fields of conventional radiation.4 For fear of late cardiac injury if IMNs were included for left-sided breast cancer patients, a large prospective population-based cohort study of internal mammary node irradiation treated right-sided patients only.5
In a retrospective review of 2,168 women treated for breast cancer from 1958 to 2001, heart dose was estimated from idealized phantom measurements.6 They found that the mean heart dose correlated with excess relative risk of coronary events by 7.4% per 100 cGy. In that period, the mean heart dose was estimated to be 6.6 Gy for women with tumors in the left breast. In a systematic review of 149 studies published during 2003 to 2013, the mean heart dose from left-sided breast radiation therapy was 5.4 Gy.7 The lowest mean heart doses were from tangential radiation with breathing control (1.3 Gy) or proton radiation (0.5 Gy), and the highest inclusion of internal mammary lymph nodes (8 Gy). Aiming to reduce the mean dose is an important goal for modern radiation therapy in order to reduce ultimate late cardiac complications. In this way, the survival improvements associated with adjuvant radiation could be further improved if excess cardiac deaths could be eliminated altogether.
Early whole-breast irradiation used photon beam 2D techniques consisting of opposed tangential beams of uniform radiation intensity across the field that could be modified with wedge compensators. The introduction of 3-dimensional computed tomography (3D CT) planning in the 1990s permitted the calculation of heart dose in a more precise manner than just observing the amount of the heart silhouette in a tangential portal film. 8 Early attempts to limit heart dose in a breast tangent would be adjusting the beam angle to avoid the heart or adding a block over the heart silhouette.9 Forward planning then developed to optimize dose heterogeneity within the target by manually creating smaller fields using custom blocking or multileaf collimation within a larger tangent — what is known as a “field-in-field” technique.10 In early experiences, such techniques of using beams of nonuniform fluence applied to a target structure were labeled as intensity-modulated radiation therapy (IMRT) but today are considered and reimbursed as 3D conformal radiation. Forward-planned tangential radiation has been shown to be superior to 2D tangential radiation using wedges in 3 prospective randomized trials for reducing desquamation, late skin telangeictasias and fibrosis.11-13 The 3D conformal tangents with forward planning with custom blocking or predefined segments can decrease the heart dose14,15 and normal tissue complication probability for late cardiac toxicity on average by 30%16 compared to using simple wedged tangents.
Prone positioning may have advantages for some women with large or pendulous breasts, or left-sided breast cancers compared to traditional supine positioning. When supine, large- or pendulous-breasted women often have a large separation, or width, between the posterior entry and exit points of the tangential radiation field. This is a cause for large dose inhomogeneity that may only be partially overcome by advances in 3D conformal or IMRT. These women may also have large skin folds particularly in the inframammary region that increase acute dermatitis and risk for moist desquamation. For left-sided women, the lateral displacement of the breast in large women may require a deeper tangent for breast coverage that increases heart dose.
Prone positioning can reduce chest wall separation, deep skin folds, dose inhomogeneity, and heart dose for a large majority of breast cancer patients.17,18 Prone is generally limited to treatment of the breast only, or breast and low axilla,19 but full regional nodal coverage of the high axilla, supraclavicular and internal mammary nodes is generally not possible in the prone position. In addition, caution is needed during simulation for patient selection — judgment of the cardiac anatomy and possible breast tangent — because in a small minority of patients, prone positioning may increase heart dose. In a comparison study of 30 left-sided patients simulated both prone and supine, prone positioning reduced heart and left anterior descending (LAD) doses in 19 patients, increased it in 8 patients, and had no effect in 3 patients.20 In a prospective study of 200 left-sided patients simulated both supine and prone, prone position was associated with an 85% reduction of in-field heart volumes compared to supine.21 This did not reach significance in small-breasted women. A benefit was seen in 85% of patients to prone positioning for the heart volume in the radiation field, but supine position was better for 15%.
IMRT describes an inverse planning technique in which beams of nonuniform fluence are created by optimizing coverage of a planning target volume (PTV). Much use of IMRT in adjuvant treatment of breast cancer has been using standard tangential beam arrangements. A benefit in dose homogeneity with inverse planned or hybrid IMRT techniques compared to forward-planned 3D conformal has been shown in some studies22,23 but not all.24 IMRT has been reported to reduce dose to heart compared to 3D in most studies25-27 but not others.28 There can be significant variation in patient anatomy so that there are overlapping ranges of heart dose for IMRT vs. 3D, and IMRT may be superior to 3D in heart dose for some patients but not all.14 There may also be a tradeoff in reduced PTV coverage with IMRT that prioritizes cardiac sparing.14,15 In some studies, an added benefit for IMRT is an overall reduced planning time and decreased dependence on dosimetrist experience compared to 3D conformal.29,30
ASTRO’s Choosing Wisely campaign advocated against the routine use of IMRT to deliver whole-breast radiation therapy. The randomized Canadian multicenter study that showed reduced acute toxicity from tangential radiation with IMRT compared to 2D tangents did include patients treated with forward-planned or inverse-planned IMRT.11 However, the Cambridge Breast IMRT trial did not show a reduction in toxicity.31 The rates of IMRT for breast cancer increased dramatically from 2001 to 2011,32,33 and this increase in IMRT usage is associated with a markedly higher cost for adjuvant radiation.32 The Radiation Therapy Oncology Group trial 1005 was a phase III trial that created a database of CT plans for approximately 2,000 patients treated with whole-breast radiation from 2011 to 2014. The trial allowed field-in-field 3D conformal or IMRT as long as preset minimum constraints could be met. A subgroup analysis of differences in mean heart dose and late toxicity outcomes will be a useful prospective, although not randomized, comparison.
Certain patient subgroups may benefit from inverse planning IMRT compared to 3D conformal. This could be an option for some cases of challenging anatomy, such as large chest wall separation causing dose inhomogeneity; left-sided cases with a large amount of heart close to the chest wall or pectus excavatum; or where internal mammary node irradiation is needed. Inverse-planned IMRT has been shown to improve dosimetric coverage, homogeneity, and high doses received by the lung and heart for patients requiring internal mammary node irradiation compared to partly wide tangents or mixed-beam plans. 26 However, the tradeoff is that the addition of nontangential beams to IMRT increases the low-dose radiation to the heart and V5 dose.34-38 IMRT should be considered and comparison plans created when 3D conformal forward planning is not able to achieve the initial desired dose goals.
There are several commercially available methods for respiratory control during radiation therapy for breast cancer. The purpose is to use an increase in lung volume and inferior displacement of the diaphragm to increase the distance between the heart and the breast/chest wall to reduce radiation dose. In one method, an active-breathing control (ABC) device is used for regulation of respiratory inspiratory volume. The other method relies on patient coaching for voluntary deep inspiration breath holding (DIBH) that is verified with either direct volume measurement or surface anatomy verification.
Studies comparing mean heart dose with free breathing vs. respiratory control are shown in Table 1. In one study, moderate DIBH with ABC in 87 of 99 (88%) patients was associated with a mean heart dose of 254 cGy compared to 423 cGy with free breathing (FB) (p < 0.001).39 In a prospective study of ABC for left-sided breast cancer, 72% of enrolled patients were ultimately treated with ABC with inability to tolerate the procedure being the predominant cause for ineligibility.40 The mean heart dose was reduced by ABC compared to FB by > 20% in 88% of patients, and the median mean heart dose was 270 cGy for FB compared with 90 cGy for ABC. Mast et al compared free breathing (FB) to DIBH plans with tangential 3D conformal and IMRT techniques.27 For the heart and LAD-region, a significant dose reduction was found with DIBH (p < 0.01). The mean heart dose for 3D vs. IMRT in 20 patients was 180 cGy compared to 150 cGy in DIBH, and 330 cGy and 270 cGy in FB, respectively (p = 0.01). In a prospective study of 17 left-sided patients, supine position with DIBH significantly reduced the volume of the heart receiving 30 Gy, the mean heart dose, and mean LAD coronary artery dose compared to supine with FB and prone positioning.41 In a study of 35 patients planned with FB or DIBH, mean dose for heart was 90 cGy vs. 250 cGy, (p < 0.0001)42 and in 75% of patients there was felt to be a benefit to DIBH. In a prospective registry of 150 patients, in which patients were selected for FB (38) or DIBH (110) at physician discretion, DIBH plans were associated with a mean heart dose of 137.6 cGy compared to 255.7 cGy with FB (p < 0.0001).43 On multivariate analysis, younger age, higher BMI, and larger change in lung volume between scans were associated with a greater change in mean heart dose between techniques.
The improvement of cardiac dose with respiratory control now seems well settled. These techniques have been shown to be clinically practical and have no significant impact on patient treatment time and throughput.43,44 Whether this will lead to clinically evident reduction in cardiac events is unknown. In one prospective study of ABC vs. FB, there was decreased dose to the left ventricle but no change in myocardial perfusion changes 6 months after treatment.45 Further research is also needed to determine how best to select patients. The IMN chain may be particularly sensitive to changes in position and dose coverage with respiratory motion,46,47 and ABC has been shown to improve heart dose, particularly in the setting of IMN irradiation.48 All patients with need for internal mammary node radiation would seem good candidates for respiratory control. However, treating all left-sided patients who may tolerate it may also lead to overutilization of resources in a significant minority of patients who may be appropriately treated with FB. Further research is needed to determine whether physicians can appropriately select patients at the time of simulation on a case-by-case basis,43 or whether objective measures may predict accurately who will benefit most from respiratory control.42
Proton radiation therapy may have dosimetric advantages compared to photons due to the property of the positively charged proton depositing the bulk of its energy in tissue in a finite range, or Bragg peak, with essentially no residual radiation beyond this depth. In clinical application to breast cancer, this could theoretically allow full breast or nodal target coverage within the Bragg peak with no dose to heart and lung posteriorly beyond the Bragg peak. Dosimetric studies have demonstrated the superiority of proton therapy in the postmastectomy radiation therapy setting with respect to low doses to organs-at-risk while maintaining superior target coverage, particularly regional nodes.49,50 In a report of 12 patients treated in a prospective clinical trial, 11 left-sided patients achieved an average mean heart dose of 44 cGy, and had 75% grade 2 acute skin toxicity (no grade 3) and only 1 grade 3 toxicity (fatigue).51 In a report of 30 patients, most treated to internal mammary nodes, the mean heart dose achieved was 1 Gy for left-sided patients.52 There was grade 2 dermatitis in 71%, moist desquamation in 29%, grade 2 esophagitis in 29%, and 1 grade 3 reconstructive complication. Proton therapy may reduce risk for cardiac toxicity of radiation compared to photon radiation by not only reducing mean heart dose, but dose to the critical coronary artery structures on the heart’s surface.53 In one study, a scanning proton technique for left-sided irradiation was associated with lower minimum, maximum, and dose to 0.2 cc of the LAD coronary artery than the best possible photon beam radiation technique (IMRT with DIBH).54
In practice, there are several limita-tions of protons. Coverage of the width of the breast and other targets in the patient requires creation of a wider spread-out Bragg peak (SOBP) that increases skin dose. Proton therapy distal range has intrinsic uncertainty that can lead to overshooting or undershooting the posterior target edge, and greater sensitivity to patient or organ motion. The potential advantage to protons is thought to be physical and not biological — protons are estimated to have a relative biologic effectiveness (RBE) of 1.1 compared to photons, which is taken into account for dose calculations by treatment planning systems. In actuality, there may be variation of proton linear energy transfer along the track length causing lower RBE in the SOBP and higher RBE at the end track that could potentially lower tumor control or increase complication probabilities compared to current planning system estimates.55 Current methods of proton techniques such as double scattering have limitations in field size, matching, and dose shaping. More advanced techniques like pencil-beam scanning and intensity-modulated proton therapy could potentially treat some of the most challenging postmastectomy radiation therapy cases, due to breast reconstruction, internal mammary node coverage, or lower skin dose, but may not be clinically deliverable with current equipment.53,56
The RADCOMP breast proton vs. photon study [NCT02603341] is being conducted on the hypothesis that proton therapy for locally advanced breast cancer reduces major cardiovascular events, is noninferior in cancer control, and improves health-related quality of life compared to photon therapy. Participants in the trial will be randomized to either proton or photon therapy. The inclusion criteria is broad: mastectomy with or without reconstruction or lumpectomy, any type of axillary surgery, any adjuvant or neoadjuvant chemotherapy, and left- or right-sided breast cancer as long as internal mammary nodes are intended to be treated.
This report has reviewed the wide variety of techniques for adjuvant breast or chest wall radiation therapy for minimizing heart dose. Field-in-field 3D conformal (forward planning) may be seen as the current minimum standard for breast patients today (Figure 1). In many cases, greater cardiac sparing can be achieved with prone positioning (Figure 2), DIBH (Figure 3), IMRT with 2 or more fixed angles (Figure 3), IMRT with arcs (Figure 4), or protons (Figure 5). One challenge to the practicing clinician is acquiring the equipment and experience to have one or more of the options available for their patients, which is subject to constraints on department staff and resources. In a large radiation therapy department with all of these potential options, or a region where referral to specialty centers is possible, another challenge is developing the experience to select patients a priori or at the time of simulation for one or the other modality. Matching the best approach for each patient’s unique target needs and anatomy is necessary instead of a one-size-fits-all approach to cardiac avoidance.
Freedman GM, Lin L. Cardiac-sparing radiation therapy for breast cancer. Appl Rad Oncol. 2016;5(3):6-11.
