Lung dose analysis in loco-regional hypofractionated radiotherapy of breast cancer

Discussion

The recent advances in technology and knowledge in radiotherapy had driven attempts of minimizing toxicity. In the particular case of breast RT, the ipsilateral lung is one of the main organ-at-risks, as its irradiation can induce pneumonitis, or fibrosis.9,10 The literature on lung toxicity in breast irradiation is very heterogeneous: different simulation techniques (conventional fluoroscopy versus CT), different treatment planning systems (2-dimensional conformal radiotherapy versus 3-dimensional conformal radiotherapy), different sites treated (chest wall/breast ± ALN/SCN/IMN) were reported. Gokula et al10 demonstrated in a recent meta-analysis that the strongest DVH parameters associated with ipsilateral lung radiation-induced pneumonitis were V_20Gy and MLD. They recommended to keep the V_20Gy < 24% and MLD < 15 Gy without compromising the required dose coverage, and to consider other RT techniques when V_20Gy >30%, or MLD >15 Gy. We also emphasized that the use of SCN, or IMN irradiation fields correlates with an increased incidence of pneumonitis. In our institution, HF dose regimen of 42.4 Gy in 16 fractions was implemented in postoperative breast RT in January 2010 and for whole breast, or chest wall, with or without supra-clavicular field. Providing the relatively long life expectation of BC patients treated by the current therapies and the potential impact of lung damage on their quality of life, it important to analyze the dose received by lung and its possible correlations with patient’s characteristics and treatment delivery parameters. Our data showed that the recommendations of Gokula et al10 for ipsilateral lung dose can be achieved even if a supra-clavicular field is used: we found mean values of 24.6% for V_20Gy and 11.9 Gy for MLD. The contribution to V_20Gy from tangential was almost double that of supra-clavicular fields (16% versus 8.6%). Adding a posterior axillary boost field did not significantly increased the lung dose (p=0.476). Contouring the regional LNs makes the planner aware in evaluating the need for a posterior axillary boost field, and helps in applying different weights of anterior supra-clavicular and posterior axillary boost fields during the plan optimization process, aiming to achieve adequate coverage for ALN and SCN without unnecessarily irradiate the lung. Using a mono-isocentric technique and field-in-field optimization is a common practice in breast RT, widely described in the literature.14-18 A single iso-center has the advantage of a simple and reproducible positioning of the patient on linac couch and is reducing the workload of radiation therapists. Field-in-field technique is a very simplified form of intensity modulated radiation therapy (IMRT), which uses few sub-fields whose shape and weight are adjusted by the planner in order to achieve a homogenous dose distribution. Yet, the plan optimization by field-in-field technique may require the use of high energy photons to achieve an adequate homogeneity of the dose, especially if large volumes are to be irradiated.17,18 Our data showed a weak, but significant correlation between V_20Gy and lung volume (p=0.043), not between MLD and lung volume (p=0.171). While these results are predictable for MLD, the correlation of V_20Gy with lung volume was expected to be relevant, as V_20Gy is defined as percentage of lung volume receiving 20 Gy or more. The volume of the left lung is usually smaller than the right one, we further proceeded to differential analysis, and found that V_20Gy significantly associated with the lung volume for left-sided cases (p=0.006), but not for the right-sided ones (p=0.283). As 43% of our patients were left-sided and 57% were right-sided, the overall effect of the relation between V_20Gy and lung volume appeared lessened. Twenty-four (16%) of our patients were treated with 6 MV, 81 (54%) with mixed energy (6 and 18 MV), and 45 (30%) with 18 MV. Lung dose significantly decreased with beam energy (p<0.0001): mean V_20Gy was 27.8% for 6 MV, 25.4% for mixed energy, and 21.2% for 18 MV. These results are not surprising. The increased lateral range of secondary electrons in low density tissue, such as lung, when using photon energies higher than 10 MV generates a build-up region in lung tumors and increases the beam penumbra, resulting in reduced dose to the periphery of the target near the beam edge. This is a well-known phenomenon, which has been widely assessed in phantom measurement and Monte Carlo calculations.19-25 DesRosiers et al19 compared 6 MV to 15 MV photons in a Monte Carlo simulation for different beam arrangements in low density media and demonstrated that for every beam arrangement 15 MV photons delivered 5-10% lower dose than 6 MV photons. In a similar study, comparing 6 MV to 18 MV for IMRT treatment for lung cancer, Madani et al24 found significant difference in target coverage between the 2 energies, and highlighted the need for highly accurate dose calculation algorithms for heterogeneous media. The performance of AAA algorithm used in our study was evaluated against Monte Carlo calculation in low density tissue by several investigators, and found to be an adequate alternative.25-31 Based on the research evidence, there is a preference of low photon energies for the radiotherapy of lung cancer, that are even required in lung cancer radiation Therapy Oncology Group (RTOG) protocols and American Association of Physicists in Medicine (AAPM) guidelines.32-35 While the lung dose sparing by high energy photons is detrimental in lung radiotherapy, it appears to be a significant advantage in breast irradiation, where the lung is an organ-at-risk. Nevertheless, the choice of appropriate beam energy in breast RT planning is determined by patient separation for tangential, and depth of LNs for supra-clavicular fields. This explains the results of our study: the lung dose was decreasing with patient separation and depth of ALN and SCLN (p<0.0001), as a large separation and deep LNs required the use of high energy in order to achieve acceptable dose coverage and homogeneity. However, multivariate analysis indicated the beam energy being the strongest factor affecting V_20Gy (p<0.0001), when compared to patient separation (p=0.169), depth of ALN (p=0.073), and SCLN (p=0.001). High energy photons will generate a characteristic skin-sparing effect, under-dosing the subcutaneous tissue, which is a possible area of clinical failure in breast RT. The effect can be easily outweighed by using a bolus.36,37

Treatment planning takes into account the individual clinical condition of the patient, ensuring adequate dose coverage to regional LNs that may not have been properly addressed surgically, or are known to be involved. The different range of dose coverage for different nodal regions are shown in Figure 1C. The aimed dose coverage of LNs had a significant effect on the lung dose: V_20Gy was increasing with ALN (p=0.001) and SCLN (p=0.003) coverage assessed by V95%. For patients receiving breast radiotherapy, the conventional set-up is supine position, on an angulated breast board. The breast board angle is chosen to align patient’s sternum with the horizontal plane, thus, eliminating the need for collimator angulation on tangential fields and reducing the irradiated lung volume in cranio-caudal direction. In our practice, the 15 degrees breast board angulation is usually selected, but in certain circumstances different angles might be used. While lower angle, such as 10 degrees were applied for large, or pendulous breast and obese patients, aiming to decrease the inframammary fold, larger angle of 20 degrees is occasionally preferred by some patients, for better comfort. The board angulation causes an increased lung volume exposed to the supra-clavicular field. The literature is lacking studies assessing the effect of breast board angulations on plan dosimetry. However, Yang et al38 demonstrated that a 90 degrees rotation of the couch along with a gantry tilt for supra-clavicular field, can outweigh the effect of the board angle, and significantly reduce the lung volume irradiated by the supra-clavicular field from 14.9-5.3%.38 Nevertheless, the technique they proposed does not allow the shield of the spinal cord and a compromise between cord avoidance and lung dose reduction is needed. Our study shows similar results: a trend between lung dose and breast board angle was noticed, favoring a low board angle (mean V_20Gy was 22.9% for 10 degrees, 24.5% for 15 degrees and 25.2% for 20 degrees). As the contribution to lung dose from supra-clavicular field is approximately one third, the benefit appears low. Only 9 (6%) patients were treated on 10 and 20 degrees board, a statistical evaluation was not performed.