Deep inspiration breath hold reduces the mean heart dose in left breast cancer radiotherapy

Most patients with breast cancer, who are treated surgically also undergo postoperative radiotherapy1,2, which has been shown to improve locoregional control, recurrence rates, and survival.3, 4, 5, 6 However, long-term population-based analyses have found that postoperative radiotherapy is associated with an increased risk of mortality due to radiation-induced cardiovascular disease ( CVD).7, 8, 9 In recent years, better radiotherapy treatment planning protocols6,10,11 and widespread use of respiratory gating techniques – which have been applied in all modern linear accelerators – have improved cardioprotection.12, 13, 14, 15, 16

Cardioprotection can be achieved through better treatment planning protocols and respiratory gating. When this latter technique is used, the radiation is delivered only during the inspiratory phase, when the heart is at its most distant point from the chest wall, thus reducing the radiation dose to the heart.8,13, 14, 15, 16 In recent years, the deep inspiration breath hold ( DIBH) technique has become increasingly common due to the growing body of evidence showing that this approach can reduce the mean heart dose (MHD) by 1–3 Gy compared to conventional techniques.17 The use of cardioprotective techniques such as DIBH is crucial in patients with left breast cancer, as these patients have a high risk of developing heart disease within 10 years of radiotherapy treatment.18 Dosimetric studies have shown that DIBH reduces cardiac dose in comparison with free-breathing ( FB) without gating. Additional data from population analyses show that MHD decreases over successive years.19,20

At our institution, we have modified the radiotherapy treatment protocols over time to reflect technological advances and a better understanding of the importance of cardioprotection. From 2014 to 2017, we gradually transitioned from treating patients with non-gated FB to gated FB, and finally to DIBH. Although some studies have compared FB without gating to DIBH12, 13, 14, 15, 16, 17, the studies analysing whether DIBH reduces the risk of cardiotoxicity in a large, real-world clinical cohort of patients are limited.19,20 Likewise, clinical data on the influence of DIBH on cardiac complications in these patients is limited.

In this context, the aim of this study was to compare differences in mean heart dose for patients with left breast cancer treated at our institution from 2014 to 2017 during which we transitioned from non-gated free-breathing (FB) radiotherapy to gated radiotherapy, and finally to DIBH.

This was a retrospective analysis of all patients (n = 2022) diagnosed and treated for breast cancer at West Pomeranian Oncology Center in Szczecin from January 1, 2014 through December 31, 2017.

Patients’ written inform consent about the study was waved because of retrospective clinical data analysis. The study was conducted according the Helsinki Declaration and the European Council Convention on Protection of Human Rights in Bio-Medicine (Oviedo 1997).

Virtually all of patients (99.6%) received postoperative radiotherapy and 1049 (51.9%) were treated for left breast cancer. During the study period, most patients were treated with conventional three-dimensional (3D) radiotherapy (n = 1513, 74.8%) or intensity-modulated radiotherapy (IMRT; n = 69, 3.4%) with free-breathing. A total of 188 patients (9.3%) underwent FB-gated radiotherapy (FB-GRT). Starting in October 2016, all new left breast cancer patients were treated with DIBH. Thus, from that point in time until the study end (2017), the DIBH technique was applied in 252 (12.5%) patients. Gated radiotherapy during FB and DIBH were applied only to left breast cancer patients.

All patients underwent CT-based 3D planning in the therapeutic position. All patients were treated on the same linear accelerator model (Artiste, Siemens Healthcare, Erlangen, Germany). Gated radiotherapy during FB procedures was performed with assistance of a respiratory gating system (AZ-733VI, Anzai Medical Co., Tokyo, Japan), which divides the normal breathing cycle into eight phases, with irradiation administered only during the inhalation phase. Patients who were able to maintain a stable breath cycle received FG-GRT if, in the clinical judgement of the treating radiation oncologist, there was a dosimetric benefit identified by any significant separation of the heart from chest wall (increase of at least 5 mm). DIBH (AlignRT system Vision RT Ltd, London, UK) was used in patients expected to benefit from this approach. Patients unable to hold their breath were not considered eligible for this procedure. The DIBH irradiation technique was used with assistance of a real-time 3D surface tracking system (AlignRT) as described elsewhere.12,21,22

Treatment planning followed institutional protocol. CTV contours were drawn according to ESTRO recommendations23 and heart contours according to Feng et al.24 Regional lymph nodes (ipsilateral axillary and supraclavicular ones) were irradiated in every patients with macrometastases in axillary lymph nodes as internal mammary lymph nodes (upper I–IV) in patients after mastectomy. Five millimetre margin was added to create PTV from CTV. Dose constraints for heart were V₂₀ < 10% (less than 10% of the organ covered be dose of 20Gy), V₄₀ < 5% for conventional fractionation and V₁₇ < 10%, V₃₅ < 5% for hypofractionated regimens.

Treatment plans were created with the Prowess Panther system for IMRT (Radiology Oncology Systems, Inc., San Diego, CA, USA) and the Oncentra Masterplan (Nucletron, Veenendaal, The Netherlands) for other techniques. All patients treated with FB-GRT or gated radiotherapy during FB and DIBH underwent 3D-RT. IMRT was used in patients with left breast cancer if the 3D conformal plan did not meet the prescribed dose constraints.

The data were obtained from the planning systems, which included the: MHD; the heart volume receiving > 40%, 60%, 80%, 100% of the defined dose ( V_40%, V_60%, V_80%, V_100%). The MHD value was expressed in Gy and V_40%, V_60%, V_80%, V_100% were the value for the absolute heart volume in cubic centimetres.

The conventional dose scheme was 50 Gy (2 Gy per fraction administered daily from Monday through Friday) to the breast/chest wall, with or without nodal irradiation. A 10–16 Gy boost to the tumor bed was prescribed for patients undergoing breast-conserving surgery (BCS). Hypofractionated schemes were 42.5–45 Gy (2.25–2.5 Gy per fraction) plus a boost of 10 Gy, or 40.05 Gy (2.67 Gy per fraction) without boost. In the subset of patients who underwent BCS, a total of 155 were given a boost dose with either intraoperative radiotherapy (n = 93) as an early boost or brachytherapy (n = 62). The boost dose was not included in the present analysis. We normalized hypofractionated plans to the conventional scheme (50 Gy in 25 fractions) and recalculated them to obtain the corrected MHD ( MHD_Fx).

Patients who received IMRT were not included in the MHD and MHD_Fx analyses, as the MHD in IMRT plans is higher than those obtained with 3D conformal radiotherapy, with a different impact on cardiac morbidity.20,25

Table 1 shows the clinical characteristics of the left breast cancer patients and treatment parameters according to year of treatment. As that table shows, there were no significant differences in baseline characteristics of the patients (e.g., body mass index, type of surgery, axillary lymph node surgery, nodal irradiation) regardless of the year. Table 1 also shows that the use of IMRT decreased over time as the number of patients undergoing gated therapy increased. Similarly, an increasing proportion of patients received hypofractionated radiotherapy over time.

Overall, MHD values ranged from 0 to 19.44 Gy, with a mean of 2.48 Gy (95% confidence interval [CI], 2.39–2.57). For patients with left breast cancer, the MHD was 3.37 Gy (range, 0.56–19.44 Gy; 95% CI 3.23–3.5) and 1.51 Gy (range 0–17.31; 95% CI 1.43–1.58) for the right side. Overall, the MHD_Fx was 2.62 (range, 0–19.44 Gy; 95% CI, 2.53–2.71). For patients with left breast cancer, the MHD_Fx was 3.52 (range, 0.66–19.44; 95% CI 3.38–3.65) and 1.62 Gy (range 0–17.31; 95% CI 1.54–1.69).

Table 2 shows the MHD and MHD_Fx by year of treatment, indicating that the proportion of patients with left breast cancer exposed to MHD and MHD_Fx values > 4 Gy decreased every year – from 40% in 2014 to 7.9% in 2017 – with a statistically significant decrease in MHD values from 2014 to 2015 and from 2016 to 2017. Similarly, the maximum MHD values fell every year from 2014 to 2017, from 19.44 Gy to 12.27 Gy to 11.07 and finally to 7.36 Gy in 2017. Figures 1 and 2 show these results graphically, indicating an increase in the proportion and number of patients who received lower MHD (p < 0.0001) and MHD _Fx (p < 0.0001).

Figure 1

Figure 2

Despite the above observation every year MHD and MHD_Fx mean values were significantly higher for left-sided breast cancer when compared to right-sided (Table 3). Additionally Table 3 shows, that MHD and MHD_Fx improved every year among those either irradiated to lymph nodes or not and in those with body mass index (BMI) either below or above 30. Every year patients with BMI < 30 were exposed to lower MHD and MHD_Fx and in 2016 and 2017 regional nodal irradiation (RNI) led to higher MHD and MHD_Fx values comparing to no RNI (Table 3).

As Table 4 shows, the mean V_40%, V_60%, V_80%, and V_100% values all decreased year over year, although this decrease was not statistically significant every year (e.g., V_100%). Notably, V_100% improved irrespective of the specific gating technique (FB-GRT or DIBH) versus FB, with the best V_100% values observed in patients treated with DIBH. There was a non-significant difference in mean V_40%, V_60%, V_80% values when comparing gated FB-GRT to non-non-gated FB. For all parameters (V_40%, V_60%, V_80%, V_100%) DIBH was significantly better than gated FB-GRT. DIBH was associated with significantly lower mean V_60% and V_80% values compared to FB. The mean V_40% was lower for DIBH than for FB, but not significantly (p = 0.0529) (Table 4).

Table 5 shows the comparison according to radiation technique (FB vs. FB-GRT vs. DIBH). DIBH was associated with the lowest values (2.1 Gy and 2.31 Gy, respectively), which were significantly lower than those observed for FB-GRT (3.28 Gy, p < 0.0001 and 3.45 Gy, p < 0.0001, respectively) and non-gated FB (3.58 Gy, p < 0.0001 and 3.69 Gy, p < 0.0001, respectively). There were no significant differences between the FB and FB-GRT groups. Note that fewer patients were exposed to high (> 4 Gy) MHD and MHD_Fx values when gated irradiation was used, particularly in the DIBH group in which only 4.8% presented a MHD ≥ 4 Gy versus 22.4% for patients treated with FB-GRT and 35.2% in FB.

Table 6 shows no difference in MHD and MHD_Fx values comparing RNI and no RNI in a group of patients without gating procedure applied. The biggest difference was observed for DIBH (MHD 2.45 vs. 1.89, p < 0.0001, MHD_Fx 2.58 vs. 2.15, p = 0.0026 respectively). MHD and MHD_Fx did not differ significantly in patients with BMI below and above 30 if DIBH was used (Table 6). The difference was significant if FB and FB-GRT was used and MHD and MHD_Fx were higher for patients with BMI above 30 (Table 6).

The present study was performed to evaluate the influence of DIBH on mean heart doses in patients with left breast cancer. Overall, the MHD in patients with left breast cancer was 3.37 Gy. Patients treated with DIBH had significantly lower MHD values than patients treated with FB or FB-GRT techniques (2.1 Gy vs. 3.48 and 3.38 Gy, respectively, p < 0.0001). The lowest MHD values were obtained in the last year of this study (2017), when nearly 85% of left breast cancer patients were treated with DIBH. Moreover, of the patients with MHD values > 4Gy, the smallest proportion was observed in the DIBH group. These data confirm that DIBH reduces the mean radiation dose to the heart in patients with left breast cancer. Drost et al. analysed studies published between 2014 and 2017, reporting MHD 3.6 Gy in left breast cancer patients and 1.7 Gy if any breathing control technique was used (19). The data from 20 sites in United States show that median MHD decreased from 2.19 Gy in 2012 to 1.65 Gy in 2015 (20). Comparable values were observed for left breast cancer patients with median MHD 1.5 Gy for gated radiotherapy.26 In our study the reported values are higher. It might be explained be differences in 3D planning systems used in different centres, as mean values for right breast cancer patients in our study are twice higher than in other studies (1.51 Gy vs. 0.7 Gy).26 Testolin et al. presented data on 280 left breast cancer patients who underwent DIBH combined with IMRT.

The mean MHD was 0.94 Gy in DIBH group and 2.14 Gy in those with no gating.27 Those values are lower than the ones we present, but on the other hand in mentioned trial only 11% of patients were after mastectomy and only 11.4% patients underwent RNI comparing to 44% and 31% in 2017 in our study. Nevertheless in our study mean MHD without RNI was 1.89 Gy.

There is a large body of evidence on the negative impact of excessive radiation doses to the heart. Darby et al.28 showed that every 1 Gy increase in MHD increases the risk of CVD-related mortality by 7.4%. Those authors estimated the risk of developing CVD according to increases in the MHD, as follows: 10% increased risk for MHD < 2 Gy; 30% for MHD 2–4 Gy; and 40% for MHD at 5–9 Gy. Taylor et al. estimated that every additional 1 Gy in MHD is associated with a 4% increase in CVD mortality29, estimating no increase in CVD mortality risk for MHD values < 4 Gy, but an increase in risk of up to 25% for doses ranging from 4–8 Gy.

In our cohort, MHD and MHD_Fx values – which indicate a lower risk of CVD – trended downwards over time as radiotherapy and gating techniques improved. In 2017, most of the left breast cancer patients in our study received MHD doses below 2–4 Gy, and none were exposed to a MHD > 8 Gy. Relevantly, the only factor that changed in this period was the introduction of DIBH irradiation in October 2016.

Sardaro et al. suggested the following planning dose constraints to achieve a low risk (< 1%) of CVD-related mortality: V_30Gy < 20 cm³, V_40Gy < 10 cm³, and V_50Gy < 2 cm³.30 These constraints depend only on the heart volume exposed to a given radiation dose and do not depend on the quality of organ contouring. The values reflecting those constraints in our analysis were V_60%, V_80%, and V_100%. In 2017, only 11 patients did not fulfil those criteria, and only 5 of those patients were treated with DIBH. In other words, 97.8% of breast cancer patients treated with DIBH radiotherapy had a less than 1% increased risk of CVD-related mortality. The largest improvement in V_80%, V_60%, V_40% values was observed in 2017, when DIBH was introduced. Data presented by Testolin et al. suggest that there is still possibility to improve the results as they presented mean V_80%, V_60%, V_40% values below 1 cm³ with combined DIBH and IMRT.27

In our cohort, gated radiotherapy during FB modestly decreased the cardiac dose compared to non-gated FB, possibly due to more stringent qualification criteria, such as the expected benefit from gated radiotherapy during FB and the patients’ ability to cooperate with the procedure. Thus, the non-gated FB group included patients whose heart was optimally located in relation to the chest wall. However, compared to gated and non-gated FB, DIBH resulted in significantly better heart sparing on nearly all parameters. These findings are consistent with other studies that have compared non-gated FB to DIBH, which have shown that DIBH decreases the MHD by 33%–66% from the initial value compared to non-gated FB.12, 13, 14, 15, 16,22,31, 32, 33, 34, 35, 36 However, those studies are limited by the type of analyses performed: the authors created plans based on CT scans obtained during FB and DIBH, and then calculated the estimated (i.e., theoretical) benefit from gated radiotherapy techniques. By contrast, we present real-world data from routine clinical practice, confirming the findings reported by Eldredge et al. in a prospective trial that demonstrated that radiotherapy with the Active Breathing Coordinator (ABC) reduced MHD values by ≥ 20% in 88% of patients.15

In 2017, 86% of left side breast cancer patients successfully underwent radiotherapy with DIBH. Comparable results presented Testolin et al., but Eldredge et al. reported that 72% patients successfully underwent radiotherapy with DIBH using a different gating system (Active Breath Coordinator System, Elekta Instrument AB, Stockholm, Sweden).15,27 Surface monitoring systems seem to be more comfortable for patients.

MHD and MHD_Fx were higher for RNI comparing to no RNI in those who underwent either FB-GRT or DIBH. DIBH offers the best sparing in both clinical scenarios. RNI leads to higher MHD20,27,³⁸, with the strongest impact if internal mammary lymph nodes are irradiated.37 In our study 33% of left breast cancer patients were irradiated to internal mammary lymph nodes. It may explain higher MHD values reported in our study comparing to other series.19,20,27

In our cohort, MHD and MHD_Fx were higher in patients with BMI > 30 if FB or FB-GRT was used but not DIBH. The correlation between BMI and MHD was also reported by Finazzi et al.37

Our results show that the DIBH technique lowers the mean heart dose in patients with left breast cancer treated with radiotherapy, minimising the risk of radiation-induced CVD. Although the clinical impact of these findings remains unknown due to the long latency period, it seems highly probable that lower radiation doses to the heart will reduce radiation-induced CVD in these patients. The data from our study, considered in the context of other published studies, suggest that DIBH should be offered to every patient with left breast cancer to reduce treatment-related morbidity and mortality.