Pectus excavatum and adjuvant radiotherapy for early stage breast cancer: balancing dose to target versus heart

Adjuvant radiotherapy (RT) significantly reduces the risk of loco-regional recurrence and increases survival after breast conserving surgery or mastectomy in high-risk early stage breast cancer patients [1]. The addition of internal mammary node (IMN) irradiation further improves the outcome of the patients [2–4].

Three-dimensional conformal radiotherapy (3D-CRT) is generally a safe and fast way to reduce toxicity and still achieve an acceptable coverage of CTV and good control rates [5]. Still, left-sided irradiation delivers a higher mean heart dose (MHD) compared to right-sided irradiation and it is associated with an increase in late cardiac morbidity and mortality [6]. In order to achieve a low heart dose, treatment planning is more complicated when there is a deformity of the chest wall, such as pectus excavatum (PE), especially with inclusion of the left-sided IMNs. Few studies have investigated patients with PE treated with RT for early stage breast cancer. In one study 3D-CRT was compared to more advanced RT techniques regarding dose to target and organs at risk (OARs) [7]. Another study evaluated 3D-CRT in ten patients with PE treated with whole breast irradiation (WBI) and compared MHDs to those in a group of breast cancer patients without PE [8].

The aim of this study was to evaluate the suitability of 3D-CRT for patients with and without lymph node positive early stage breast cancer and PE.

Material and methods

From 2013 to 2017, 18 consecutive women with PE causing a challenge in treatment planning were prospectively identified among approximately 2500 women with early stage breast cancer referred to our institution for adjuvant RT. There was an indication for loco-regional RT in nine women (seven left-sided and two right-sided), while the remaining received WBI (four left-sided and five right-sided). All patients had PE defined as a Haller index >2 measured on the treatment planning computed tomography (CT) [7].

Planning CT scans without intravenous contrast with the patient in treatment position using a slice thickness of 3 mm were acquired. Contouring and dose planning was performed with the Varian Eclipse (AAA dose algorithm) radiotherapy planning system. Targets were delineated and named according to the ESTRO guideline [9,10] and contouring was verified by one oncologist (BVO). The nodal target in patients treated with loco-regional RT included a clinical target volume (CTVn) of level 2–4, interpectoral and IMN and in patients with extensive nodal disease also CTVn_L1. The CTVn IMN included intercostal spaces 1–3. Since it may be challenging to achieve optimal dose to CTVn IMN in PE patients, dose to nodal targets was reported separated into dose to CTVn_IMN and dose to a combined volume of CTVn_L(1), 2, 3, 4, and interpectoral nodes, respectively (CTVn_L1–L4). The planning target volume (PTV) was defined as CTV +5 mm. OARs included heart, lungs and contralateral breast. The treatment technique was 3D-CRT for all patients but one treated with intensity modulated RT (IMRT). Dose distribution to CTVp_breast/chest wall was recommended 95–105%, whilst nodal volumes were to be covered by 90–105% dose according to Danish Breast Cancer Group (DBCG) guidelines (www.dbcg.dk).

Initially, a target prioritized 3D-CRT plan (TPP) was created for each patient ensuring full target coverage and max 25 or 35% of the ipsilateral lung for WBI or loco-regional irradiation received 17 Gy or more, respectively. DBCG guidelines for heart constraints recommended max 10% of the heart received 17 Gy or more and max 5% received 35 Gy or more. However, in the TPP no attempts were made to minimize dose to the heart. For all patients, a second OARs prioritized plan (OARPP) was made for each patient by modifying the TPP in order to respect the constraints of OARs and avoiding large doses to the contralateral breast. Thus, a compromise on dose to target was accepted. This plan was approved for treatment and used in the clinic. The prescribed treatment dose was 40 Gy in 15 fractions, five fractions per week for all patients.

Dose-volume histograms (DVHs) were calculated for all delineated volumes. The relative volume irradiated to a minimum absolute dose x Gy, V_x Gy or a minimum relative to referred dose x%, V_x% was determined from the DVH. Quantile–Quantile plots were used to examine the distribution of variables. For normally distributed variables, mean and 95% confidence intervals (CIs) were calculated. Means were compared with Student’s paired t-test. For non-normally distributed variables, median and range were reported, and the Wilcoxon signed-rank test was used for comparison. Differences were considered statistically significant for p < .05. Analyses were carried out in STATA version 11.2 (Stata Corp., College Station, TX, USA).

Results

The coverage of CTVp_breast/chest wall and CTVn L1–L4 fulfilled the dose constraints in both the TPP and OARPP. In the TPP, it was possible to obtain an acceptable dose coverage of CTVn IMN. In the OARPP the dose coverage to CTVn IMN decreased to 88% (range, 20%; 97%) from 99% (range, 71%; 100%) in the TPP (p = .04) (Table 1). For patients treated with left-sided loco-regional RT, the coverage of PTVn IMN in the TPP was 83% (range, 68%; 87%) and in the OARPP 75% (range, 32%; 87%) (p =.09).

Table 1. Relative target volumes in the TPP and OARPP covered by 90% of the prescribed dose.

Target	n	TPP Median (range)	OARPP Median (range)
V90 (%)
CTVn_L1–L4, right	2	94 (88;99)	93 (89;98)
CTVn_IMN, right	2	96 (93;99)	94 (92;97)
CTVn_L1–L4, left	7	100 (76;100)	95 (87;99)
CTVn_IMN, left	7	99 (71;100)	88 (20;97)
V90 (%)
PTVn_L1–L4, right	2	88 (85;92)	89 (87;91)
PTVn_IMN, right	2	77 (68;86)	87 (73;100)
PTVn_L1–L4, left	7	92 (63;99)	90 (74;95)
PTVn_IMN, left	7	83 (68;87)	75 (32;87)

TPP: target prioritized plan; OARPP: organs at risk prioritized plan; V_x, Relative volume V irradiated to a minimum x % of prescribed dose; CTV: clinical target volume; PTV: planning target volume; L1–L4: levels 1–4 including the interpectoral level; IMN: internal mammary node.

In left-sided loco-regional treatment plans the average MHD was 3.0 Gy (95% CI [1.9 Gy; 4.2 Gy]) in the TPP compared to 1.7 Gy (95% CI [0.9 Gy; 2.4 Gy]) (p < .01) in the OARPP (Table 2).

Table 2. Dose parameters of the organs at risk in the TPP and OARPP.

Organ at risk	n	TPP Mean [95% CI]Or median (range)	OARPP Mean [95% CI]Or median (range)
Heart
MHD (Gy), right, WBI	5	0.4 [0.2;0.5]	0.3 [0.2;0.4]
MHD (Gy), right, loco-reg	2	1.5 [−112;14.2]	3.7 [−36.4;43.7]
MHD (Gy), left, WBI	4	4.3 [−1.3;9.9]	1.0 [0.7;1.3]
MHD (Gy), left, loco-reg	7	3.1 [1.9;4.2]	1.7 [0.9;2.4]
Ipsilateral Lung
V17 Gy (%), right, WBI	5	23 [19;26]	21 [14;27]
V17 Gy (%), right, loco-reg	2	31 [−39;100]	25 [24;27]
V17 Gy (%), left, WBI	4	21 [12;29]	16 [11;22]
V17 Gy (%), left, loco-reg	7	32 [29;346]	28 [23;34]
Contralateral breast
V5 Gy (%), right, WBI	5	1 (0;10)	0 (0;5)
V5 Gy (%), right, loco-reg	2	31 (1;66)	5 (0;10)
V5 Gy (%), left, WBI	4	0 (0;21)	0 (0;0)
V5 Gy (%), left, loco-reg	7	0 (0;7)	1 (0;6)

TPP: target prioritized plan; OARPP: organs at risk-prioritized plan; CI: confidence interval; MHD: mean heart dose; WBI: whole breast irradiation; loco-reg, loco-regional RT; V_x Gy: percentage of a relative volume V irradiated with x Gy. All treatments were based on 40 Gy in 15 fractions, 2.67 Gy pr. fraction.

Discussion

The aim of this study was to evaluate the suitability of 3D-CRT in patients with PE regarding dose to the heart. In both treatment plans constraints to ipsilateral lung were respected, but the median V_17 Gy was generally lower in the OARPP versus TPP (Table 2). The challenge for clinically acceptable treatment planning was seen in patients who received left-sided loco-regional RT. In the TPP acceptable doses to targets were reached at the cost of higher doses to the heart and ipsilateral lung. In the OARPP, the dose coverage of CTVn IMN was compromised in order to lower the heart dose. Likewise, the coverage of the PTVn IMN was reduced which made the OARPP less robust to daily fractionated RT.

The lower dose to the IMNs observed in the OARPP in our study may increase the risk of distant failure. Due to a higher MHD in left-sided patients, especially with IMN RT, a long-term side effect is the risk of developing ischemic heart disease. Thus, the challenge is the balance between optimal dose coverage of the IMNs and a potential long-term risk of radiation-induced heart disease [6]. It has been estimated in high risk patients that the balance between benefit and harm from IMN RT is in favor of including the IMNs in the RT fields [11]. However, for one left-sided patient in our study referred to loco-regional RT, we resigned from IMN RT in the OARPP as a clinical compromise to minimize dose to the heart. In another patient treated with right-sided loco-regional irradiation a high dose to contralateral breast was observed with 3D-CRT (V_5% of 66 Gy). Thus, the OARPP was based on IMRT although it caused a higher MHD (6.8 Gy). Patients treated with left-sided WBI showed quite high MHDs in the TPP. Three out of four patients had a severe PE but still, it shows that women treated with RT for a left-sided breast cancer may receive a higher MHD compared to right-sided RT.

To our knowledge this is the largest study to systematically evaluate the suitability of 3D-CRT for treatment planning in a group of breast cancer patients with PE. One study evaluated 10 patients with PE and treated with WBI with 3D-CRT [8]. The average MHD for PE patients was 3.0 Gy compared to 2.2 Gy for patients without PE. The Haller index was associated with MHD, and the mean Haller index and MHD found in that study were comparable to the results in our study. However, we did not find a correlation between the Haller index and MHD, presumably because we had only seven patients with left-sided loco-regional RT.

Other studies compared 3D-CRT to more modern RT techniques such as helical tomotherapy, IMRT and volumetric modulated arc therapy [7,12–14]. All these studies reported a higher MHD using 3D-CRT compared to the more advanced techniques, which furthermore improved dose coverage to PTVp breast/chest wall and reduced dose to ipsilateral lung but at the cost of more dose to the contralateral breast and lung. The dose bath to OARs delivered with modern techniques for RT may increase the risk of second cancer [15].

The strength of our study is the relatively large number of consecutive patients, however, since PE is rare only 18 patients have been reported.

In conclusion, this study showed that in some patients with PE it was possible to obtain an acceptable dose coverage to targets without exceeding the constraints to OARs by using 3D-CRT. However, patients with a severe chest wall deformity require a special attention, and in some cases 3D-CRT may not be the best choice of treatment.

Disclosure statement

No potential conflict of interest was reported by the authors.

Additional information

Funding

BVO was supported by the Danish Cancer Society.