1. Patients and target delineation
This study included 37 left breast cancer patients who underwent modified radical mastectomy, with or without axillary clearance, and were clinically staged as II or III. The median age of the patients was 53 years (range, 30–66 years), and their median body mass index (BMI) was 22.3 (range, 16.2–29.5). All patients exhibited a Haller index of less than 3.5 (range, 1.9–3.4), indicating the absence of chest-wall deformities. Computed tomography (CT) scans were performed using a Brilliance Big Bore CT (Philips, NL) with 3 mm thick slices. The patients were positioned supine, with scans extending from the mandible to the lower margin of the liver. This retrospective study received approval from the institutional review board, and the requirement for informed consent was waived.
The clinical target volumes (CTVs) were delineated by two senior clinical oncologists, adhering to RTOG guidelines. The CTVs included the left chest-wall (CW), axillary lymph-node region (AXN), internal mammary lymph-node region (IMN), supraclavicular fossa (SCF), and infraclavicular fossa (ICF). The planning target volume (PTV) was established by extending the CTVs by a 5-mm margin, confined within the patient’s body contour. The organs at risk (OARs) outlined were the ipsilateral and contralateral lungs, heart, left anterior descending coronary artery (LAD), contralateral breast, esophagus, spinal cord, left humeral head, and thyroid. Anatomical cardiac markers, such as the anterior interventricular and left atrioventricular grooves, were used for accurate determination of LAD volumes.
2. Treatment planning
The treatment planning for all patients utilized the Eclipse system (v13.5, Varian, US), overseen by the same experienced physicist. Dose calculation was conducted using the Anisotropic Analytical Algorithm (AAA) with a calculation grid of 2.5 mm. The LINAC model, Trilogy (Varian, US), was equipped with the Millennium 120 MLC and used a photon energy of 6 MV. To ensure adequate skin dose on the chest wall, a 5-mm virtual bolus was applied. Figure 1 depicts the beam orientations for the CW in the three RT techniques.
3D-CRT: The 3D-CRT plan incorporated 4–5 fields in field beams, with up to 6 subfields per beam. Two tangential beams for the chest wall and two oblique beams for the SC region were standard. An additional 0° beam, wide enough to cover the IMN volume, was used if IMN coverage was insufficient.
IMRT: The IMRT plan involved a greater number of beams. Chest wall target volume segmentation facilitated the use of three pairs of semi-blocked short tangential beams, aimed to reduce heart and lung exposure. Beams at 0°, 40°, and 320° angles targeted the SC region. An anterior beam was included to enhance IMN dose coverage. To prevent dose inconsistencies near segmented areas, beams were overlapped or connected in pairs. The IMRT technique employed sliding window delivery with a maximum dose rate of 600 MU/min.
VMAT: The VMAT plan used two pairs of semi-blocked short tangential arcs for PTV and a single short arc for SC irradiation (30° to 330°). Collimator angles and jaw positions were tailored to each beam based on the patient’s anatomy in the beam’s eye view (BEV), ensuring PTV coverage while minimizing exposure to sensitive organs. Each arc contained 98 control points, with a dose rate capped at 600 MU/min.
The prescribed dose was 50 Gy delivered in 25 fractions. For comparative purposes, all plans were normalized to ensure 50 Gy encompassed 95% of the PTV volume. Dose objectives and OAR constraints specific to our institution are presented in Table 1.
Table 1
Dose objectives and constraints for treatment planning
Structure | Parameter | Objective |
PTV | V50Gy(%) | = 95% |
| D2%(Gy) | < 55Gy |
| D98%(Gy) | > 45Gy |
PTVIMN | V50Gy(%) | ≥ 80% |
I_lung | V5Gy(%) | < 60% |
| V20Gy(%) | < 30% |
| V30Gy(%) | < 20% |
| Dmean(Gy) | < 15Gy |
C_lung | V5Gy(%) | < 25% |
| Dmean(Gy) | < 5Gy |
Heart | V25Gy(%) | < 20% |
| Dmean(Gy) | < 8Gy |
C_breast | Dmax(Gy) | < 40Gy |
| Dmean(Gy) | < 5Gy |
Esophagus | Dmean(Gy) | < 30Gy |
Thyroid | Dmean(Gy) | < 30Gy |
Left humeral head | Dmean(Gy) | < 20Gy |
Spinal cord | Dmax(Gy) | < 40Gy |
3. Plan evaluation
3.1. Dosimetric parameters
The dose-volume histogram (DVH) data from all plans were analyzed for target coverage and OARs sparing effectiveness. For the PTV, metrics such as D2%, D98%, conformal index (CI), and homogeneity index (HI) were assessed. CI calculations followed Paddick’s formula23: \(\text{C}\text{I}={\text{T}\text{V}\text{P}\text{V}}^{2}/\left(\text{T}\text{V}\times \text{P}\text{V}\right)\), where \(\text{T}\text{V}\text{P}\text{V}\) represents the PTV volume receiving the prescribed dose, \(\text{T}\text{V}\) is the PTV volume, and \(\text{P}\text{V}\) is the volume covered by the prescribed isodose. HI was defined as \(\text{H}\text{I}=\left({\text{D}}_{5\text{%}}-{\text{D}}_{95\text{%}}\right)/{\text{D}}_{\text{m}\text{e}\text{a}\text{n}}\), with \({\text{D}}_{5\text{%}}\) and \({\text{D}}_{95\text{%}}\) indicating the doses delivered to 5% and 95% of the target volume, respectively. Additionally, the V50Gy for PTVIMN was assessed separately. For evaluating OAR exposure, parameters such as Dmean, Dmax, and Vx (volume of the OAR receiving at least x Gy) were analyzed, based on the specific organ.
3.2. NTCP calculation
The NTCP model, originating in the 1980s, describes the correlation between radiation dose/volume and clinical toxicity.19,20 In our study, the Lyman-Kutcher-Burman model was employed for calculating NTCP for ipsilateral pneumonia (grade ≥ 2) and radiation esophagitis (grade ≥ 2),24,25 while the Poisson-LQ model was used for NTCP estimation of radiation-induced cardiac mortality.26 The Eclipse TPS incorporated the biological evaluation module used for these NTCP calculations.
3.3. EAR calculation
The specific OAR’s SCR caused by RT can be assessed using Schneider’s full dose-response EAR model.[22, 27] The EAR was calculated using the following formula:
$${EAR}^{org}=\frac{1}{{V}_{t}}\sum _{i}V\left({D}_{i}\right){\beta }_{EAR}RED\left({D}_{i}\right)\mu \left(agex,agea\right)$$
1
,
where \({\beta }_{EAR}\) represents the initial slope, \(V\left({D}_{i}\right)\) is the organ volume receiving dose \({D}_{i}\), and \({V}_{t}\) is the total organ volume. The function \(\mu \left(agex,agea\right)\) modifies based on population variables and is defined by Eq. (2):
$$\mu \left(agex,agea\right)={e}^{\left[\gamma e\left(agex-30\right)+{\gamma a}^{ln\left(agea/70\right)}\right]}$$
2
,
where \(agex\) is the patient’s irradiated age and \(agea\) is the attained age. The constants \(\gamma e\) and \(\gamma a\) are age modifying parameters. The term \(RED\) in Eq. (1) refers to the risk equivalent dose model, incorporating cell killing and fractionation effects, expressed as Eq. (3):
$$RED\left({D}_{i}\right)=\frac{{e}^{-{\alpha }^{\text{’}}{D}_{i}}}{{\alpha }^{\text{’}}R}\left(1-2R+{R}^{2}{e}^{{\alpha }^{\text{’}}{D}_{i}}-{\left(1-R\right)}^{2}{e}^{-\frac{{\alpha }^{\text{’}}R}{1-R}{D}_{i}}\right)$$
3
,
where \(\text{R}\) denotes the repopulation/repair parameter between dose fractions, and \({{\alpha }}^{\text{’}}\) is calculated using Eq. (4):
$${\alpha }^{\text{’}}=\alpha +\beta d=\alpha +\beta {D}_{i}/{D}_{T}{d}_{T}$$
4
,
where \({D}_{T}\) and \({d}_{T}\) represent the prescribed and fractionation doses to the target volume, respectively. This study investigated the EAR per 10,000 persons-years at 70 years following exposure at 30 years for ipsilateral and contralateral lungs, contralateral breast, esophagus, and thyroid. The biological parameters for these calculations were sourced from several studies,17,22,28–30 as detailed in Table 2.
Table 2
Biological parameters for EAR calculation
OAR | \({\beta }_{EAR}\)(10000 PY/Gy) | \(\gamma e\) | \(\gamma a\) | \(\alpha\) | \(\alpha \beta\)(Gy) | \(R\) |
I_lung | 7.5 | 0.002 | 4.23 | 0.042 | 3 | 0.83 |
C_lung | 7.5 | 0.002 | 4.23 | 0.042 | 3 | 0.83 |
C_breast | 9.2 | -0.037 | 1.7 | 0.044 | 3 | 0.15 |
esophagus | 0.58 | -0.002 | 1.9 | 0.026 | 3 | 0.81 |
thyroid | 1.2 | -0.046 | 0.6 | 0.033 | 3 | 0.5 |
4. Statistical analysis
Statistical analyses were conducted using SPSS (v25, IBM, US) to evaluate differences in dosimetric and radiobiological data, with results presented as mean ± SD. The Shapiro–Wilk test was used to assess data normality within each group, considering the small patient numbers (non-normality was inferred if p < 0.05). The Wilcoxon signed ranks test was employed for non-normally distributed data, while the paired samples t-test was utilized for normally distributed data, comparing three techniques. The level of statistical significance was set at p < 0.05.