This study was approved by the Ethics Board of the Affiliated Hospital of Qingdao University. Patients diagnosed with breast cancer between January 1, 2014, and December 31, 2019, who were treated with unilateral mastectomy and either local (chest wall) or locoregional (chest wall and regional nodal) radiation therapy were included. The inclusion criteria in this study were: (1) a diagnosis of breast cancer confirmed from pathological specimens; (2) patients could be treated with neoadjuvant chemotherapy (ypT0-4, N0-3) or adjuvant chemotherapy (pT1-4, pN0-3); (3) received modified radical mastectomy and routine level I/II axillary lymph node dissection; (4) received post-mastectomy IMRT. Meanwhile, the exclusion criteria were: (1) with M1 disease; (2) received breast reconstruction; (3) treated with 2D technology or 3D-CRT; (4) received radiation therapy after chest wall recurrence; (5) treatment termination due to personal or financial reasons. The decision of bolus application was selectively made for high-risk patients, based on the professional judgment of our oncologists. Patients with skin involvement applied bolus, which was a consensus among us oncologists. The other high-risk factors were collected through a survey conducted prior to our analysis. The factors that influenced oncologists to choose bolus were listed in Table 1 for further analysis using PSM.

The detailed patients selecting process is shown in Fig. 1.

Consort diagram showing the flowchart of this study. 3DCRT: three-dimensional conformal radiation therapy, IMRT: intensity modulated radiotherapy, PSM: propensity score matching

All patients recruited for this study underwent a modified radical mastectomy and received routine level I/II axillary lymph node dissection. According to standard treatment guidelines, patients with node-positive breast cancer are typically given an anthracycline-containing regimen followed by taxane administration as perioperative systemic therapy. Patients with positive hormone receptors were given adjuvant hormone therapy, with tamoxifen or aromatase inhibitors being administered based on menopausal status. Patients with HER2 overexpression were treated with trastuzumab.

The majority of our patients (98.68%) utilized the filed-in-filed (FIF) technique, while a small percentage (1.32%) applied volumetric modulated arc therapy (VMAT). All patients received a prescription dose of 5000 cGy, divided into 25 fractions, using 6MV X-ray radiation. Patients with node-positive disease underwent locoregional radiation therapy following mastectomy. Contouring of the clinical target volume (CTV) was based on the Radiation Therapy Oncology Group (RTOG) guidelines. The CTV was expanded by 2–5 mm non-uniformly to create the planning target volume (PTV). In cases where the primary tumor is located in the central or medial part of the breast, radiotherapy of internal mammary nodes was considered. Treatment planning for IMRT plans were performed on the Eclipse treatment planning system (TPS) version (Varian Medical Systems, Palo Alto, CA, USA). Each IMRT plan was optimized using the photon Optimization Method (PO), and the resulting dose distribution was evaluated using the Anisotropy Analysis Algorithm (AAA) with a grid square resolution of 2.5 mm. The plan involved employing 5 to 8 radiation fields, with the collimator bar rotation angle ranging from 0° to 30°. The optimization process took into account the uniformity and dose coverage of the PTV. All plans achieved clinically acceptable coverage of the modified PTV (mPTV), with mD95% of the target area received a prescription dose. Furthermore, all risk factors met the clinical goals. Planning approach was similar for both groups. These planning qualities have been estimated according to the ‘RATING’ score sheet [11].

During the radiotherapy process, patients were positioned supine on a breast board with both arms abducted above their head. To immobilize their chests, a thermoplastic shell (YC-M10-501AHW, Shenzhen Teng Unisoft Technology CO., LTD) extending from the supraclavicular fossa to the bottom of the ribcage was used [12] (Fig. 2A). For patients with high-risk factors selected by oncologists, a 5 mm bolus was placed daily between the chest wall and thermoplastic sheet (Fig. 2B). No patients received a chest wall boost.

CT scan images with and without bolus application. A Thermoplastic shell was used to immobilize the patient's position during the scan; B Bolus was applied between the thermoplastic shell and chest wall. : thermoplastic shell; : bolus between the thermoplastic shell and chest wall

The primary endpoints were local recurrence (LR) and locoregional recurrence (LRR), while the secondary outcomes were disease free survival (DFS), overall survival (OS), and radiation related adverse events, including acute skin toxicity and late radiation pneumonia. LR was defined as recurrence within the ipsilateral skin and/or subcutaneous tissue, while LRR was defined as recurrence within the ipsilateral skin and/or subcutaneous tissue and the regional nodes. The term 'relapse' was used to describe the initial occurrence of either locoregional or distant tumor metastasis. DFS was calculated as the time from the date of diagnosis to the occurrence of tumor relapse. OS was defined as survival time from time of diagnosis. All of the patients were followed until January 31, 2023.The follow-up methods included outpatient reexamination, telephone follow-up, and medical record review.

The occurring time of skin side effect including dry or moist desquamation and the degree of erythema were assessed. Based on the RTOG classification [13], the primary distinction between grades 2 and 3 was the presence of moist desquamation and tenderness, whereas grade 4 was characterized by necrosis, ulceration, or bleeding. During the treatment of our patients, we regularly inspected their chest wall and recorded observations of their skin condition. For statistical analysis, we collected the most severe skin toxicity recorded during radiotherapy. During the follow-up phone calls, patients were asked to describe their skin condition, with a specific focus on keywords such as dry peeling, edema, wet peeling, or ulcer. These descriptions, along with the medical records, were evaluated by two doctors simultaneously to ensure accurate assessment. Treatment interruption and early termination due to side effects and other reasons were collected through the medical records and the treatment scheduling system, confirmed by follow-up phone calls.

The severity of radiation pneumonia was assessed based on the Radiation Therapy Oncology Group (RTOG) criteria. Heart and lung dosimetry were calculated using the treatment planning system (TPS). The chest CT films or X-rays of the patients were compared before and after radiotherapy to determine the presence of radiation-induced lung injury.

To minimize potential selection bias and confounders, we used propensity score matching (PSM) to control for differences in baseline characteristics. Patients in the entire cohort were matched at a 1:1 ratio using a caliper of width equal to 0.2 without replacement, simulating random allocation using SPSS. Clinicopathological characteristics and treatments data between two groups (bolus versus no-bolus) were compared using Pearson’s chi-square tests. Cumulative incidence was used to evaluate LR and LRR, and Kaplan–Meier survival was applied to analyze breast cancer DFS and OS. Additionally, Pearson’s chi-square was used to analyze skin toxicity and late radiation pneumonia, while the independent t-test was used to analyze the differences in OAR dosimetry. All tests were considered two-sided, and a statistically significant P-value was determined to be less than 0.05. The statistical analyses were performed using SPSS (version 27).