Patient selection

Locally advanced post-mastectomy breast cancer patients who received adjuvant RT between 1 January 2018 and 31 December 2020 were reviewed for this study. Twenty-five patients were randomly selected as fulfilling the following inclusion criteria: (1) patients with left-sided CW, (2) patients with SCL and IMN nodal stations and (3) patients simulated in a deep inspiration breath-hold (DIBH) setting. Exclusion criteria were (1) patients with bilateral mastectomy and (2) patients with axillary nodes. The median age of these patients was 49 years. The institutional scientific and ethics board has approved this study and informed consent was obtained from each patient.

Simulation and contouring

For simulation, patients were placed in a supine position using a customised vacuum bag. A planning computed tomography (CT) scans had been generated with 2.5-mm slice thickness by Discovery IQ CT scanner (GE Healthcare, Chicago, IL, USA). Clinical target volumes (CTVs) of CW, SCL and IMN nodal regions were delineated. The CW PTV (PTV_CW), SCL PTV (PTV_SCL) and IMN PTV (PTV_IMN) were created by giving margin to the CTVs around 5 mm. The PTV_SCL was brought in 5 mm from the skin surface, while the PTV_CW included the skin region. A wax bolus material of 5 mm thickness was placed on the body surface beside the PTV_CW to achieve adequate dose coverage in the near skin region. The OARs delineated were the heart, left anterior descending coronary artery (LAD), right coronary artery (RCA), IL, contralateral lung, CB, spinal cord, oesophagus, trachea, coeliac plexus and gastroesophageal junction (GEJCP), and NT, defined as the body volume minus PTVs were delineated.

Treatment planning

For each patient, 3DCRT+IMRT, 3DCRT+VMAT and IMRT+VMAT hybrid treatment plans were created in the Eclipse treatment planning system using 6, 10, 15 MV photon beams of Truebeam linear accelerator (Varian Medical Systems, Palo Alto, CA, USA). A hypofractionated dose prescription of 40·5 Gy (2·7 Gy per fraction) in 15 fractions was used for all PTVs. The dose weightings utilised were 70% of the prescription dose for the base-dose component and 30% of the prescription dose for the hybrid component. ^{Reference Balaji, Yadav and BalajiSubramanian12} Table 1 presents the clinical dose constraints of OARs to be considered in planning. Both the IMRT and VMAT plan optimisations were performed by the photon optimiser (PO) algorithm. Volume dose was computed by analytical anisotropic algorithm (AAA) using a dose grid of 2·5 mm. All individual plans were normalised in such a way that the plans deliver the PTVs mean dose equal to the prescribed dose.

Table 1. Clinical dose constraints for organs at risk

LAD, left anterior descending coronary artery; RCA, right coronary artery; IL, ipsilateral lung; CL, contralateral lung; CB, contralateral breast; GEJCP, coeliac plexus and gastroesophageal junction; SC, spinal cord; NT , normal tissue; Gy, Gray; V_XGy, volume receiving X Gy dose; D_Mean, mean dose; D_Max, maximum dose.

For the 3DCRT, IMRT and VMAT plans, an isocentre was positioned at the PTV_CW and PTV_SCL junction and axially on ribs to minimise beam divergence into the IL and heart (Figure 1). Two tangential half-beam blocked fields for the PTV_CW, PTV_IMN and one anterior oblique, one posterior oblique, half-beam blocked fields for the PTV_SCL were used for the 3DCRT plan as shown in Figure 1. Varied X-ray photon energies were utilised for 3DCRT fields depending on the skin to PTV surface distance along the beam direction, according to Balaji et al. ^{Reference Balaji, Yadav and BalajiSubramanian12} The 3DCRT plans were dose-computed for 1·9 Gy (approximately 70% of prescription dose). Similarly, the IMRT base-dose plans were created with the same beam arrangements as the 3DCRT plans. In the IMRT, each half-beam blocked field was allowed to over travel 1 cm beyond the isocentre for better modulation in field junction. Fixed jaw setting was enabled during optimisation.

For the hybrid component, IMRT and VMAT plans were generated with the remaining prescription dose of 0·8 Gy (approximately 30% of the PTVs prescription dose). The IMRT plans consisted of five fields with equally separated gantry angles 300°, 355°, 50°, 105° and 160° as shown in Fig. 2. The collimator angles were set at ± 5°. The VMAT plans utilised two coplanar partial arcs and two tangential arcs as shown in Figure 2. Arc1 rotated clockwise from 300° to 50° with 10° collimator angle, and arc2 rotated clockwise from 50° to 160° with 350° collimator angle. The tangential arc3 rotated counterclockwise from 160° to 110° with 345° collimator angle and arc4 rotated counterclockwise from 350° to 300° with 15° collimator angle. Both the IMRT and VMAT plans employed 6 MV photon beams. While doing optimisation, the corresponding base-dose 3DCRT and IMRT plans were enabled. After dose computation and normalisation, plan sum of the 3DCRT+IMRT, 3DCRT+VMAT and IMRT+VMAT hybrid plans were created. The entire planning and dosimetric evaluation process were executed by an experienced medical physicist to avoid any planner variability.

Figure 1. (a) Location of isocentre and (b) field arrangements in base-dose 3DCRT and IMRT plans (four fields).

Figure 2. Field arrangements in (a) IMRT plan (five fields) and (b) VMAT plan (two partial and two tangential arcs).

Dosimetric evaluation

For the PTVs quality comparison, dosimetric indices like coverage index (COI) and uniformity index (UI) were calculated for PTV_CW, PTV_SCL and PTV_IMN. Conformity index (CI) and gradient index (GI) were appraised for the combined volume of all PTVs. The COI was defined as:

$$\;{\rm{COI}} = \;{{{{\rm{D}}_P}} \over {{{\rm{D}}_{95\% }}}},$$

where D_P is the prescription dose and D_95% is the dose received by 95% of the PTV. A COI value nearer to 1 indicates superior dose coverage to PTV. The UI was calculated as:

$${\rm{UI}} = \;{{{{\rm{D}}_{2\% }}} \over {{{\rm{D}}_{98\% }}}},$$

where D_2% and D_98% are the doses received by 2% and 98% of the PTV, respectively. A UI value close to 1 specifies a more homogeneous dose distribution to PTV. The CI was defined as:

$${\rm{CI}} = \;{{{{\rm{V}}_{PTV}}} \over {{{\rm{V}}_{PTVref}}}}\; \times \;{{{{\rm{V}}_{ref}}} \over {{{\rm{V}}_{PTVref}}}},$$

where V_PTV is the volume of PTV, V_PTVref is the reference isodose (95%) volume within the PTV and V_ref is the volume of reference isodose (95%). A CI value close to 1 designates a superior conformal dose plan. ^{Reference Nakamura, Verhey and Smith17} The GI was calculated as:

$${\rm{GI}} = {\rm{ }}{{{{\rm{V}}_{50\% }}} \over {{{\rm{V}}_{PTV}}}},$$

where V_50% is 50% isodose volume and V_PTV is the volume of PTV. A GI value close to 1 specifies a better dose fall-off plan. ^{Reference Ohtakara, Hayashi and Hoshi18} The dose-volume parameters evaluated for the OARs comparison were listed in Table 1. ^{Reference Blom, Wennberg, Svane, Bylund and Lind19–Reference Lazzari, Terlizzi, Leo and Silvano23} In addition, total MU and treatment time (TT) were noted to assess the delivery efficiency. A simple scoring method was utilised to calculate an overall score that incorporates all dosimetric parameters evaluated. The overall score was calculated as:

$${\rm{Overall}}\;{\rm{Score}} = {\rm{ }}{{\sum _{i{\mkern 1mu} = {\mkern 1mu} 1}^n({{{{\rm{A}}_i}} \over {{{\rm{M}}_i}}})} \over {\rm{n}}},$$

where A _i is the achieved value of an i^th dosimetric parameter of a particular plan and M _i is the mean value of the i^th dosimetric parameter of all three plans and n is the number of dosimetric parameters evaluated. The overall score close to 0 designates a superior plan.

Statistical analysis

For statistical analyses, a non-parametric Wilcoxon signed-rank test was executed for a paired group of plan comparisons. All statistical tests were two-tailed, and a threshold value of p < 0·05 indicated statistically significant results.