Comprehensive nodal breast VMAT: solving the low‐dose wash dilemma using an iterative knowledge‐based radiotherapy planning solution

Abstract

Abstract Introduction Aimed to develop a simple and robust volumetric modulated arc radiotherapy (VMAT) solution for comprehensive lymph node (CLN) breast cancer without increase in low‐dose wash. Methods Forty CLN‐breast patient data sets were utilised to develop a knowledge‐based planning (KBP) VMAT model, which limits low‐dose wash using iterative learning and base‐tangential methods as benchmark. Another twenty data sets were employed to validate the model comparing KBP‐generated ipsilateral VMAT (ipsi‐VMAT) plans against the benchmarked hybrid (h)‐VMAT (departmental standard) and bowtie‐VMAT (published best practice) methods. Planning target volume (PTV), conformity/homogeneity index (CI/HI), organ‐at‐risk (OAR), remaining‐volume‐at‐risk (RVR) and blinded radiation oncologist (RO) plan preference were evaluated. Results Ipsi‐ and bowtie‐VMAT plans were dosimetrically equivalent, achieving greater nodal target coverage (P < 0.05) compared to h‐VMAT with minor reduction in breast coverage. CI was enhanced for a small reduction in breast HI with improved dose sparing to ipsilateral‐lung and humeral head (P < 0.05) at immaterial expense to spinal cord. Significantly, low‐dose wash to OARs and RVR were comparable between all plan types demonstrating a simple VMAT class solution robust to patient‐specific anatomic variation can be applied to CLN breast without need for complex beam modification (hybrid plans, avoidance sectors or other). This result was supported by blinded RO review. Conclusions A simple and robust ipsilateral VMAT class solution for CLN breast generated using iterative KBP modelling can achieve clinically acceptable target coverage and OAR sparing without unwanted increase in low‐dose wash associated with increased second malignancy risk.

Keywords

• Breast
• cancer
• knowledge‐based planning
• lymph nodes
• radiotherapy
• simultaneous integrated boost (SIB)