Deep learning based MLC aperture and monitor unit prediction as a warm start for breast VMAT optimisation

Objective. Automated treatment planning today is focussed on non-exact, two-step procedures. Firstly, dose-volume histograms (DVHs) or 3D dose distributions are predicted from the patient anatomy. Secondly, these are converted in multi-leaf collimator (MLC) apertures and monitor units (MUs) using a generic optimisation to obtain the final treatment plan. In contrast, we present a method to predict volumetric modulated arc therapy (VMAT) MLC apertures and MUs directly from patient anatomy using deep learning. The predicted plan is then provided as initialisation to the optimiser for fine-tuning.Approach. 148 patients (training: 101; validation: 23; test: 24), treated for right breast cancer, are replanned to obtain a homogeneous database of 3-arc VMAT plans (PTVBreast: 45.57 Gy; PTVBoost: 55.86 Gy) according to the clinical protocol, using RapidPlanTMwith automatic optimisation and extended convergence mode (clinical workflow). Projections of the CT and contours are created along the beam's eye view of all control points and given as input to a U-net type convolutional neural networks (CNN). The output are the MLC aperture and MU for all control points, from which a DICOM RTplan is built. This is imported and further optimised in the treatment planning system using automatic optimisation without convergence mode, with clinical PTV objectives and organs-at-risk (OAR) objectives based on the DVHs calculated from the imported plan (CNN workflow).Main results. Mean dose differences between the clinical and CNN workflow over the test set are 0.2 ± 0.5 Gy atD95%and 0.6 ± 0.4 Gy atD0.035ccof PTVBreastand -0.4 ± 0.3 Gy atD95%and 0.7 ± 0.3 Gy atD0.035ccof PTVBoost. For the OAR, they are -0.2 ± 0.2 Gy forDmean,heartand 0.04 ± 0.8 Gy forDmean,ipsilateral lung. The mean computation time is 60 and 25 min respectively.Significance. VMAT optimisation can be initialised by MLC apertures and MUs, directly predicted from patient anatomy using a CNN, reducing planning time with more than half while maintaining clinically acceptable plans. This procedure puts the planner in a supervising role over an AI-based treatment planning workflow.

Gating and tracking, 4D in thoracic tumours

The limited ability to control for a tumour's location compromises the accuracy with which radiation can be delivered to tumour-bearing tissue. The resultant requirement for larger treatment volumes to accommodate target uncertainty restricts the radiation dose because more surrounding normal tissue is exposed. With image-guided radiation therapy (IGRT), these volumes can be optimized and tumouricidal doses may be delivered, achieving maximum tumour control with minimal complications. Moreover, with the ability of high precision dose delivery and real-time knowledge of the target volume location, IGRT has initiated the exploration of new indications in radiotherapy such as hypofractionated radiotherapy (or stereotactic body radiotherapy), deliberate inhomogeneous dose distributions coping with tumour heterogeneity (dose painting by numbers and biologically conformal radiation therapy), and adaptive radiotherapy. In short: "individualized radiotherapy". Tumour motion management, especially for thoracic tumours, is a particular problem in this context both for the delineation of tumours and organs at risk as well as during the actual treatment delivery. The latter will be covered in this paper with some examples based on the experience of the UZ Brussel. With the introduction of the NOVALIS system (BrainLAB, Feldkirchen, Germany) in 2000 and consecutive prototypes of the ExacTrac IGRT system, gradually a hypofractionation treatment protocol was introduced for the treatment of lung tumours and liver metastases evolving from motion-encompassing techniques towards respiratory-gated radiation therapy with audio-visual feedback and most recently dynamic tracking using the VERO system (BrainLAB, Feldkirchen, Germany). This evolution will be used to illustrate the recent developments in this particular field of research.

An overview of volumetric imaging technologies and their quality assurance for IGRT

Image-guided radiation therapy (IGRT) aims at frequent imaging in the treatment room during a course of radiotherapy, with decisions made on the basis of this information. The concept is not new, but recent developments and clinical implementations of IGRT drastically improved the quality of radiotherapy and broadened its possibilities as well as its indications. In general IGRT solutions can be classified in planar imaging, volumetric imaging using ionising radiation (kV- and MV- based CT) or non-radiographic techniques. This review will focus on volumetric imaging techniques applying ionising radiation with some comments on Quality Assurance (QA) specific for clinical implementation. By far the most important advantage of volumetric IGRT solutions is the ability to visualize soft tissue prior to treatment and defining the spatial relationship between target and organs at risk. A major challenge is imaging during treatment delivery. As some of these IGRT systems consist of peripheral equipment and others present fully integrated solutions, the QA requirements will differ considerably. It should be noted for instance that some systems correct for mechanical instabilities in the image reconstruction process whereas others aim at optimal mechanical stability, and the coincidence of imaging and treatment isocentre needs special attention. Some of the solutions that will be covered in detail are: (a) A dedicated CT-scanner inside the treatment room. (b) Peripheral systems mounted to the gantry of the treatment machine to acquire cone beam volumetric CT data (CBCT). Both kV-based solutions and MV-based solutions using EPIDs will be covered. (c) Integrated systems designed for both IGRT and treatment delivery. This overview will explain some of the technical features and clinical implementations of these technologies as well as providing an insight in the limitations and QA procedures required for each specific solution.