AUTOMATIC BRAIN METASTASES PLANNING Clinical White Paper Radiosurgery has become the first-line treatment option for patients with multiple brain metastases, prolonging overall survival for younger patients when salvage treatment options are readily available1. Treatment decisions are no longer based on the number of metastases but rather on the cumulative tumor volume2, with hypo-fractionated radiosurgery regimens a viable alternative for patients with large disease burden3-6. Traditional radiosurgery approaches target each metastasis individually; prioritizing superior dose gradients over treatment efficiency as the total treatment time scales linearly with the number of metastases. Brainlab Automatic Brain Metastases Planning advances the radiosurgery legacy by targeting up to ten metastases at once; optimizing treatments both in terms of plan quality and delivery efficiency. Figure 1: Automatic Brain Metastases Planning details for Patient 4 with 9 lesions. The Assuta Medical Center located just north of Tel Aviv, along the popular Mediterranean coastline of Israel, has been a radiosurgery center of excellence for many years. Equipped with Novalis Radiosurgery and a high definition large field multileaf collimator, this center has been pioneering the clinical implementation of Brainlab Automatic Brain Metastases Treatment Planning. Starting treatments in summer 2014, they retrospectively selected ten early cases treated at their institution for review of the entire treatment planning and delivery process. The detailed characteristics of the selected patient population can be found in Table 1. Ten patients over fifty years of age were selected with two to nineteen brain metastases; seven related to a primary lung cancer, two associated with breast cancer and one resulting from melanoma. The total accumulated treated volumes range from 0.15 cc to 19.78 cc. All lesions were identified on 1 mm T1 contrast-enhanced magnetic resonance imaging slices rigidly fused to 0.6 mm computed tomography slices. Treatment planning starts with importing all DICOM data into the Elements treatment planning platform, fusing the various data sets and contouring the gross tumor volumes of each target on the magnetic resonance data sets utilizing the Smartbrush Element. All critical organs are automatically segmented based on a universal atlas approach and can be consulted in a dedicated DICOM Viewer. Automatic Brain Metastases Planning takes the target volumes and models the patient outer contour and underlying table top for correct dose calculations. A list of numerous treatment templates matching diverse hospital practices can be defined upfront with the most appropriate template chosen for each patient. Treatment templates contain various information: possible enlargement of the tumor volumes to encompass biological and treatment uncertainties, different prescription doses in function of individual tumor volumes or sizes, table and gantry angle preferences as a guide for the subsequent automatic optimization of the arcs.

Table 1: Patient Characteristics Automatic Brain Metastases Planning determines the isocenter position as the average position of the centers of mass of each individual target. The software uses the predefined table and gantry angles to create a set of two independent arcs per table angle. Depending on the hemisphere where the isocenter is located, the arcs are mirrored around the sagittal plane. The start and stop angles of the arc are first set to the template values and automatically modified during optimization. The leaf positions are shaped to conform the targets, with an additional margin...

Table 2: Treatment Characteristics Inverse Paddick Conformity Index High quality imaging for localization and simulation is a necessity for reliable image fusion. For the localization images, ExacTrac accredits superior contrast resolution in optimal exposure conditions to the unique configuration of the X-Ray system. The fixed configuration of X-Ray tubes and flat-panel detectors eliminates any potential spatial uncertainty caused by mechanical movement. Furthermore, the large source-to-detector distance reduces the solid angle of the radiation beam and hence reduces potential geometric distortion....