Physics Contribution Sparing Healthy Tissue and Increasing Tumor Dose Using Bayesian Modeling of Geometric Uncertainties for Planning Target Volume Personalization Alan Herschtal, BE,* ,y Luc Te Marvelde, PhD,* Kerrie Mengersen, PhD, z Farshad Foroudi, FRANZCR, x,jj Thomas Eade, FRANZCR, {,# Daniel Pham, MMRT,** Hannah Caine, BMRS, { and Tomas Kron, PhD jj,yy *Department of Biostatistics and Clinical Trials, Peter MacCallum Cancer Centre, Melbourne, Australia; y Faculty of Health, Arts and Design, Swinburne University of Technology, Melbourne, Australia; z School of Mathematical Sciences, Science and Engineering Faculty, Queensland University of Technology, Brisbane, Australia; x Department of Radiation Oncology, Peter MacCallum Cancer Centre, Melbourne, Australia; jj The Sir Peter MacCallum Department of Oncology, Faculty of Medicine, Dentistry and Health Sciences, University of Melbourne, Melbourne, Australia; { Northern Sydney Cancer Centre, Radiation Oncology Department, Royal North Shore Hospital, St. Leonards, Sydney, Australia; # Northern Clinical School, University of Sydney, and Departments of **Radiation Therapy and yy Physical Sciences, Peter MacCallum Cancer Centre, Melbourne, Australia Received Aug 18, 2014, and in revised form Jan 7, 2015. Accepted for publication Jan 27, 2015. Summary Traditionally, clinical target volume (CTV) to planning target volume (PTV) margin widths are held constant for all patients for the duration of treatment. However, recently it has been found that random errors, which inform margin widths, vary markedly between patients. Using this finding, we Objective: To develop a mathematical tool that can update a patient’s planning target volume (PTV) partway through a course of radiation therapy to more precisely target the tumor for the remainder of treatment and reduce dose to surrounding healthy tis- sue. Methods and Materials: Daily on-board imaging was used to collect large datasets of displacements for patients undergoing external beam radiation therapy for solid tu- mors. Bayesian statistical modeling of these geometric uncertainties was used to opti- mally trade off between displacement data collected from previously treated patients and the progressively accumulating data from a patient currently partway through treatment, to optimally predict future displacements for that patient. These predictions were used to update the PTV position and margin width for the remainder of treat- ment, such that the clinical target volume (CTV) was more precisely targeted. Reprint requests to: Alan Herschtal, BE, 2/10 St Andrew’s Pl., E. Melbourne 3002, Australia. Tel: (þ613) 9656-3639; E-mail: Alan. Herschtal@petermac.org Supported by an Australian Government National Health and Medical Research Council (NHMRC) funding grant, no. 1023031. Conflict of interest: none. Supplementary material for this article can be found at www.redjournal.org. Int J Radiation Oncol Biol Phys, Vol. 92, No. 2, pp. 446e452, 2015 0360-3016/$ - see front matter Ó 2015 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.ijrobp.2015.01.034 Radiation Oncology International Journal of biology physics www.redjournal.org