Page 55 - FiM_03_2016

ABSTRACT
ON THE MAGNITUDE AND DETECTABILITY OF VMAT DELIVERY ERRORS
Alessandro Scaggion
Università degli Studi di Torino
Scuola di Specializzazione in Fisica Medica (A.A. 2014-2015) Direttore: Roberto Cirio
Relatore: Vincenzo Monaco
Istituto Oncologico Veneto IOV – IRCCS, Padova Correlatore:Marta Paiusco
INTRODUCTION
In this work two different aspects of the mechanical errors affecting the delivery of Volumetric Modulated Arc Therapy (VMAT) radiotherapy treatments are investigated. Firstly a comprehensive study of the different mechanical errors is presented. The log files recorded from more than 2000 clinical treatments, collected during two years of clinical activity, are statistically analyzed to assess the delivery accuracy and the long term stability of VMAT treatments. Part of the collected data are then used to simulate the pretreatment QA procedure of VMAT plans: different VMAT treatments were virtually delivered on analytical phantoms with different geometries. This procedure allowed to verify how a common QA metric and its predictive power are influenced by the geometry of the dosimetry system.
MATERIALS ANS METHODS
From May 2013 to March 2015, 179 RapidArc clinical treatments were investigated. All treatments were delivered with a UNIQUE machine equipped with a Millennium 120 MLC. Every 50 ms the UNIQUE control system verifies each RapidArc element and compares expected and actual values. Two log files are generated for every delivery. The first is a Linac file recording, at each control point, the delivered cumulative MUs versus gantry angle; the second registers the actual versus the planned MLC positions. These files were collected during patient treatment sessions nearly twice a week for a total of total of 2230 clinical delivery sessions. The performance of UNIQUE and its long-term stability have been investigated using information derived from the log files: MLC motion errors, gantry rotation errors and MU delivery errors.
Tesi di Specialità
47