Premium
Sci—;Wed PM: Delivery—04: Real‐Time MR Tumour Tracking and Dose Adaptation
Author(s) -
Yun J,
Robinson D,
MacKenzie M,
Rathee S,
Murray B,
Fallone BG
Publication year - 2009
Publication title -
medical physics
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.473
H-Index - 180
eISSN - 2473-4209
pISSN - 0094-2405
DOI - 10.1118/1.3244096
Subject(s) - contouring , imaging phantom , collimator , computer science , radiation therapy , tracking (education) , medical imaging , nuclear medicine , computer vision , medical physics , artificial intelligence , medicine , physics , radiology , computer graphics (images) , optics , psychology , pedagogy
Treating mobile tumours (e.g. lung tumours) with radiation is a difficult task. In current practice, radiotherapy plans are generated that irradiate extended regions surrounding the tumour in order to encompass the entire path of tumour motion. Doing so, however, can result in unacceptable medical complications or restricted radiation deliveries. Addressing these problems, researchers at our centre initiated the Advanced Real‐Time Adaptive Radiation Therapy ( ( ART ) 2 ) project with a novel design of a Linac‐MR device which can provide intra‐fractional real time MR imaging. Using this system, our objective is to develop and evaluate an algorithm that can track the tumour movement in real‐time and adapt the radiation delivery accordingly, depending on the shape, position, and surrounding anatomy of the tumour. An auto‐contouring algorithm was developed to identify the shape and location of tumours. A MR compatible motion phantom was built to simulate the motion of tumours and to perform dosimetric measurements. To conform the radiation beam in real‐time, a Multi‐Leaf Collimator (MLC) controller was developed. To compensate for the time delay between MR imaging and irradiating instances, a motion prediction algorithm was developed and optimized. The auto‐contouring algorithm showed successful results in the 4D MR images of thoracic region and phantom. The motion prediction algorithm was evaluated by various means, and in most cases the position of tumour was correctly predicted more than 95% of the total treatment time. Integrated with these algorithms, our MLC system fully demonstrated its real‐time tracking capability.