Premium
WE‐C‐204B‐07: Real‐Time MRI for Soft‐Tissue‐Based IGRT of Moving and Deforming Lung Tumors
Author(s) -
Sawant A,
Pauly K Butts,
Alley M,
Vasanawala S,
Loo B,
Joshi S,
Hinkle J,
Keall P
Publication year - 2010
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 - 10.1118/1.3469379
Subject(s) - steady state free precession imaging , centroid , nuclear medicine , voxel , magnetic resonance imaging , image guided radiation therapy , medicine , radiation therapy , radiology , artificial intelligence , computer science
Purpose : Integrated MRI+Linacs can potentially provide real‐time soft‐tissue‐based image‐guidance for lung cancer IGRT. Towards this, we investigate guidance strategies using prospective rapid lung MRI coupled with deformable image registration. Method and Materials : All experiments were performed on a 1.5 T MRI scanner, using a 4‐channel cardiac coil, under free‐breathing conditions, without extrinisic contrast. A balanced steady‐state free precession (b‐SSFP) imaging sequence was optimized for prospective imaging and reconstruction. Two lung cancer patients (Pt#l: 4 cm tumor, right lower lobe, Pt#2: 6 cm tumor, left upper lobe) were imaged. A viscous fluid‐flow‐based deformable registration was applied to each MRI time series in order to determine motion trajectories of voxels within the field of view. These trajectories were used to characterize: (i) motion of the tumor centroid. (ii) relative trajectories of the tumor centroid and the diaphragmrelative trajectories of different points on the tumor — characterizing tumor rotation/deformation. Results : The modified b‐SSFP sequence yielded acquisition times of ∼0.16s and ∼1.5s for 2D and 3D acquisition, respectively. Tumor trajectory analysis: (i) significant cycle‐to‐cycle variation in tumor motion was observed in both patients (ii) For Pt#l, the tumor centroid showed good correlation with diaphragmatic motion. For Pt#2, this correlation was relatively poor (iii) Pt#l did not exhibit significant tumor rotation/deformation. In Pt.#2, the trajectories of two points on the tumor showed maximum deviations of ∼8 mm (superior‐inferior) and 3.4 mm (anterior‐posterior), indicating non‐negligible rotation/deformation likely due to the influence of the adjacent cardiac wall. Conclusion : To our knowledge, this is the first demonstration of MRI for real‐time imaging of lung cancer. The incorporation of these strategies into MRI+Linacs offers image‐guidance capabilities that are not possible using current techniques: (i) soft‐tissue‐based rather than surrogate‐based monitoring (ii) no fiducial implantation or imaging dose (iii) arbitrary slice selection and (iv) ability to monitor complex motion.