PO-0902: Identifying the Dominant Prostate Cancer Focal Lesion Using 3D Image Texture Analysis
D. Montgomery,K. Cheng,Y. Feng,D. B. McLaren,S. McLaughlin,W. Nailon
DOI: https://doi.org/10.1016/s0167-8140(17)31339-7
IF: 6.901
2017-01-01
Radiotherapy and Oncology
Abstract:Material and MethodsTen patients who were referred for brain metastasis radiosurgery were analysed in this study.A planning CT (1 mm slice thickness), a contrast-enhanced T1 3D MRI scan (1.5T, 1 mm isotropic voxel size, surface coils) with patient immobilized in a 3-point thermoplastic shell (mask-MR) and a contrast-enhanced T1 3D MRI scan (1.5T, 1 mm isotropic voxel size, multi-channel head coil) without immobilization mask (no mask-MR) were acquired.First, a clinician stated which of the MRI scans had superior quality, to assure that the no-mask MR had at least the same image quality compared to the clinically used mask-MR.Then, the two MRIs were registered independently to the planning CT by a normalized mutual information algorithm which was restricted to rigid registration.The GTV was delineated by 3 clinicians on 1) mask-MR and 2) no mask-MR.The brain stem, chiasm and right eye were delineated by one clinician.Furthermore, 8 well-defined landmarks were marked by an observer in both scans.Residual registration errors were estimated for both MRIs by measuring the absolute coordinate differences in the three orthogonal directions between the set of landmarks on both imaging series after registration.Moreover, the absolute differences in the centres-of-gravity coordinates of GTV (median of 3 observers), brain stem, chiasm and right eye on mask-MR and no mask-MR were compared. ResultsThe no mask-MR image quality was found to be superior in 9 of the 10 patients.The average coordinate difference between mask-MR and no mask-MR for all landmarks along the three orthogonal directions were within 0.5 mm (table 1).Similar results were found for the coordinates of the centre-of-gravity of all delineated OARs and GTV.Deviations in OAR registration > 1mm could be attributed to variations in delineation (figure 1).Only in one case, a registration error was observed.All GTV deviations were within 1mm. ConclusionThe registration of MRIs obtained with or without immobilization mask to a planning-CT generally differs less than the MRI resolution (1 mm isotropic).Therefore, immobilization of the head during MRI for patients undergoing radiotherapy of brain metastasis is not necessary.However, to guarantee high accuracy of image registration when omitting an immobilization device during MRI, more attention should be paid to the quality of MR-CT fusion.Furthermore, consecutive MR images should be matched separately to CT, to correct for intra-scan motion.We foresee two benefits of scanning without mask.Firstly, the patient comfort during the MRI scan sessions will be improved.Secondly, omission of the immobilization mask permits the use of a multi-channel head coil which results in higher image quality.Moreover, using a head coil allows for introduction of MRI techniques which require high signal-to-noise ratios or acceleration (e.g.DWI and FLAIR).