March 2013 - Institute of Physics and Engineering in Medicine

NEWS BY USMAN I. LULA AND RICHARD AMOSMRI simulator: is it time for an additional standardimaging tool in radiotherapy treatment planning?IMAGINGCT is still the backbone ofradiotherapy treatment planning(RTP) as it provides superb spatialreproducibility and information onelectron density. A major limitationof CT is its poor ability tocharacterize different tissues andthus MRI, with its excellent softtissue contrast, has historicallybeen incorporated into the targetdefinition. Integrating MRI data fortarget delineation requiressophisticated processes of imageco-registration. Increasingly,improved precision of radiotherapydelivery technologies seek tightermargins in GTV definitions andconsequently the need to use MRIsimulation to its full extent.Co-registration of MRI with CTrequires the application of aparticular transformation modelonto the secondary dataset. Acommon model is ‘affine’transformation, used when there isglobal gross overall distortion.Currently available RTP softwareuses various approaches to coregisterany two imaging modalities(table 1).MRI is useful in outlining criticalstructures where tooth fillings orprosthetic implants severelydegrade the CT image quality. Forbrain neoplasms, it has beenreported that MRI markedlyincreased the apparentmacroscopic tumour volume fromthat seen on CT alone. IMRTplanning with CT/MRI coregistrationfor nasopharyngealcarcinomas significantly improvedcoverage of targets and sparing ofcritical structures. MRI has beenfound to be superior for localisationof the prostatic apex. Moreover, thedose delivered to the rectal wall andbulb of the penis was significantlyreduced with treatment plansbased on MRI-delineated prostate,and that tumour volumes weresmaller, shorter and more distalfrom the anal sphincter than CTbasedvolumes. This resulted in areduced dose to Organs At RiskTABLE 1Author Year Method AnatomicalsiteAlpert et al. (ref. 13) 1990 Matching moments method Brain 1.0Turkington et al. (ref. 14) 1993 Maximum gradient technique Brain 2.0Van Herk and Kooy (ref. 15) 1994 Contours chamfer matching Brain 1.0Kagawa et al. (ref. 8) 1997 3 anatomical landmarks Prostate 0.9 (±0.6)Sannazzari et al. (ref. 16) 2002 9–15 anatomical landmarks Prostate 1.5Krempien et al. (ref. 10) 2003 Mutual information H&N, GYN 1.8 (±0.9)Brock et al. (ref. 12) 2010 Various deformable Various 0.4–6.2Table 1: Reported MRI to CT co-registration errors for various used co-registration methods. Table kindlysupplied by Slobodan Devic, Department of Radiation Oncology, Jewish General Hospital, McGill University,Montreal, Quebec H3T 1E2, Canada. © 2012 American Association of Physicists in Medicine, ‘MRIsimulation for radiotherapy treatment planning’, Slobodan Devic, Med Phys 2012; 39 (11)TABLE 2(OAR) and facilitated dose escalation.Factors that could seriouslyundermine the use of MRI outsidethe head region is geometric MRimage distortions and anatomicalshape changes of target and OARdespite the use of image distortioncorrection techniques anddeformable registration methods.Contemporary radiotherapydelivery technologies allow for veryhigh dose gradient dosedistributions to be delivered with aspatial resolution approaching amicroscopic scale. These advancesdemand better accuracy in definingnot only the tumour but also theOAR to improve clinical outcomes.Ideally, the acquisition time for anMRI simulator must be brought to aminimum whilst achieving therequired image quality and spatialresolution (table 2).The multitude of benefits ofReported coregistrationerror (mm)Anatomical site Acquisition sequence ReferenceBrainHead and neckBreastT1 3D gradient echoPost Gd-T1 standard spin-echoProton density fluid-attenuated inversion recoveryT2 FLAIRPost Gd-T1 standard spin-echoT2-weighted sequence with fat saturationT1 3D gradient echo (pre- and post-contrast)T1 inversion recovery (STIR) sequencesT2-weighted 3D fast spin echo (XETA)T1-weighted turbo spin echoT1 3D gradient echoGYN Turbo spin echo T2 (TSE T2)T2-weighted fast spin echo (FSE)T1 3D gradient echoProstateRectumT2-weighted fast spin echo (FSE)T1 3D gradient echoT2-weighted fast spin echo (FSE)T1 and T2 short tau inversion recovery (STIR)T1 3D gradient echo20, 592222602425602628296043606034, 36, 376039, 406160Table 2: Overview of MRI sequences used for target definition in radiotherapy treatment planning process.Table kindly supplied by Slobodan Devic, Department of Radiation Oncology, Jewish General Hospital,McGill University, Montreal, Quebec H3T 1E2, Canada. © 2012 American Association of Physicists inMedicine, ‘MRI simulation for radiotherapy treatment planning’, Slobodan Devic, Med Phys 2012; 39 (11)using MRI in the RTP processhighlights the need to incorporatean MRI simulator as a standardimaging modality in radiotherapy.An optimal setup would be theinclusion of both a CT simulatorand an MRI simulator within adepartment.‘ MORE INFORMATIONThis work was recently published inMed Phys 2012; 39(11).http://dx.doi.org/10.1118/1.4758068SCOPE | MARCH 2013 | 07