CT Perfusion Imaging CT Perfusion Imaging Ivan Yeung, Ph.D. Dept of Radiation Physics, Princess Margaret Hospital Dept of Radiation Oncology, University of Toronto University of Toronto
st Commercial CT Unit 1 st 1 Commercial CT Unit Commercial CT Unit Commercial CT Unit • In 1972, Sir Godfrey Hounsfield introduced the first EMI In 1972, Sir Godfrey Hounsfield introduced the first EMI In 1972, Sir Godfrey Hounsfield introduced the first EMI In 1972, Sir Godfrey Hounsfield introduced the first EMI CT scanner. CT scanner. • It takes 5 min to scan and 2.5 hours to reconstruct. It takes 5 min to scan and 2.5 hours to reconstruct. • Designed as a revolutionary Designed as a revolutionary anatomical anatomical imaging tool. imaging tool. EMI scanner PMH 50 th Anniversary Conference 2 2
Development of CT Development of CT Development of CT Development of CT • CT was quickly development, during the 80’s, CT CT was quickly development, during the 80’s, CT has become a work horse in almost every radiology has become a work horse in almost every radiology y y gy gy department. department. • Due to presence of cables, the fastest CT scan was Due to presence of cables, the fastest CT scan was done every 4 done every 4-6 seconds done every 4 done every 4-6 seconds. 6 seconds 6 seconds. • First attempt to use CT to measure function First attempt to use CT to measure function – – tissue tissue perfusion perfusion – – is with stable Xenon is with stable Xenon- -133. 133. • ‘Messy’ experiment: patient breathes in Xenon, it ‘Messy’ experiment: patient breathes in Xenon, it diffuses from lung to blood, then carry to tissue diffuses from lung to blood, then carry to tissue • Nice property Nice property – high diffusion coefficient across Nice property Nice property high diffusion coefficient across high diffusion coefficient across high diffusion coefficient across capillary. capillary. • Can be described by one Can be described by one- -compartment model. compartment model. Hence less computational demand and sparse Hence less computational demand and sparse Hence less computational demand and sparse Hence less computational demand and sparse sampling is ok. sampling is ok. PMH 50 th Anniversary Conference 3 3
One One-Compartment Model One One Compartment Model Compartment Model Compartment Model PMH 50 th Anniversary Conference 4 4
Slip Ring Technology (mid 90’s) Slip Ring Technology (mid 90 s) Slip Ring Technology (mid 90’s) Slip Ring Technology (mid 90 s) • Power and data transfer via Power and data transfer via contact around gantry contact around gantry • Gantry moves in one Gantry moves in one di di direction direction ti ti • Gantry can move up to 1/3 Gantry can move up to 1/3 sec per rotation sec per rotation sec per rotation sec per rotation • High temporal resolution High temporal resolution • Data able to handle more Data able to handle more Data able to handle more Data able to handle more complex model complex model – – exchange exchange of contrast agent of contrast agent PMH 50 th Anniversary Conference 5 5
Distributed Model Distributed Model Distributed Model Distributed Model • F F – – Perfusion, blow flow through the capillary Perfusion, blow flow through the capillary • PS – PS – Leakiness from intravascular to extravascular space Leakiness from intravascular to extravascular space • MTT MTT – – Time to take bolus to travel across the capillary Time to take bolus to travel across the capillary PMH 50 th Anniversary Conference 6 6
Perfusion CT Perfusion CT contrast injector AIF(t) Tissue(t) Model PMH 50 th Anniversary Conference 7 7
Tracer Kinetics Models Tracer Kinetics Models Tracer Kinetics Models Tracer Kinetics Models AIF(t) Tissue(t) Model GE Perfusion … GE P f i PMH 50 th Anniversary Conference 8 8
Non Non Non linear Curve Fitting Non-linear Curve Fitting linear Curve Fitting linear Curve Fitting AIF(t) ( ) Tissue(t) ssue( ) Model Model F 1 , PS 1 , Vb 1 …. F PS Vb Measured F 2 , PS 2 , Vb 2 …. Fitted F 3 PS 3 Vb 3 F 3 , PS 3 , Vb 3 …. . F n , PS n , Vb n … n , n , n PMH 50 th Anniversary Conference 9 9
Patients received Sorafenib + Rad Patients received Sorafenib + Rad Pt # 1 Pt # 1 Baseline Day 7 Day 21 CT 0.022 0.02 0.018 0.016 0.014 Perfusion 0.012 0.01 Map 0.008 0.006 0.004 0.002 0.002 PMH 50 th Anniversary Conference 10 10
Patients received Sorafenib + Rad Patients received Sorafenib + Rad Patients received Sorafenib + Rad Patients received Sorafenib + Rad Pt # 2 Pt # 2 Baseline Day 7 Day 21 CT 0.05 0.04 Perfusion 0.03 Map 0.02 0.01 0 PMH 50 th Anniversary Conference 11 11
Preclinical Cervix Tumor xenograft in mice Preclinical Cervix Tumor xenograft in mice Preclinical Cervix Tumor xenograft in mice Preclinical Cervix Tumor xenograft in mice Perfusion Perfusion High Low PMH 50 th Anniversary Conference 12 12
Perfusion CT on Mouse 1 (anti Perfusion CT on Mouse 1 (anti- -vascular drug) vascular drug) Blood Flow ood o Enhanced CT Enhanced CT Blood Volume Blood Volume MTT MTT PMH 50 th Anniversary Conference 13 13
Advancement in Perfusion CT Advancement in Perfusion CT Advancement in Perfusion CT Advancement in Perfusion CT • True 4D CT True 4D CT – True 3D volume scan is possible True 4D CT True 4D CT True 3D volume scan is possible True 3D volume scan is possible True 3D volume scan is possible • Dose reduction schemes: intermittent X Dose reduction schemes: intermittent X- -ray on, low ray on, low dose CT, less frequent sampling dose CT, less frequent sampling dose CT, less frequent sampling dose CT, less frequent sampling • Optimization of protocol Optimization of protocol • Limited speed cone • Limited speed cone Limited speed cone beam CT perfusion Limited speed cone-beam CT perfusion beam CT perfusion beam CT perfusion PMH 50 th Anniversary Conference 14 14
Dynamic Contrast Enhanced Perfusion CT Dynamic Contrast Enhanced Perfusion CT vs Contrast CBCT vs Contrast CBCT C C t t t CBCT t CBCT Multiple CT images at CT Acquisition over time different time different time Single contrast enhanced Acquisition over time CBCT CBCT image PMH 50 th Anniversary Conference 15 15
DCE DCE CBCT DCE-CBCT DCE CBCT CBCT • Assume each voxel behaves as a single compartment with its wash Assume each voxel behaves as a single compartment with its wash- -in in and washout behaviour and washout behaviour t 2 2 2 = − α − α + α − kt 2 t t t C ( t ) Ie [ t e e e ] α α α b 2 2 where Cb(t) where Cb(t) – – contrast enhancement in each voxel contrast enhancement in each voxel t t – – time time I, α , k I, , , k – , – unknown parameters unknown parameters p t = • t=0s, 0 o 1 s Reconstruct the kinetics behavior Reconstruct the kinetics behavior , 3 o according to the formula for each according to the formula for each t=2s, voxel by optimization to determine the voxel by optimization to determine the voxel by optimization to determine the voxel by optimization to determine the , 6 o 3 unknown parameters (I, α , k) with 3 unknown parameters (I, , k) with projection data projection data • 640 projections were obtained in 2 640 projections were obtained in 2 min. The min. The net net (i.e. contrast (i.e. contrast – – baseline) baseline) contrast projections were used for contrast projections were used for optimization in Matlab for a volume of optimization in Matlab for a volume of 128x128x40. 128x128x40. PMH 50 th Anniversary Conference 16 16
Results: CT and Results: CT and Estimated Estimated CBCT Tissue CBCT Tissue E h Enhancement and Perfusion Enhancement and Perfusion E h t t d P d P f f i i CT CBCT N t Net Enhancement Perfusion PMH 50 th Anniversary Conference 17 17
Conclusions Conclusions Conclusions Conclusions • CT originally designed for anatomy CT originally designed for anatomy CT originally designed for anatomy CT originally designed for anatomy • Functional measurement (F, PS, MTT) can be Functional measurement (F, PS, MTT) can be performed for clinical and preclinical studies performed for clinical and preclinical studies performed for clinical and preclinical studies performed for clinical and preclinical studies • On On- -going Advancement in perfusion CT going Advancement in perfusion CT PMH 50 th Anniversary Conference 18 18
Acknowledgement Acknowledgement Acknowledgement Acknowledgement Lab members (past/present) Lab members (past/present) (p (p p p ) ) Collaborators Collaborators • • • Sunmo Kim Sunmo Kim Richard Hill Richard Hill Masoom Haider Masoom Haider • • • Brian Lim Brian Lim Mike Milosevic Mike Milosevic Doug Moseley Doug Moseley g g y y • • • Richard Clarkson Richard Clarkson Rob Bristow Rob Bristow Young Young- -Bin Cho Bin Cho • • • Qiulin Tang Qiulin Tang g Anthony Fyles Anthony Fyles y y y y Ting-Yim Lee Ting g Yim Lee • • Bill Qian Bill Qian David Hedley David Hedley • David Jaffray David Jaffray y • Jeff Siewerdsen Jeff Siewerdsen Funding Sources Funding Sources • • Terry Fox Foundation Terry Fox Foundation Young Young-Bin Cho g Bin Cho • Canadian Institute of Health Research Canadian Institute of Health Research PMH 50 th Anniversary Conference 19 19
Recommend
More recommend
