8/1/2018. Optimization Strategies for Pediatric CT Imaging. Educational Objectives. Brain CT

Transcription

1 Image Quality Radiation Dose Contrast Dose Sedation Optimization Strategies for Pediatric CT Imaging Samuel Brady, M.S. Ph.D. DABR facebook.com/cincykidsrad Educational Objectives Pediatric CT Protocol Development Brain & Body CT Brain CT Brain CT (one of the more tricky protocols to optimize) Two biggest limitations Skull (especially the posterior fossa) Minimum inherent differentiation between structures in brain Beam Hardening Gray matter 34 HU White matter 26 HU 1

2 Brain CT The head grows logarithmically In diameter & bone density Kleinman, et al. AJR 21; 194: Brain CT Cone angle truncation The head grows logarithmically In length (S/I) Volumetric CT (16 mm) is limited by cone artifact (~14 mm) Brain CT Established 5 protocols -6 mon 7-24 mon 2-5 years 6-13 years > 13 years Kleinman, et al. AJR 21; 194:

3 Brain CT 1 kv: better gray/white differentiation CTDI: dose and noise balanced Fixed ma: not sensitive to misalignment issues with TCM Speed: volume mode < 2 yr old Protocol kv ma Rot (s) Coll (mm) CTDIvol (mgy) -6 m (V) x m (V) x m-5y (H) x y (H) x y (H) x mon 1 yr 5 yr 12 yr 2 yr Body CT Proper protocol Keys to success: Organ dose modulation Simplify protocol tree Iterative reconstruction Metal Artifact reduction software Myth: you need lots of protocol variations for pediatrics 5 will cover the range from to > 1 kg Acquisition Time Faster acquisition may lead to lower sedation rates E.g. 5 yr; 3 cm scan length Pitch sec rotation Total exposure time =.7 sec Single acquisition ~.28 sec Great for head, chest, cardiac, or limited FOV (e.g., kidneys) 3

4 CT # (HU) Acquisition Time Faster acquisition may lead to lower sedation rates E.g. 5 yr; 3 cm scan length Pitch sec rotation Total exposure time =.7 sec Single acquisition ~.28 sec Great for head, chest, cardiac, or limited FOV (e.g., kidneys) Can almost freeze even cardiac motion Body CT Proper protocol Organ dose modulation Iterative reconstruction Weight/size based adjustments kvp Metal Artifact reduction software Develop good contrast timing schemes Relative Contrast 5 Soft Tissue 4 Fat 3 Bone Tube Potential (kv) Why decrease kv? Enhance Iodine or bone contrast 12 vs. 8 kv = 42% contrast Soft tissue vs. bone 12 vs. 8 kv ~ 1% contrast 12 kv 1 kv 4

5 CNR/CTDIvol CNR/CTDIvol CNR/CTDIvol x1.5 cm Phantom-NB kv 18.5x24 cm Phantom-12YO 2 mg/cc 5 mg/cc 1 mg/cc kv 2 mg/cc 5 mg/cc 1 mg/cc x18 cm Phantom-5YO x32.5 cm Phantom-Adult 2 mg/cc 5 mg/cc 1 mg/cc kv kv 2 mg/cc 5 mg/cc 1 mg/cc Why decrease kv? Lower IV/oral contrast dose Compensate contrast intensity w/ lower kv Only valid in small patients How to decrease kv clinically? Don t match noise, match CNR! Calculate target CNR in phantom mas 2 mas 1 = kv1 α kv2 mas2 ln mas1 α = ln kv1 kv ln 25 = ln 12 8 = 3.8 CNR = 7.5 CNR = ma.5 sec 12 kv 2 mgy ~5 yr old phantom 23 ma.5 sec 8 kv 2 mgy *Bushberg et al, The essential physics of medical imaging 3 rd Ed, LWW Apply α value towards patient protocols Use avg TCM ma value α mas 2 = mas 1 kv1 kv2 e.g., mas 85 mas 1 CNR improves with lower kv even though noise increases Noise can be higher at lower kv Same Patient 12 kv 1 kv 31% dose reduction Fletcher, AAPM 21 Yu et al. Med Phys 57(1) 21; *Bushberg et al, The essential physics of medical imaging 3 rd Ed, LWW 5

6 Dose and Image Quality Optimization-CT 12 kv; 1.81 mgy 1 kv; 2.3 mgy Why decrease kv? Caveat: you can get good dose reduction at 12 kv Rule of thumb (trunk) Routine 8 kv 15 cm d w dia. (CCHMC, <15 kg) 36 cm lat dim (Mayo)* Routine 1 kv 25 cm dia./d w (CCHMC, <71 kg) 41 cm lat dim (Mayo)* 12 kv; 1.81 mgy 1 kv; 2.75 mgy *Fletcher, AAPM 21 Yu et al. Med Phys 57(1) 21; Why increase kv? Penetration Larger patients, generally, require higher kv (i.e., 12 kv) Brain imaging ~> 4 yr old (use 12 kv) Focal spot consideration For high resolution (small structure imaging) use smallest focal spot Small focal spot (~.6 mm) is available ~ 35 ma Body CT Proper protocol Organ dose modulation Iterative reconstruction Weight/size based adjustments kvp Metal Artifact reduction software Develop good contrast timing schemes 6

7 Hepatic Enhancement (HU) IV Contrast Develop good contrast timing schemes Poor contrast management may mask pathology Scanned too early Portal Venus Phase Popular approaches Bolus tracking Fixed time IV Contrast Constant injection duration.5 cc/sec 1 cc/sec 1.5 cc/sec 2 cc/sec 2.5 ccsec Time after the start of injection (sec) Bae; Radiology (21) 256(1):32 61 T delay = ID+T ARR -5- ½ T ScanDuration (Arterial phase) T delay = ID+T ARR +25- ½ T ScanDuration (PV phase) T ARR for liver parenchyma is 2x Abd Aorta or ~15 sec Protocol (kg) ID: injection delay ID Iodine (sec) Volume (cc); Rate (cc/sec) T peak enhancement (sec) (ID iodine +1/3) (Arterial/PV) <3 ~ / / /7 > /9 > /9 Dose Dose and Image Quality Optimization-CT Proper protocol Proper set up Organ dose modulation Iterative reconstruction Metal Artifact reduction software 25-3% Anterior Reduction Lungren, et al, AJR (212) 199:W65 W73 7

8 Dose and Image Quality Optimization-CT TCM TCM + ODM Dose and Image Quality Optimization-CT Proper protocol Dose Reduction 2-7% Proper set up Organ dose modulation Iterative reconstruction Metal Artifact reduction software IR #1 IR # 2 Dose and Image Quality Optimization-CT Iterative reconstruction limitations: Plastic or soft look No matter the level of iterative recon there will be losses in edge definition Mostly affecting soft tissue kernels Irecon Off Irecon On Irecon Off Irecon On 8

9 Future Directions of CT Reconstruction Deep learning models 1 s of CT images used to train a reconstruction algorithm Reconstruction will only be as good as the data used to train Must include adults & peds Performs well in sinogram space at filling in missing data e.g., fixing metal artifact e.g., cleaning up noisy data Future Directions of CT Reconstruction Low dose reconstruction Full dose ~23-43 ma ¼ dose ~6-11 ma Recon time 1.6 sec/slice Total: 2-3 min (MATLAB) 1 mm thin slices Kang, et al. Med Phys 44 (1) 217 e36-e375 Use metal artifact reduction software Increasing kv and mas does NOT improve metal artifacts! Original Dual energy Single energy Canon, SEMAR white paper 9

10 Thank you 1