Dual-Energy Imaging of Bone Edema on a Dedicated Multi-Source Cone-Beam CT System for the Extremities W Zbijewski, 1 A Sisniega, 1 JW Stayman, 1 N Packard, 2 J Yorkston, 2 G Thawait, 3 S Demehri, 3 J Fritz, 3 JH Siewerdsen 1,3 1 Dept.of Biomedical Engineering, 2 Carestream Health 3 Russel H. Morgan Dept. of Radiology, The I-STAR Laboratory Imaging for Surgery, Therapy, and Radiology http://www.jhu.edu/istar/ Collaborators C Bingham, S Ghazarian (JHU Rheumatology) M Mahesh (JHU Radiology) Funding Support NIH 2R01-CA-112163 NIH 1R21-AR-062293 NIH 1R01-EB-018896
Dedicated Extremities CBCT Extremities CBCT Standing and sitting scans Dose ~10 mgy High spatial resolution (~0.5 mm) Soft-tissue contrast resolution Flexible platform Three-Source Configuration Custom fixed-anode unit Three x-ray tubes arranged axially Increased Field-of-View Reduced cone beam artifacts New method for Dual-Energy CBCT Obviates need for double scan Dedicated Extremities CBCT Extremities CBCT Standing and sitting scans Dose ~10 mgy High spatial resolution (~0.5 mm) Soft-tissue contrast resolution Flexible platform Three-Source Configuration Custom fixed-anode unit Three x-ray tubes arranged axially Increased Field-of-View Reduced cone beam artifacts New method for Dual-Energy CBCT Obviates need for double scan 1 source 3 source
Dedicated Extremities CBCT Extremities CBCT Standing and sitting scans Dose ~10 mgy High spatial resolution (~0.5 mm) Soft-tissue contrast resolution Flexible platform Dual-Energy (DE) Imaging Arthrography Iodine / Bone / Soft Tissue Gout Uric Acid / Bone / Soft Tissue Bone Mineral Density (BMD) Bone Edema (BME) Fat / Bone / Soft Tissue Zbijewski et al, Med Phys (2014) DE Imaging of Bone Edema Bone Edema (BME) Increased fluid (water) content Decreased fat (yellow marrow) content Arthritis, trauma, metastases Detectable in T2-weigthed MRI CBCT T2 MRI BME DE Imaging of BME Challenging in conventional CT Partial volume effect from trabeculae Dual-Energy Virtual calcium subtraction Water + marrow maps CT DE CT BME T2 MRI Pache et al, Radiology (2010)
DE with Three-Source CBCT Imaging Setup Emulates extremities CBCT 10:1 antiscatter grid LE: 60 kvp HE: 105 kvp (with or w/o high-z filter) 3-Source DE Firing Pattern Central source (2): LE 360 projections Axial sources (1 + 3): HE 180 projections each Only oblique rays at central slice Image-based DE Compared to double scan DE at equivalent dose SDD ~56 cm SAD ~43 cm HE 3 6 cm LE 2 HE 1 DE with Three-Source CBCT Imaging Setup Emulates extremities CBCT 10:1 antiscatter grid LE: 60 kvp HE: 105 kvp (with or w/o high-z filter) 3-Source DE Firing Pattern Central source (2): LE 360 projections Axial sources (1 + 3): HE 180 projections each Only oblique rays at central slice Image-based DE Compared to double scan DE at equivalent dose =1 o =1 0 SDD ~56 cm SAD ~43 cm HE 3 6 cm LE 2 HE 1
Image Reconstruction in 3-Source DE CBCT Penalized Likelihood (PL) ˆ arg max L ; y R Forward model L(µ; y) Edge-preserving Huber penalty R Low Energy reconstruction L(µ; y) involves LE 2 projections High Energy Reconstruction L(µ; y) combines HE 1 and HE 3 projections LE 2 central slice 0.03 0.018 Image Reconstruction in 3-Source DE CBCT Penalized Likelihood (PL) HE 3 ˆ arg max L ; y R Forward model L(µ; y) Edge-preserving Huber penalty R Low Energy reconstruction L(µ; y) involves LE 2 projections HE 1 High Energy Reconstruction L(µ; y) combines HE 1 and HE 3 projections HE3 reconstruction HE1 reconstruction Combined reconstruction 0.03 0.018
Image-based DE Three-material decomposition Water + Fat/ + Cortical Bone Volume preservation constraint Narrow separation between Fat and Water Artifact Correction Fast GPU Monte Carlo scatter correction ~5 min correction per scan DE Decomposition Calibration of LE and HE Water and : ROIs of pure material Bone: Measurements over a range of BMDs + linear fit μ LE μ HE 0.03 0.018 Sisniega et al (PMB 2015) 0.3 1 Vol. Frac. 0.3 1 Vol. Frac. 0 0.12 Vol. Frac. Image-based DE Three-material decomposition Water + Fat/ + Cortical Bone Volume preservation constraint Narrow separation between Fat and Water Artifact Correction Fast GPU Monte Carlo scatter correction ~5 min correction per scan DE Decomposition Calibration of LE and HE Water and : ROIs of pure material Bone: Measurements over a range of BMDs + linear fit μ LE μ HE (µ ) HE LE: 60 kvp HE: 105 kvp fat (µ ) LE water bone 0.03 0.018 Sisniega et al (PMB 2015) 0.3 1 Vol. Frac. 0.3 1 Vol. Frac. 0 0.12 Vol. Frac.
Base Materials Edema Inserts Base Materials Edema Inserts Tissue-Mimicking Materials Cortical Bone=Dipotassium Phosphate K 2 HPO 4 /Fat=Ethanol Both soluble in water Stable mixtures up to ~ fat Edema: Decreasing fraction of ethanol Imaging Experiments ~10 cm diameter phantom HE: 105 kvp, 0.1 mas/frame, ~7 mgy CTDI LE: 60 kvp, 0.8 mas/frame, ~7 mgy CTDI Experimental Study Water Volume Composition 100% H 2 O 100% C 2 H 6 O K 2 HPO 4 Cortical Bone 100% K 2 HPO4 75 mg/ml K 2 HPO 4 in H 2 O Ethanol 100 mg/ml NIST K 2 HPO Bone 4 in H 2 0 Water NIST Fat 90% 75% 60% Goodsitt et al, Investig. Radiol (1987) Tissue-Mimicking Materials Cortical Bone=Dipotassium Phosphate K 2 HPO 4 /Fat=Ethanol Both soluble in water Stable mixtures up to ~ fat Edema: Decreasing fraction of ethanol Imaging Experiments ~10 cm diameter phantom HE: 105 kvp, 0.1 mas/frame, ~7 mgy CTDI LE: 60 kvp, 0.8 mas/frame, ~7 mgy CTDI Experimental Study 0.02 0.026 Water Volume Composition 100% H 2 O 100% C 2 H 6 O Cortical Bone 100% K 2 HPO4 75 mg/ml K 2 HPO 4 in H 2 O 100 mg/ml K 2 HPO 4 in H 2 0 90% 75% 60% Goodsitt et al, Investig. Radiol (1987)
Water Vol. Fraction Vol. Fraction Cortical Bone Vol. Fraction FBP reconstruction Three-Material DE Decomposition Narrow beam double scan DE Water Bone 0.02 0.026 0.75 1 0.04 0.25 0 0.05 Three-Material DE Decomposition Narrow beam double scan DE Nominal Nominal Nominal
Water Vol. Fraction Vol. Fraction Cortical Bone Vol. Fraction W. Zbijewski SPIE Medical Imaging 2015 FBP reconstruction Three-Material DE Decomposition Narrow beam double scan DE Water Bone 0.02 0.026 0.75 1 0.04 0.25 0 0.05 Three-Material DE Decomposition Narrow beam double scan DE
Water Vol. Fraction Vol. Fraction Cortical Bone Vol. Fraction FBP reconstruction Full Beam Three-Material DE Decomposition Full beam + scatter correction Water Bone Full Beam Scatter Correction 0.02 0.026 0.75 1 0.04 0.25 0 0.05 Three-Material DE Decomposition Full beam + scatter correction
Accuracy of Bone Volume Fraction Relative difference with narrow beam DE Full Beam Scatter Correction Rel. diff.=0.04 Rel. diff. =0.06 Rel. diff. =0.01 Rel. diff.=0.03 Rel. diff.=0.01 Rel. diff. =0.01 Rel. diff.=0.12 Rel. diff.=0.03 Rel. diff.=0.01 Rel. diff.=0.09 0 0.05 Bone vol. frac. 0 0.05 Bone vol. frac. Bone and Fat Fractions in Cadaveric Knee LE: 60 kvp, 7 mgy HE: 105 kvp, 7 mgy PL reconstruction 105 kvp Three-Source DE 0.016 0.024 0.1 1 0 0.12 Fat vol. frac. Bone vol. frac.
Bone and Fat Fractions in Cadaveric Knee LE: 60 kvp, 7 mgy HE: 105 kvp, 7 mgy PL reconstruction 105 kvp Three-Source DE frac.=0.74±0.13 Bone frac.=0.059±0.008 frac.=0.88±0.08 Bone frac.=0.064±0.005 frac.=0.65±0.16 Bone frac.=0.034±0.004 frac.=0.48±0.18 Bone frac.=0.027±0.006 0.016 0.024 0.1 1 0 0.12 Fat vol. frac. Bone vol. frac. Bone and Fat Fractions in Cadaveric Knee LE: 60 kvp, 7 mgy HE: 105 kvp, 7 mgy PL reconstruction frac.=0.95±0.03 Bone frac. =0.032±0.005 Three-Source DE Fat vol. frac. 0.016 0.02 0.024 0.1 0 0.55 0.6 1 Bone vol. frac. 0.12 frac.=0.85±0.09 Bone frac.=0.029±0.006
Conclusions Fat/ Three-Source CBCT CBCT: adaptable platform for new imaging configurations Improved image quality (artifact and field of view) Detection of Edema in DE CBCT volume fraction changes 20% Requires high accuracy in HU Scatter correction (even with grid) (fast MC-GPU) 3-source DE performance comparable to double scan Reduced imaging time compared to double scan Bone Quantitative CBCT Bone Imaging DE for BMD measurement <15% of value High spatial resolution - bone microarchitecture Comprehensive assessment of bone health