Materials and Methods Introduction Fabrication
Production of Hydroxyapatite/polymer Blend Standards Used for #P198 Cone Beam CT Calibration. R. Wang,* K. Varghai, S. Eppell, Andrew Levite, Case Western Reserve University Schools of Dental Medicine and Engineering
Introduction Objectives: Calibration of a CBCT scanner is important to quantitatively analyze bone quality that is closely related to the insertion torque value during implant placement. The purpose of the study was to fabricate small size of standards with known densities to calibrate CBCT images in the region of interest (ROI) to allow clinicians to quantitatively asses bone quality with CBCT scans
Materials and Methods A twin-screw extruder was used to fabricate three small standards (4mmx2mmx1.2mm) with known mass density based on the combination of hydroxyapatite (HA) and highdensity polyethylene (HDPE), 27%, 45% and 68% by weight of HA in HDPE. Weight percentage of HA were chosen so that HU of the standards span the range of HU of trabecular and cortical bone (700-3000). The three standards were secured on the lingual aspect of the first mandibular molar and the roof of the mouth of a dry skull. The skull was imaged using two CBCT scanners. After imaging, the average volumetric HU for each standard was collected using customized Matlab software. A calibration curve was generated relating the grey level number to the corresponding HU for each standard. The mean HU for the ROI was calculated using the slope and intercept from the linear fit.
Fabrication Step 1:
Method Testing Inveon Micro PET/CT Scanner (Pilot): A 40 mm x 60 mm polyurethane saw bone block (Density = 0.80 g/cc) was imaged inside an acrylic container (89 mm and height of 102 mm) filled with water. A small cylindrical vial (8.0 mm diameter, 32.0 mm tall) filled with a lead nitrate solution in distilled water was imbedded in the polyurethane block for each scan. This was done to outline a clear VOI in the block. The water and polyurethane block were meant to act as a surrogate for soft tissue and trabecular bone, respectively. Two scans were performed with the calibration standards fixed onto the polyurethane block. Calibration standards were taped onto a thin plastic board ~0.5 cm apart from each other. The samples were imbedded into a dental X-ray film pouch and taped onto the polyurethane block with waterproof tape. In the first scan, the pouch containing the standards was fixed onto the block 1.25 cm away from the ROI (marked by the imbedded Lead (II) Nitrate vial). The second scan involved taping the standards onto the block 3.25 cm away from the ROI. Testing Carestream 9300 CBCT scanner (Clinical): Standards were placed in a human skull first at the top of the oral cavity, then placed closer to the chosen region of interest. This region of interest was between the first molar and the Mental Forman. The skull was placed in a bucket of water to simulate exomass during imaging. The images were then analyzed using the Matlab code, and grayscale values of the standards and region of interests were recorded. The test was repeated three times in order to ensure reproducibility of the results.
Testing: Inveon Micro PET/CT Scanner 1.25 cm
Testing: Carestream 9300 CBCT scanner Standards
Standards
3.25 cm
Testing: Inveon Micro PET/CT Scanner
Testing: Carestream 9300 CBCT scanner
Table 1: Mean grey level measurements for four standards, Average GSV For each standard is represented when placed both 1.25 cm and 3.25 cm away from the ROI. Standards containing 27% by mass HA, 45% by mass HA and 68% by mass HA are represented by 27%, 45% and 68%, respectively. Known HU represents the calculated HU for the standards using the lookup tables form the NIST website.
Table 2: Mean grey level measurements for four standards. Average GSV for each standard is represented when placed both on the mandible near the ROI and on the roof of the oral cavity. Standards containing 27% by mass HA, 45% by mass HA And 68% by mass HA are represented by 27%, 45% and 68%, respectively. Known HU represents calculated HU for the standards using the lookup tables form the NIST website.
Standard 1.25 cm Avg. 3.25 cm Avg. Known HU 27% (GSV) 2,963 -4,042 723 45% (GSV) 8,936 2,088 1,447 68% (GSV) 22,885 18,421 2,728 Aluminum (GSV) 25,349 21,438 4,640
Standard 1.25 cm Avg. 3.25 cm Avg. Known HU 27% (GSV) 1,295 1,188 723 45% (GSV) 1,517 1,344 1,447 68% (GSV) 1,976 1,684 2,728 Aluminum (GSV) 2,748 2,536 4,640
7000
7000
6000
6000
y = 0.1291x + 1161.1 R² = 0.8676
5000 4000 3000 2000
y = 0.1474x + 168.27 R² = 0.8623
1000 0 -10,000
0
10,000 20,000 Average Grey Level
30,000
1.25 cm 3.25 cm Linear (1.25 cm) Linear (3.25 cm)
Hounsfield Unit
Hounsfield Unit
Method
y = 2.8181x - 2372.4 R² = 0.9764
5000 4000 3000 2000
y = 2.6711x - 2647.9 R² = 0.9975
1000
Near Far Linear (Near) Linear (Far)
0 40,000
0
1,000
2,000 3,000 Average Grey Level
4,000
Results and Discussion Standards: Average volumetric HU measurements revealed that the three standards contained HU numbers of 723, 1447 and 2728 HU. Clinical Scans: For the three trials where the standards were at the top of the oral cavity, the Hounsfield units of the regions of interest were 4898, 5495, and 4526, which have a range of 969 and average out to a value of 4973. For the three trials where the standards were adjacent to the ROI, the HU of the ROIs were 4258, 4610, and 4761, which have a range of 503 and average out to a value of 4543.
Table 3: Length, width, and thickness of each standard as well as the HU measurement at 40keV.
Measurement
27% HA
45% HA
68% HA
Length
3.9 ± .02 mm
4.8 ± .11 mm
4.1 ± .02 mm
Thickness
1.8 ± .01 mm
1.6 ± .03 mm
1.3 ± .02 mm
746 ± 6 HU
1,502 ± 40 HU
2,778 ± 47 HU
Hounsfield Units (40 keV)
Table 4: Grey level measurements for a single ROI in three different Clinical CBCT trials. Near Measured (GSV) represent grey levels measured for a ROI when standards were placed near the ROI. Roof measured (GSV) represent grey levels measured for a ROI when standards were placed on the roof of the oral cavity. Corrected HU represents calculated HU for the ROI using the linear fit equation.
Trial 1 2 3 Mean
Near Measured (GSV) 2,792 2,716 2,580 2,696
Near Corrected (HU) 4,258 4,610 4,761 4,543
Roof Measured Roof Corrected (GSV) (HU) 2,597 4,898 2,595 5,495 2,586 4,526 2,593 4,973
Conclusion
A positive correlation between the standards’ grey levels and HA mass percentage was observed. Data collected from the two CBCT scans revealed that the grey level for each standard changes significantly as a function of their locations. We can also conclude that the proximity of the standards to the region of interest does in fact produce a large difference in the calibrated HU of the region of interest. In this case, shifting the standards about 2.25 cm further away from the region of interest changed the HU value of the ROI by 9.465%, and also increased the range of HU values.
Introduction Objectives: Calibration of a CBCT scanner is important to quantitatively analyze bone quality that is closely related to the insertion torque value during implant placement. The purpose of the study was to fabricate small size of standards with known densities to calibrate CBCT images in the region of interest (ROI) to allow clinicians to quantitatively asses bone quality with CBCT scans
Materials and Methods A twin-screw extruder was used to fabricate three small standards (4mmx2mmx1.2mm) with known mass density based on the combination of hydroxyapatite (HA) and highdensity polyethylene (HDPE), 27%, 45% and 68% by weight of HA in HDPE. Weight percentage of HA were chosen so that HU of the standards span the range of HU of trabecular and cortical bone (700-3000). The three standards were secured on the lingual aspect of the first mandibular molar and the roof of the mouth of a dry skull. The skull was imaged using two CBCT scanners. After imaging, the average volumetric HU for each standard was collected using customized Matlab software. A calibration curve was generated relating the grey level number to the corresponding HU for each standard. The mean HU for the ROI was calculated using the slope and intercept from the linear fit.
Fabrication Step 1:
Method Testing Inveon Micro PET/CT Scanner (Pilot): A 40 mm x 60 mm polyurethane saw bone block (Density = 0.80 g/cc) was imaged inside an acrylic container (89 mm and height of 102 mm) filled with water. A small cylindrical vial (8.0 mm diameter, 32.0 mm tall) filled with a lead nitrate solution in distilled water was imbedded in the polyurethane block for each scan. This was done to outline a clear VOI in the block. The water and polyurethane block were meant to act as a surrogate for soft tissue and trabecular bone, respectively. Two scans were performed with the calibration standards fixed onto the polyurethane block. Calibration standards were taped onto a thin plastic board ~0.5 cm apart from each other. The samples were imbedded into a dental X-ray film pouch and taped onto the polyurethane block with waterproof tape. In the first scan, the pouch containing the standards was fixed onto the block 1.25 cm away from the ROI (marked by the imbedded Lead (II) Nitrate vial). The second scan involved taping the standards onto the block 3.25 cm away from the ROI. Testing Carestream 9300 CBCT scanner (Clinical): Standards were placed in a human skull first at the top of the oral cavity, then placed closer to the chosen region of interest. This region of interest was between the first molar and the Mental Forman. The skull was placed in a bucket of water to simulate exomass during imaging. The images were then analyzed using the Matlab code, and grayscale values of the standards and region of interests were recorded. The test was repeated three times in order to ensure reproducibility of the results.
Testing: Inveon Micro PET/CT Scanner 1.25 cm
Testing: Carestream 9300 CBCT scanner Standards
Standards
3.25 cm
Testing: Inveon Micro PET/CT Scanner
Testing: Carestream 9300 CBCT scanner
Table 1: Mean grey level measurements for four standards, Average GSV For each standard is represented when placed both 1.25 cm and 3.25 cm away from the ROI. Standards containing 27% by mass HA, 45% by mass HA and 68% by mass HA are represented by 27%, 45% and 68%, respectively. Known HU represents the calculated HU for the standards using the lookup tables form the NIST website.
Table 2: Mean grey level measurements for four standards. Average GSV for each standard is represented when placed both on the mandible near the ROI and on the roof of the oral cavity. Standards containing 27% by mass HA, 45% by mass HA And 68% by mass HA are represented by 27%, 45% and 68%, respectively. Known HU represents calculated HU for the standards using the lookup tables form the NIST website.
Standard 1.25 cm Avg. 3.25 cm Avg. Known HU 27% (GSV) 2,963 -4,042 723 45% (GSV) 8,936 2,088 1,447 68% (GSV) 22,885 18,421 2,728 Aluminum (GSV) 25,349 21,438 4,640
Standard 1.25 cm Avg. 3.25 cm Avg. Known HU 27% (GSV) 1,295 1,188 723 45% (GSV) 1,517 1,344 1,447 68% (GSV) 1,976 1,684 2,728 Aluminum (GSV) 2,748 2,536 4,640
7000
7000
6000
6000
y = 0.1291x + 1161.1 R² = 0.8676
5000 4000 3000 2000
y = 0.1474x + 168.27 R² = 0.8623
1000 0 -10,000
0
10,000 20,000 Average Grey Level
30,000
1.25 cm 3.25 cm Linear (1.25 cm) Linear (3.25 cm)
Hounsfield Unit
Hounsfield Unit
Method
y = 2.8181x - 2372.4 R² = 0.9764
5000 4000 3000 2000
y = 2.6711x - 2647.9 R² = 0.9975
1000
Near Far Linear (Near) Linear (Far)
0 40,000
0
1,000
2,000 3,000 Average Grey Level
4,000
Results and Discussion Standards: Average volumetric HU measurements revealed that the three standards contained HU numbers of 723, 1447 and 2728 HU. Clinical Scans: For the three trials where the standards were at the top of the oral cavity, the Hounsfield units of the regions of interest were 4898, 5495, and 4526, which have a range of 969 and average out to a value of 4973. For the three trials where the standards were adjacent to the ROI, the HU of the ROIs were 4258, 4610, and 4761, which have a range of 503 and average out to a value of 4543.
Table 3: Length, width, and thickness of each standard as well as the HU measurement at 40keV.
Measurement
27% HA
45% HA
68% HA
Length
3.9 ± .02 mm
4.8 ± .11 mm
4.1 ± .02 mm
Thickness
1.8 ± .01 mm
1.6 ± .03 mm
1.3 ± .02 mm
746 ± 6 HU
1,502 ± 40 HU
2,778 ± 47 HU
Hounsfield Units (40 keV)
Table 4: Grey level measurements for a single ROI in three different Clinical CBCT trials. Near Measured (GSV) represent grey levels measured for a ROI when standards were placed near the ROI. Roof measured (GSV) represent grey levels measured for a ROI when standards were placed on the roof of the oral cavity. Corrected HU represents calculated HU for the ROI using the linear fit equation.
Trial 1 2 3 Mean
Near Measured (GSV) 2,792 2,716 2,580 2,696
Near Corrected (HU) 4,258 4,610 4,761 4,543
Roof Measured Roof Corrected (GSV) (HU) 2,597 4,898 2,595 5,495 2,586 4,526 2,593 4,973
Conclusion
A positive correlation between the standards’ grey levels and HA mass percentage was observed. Data collected from the two CBCT scans revealed that the grey level for each standard changes significantly as a function of their locations. We can also conclude that the proximity of the standards to the region of interest does in fact produce a large difference in the calibrated HU of the region of interest. In this case, shifting the standards about 2.25 cm further away from the region of interest changed the HU value of the ROI by 9.465%, and also increased the range of HU values.
