Quantitative imaging of bone material analogues for PET/MRI phantoms

Dharshan Chandramohan; Peng Cao; Misung Han; Hongyu An; Kathryn Fowler; Karl Stupic; Kathryn Keenan; John Sunderland; Paul Kinahan; Richard Laforest; Thomas Hope; Peder Larson

Objectives: Quantitative accuracy in PET/MRI studies is crucial for comparison between scans acquired at different times and on different equipment, and it is also an important factor in establishing benchmarks for clinical trials. Of the different physiological tissues, bone offers a unique challenge because it has a little to no signal in MRI due to its rapid relaxation rate but has the highest attenuation. In order to quantify that effect and to characterize it in a controlled manner phantoms with appropriate bone-mineral analogues must be developed. The purpose of this study is to investigate various potential formulations for bone mineral analogues by precisely varying doping of the material to optimize MR relaxation properties and density on CT. Methods: Plaster was used as the basis of our bone-mineral analogue. When plaster sets it forms the solid crystalline mineral calcium sulfate dihydrate, which has a similar calcium content to bone and has been used historically as a bone substitute material in orthopedic surgery. Plaster was also chosen because it is easy to dope the water used to prepare the samples. 17 samples were prepared with 0 to 0.75% by mass Iohexol and 0 to 0.8% by mass Gadodiamide. Quantification of T2[asterisk] was done using a 3D ultrashort TE sequence with 12 echo times ranging from 24 to 3600 μsec. T1 values were quantified from two ultrashort TE scans with flip angles of 8 and 44 degrees respectively. All MR scans were acquired at 3T with an acquisition matrix of 110x110x100 and 2mm isotropic voxels. Hounsfield Units were derived from a CT scan using a standard cervical spine protocol with a 120 kVp beam. Results: T1, T2[asterisk] and HU values for pure undoped plaster are 546.97 ms, 1030.10 μs, and 1376 HU respectively. Doped T1 values plunge significantly and the estimated relaxivity of Gadodiamide with respect to percent by mass in plaster is 0.097 s-1; this likely underestimates the slope as even 0.1% plaster by mass drops the T1 by an order of magnitude relative to undoped plaster. T1 values for doped plaster range from 37.78 ms for plaster with 0.1% Gadodiamide to 11.33 ms for plaster with 0.8% Gadodiamide. The corresponding T1 values in human bone are in the 200-250 ms range, which suggests that the ideal Gadodiamide doping is between 0 and 0.1% by mass. Doped T2[asterisk] values also drop significantly, with a minimum T2[asterisk] of 389.68 μs for 0.8% Gadodiamide and maximum T2[asterisk] of 897.69 μs for 0.2% Gadiodiamide. Human values of T2[asterisk] for bone fall within this range, with a target range of 300-500 μs. 0.4% to 0.8% Gadodiamide doping covers this range, however, the corresponding T1 values are an order of magnitude too small. The estimated T2[asterisk] relaxivity of Gadodiamide with respect to percent by mass in plaster is 0.002 s-1. It is worth noting that Iohexol doping does not affect the relaxation parameters in any significant way. Human HU values in bone range from 700-3000. Doping with Gadodiamide and Iohexol increases average HU slightly; overall density increases with total concentration of chelate (Gadodiamide + Iohexol). The minimum density for doped plaster was 1140.41 HU (0.1% Gadodiamide, 0.0% Iohexol), and the maximum was 1438.53 HU (0.8% Gadodiamide, 0.75% Iohexol). Density quantification in plaster may have been slightly affected by other factors in casting the samples, such as air bubbles that may have formed inside the containers as the plaster set. Conclusions: Overall, doping with Gadodiamide reduces T1 below the physiological range of values for bone, optimally shortents T2[asterisk], and has a similar effect as Iohexol on HU values. Iohexol does not affect MR relaxation times for plaster in any significant way independently of Gadodiamide concentration. In future work, other T1 and T2[asterisk] shortening agents could be substituted for Gadodiamide and different techniques with respect to pouring the plaster could be used to modulate the density. The quantification of MR and attenuation parameters of materials with different doping agents is a crucial step towards developing accurate PET/MR phantoms for quantitative validation.

Quantitative imaging of bone material analogues for PET/MRI phantoms

View Online

Abstract

Details

Metrics