Investigation of Factors Affecting the Accuracy of a Dual Energy Quantitative CT Method for Estimating Regional Bone Marrow Cellularity
A Shenoy*, M Goodsitt, J Shen, M Schipper, Y Dewaraja, Univ Michigan, Ann Arbor, MITH-A-103-1 Thursday 8:00AM - 9:55AM Room: 103
Purpose: To evaluate a 3-equation 3-unknown Dual Energy Quantitative CT technique for determining region specific variations in bone marrow cellularity for improved red marrow dose estimation in radionuclide therapy.
Methods:The DEQCT method was applied to 80/140kVp images of patient simulating lumbar phantoms of three different sizes (small,medium,large). External calibration rods of bone, marrow and fat simulating material were placed beneath the phantoms. Similar internal calibration inserts were placed at a vertebral location within the phantoms. Six test inserts of known volume fractions of bone, fat and red marrow (Red marrow range:0.15-0.74) were also scanned. External to internal calibration correction factors were derived. The effects of phantom size, slice thickness, dose, spongiosa region segmentation(1,4(2x2),9(3x3)) and calibration method on the accuracy of the calculated red marrow volume fractions (cellularity) were evaluated. The method was also applied to six patients.
Results:For both the small and medium sized phantoms using internal calibration, RMS errors(RMSEs) in the estimated volume fractions of red marrow were 0.03, 0.04 and 0.05 for partitioning of a 23.4mmx23.4mm ROI into 1 region, 4 regions and 9 regions, respectively. Corresponding results for a large sized body phantom were significantly poorer (0.06,0.09,0.10). Reducing slice thickness to 2.5mm(vs. 5mm) increased RMSEs of the marrow volume fractions by 0.01 to 0.06. Similar increases in RMSEs were obtained for reduction in dose(mAs). Results for external calibrations exhibited large errors that were reduced to those of internal calibration by application of external to internal correction factors. Application of these factors to the 6 patient lumbar vertebrae yielded cellularity values that differed by 29% on average from biopsy values at the iliac crest, potentially due to regional variability.
Conclusion:Phantom studies were promising, but show poorer results for large bodies, thinner slices, lower doses and smaller sub-regions. Future validation studies in human cadaver vertebrae are needed.
Funding Support, Disclosures, and Conflict of Interest: National Institutes of Health (NIH) under Grant 2RO1 EB001994