Development of a thresholding algorithm for calcium classification at multiple CT energies

10th International Seminar on Medical Physics; 1st AMDI International Oncology Symposium

物理学;医药卫生

Ng, L.Y.^1 ; Alssabbagh, M.^1 ; Tajuddin, A.A.^1,2 ; Shuaib, I.L.^1 ; Zainon, R.^1

Advanced Medical and Dental Institute, Universiti Sains Malaysia, Kepala Batas, Pulau Pinang

13200, Malaysia^1

School of Physics, Universiti Sains Malaysia, Pulau Pinang

11800, Malaysia^2

关键词: Calcium concentration;    Chloride solutions;    Iron nitrates;    Low concentrations;    National institute of healths;    Slice thickness;    Thresholding algorithms;    Thresholding methods;   

Others  :  https://iopscience.iop.org/article/10.1088/1742-6596/851/1/012010/pdf
DOI  :  10.1088/1742-6596/851/1/012010

学科分类：卫生学

来源: IOP

【 摘 要 】

The objective of this study was to develop a thresholding method for calcium classification with different concentration using single-energy computed tomography. Five different concentrations of calcium chloride were filled in PMMA tubes and placed inside a water-filled PMMA phantom (diameter 10 cm). The phantom was scanned at 70, 80, 100, 120 and 140 kV using a SECT. CARE DOSE 4D was used and the slice thickness was set to 1 mm for all energies. ImageJ software inspired by the National Institute of Health (NIH) was used to measure the CT numbers for each calcium concentration from the CT images. The results were compared with a developed algorithm for verification. The percentage differences between the measured CT numbers obtained from the developed algorithm and the ImageJ show similar results. The multi-thresholding algorithm was found to be able to distinguish different concentrations of calcium chloride. However, it was unable to detect low concentrations of calcium chloride and iron (III) nitrate with CT numbers between 25 HU and 65 HU. The developed thresholding method used in this study may help to differentiate between calcium plaques and other types of plaques in blood vessels as it is proven to have a good ability to detect the high concentration of the calcium chloride. However, the algorithm needs to be improved to solve the limitations of detecting calcium chloride solution which has a similar CT number with iron (III) nitrate solution.