Abstract
Background
Serial measurement of left ventricular ejection fraction (LVEF) is used to monitor patients undergoing potentially cardiotoxic chemotherapy. MRI is the gold standard, but due to cost and availability, nuclear medicine (MUGA) scans are often preferred. CT is cheap, fast and correlates well with MRI, but has not been adopted due to concerns over radiation exposure. Therefore, we investigated how much CT radiation dose can be lowered while preserving LVEF accuracy,
Methods
Cardiac CT images of the full cardiac cycle were acquired at high dose. End-diastolic and -systolic phases were segmented and converted to STL-format. 3D-printed models were produced in myocardium attenuation-equivalent material, filled with contrast media, and placed in an anthropomorphic chest phantom. Scanning was performed at different mAs, kVp levels and slice thicknesses with dose recorded. Ventricular volumes and LVEF was calculated. 40 clinical scans were used for simulated dose reduction using MATLAB. Bland-Altman analysis of LVEF with full-dose scans was performed. Correlation between simulated and actual dose reduction was examined.
Results (preliminary)
Phantom and simulation scans were readable for semiautomatic measurement on clinical workstations. Simulated and phantom scan image noise showed high correlation with preserved Hounsfield values. At CNR values down to 5, LV volume varied less than +/- 5% compared to full-dose scans. This was achievable at a dose of about 50 mGycm (1 mSv) using low kVp’s and slice thickness ≥1 mm (MUGA dose 4-8 mSv).
Conclusion
Simulation and 3D-printed phantom scans is a feasible way of exploring dose saving potential in functional cardiac CT.
Serial measurement of left ventricular ejection fraction (LVEF) is used to monitor patients undergoing potentially cardiotoxic chemotherapy. MRI is the gold standard, but due to cost and availability, nuclear medicine (MUGA) scans are often preferred. CT is cheap, fast and correlates well with MRI, but has not been adopted due to concerns over radiation exposure. Therefore, we investigated how much CT radiation dose can be lowered while preserving LVEF accuracy,
Methods
Cardiac CT images of the full cardiac cycle were acquired at high dose. End-diastolic and -systolic phases were segmented and converted to STL-format. 3D-printed models were produced in myocardium attenuation-equivalent material, filled with contrast media, and placed in an anthropomorphic chest phantom. Scanning was performed at different mAs, kVp levels and slice thicknesses with dose recorded. Ventricular volumes and LVEF was calculated. 40 clinical scans were used for simulated dose reduction using MATLAB. Bland-Altman analysis of LVEF with full-dose scans was performed. Correlation between simulated and actual dose reduction was examined.
Results (preliminary)
Phantom and simulation scans were readable for semiautomatic measurement on clinical workstations. Simulated and phantom scan image noise showed high correlation with preserved Hounsfield values. At CNR values down to 5, LV volume varied less than +/- 5% compared to full-dose scans. This was achievable at a dose of about 50 mGycm (1 mSv) using low kVp’s and slice thickness ≥1 mm (MUGA dose 4-8 mSv).
Conclusion
Simulation and 3D-printed phantom scans is a feasible way of exploring dose saving potential in functional cardiac CT.
| Original language | English |
|---|---|
| Publication date | 21. Aug 2021 |
| Publication status | Published - 21. Aug 2021 |
| Event | International society of radiographers and radiological technologists : Dublin 2021 - Convention Centre Dublin, Dublin, Ireland Duration: 20. Aug 2021 → 22. Aug 2021 |
Conference
| Conference | International society of radiographers and radiological technologists |
|---|---|
| Location | Convention Centre Dublin |
| Country/Territory | Ireland |
| City | Dublin |
| Period | 20/08/2021 → 22/08/2021 |