Purpose: To investigate the radiation dose reduction and image quality of ultra-high-resolution (UHR) temporal bone imaging with spectral shaping on a research photon-counting-detector (PCD) CT system.
Methods: A head phantom and a fresh cadaveric head were scanned on a 128-slice commercial dual source CT system (Definition Flash, Siemens Healthcare) using clinical temporal bone protocols (120 kV, UHR). The same phantom and cadaveric head were then scanned on a PCD-CT system, using 100 kV with additional 0.4 mm tin filtration (Sn100 kV), UHR mode and CTDIvol of 10.1 mGy. With IRB approval, a temporal bone patient was recruited after his clinical CT exam and was then scanned on the PCD-CT using UHR mode, Sn100 kV at 10.1 mGy. All image reconstructions followed on-site clinical protocols and were matched across both systems. Image noise was measured as the standard deviation of CT number from a uniform soft tissue region.
Results: For head phantom, image noise of clinical CT (140 HU, 49.0 mGy) was higher than that of low dose PCD-CT scans (127 HU, 10.1 mGy). For the cadaveric head, image noise of clinical CT (136 HU, 48.0 mGy) was close to that of low dose PCD scans (132 HU, 10.1 mGy). For the in vivo patient, image noise of clinical CT (143 HU, 65.7 mGy) was also comparable to that of low dose PCD scans (147 HU, 10.1 mGy). Phantom, ex vivo cadaver and in vivo patient results demonstrated an overall 80~84% dose reduction using PCD Sn100 kV to achieve the same noise level as the clinical temporal bone scans.
Conclusion: The proposed method using additional tin filtration and lower kV on a PCD-CT system could achieve dose reduction by a factor of 5-6 for temporal bone imaging while maintaining diagnostic image quality, compared to the state of art clinical CT system.