1. Evaluation of spatial resolution on the longitudinal (z-axis)

  Fig. 7 shows SSPz and MTFon the z-axis. FWHM and FWTM for 0.5 mm×64 DAS were shorter than those for 1.0 mm×32 DAS. MTF in 0.5 cycles/mm for 0.5 mm×64 DAS was shorter than for 1.0 mm×32 DAS. These results show that spatial resolution on the z-axis for 0.5 mm×64 DAS is superior to that for 1.0 mm×32 DAS.

Fig. 7: FWHM and FWTM for 0.5 mm×64 DAS were shorter than those for 1.0 mm×32 DAS. MTF in 0.5 cycles/mm for 0.5 mm×64 DAS was shorter than for 1.0 mm×32 DAS.

2. Evaluation of image noise and streak artifacts

  Fig. 8 shows SD in Gaussian distribution and location parameter in Gumbel distribution for CT images obtained with tube current of 15, 20, 30 mA and the reconstruction algorithm of FBP, reconstruction kernel of FC11 and slice thickness of 1 mm. This result shows that SD and the location parameter increased depending on the reduction in mAs.

Fig. 8: SD and location parameter for CT images obtained with tube current of 15, 20, 30 mA.

  Fig. 9 shows SD in Gaussian distribution and location parameter in Gumbel distribution for CT images obtained with various reconstruction algorithm and reconstruction kernel and tube current of 30 mA and slice thickness of 1 mm. This result shows that SD and the location parameter for images obtained with AIDR3D were less than half of those with FBP. SD and the location parameter also decreased in the order of AIDR3D_weak, AIDR3D_mild, AIDR3D_standard and decreased with smoother reconstruction kernel.

Fig. 9: SD and location parameter for CT images obtained with various reconstruction algorithm and reconstruction kernel.

  Fig.10 shows SD in Gaussian distribution and location parameter in Gumbel distribution for CT images obtained with slice thickness of 1 mm and 5 mm, tube current of 30 mA, the reconstruction algorithm of FBP and reconstruction kernel of FC11. This result shows that SD and the location parameter for images obtained with slice thickness of 1mm were less than half of those in slice thickness of 5 mm.

Fig. 10: SD and location parameter for CT images obtained with slice thickness of 1 mm and 5 mm.

  Fig. 11 shows scan and reconstruction conditions of low-dose lung cancer CT screening used in another hospital in Japan and also show SD and location parameter in the reference condition. Fig. 11 also shows various combinations of tube current, reconstruction algorithm and kernel and slice thickness of 1 mm, which provides almost same SD and location parameters as those in the reference condition. The reason for slice thickness of 1 mm for low-dose lung cancer CT screening in this hospital is that a computer aided diagnosis (CAD) system will be used for CT screening and CT images of slice thickness of 1 mm will be used for the system.

Fig. 11: Lung cancer CT screening conditions providing almost same SD and location parameters as those with the reference condition.

3. Radiation dose

  CTDI_vol displayed on CT console for lung cancer CT screening were 0.5-1.1 mGy in tube current of 15-30 mA. Organ doses in 15-30 mA are summarized in Fig. 12. Doses for thyroid, lung, esophagus and breast in 30 mA were 1.1–2.3 mGy. Effective doses in 15-30 mA were estimated to be 0.4-0.9 mSv.

Fig. 12: Organ doses for lung cancer CT screening in tube current of 15-30 mA.