Low-dose CT urography using deep learning image reconstruction: a prospective study for comparison with conventional...

Area of Science:

• Radiology
• Medical Imaging
• Artificial Intelligence in Medicine

Background:

• Conventional CT urography (C-CTU) involves significant radiation exposure.
• Deep learning image reconstruction (DLIR) offers potential for dose reduction in CT imaging.
• Adaptive statistical iterative reconstruction with V (ASIR-V) is a current standard for CT image quality enhancement.

Purpose of the Study:

• To compare the image quality of low-dose CT urography (LD-CTU) utilizing DLIR against C-CTU employing ASIR-V.
• To evaluate radiation dose reduction achieved with DLIR in CT urography.
• To assess image quality parameters including artifacts, noise, and sharpness.

Main Methods:

• A prospective, single-institutional study comparing LD-CTU with DLIR (100kV, NI 23) against C-CTU with ASIR-V (100kV, NI 10).
• Two radiologists independently assessed image quality, artifacts, noise, and sharpness.
• Quantitative analysis included mean CT attenuation, signal-to-noise ratio (SNR), and contrast-to-noise ratio (CNR).

Main Results:

• The LD-CTU with DLIR group received a significantly lower effective radiation dose (2.01 mSv) compared to the C-ASIR-V group (6.9 mSv) (p < 0.001).
• Overall image quality in the LD-DLIR group was rated as good and comparable to the C-ASIR-V group.
• Both groups demonstrated adequate and similar attenuation values, SNR, and CNR in most urinary tract segments.

Conclusions:

• DLIR enables a 71% reduction in radiation dose for CT urography while maintaining diagnostic image quality comparable to conventional methods.
• This feasibility study demonstrates the potential of DLIR to improve patient safety by lowering radiation exposure in CT urography.
• LD-CTU with DLIR is a viable alternative to C-CTU with ASIR-V, offering significant dose savings.