Iterative reconstruction algorithms of computed tomography for the assessment of small pancreatic lesions: Phantom study

摘要

OBJECTIVE: To evaluate the image quality and radiation dose reduction of iterative reconstruction (IR) used for computed tomographic (CT) scanning of small pancreatic lesions. METHODS: An anthropomorphic pancreas phantom with 16 small lesions was scanned using 4 kinds of CT scanners with different tube current-time products (75-250 mAs). The CT images were reconstructed using filtered back projection (FBP) and the relevant IR of each vendor (GE Healthcare, Philips Healthcare, Siemens Healthcare, Toshiba Medical Systems). The image qualities, dose reduction rate (in percent), and figure of merit (FOM) were evaluated in comparison with the reference images (250 mAs, FBP). RESULTS: Image noise was markedly improved with the IR; therefore, a 36 to 60% dose reduction was possible. As a result, the final CT dose index volume can be diminished to 7.05 to 11.40 mGy with the IR algorithms. The IR demonstrated 1.52 to 7.84 times higher FOM than that of FBP. Particularly, an advanced fully IR showed outstanding results of FOM (6.06-7.84 times). CONCLUSIONS: Because IR can reduce image noise while maintaining image quality for the delineation of small pancreatic lesions, it can be used for pancreatic imaging with substantial radiation dose reduction.