PURPOSE: To prospectively compare detection and reader confidence of pancreatic lesions using a standard multi-detector computed tomography (MDCT) imaging protocol to a dual-energy computed tomography (DECT) imaging protocol with additional virtual non-contrast series. METHODS: 60 subjects imaged for suspected or known pancreatic lesions were included. Subjects underwent pancreatic MDCT including non-contrast, pancreatic phase, and portal venous phase (PVP). The PVP was performed in dual energy mode. Virtual non-contrast and blended 120 kVp weighted images were created from the DECT data. Overall noise and absolute attenuation differences of pancreatic lesions and normal pancreatic tissue were measured. Images were read by two staff radiologists blinded to the underlying diagnosis. MDCT and DECT scans were reviewed separately to evaluate image quality and level of confidence in diagnosis of a pancreatic lesion. RESULTS: Image quality was ranked excellent for 90 % and 95 % of the 120 kVp studies and 93 % and 95 % of the 100 kVp studies by readers 1 and 2, respectively. VNC was ranked sufficient quality or better by both readers. Average attenuation difference was 74 HU (120 kVp) and 71 HU (100 kVp). Average noise was 11.31 HU (120 kVp) and 15.89 HU (100 kVp). No lesions were missed by either approach. There was increased confidence in diagnostic interpretation in 14 % (± 9 % [95 % CI]) and 9 % (± 7 % [95 % CI]) of DECT scans compared to MDCT. CONCLUSIONS: DECT compared to MDCT pancreatic imaging leads to increased reader confidence with identical diagnostic sensitivity for pathologically proven cases. This approach could be implemented as a single phase acquisition study with calculated VNC leading to a significant dose savings to the patient.
PURPOSE: To prospectively compare detection and reader confidence of pancreatic lesions using a standard multi-detector computed tomography (MDCT) imaging protocol to a dual-energy computed tomography (DECT) imaging protocol with additional virtual non-contrast series. METHODS: 60 subjects imaged for suspected or known pancreatic lesions were included. Subjects underwent pancreatic MDCT including non-contrast, pancreatic phase, and portal venous phase (PVP). The PVP was performed in dual energy mode. Virtual non-contrast and blended 120 kVp weighted images were created from the DECT data. Overall noise and absolute attenuation differences of pancreatic lesions and normal pancreatic tissue were measured. Images were read by two staff radiologists blinded to the underlying diagnosis. MDCT and DECT scans were reviewed separately to evaluate image quality and level of confidence in diagnosis of a pancreatic lesion. RESULTS: Image quality was ranked excellent for 90 % and 95 % of the 120 kVp studies and 93 % and 95 % of the 100 kVp studies by readers 1 and 2, respectively. VNC was ranked sufficient quality or better by both readers. Average attenuation difference was 74 HU (120 kVp) and 71 HU (100 kVp). Average noise was 11.31 HU (120 kVp) and 15.89 HU (100 kVp). No lesions were missed by either approach. There was increased confidence in diagnostic interpretation in 14 % (± 9 % [95 % CI]) and 9 % (± 7 % [95 % CI]) of DECT scans compared to MDCT. CONCLUSIONS: DECT compared to MDCT pancreatic imaging leads to increased reader confidence with identical diagnostic sensitivity for pathologically proven cases. This approach could be implemented as a single phase acquisition study with calculated VNC leading to a significant dose savings to the patient.
Authors: Naveen M Kulkarni; Lorenzo Mannelli; Marc Zins; Priya R Bhosale; Hina Arif-Tiwari; Olga R Brook; Elizabeth M Hecht; Fay Kastrinos; Zhen Jane Wang; Erik V Soloff; Parag P Tolat; Guillermo Sangster; Jason Fleming; Eric P Tamm; Avinash R Kambadakone Journal: Abdom Radiol (NY) Date: 2020-03
Authors: Elizabeth George; Jeremy R Wortman; Urvi P Fulwadhva; Jennifer W Uyeda; Aaron D Sodickson Journal: Br J Radiol Date: 2017-09-22 Impact factor: 3.039