Prospective Estimation of Diagnostic Performance and Radiation Dose for Individual CT Scans
x tian1*, C Smitherman2, O Christianson3, D Frush4, E Samei5, (1) ,,,(2) Duke University, Durham, NORTH CAROLINA, (3) ,Durham, NC, (4) Duke University Medical Center, Durham, NC, (5) Duke University Medical Center, Durham, NCTU-C-103-7 Tuesday 10:30AM - 12:30PM Room: 103
Purpose: To prospectively determine size- and task-based image quality indices and to quantify the relationship between image quality and radiation dose as a patient-specific basis for abdominopelvic CT scans.
Methods:Model observer metric of detectability index was derived as a surrogate of task-based diagnostic accuracy. The detectability index incorporated the noise (NPS), resolution (MTF), and the exam indication (task function). The noise (NPS) was characterized as a function of patient size and scanner characteristics (kVp, mAs, slice thickness, and reconstruction algorithm). The resolution (MTF) was modeled as task-based MTF, a function of contrast and noise-level corresponding to the clinical feature of interest. The NPS and MTF were measured using a variable-size, multi-inserts phantom on two CT scanner models (GE HD 750 and Siemens Definition Flash). To assess the radiation burden, patient organ dose, effective dose, and a risk index were estimated based on a validated Monte Carlo simulation program and a library of 100 computational XCAT phantoms. The relationship between radiation dose and detectability was determined to assess the dose efficiency specific for each patient and each exam.
Results:Detectability index for a 6 mm liver lesion were derived for patients with sizes range from 16 cm to 40 cm. The relationship between detectability index and patient-specific dose were determined for individual XCAT models.
Conclusion:Prospectively determining patient- and task-based image quality enables optimized diagnosis performance with the minimum amount of radiation dose. The quantitative relationship between image quality and radiation dose can be the basis of dose/quality optimization by enabling the operator to select an operating point for a particular exam before it is initiated.