Click here to


Are you sure ?

Yes, do it No, cancel

Minimum Detectable Contrast, Dose, Lesion Size and Adaptive Statistical Iterative Reconstruction (ASIR) Blending Fraction- a Quantitative Model

Y Zhou*, A Scott , C Kim , J Nute , C Lee , Cedars-Sinai Medical Center, Los Angeles, CA


(Tuesday, 7/31/2018) 3:45 PM - 4:15 PM

Room: Exhibit Hall | Forum 8

Purpose: To make task specific dose optimization possible, we developed a quantitative form relating the minimum detectable contrast (MDC) to the size specific dose, lesion size and adaptive statistical iterative reconstruction (ASIR) blending fraction.

Methods: The MDC was statistically defined as the separation of the lesion foreground from the background in noise distribution with a confidence level of 95%. The functional dependence of the MDC on dose, lesion size and ASIR blending fraction was obtained using three steps. First, with helical scans on the adult abdomen phantom using a GE 750 HD, the MDC was fitted against the ASIR blending fraction (10 –100%) at eleven dose levels (size specific dose: 2 –38 mGy) and six lesion sizes (2- 10 mm). Second, the resultant fit parameters were themselves fitted against the dose for each lesion size. Finally, the second group of fit parameters were fitted against the lesion size.

Results: The MDC versus ASIR blending fraction data was found to fit a linear equation (R^2 > 0.984) with the intercept and slope varying with the dose and lesion size. For each lesion size, the intercept and slope were fitted against the dose with a power law form (R^2 > 0.962 and R^2>0.905, respectively). The power law proportionality and index were further fitted against the lesion size by another power law form and a linear equation, respectively (R^2 > 0.998 and R^2 > 0.922). The final equation for MDC was compared with the MDC data. The standard error of estimate, defined as the root mean square of the percentage errors over all the data points, was found to be smaller than 6.8%.

Conclusion: The quantitative form relating the minimum detectable contrast to the dose, lesion size and ASIR blending was obtained. It can be of help for task specific dose optimization.


Not Applicable / None Entered.


Not Applicable / None Entered.

Contact Email