A Method for Deriving Exam-Specific Target Exposure Indices (TEI) in Computed Radiography as a Function of a Reference TEI
K Hulme*, V Kubicki, F Dong, W Davros, P Johnson, The Cleveland Clinic, Cleveland, OHTU-A-116-4 Tuesday 8:00AM - 9:55AM Room: 116
Determining appropriate target exposure index (TEI) values is one of the major barriers in the clinical efficacy of the IEC deviation index (DI). The goal of this work was to isolate the impact of exam group on TEI and thus derive TEI-subgroups for Agfa CR that would enable target values to be more easily established from a single reference value.
Data were exported from 13 Agfa NX workstations (software versions 8600 and 8700SU1) at 8 sites affiliated with Shimadzu RadSpeed radiographic rooms (generator UD150V-40) using Agfa CR. XML files were imported into MATLAB and 147,619 dose statistics were analyzed. The DI for TEI₍ref₎=400 (DI₍ref₎ = 10*log10(EI/TEI₍ref₎)) was calculated for each data point. The distribution in DI₍ref₎ was normal for each exam group. Variance was independent of TEI₍ref₎. The mean -DI₍ref,k,i₎- and standard deviation -SD(DI₍ref,k,i₎)- were calculated for each workstation i and exam group k. DI₍ref,k₎ was determined from the average of DI₍ref,k,i₎ weighted by number of exams acquired for that exam group on each workstation. A bin width of ΔDI₍ref₎=1 was used to create TEI subgroups.
Exam groups were binned into seven subgroups centered on DI₍ref,k₎ =[-2,-1,0,1,2,3,4] corresponding to linear scaling factors K₍ref,k₎=[0.6,0.8,1.0,1.3,1.6,2.0,2.5]. Standard error of the mean was <0.1 for all DI₍ref,k₎ except one, which would have allowed for finer sub-group binning. SD(DI₍ref,k₎) ranged from 0.75 to 2.68, and was not correlated with DI₍ref,k₎ (r²=0.003).
Little guidance exists for the establishment of TEI tables as function of anatomy and view. The described methodology demonstrates how TEI values can be derived from a single reference value using measured scaling factors (TEI₍k₎=TEIref*K₍ref,k₎). Seven TEI subgroups were established for Agfa CR. Vendor-specific volume of interest algorithms and detector type will likely affect measured scaling factors. Determination of an appropriate TEI₍ref₎ requires further evaluation.