Evaluation of digital radiography practice using exposure index tracking

Abstract

Some digital radiography (DR) detectors and software allow for remote download of exam statistics, including image reject status, body part, projection, and exposure index (EI). The ability to have automated data collection from multiple DR units is conducive to a quality control (QC) program monitoring institutional radiographic exposures. We have implemented such a QC program with the goal to identify outliers in machine radiation output and opportunities for improvement in radiation dose levels. We studied the QC records of four digital detectors in greater detail on a monthly basis for one year. Although individual patient entrance skin exposure varied, the radiation dose levels to the detectors were made to be consistent via phototimer recalibration. The exposure data stored on each digital detector were periodically downloaded in a spreadsheet format for analysis. EI median and stan-dard deviation were calculated for each protocol (by body part) and EI histograms were created for torso protocols. When histograms of EI values for different units were compared, we observed differences up to 400 in average EI (representing 60% difference in radiation levels to the detector) between units nominally cali-brated to the same EI. We identified distinct components of the EI distributions, which in some cases, had mean EI values 300 apart. Peaks were observed at the current calibrated EI, a previously calibrated EI, and an EI representing computed radiography (CR) techniques. Our findings in this ongoing project have allowed us to make useful interventions, from emphasizing the use of phototimers instead of institutional memory of manual techniques to improvements in our phototimer calibration. We believe that this QC program can be implemented at other sites and can reveal problems with radiation levels in the aggregate that are difficult to identify on a case-by-case basis.

Figure 1

EI distributions for abdomen studies for units A1, A2, B1, and B2 for datasets 2014 (left) and 2015 (right). Units A1 and B1 are the integrated units while A2 and B2 are the retrofit units, from departments A and B, respectively. Distributions are normalized to the number of exams so that the area under each curve is unity. The 2014 dataset is prior to the phototimer recalibration and the 2015 dataset is after.

Figure 2

EI distributions for chest studies for units A1, A2, B1, and B2 for datasets 2014 (left) and 2015 (right). Units A1 and B1 are the integrated units while A2 and B2 are the retrofit units from departments A and B, respectively. Distributions are normalized to the number of exams so that the area under each curve is unity. The 2014 dataset is prior to the phototimer recalibration and the 2015 dataset is after.

Figure 3

Before‐and‐after EI comparisons for one X‐ray unit. Distribution of EI values for abdomen protocol exams (left) and chest protocols (right) for unit A2 (department A, retrofit Carestream). Both 2014 and 2015 distributions are normalized to the number of exams so that the area under each curve is unity. The phototimer recalibration between datasets 2014 and 2015 has been ineffective in shifting the EI distribution from an EI target of 1600 to a target EI of 1400 in both cases.

Figure 4

Timeline of EI targets for retrofit units. The 2014 dataset comes from the target $\text{EI}=1600$, while the 2015 dataset comes from target $\text{EI}=1400$.

Figure 5

Evidence of lowered radiation dose after intervention. Average EI value per month for abdomen protocol exams for units A1 and A2 (integrated and retrofit units, respectively). The QC team worked with a lead technologist in April to start tracking of EI values by technologists, and an in‐service was delivered in June for lowering the average EI. The EI control is 1400 and is drawn in a dashed line.