Room 201
Purpose: To determine a methodology to systematically implement water equivalent diameter (WED) based patient dose thresholds for adult abdomen and pelvis CT scans in an enterprise dose tracking software (Radimetrics) to monitor individual case dose appropriateness.
Methods: Data was gathered and analyzed, with Python, from a large hospital enterprise and compared to the ACR Dose Index Registry (DIR) published achievable dose (AD) and diagnostic reference levels (DRL) for abdomen and pelvis with contrast protocols. After confirming an agreement with these statistical values, enterprise data was split into the WED-based size groups as set by the ACR DIR. Statistical analysis was done to determine the percentile that would produce an equivalent number of alerts as the previous methodology. The cases that caused alerts from the previous threshold were subjected to the size specific thresholds to eliminate alerts with appropriate dose with respect to patient size.
Results: When compared to the ACR DIR AD and DRL our enterprise data had a percent error of -3.8% and -7.9% respectively. Setting the new threshold at the size group specific 97th percentile CTDIvol, produced a comparable total number of alerts, 7% fewer alerts, with the previous single threshold but had a distribution that was more reflective of the shape of the whole data set. Analyzing the current alerts with the new thresholds showed that 74.8% of the current alerts would be eliminated.
Conclusion: Setting the alert threshold to the 97th percentile CTDIvol value for each size group will result in a much more effective alert notification system across the hospital enterprise. This WED based alert methodology will better address patients with lower WED and allow for more meaningful utilization of the alerts. The Python based application developed to determine the size based alert thresholds can seamlessly be applied to other adult and pediatric protocols.