Evaluation of the potential for radiation-induced skin damage during interventional neuroradiological procedures using automated dosimetry

Data on the entry skin exposure from 522 interventional neuroradiology procedures have been gathered using an automated dosimetry system (ADS). Five classes of procedures are studied: carotid stenosis procedures (94 cases), cerebral angiograms (160 cases), embolizations (94 cases), nerve blocks/injections (105 cases) and lumbar procedures (101 cases). The data from each procedure is converted to entry skin dose and compared to the recommended thresholds for monitoring from the FDA/CDRH of 1 Gy and to thresholds for radiation induced skin damage presented in the literature. The effects of movement of the radiation field outside of the primary site, of overlap of fields within the primary site, and of overlap of the posterior-anterior and lateral fields as modifying factors for estimates are determined. Recommended criteria for not monitoring a class of procedure; namely that the 95th percentile of the cases in a given class be below the threshold for any skin damage (2 Gy) and that no cases be recorded that cause permanent damage (6 Gy) are presented. Embolization procedures and cerebral angiograms require monitoring under these criteria. A model for estimating the probability for doses leading to skin effect for a given class is derived from the modified ADS estimates and is found to follow an exponential distribution of the form y(x)=a•e −a•x, where a is a constant for a given class and view and y is the probability of a case having an ESD greater than x. The values of a are presented for both modified and unmodified ADS estimates. Cumulative expressions of this distribution are presented. The hypothesis that knowledge of the fluoroscopy time and the number of digital frames can be used to predict ESD for a given class is then examined. While this information is not useful in individual cases because of the high coefficient of variation, cumulative use of the ESE estimates of the ADS allow the estimation of the probability that a procedure is less than a given ESD from a combination of fluoroscopy time and number of frames. This is a significant improvement in implementing the analysis of risk for a procedure.