Guild Jeffrey, Takacs Elizabeth, Kircher Stephen, Arbique Gary, Zimmern Philippe E
UT Southwestern Medical Center, Dallas, TX, USA.
Neurourol Urodyn. 2009;28(5):385-9. doi: 10.1002/nau.20678.
Based on a prior study, over 60% of the average radiation dose per cystourethrography (CU) originated from lateral radiograph images (RI). This analysis focuses on the feasibility of decreasing the number of RI without compromising study interpretation.
Following IRB approval, all RIs from the initial study in women with incontinence and/or bladder prolapse were assessed at random for reading consistency (inter-rater reliability (IRR)) and reliability (intra-rater reliability (IaRR)) by four independent reviewers. Interpretation guidelines on urethral readability were established and 15 repeated RIs tested for IaRR. Two months later, the reviewers selected the best two RIs for each CU with the RI presented in a random order.
From 88 CU exams providing 304 RIs, good IRRs for readability (kappa=0.82, ICC=0.80) and interpretation (kappa=0.60, ICC=0.57) were found between reviewers. Mean IaRR for readable versus unreadable categories was 87% and among readable images was 95% for interpretation. The RIs judged best had a high IRR (kappa=0.83, ICC=0.83) and were predominantly from the early and middle phases of the void.
A high level of agreement was found for urethral readability and interpretation between reviewers of varying experience using pre-established guidelines. No more than 3 RIs were required for study interpretation and those views were from the early to middle phase of voiding, thus allowing a radiation dose reduction without compromising the quality of the CU exam. Neurourol. Urodynam. 28:385-389, 2009. (c) 2008 Wiley-Liss, Inc.
