The Reliability of a Grading System for Digital Subtraction Imaging Quality During Cervical Transforaminal Epidural Steroid Injection

Objective. Digital subtraction imaging (DSI) decreases the risk of intravascular injection during cervical transforaminal epidural steroid injection (CTFESI); however, sequence acquisition and interpretation are operator-dependent skills. This study tests the reliability of a grading system to determine adequate DSI during CTFESI. Setting. Academic tertiary medical center. Methods. A grading scheme for adequate DSI quality during CTFESI was created by the study authors based on patient positioning, mask image, and volume of contrast injected. The inter-rater and intrarater reliability values of this grading scheme were tested using 50 DSI images evaluated by three raters during two distinct sessions separated by four weeks. Based on a power analysis, a sample of 50 scans was sufficient to detect significant correlations. Inter-rater reliability was determined by percent agreement between graders for dichotomized categories of "quality of DSI is adequate for safe C-TFESI" vs "quality of DSI is inadequate for safe C-TFESI." The percentage of agreement was reported, along with Gwet's agreement coefficient (AC). The intrarater (pre/post) correlation was assessed using Yule's Q statistics. Results. Correlation coefficients were interpreted as follows: 0.00-0.19 "very weak,"0.20-0.39 "weak," 0.40-0.59 "moderate," 0.60-0.79 "strong," and 0.80-1.00 "very strong." Interrater reliability analyses demonstrated that the patient position category had "very strong" agreement, contrast volume had "strong" agreement, and mask image had "moderate" agreement. The overall inter-rater reliability was "moderate." All of the raters demonstrated "very strong" intrarater reliability. Conclusions. The proposed grading system for adequate-quality DSI during CTFESI showed overall "moderate" and "very strong" inter- and intrarater reliability, respectively. This scheme provides an objective measure of DSI quality for CTFESI. Refinement is needed to improve the reliability of this scheme.