BACKGROUND: The Chicago Classification (CC) facilitates interpretation of high-resolution manometry (HRM) recordings. Application of this adult based algorithm to the pediatric population is unknown. We therefore assessed intra and interrater reliability of software-based CC diagnosis in a pediatric cohort.
METHODS: Thirty pediatric solid state HRM recordings (13M; mean age 12.1 ± 5.1 years) assessing 10 liquid swallows per patient were analyzed twice by 11 raters (six experts, five non-experts). Software-placed anatomical landmarks required manual adjustment or removal. Integrated relaxation pressure (IRP4s), distal contractile integral (DCI), contractile front velocity (CFV), distal latency (DL) and break size (BS), and an overall CC diagnosis were software-generated. In addition, raters provided their subjective CC diagnosis. Reliability was calculated with Cohen's and Fleiss' kappa (κ) and intraclass correlation coefficient (ICC).
KEY RESULTS: Intra- and interrater reliability of software-generated CC diagnosis after manual adjustment of landmarks was substantial (mean κ = 0.69 and 0.77 respectively) and moderate-substantial for subjective CC diagnosis (mean κ = 0.70 and 0.58 respectively). Reliability of both software-generated and subjective diagnosis of normal motility was high (κ = 0.81 and κ = 0.79). Intra- and interrater reliability were excellent for IRP4s, DCI, and BS. Experts had higher interrater reliability than non-experts for DL (ICC = 0.65 vs ICC = 0.36 respectively) and the software-generated diagnosis diffuse esophageal spasm (DES, κ = 0.64 vs κ = 0.30). Among experts, the reliability for the subjective diagnosis of achalasia and esophageal gastric junction outflow obstruction was moderate-substantial (κ = 0.45-0.82).
CONCLUSIONS & INFERENCES: Inter- and intrarater reliability of software-based CC diagnosis of pediatric HRM recordings was high overall. However, experience was a factor influencing the diagnosis of some motility disorders, particularly DES and achalasia.