Can we improve accuracy and reliability of MRI interpretation in children with optic pathway glioma? Proposal for a reproducible imaging classification
| Autor: | Xavier Rialland, Josué Rakotonjanahary, Julien Lambron, Emilie De Carli, Eric Frampas, Didier Loisel, Frédérique Toulgoat, Matthieu Delion |
|---|---|
| Rok vydání: | 2015 |
| Předmět: |
Male
Optic Nerve Glioma medicine.medical_specialty Sensitivity and Specificity 030218 nuclear medicine & medical imaging Interpretation (model theory) 03 medical and health sciences 0302 clinical medicine Cohen's kappa Image Interpretation Computer-Assisted Medicine Humans Radiology Nuclear Medicine and imaging Medical physics Reliability (statistics) Neuroradiology Observer Variation medicine.diagnostic_test business.industry Reproducibility of Results Magnetic resonance imaging medicine.disease Image Enhancement Magnetic Resonance Imaging Pediatric Radiology Child Preschool Female Neurology (clinical) Optic nerve glioma Cardiology and Cardiovascular Medicine business Optic pathway glioma Nuclear medicine 030217 neurology & neurosurgery Algorithms |
| Zdroj: | Neuroradiology. 58(2) |
| ISSN: | 1432-1920 |
| Popis: | Magnetic resonance (MR) images from children with optic pathway glioma (OPG) are complex. We initiated this study to evaluate the accuracy of MR imaging (MRI) interpretation and to propose a simple and reproducible imaging classification for MRI. We randomly selected 140 MRIs from among 510 MRIs performed on 104 children diagnosed with OPG in France from 1990 to 2004. These images were reviewed independently by three radiologists (F.T., 15 years of experience in neuroradiology; D.L., 25 years of experience in pediatric radiology; and J.L., 3 years of experience in radiology) using a classification derived from the Dodge and modified Dodge classifications. Intra- and interobserver reliabilities were assessed using the Bland–Altman method and the kappa coefficient. These reviews allowed the definition of reliable criteria for MRI interpretation. The reviews showed intraobserver variability and large discrepancies among the three radiologists (kappa coefficient varying from 0.11 to 1). These variabilities were too large for the interpretation to be considered reproducible over time or among observers. A consensual analysis, taking into account all observed variabilities, allowed the development of a definitive interpretation protocol. Using this revised protocol, we observed consistent intra- and interobserver results (kappa coefficient varying from 0.56 to 1). The mean interobserver difference for the solid portion of the tumor with contrast enhancement was 0.8 cm3 (limits of agreement = −16 to 17). We propose simple and precise rules for improving the accuracy and reliability of MRI interpretation for children with OPG. Further studies will be necessary to investigate the possible prognostic value of this approach. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|
Full text from SpringerLink