Inter-observer variability in the delineation of pharyngo-laryngeal tumor, parotid glands and cervical spinal cord: comparison between CT-scan and MRI

BACKGROUND AND PURPOSE: The goal of this study was to compare the inter-observer variability between CT and MRI for the delineation of pharyngo-laryngeal SCC, parotid glands and spinal cord. PATIENTS AND METHODS: Twenty pharyngo-laryngeal tumors were delineated by five observers on CT and MRI, using consistent delineation guidelines. Spinal cords and parotid glands were also delineated on CT and MRI by three observers. Mean GTVs and coefficients of variation were calculated for each observer and compared using ANOVA and its derived Pearson intra-class coefficient (R). For GTVs, a mismatch analysis (ratio between intersection and union volumes) was also performed. RESULTS: Regarding oropharyngeal GTVs (n=10), no significant difference was observed between observers either with CT (33.9, 31.1, 32, 34 and 34.7 ml, five observers, P=0.47) or with MRI (30.5, 29.4, 30.1 and 31.5 ml, four observers, P=0.59). CVs (13.6 vs 12.9%), (0.98 vs 0.99) and mismatches (0.43 vs 0.42) between CT and MRI did not significantly differ. Regarding laryngeal-hypopharyngeal GTVs (n=10), no significant difference was observed between observers either on CT (18.1, 20.7, 20.9, 19.3 and 21.9 ml, five observers, P=0.29) or on MRI (19.3, 21.5, 20, 22.1 and 21.8 ml, five observers, P=0.16). CVs (20.2 vs 13.8%), (0.94 vs 0.94) and mismatches (0.31 vs 0.41) were comparable. Regarding OARs, a small but significant difference in mean parotid volume was observed between observers (P