AIMS:Radiotherapy treatment planning for head and neck cancers (HNCs) is usually based on contrast-enhanced computed tomography (CECT). However, soft-tissue contrast is better evident in magnetic resonance imaging (MRI). The study evaluates the gross tumour volumes (GTVs) delineated on CECT vs MRI along with their Dice similarity coefficients (DSCs) and resultant impact on the dose-volume histogram (DVH) parameters, conformity index (CI), and homogeneity index (HI) during intensity-modulated radiotherapy (IMRT) planning in HNCs.
MATERIAL AND METHODS:This prospective study enrolled 50 consecutive HNC patients. Following CECT and MRI simulations, GTVp (primary) and GTVn (node) were delineated independently on these co-registered images. Corresponding MRI volumes were then copied onto co-registered CECT images and IMRT plans were generated on the CECT-defined planning target volume (PTV) of primary and nodes (PTVp+n).
RESULTS:The GTVp, GTVn, and GTVp+n observed on MRI were significantly larger than the corresponding GTVs defined on CECT (all P < .001). The DSC of GTVp, GTVn, and GTVp+n was inversely correlated with the corresponding % differences of GTVp (r = -0.49, P < .001), GTVn (r = -0.41, P = .021), and GTVp+n (r = -0.73, P < .001) between CECT and MRI. The mean DSCs of GTVp, GTVn, GTVp+n, and PTVp+n were 0.78, 0.32, 0.67, and 0.78, respectively. This led to significant differences in CI and HI (both P < .001), as well as other DVH parameters (D2, D50, D95, D98, V95, and V100, all P < .001) between CECT- and MRI-defined PTVp+n.
CONCLUSION:The GTVs and PTVp+n defined on MRI were significantly greater than those depicted on CECT, resulting in significant differences in DSC, DVH parameters, CI, and HI. Thus, IMRT planning for HNCs based on CECT-defined PTV appears inappropriate. The study emphasises the importance of accurate delineation to ensure adequate coverage of the target volume and the potential benefit of MRI in this regard.