Nasopharyngeal carcinoma (NPC) is a type of malignant head and neck tumor. Imaging is needed to assess the extent of the tumor for staging and delineating the target region of radiotherapy. In MRI, contrast enhanced T1 has advantages in NPC diagnosis and staging due to the improved contrast noise ratio between tumor and surrounding tissue (such as fat and muscle) ¹, it is also the only MR examination recommended by national comprehensive cancer network (NCCN) for NPC. However, contrast agent administration is associated with the danger of allergy and nephrogenic systemic fibrosis (NSF). Arterial spin labeling (ASL) uses RF labeled blood as bolus, and produces contrast free perfusion images. However, ASL itself lacks necessary anatomical information for diagnosis as compared to contrast enhanced T1. The fat suppressed (FS)-T2/ASL fusion image overcomes this limitation. In this work, we investigate if FS-T2/ASL fusion images may be a viable alternative to enhanced T1 when contrast administration is not feasible.21 patients with untreated NPC confirmed by pathology received MRI before nasopharyngoscope biopsy and treatment. MR scans were performed on a 3.0T whole body scanner (GE Discovery 750) equipped with an 8-chanel head and neck coil. Consent forms were obtained prior to the scans. Enhanced T1, fat suppressed T2 weighted and ASL were performed with matching spatial coverage. In ASL, three post labeling delays (PLD) of 1.0s, 1.5s and 2.0s were used. Fusion images of FS-T2/ASL were performed using vendor supplied software based on the scan spatial locations of each slices. The tumor regions based on different contrasts were delineated manually by two experienced radiologists and their volumes were used as the metric for assessing the consistency between FS-T2/ASL and enhanced T1. The comparison of volume measurements included：1) the inter-series correlation and consistency (ICC), which was the comparison between the measurements made using enhanced T1 and those made using FS-T2 alone and the FS-T2/ASL fusion images respectively by same observer, 2) the inter-observers consistency, which was the comparison between the measurements by the two observers based on the same image contrast. Spearman correlation coefficients were performed to determine the correlation. The consistency was evaluated using the Bland Altman plots, which were made of the mean values against the difference values, with the 95% limits of agreement calculated as the mean difference plus or minus 1.96 times the standard deviation of the difference.A typical set of FS-T2, FS-T2/ASL and enhanced T1 images are shown in Fig.1. It is seen that FS-T2 image (Fig.1a) exhibit relatively poor delineation of the tumor region as compared to ASL (Fig.1b) or enhanced T1 (Fig.1d) as arrowed. ASL image although delineates the high perfusion regions well, the matching anatomical information is completely missing. In comparison, FS-T2/ASL fusion images (Fig.1c) offers a much consistent description to enhanced T1 image. Statistically, the volumes measured by two observers using enhanced T1 and those using FS-T2, FS-T2/ASL (PLD of 1.0s, 1.5s, 2.0s) fusion images showed significant correlation. It is seen that the correlation coefficients of FS-T2 /ASL (0.921 ~ 0.973) were higher than those of FS-T2 alone based on both observers; and the correlation coefficient of FS-T2 /ASL (PLD = 1.0s) from observer 1 was the highest (0.973) while that of FS-T2 from observer 2 was the lowest (0.896) (Fig.2). The superior performance of FS-T2/ASL (PLD = 1.0s) can also be seen in the Bland-Altman plots, where the difference between measurements by FS-T2 and enhanced T1, measurements by FS-T2/ASL and enhanced T1 are shown (Fig.3). It is seen that by two observers the variations of the measurements of FS-T2/ASL are much smaller than those of FS-T2 alone.ASL is able to provide good contrast between tumor and non-tumor regions ², but lacks the necessary anatomical information of the lesions. FS-T2 alone has difficulty in in defining the tumor extents as although NPC shows hyper-intense signal to muscle, it has near iso-intense signal level to normal nasopharyngeal mucosa and cerebral cortex. FS-T2/ASL fusion image provides both anatomical and functional perfusion information, and hence offers very consistent information as enhanced T1 image, while being contrast free. As demonstrated by this study, FS-T2/ASL may be a viable alternative to enhanced T1 when contrast administration is infeasible.