Objective To investigate CT and MRI manifestations of uveal melanoma as well as their diagnostic value.Methods 33 cases of uveal melanoma proved by operation and pathology were studied retrospectively.CT and MRI scanning were performed in 15 and 18 patients respectively.Results The round occupying lesions within eyeball were the characteristic imaging finding of uveal melanomas, most of them were found in the location of choroid membrane. These lesions appeared as homogenous density on CT and short T 1 and short T 2 intensity on MRI. MRI was superior to CT because it could provide more accurate information such as location, shape,extension,and differentiation of the complication (retina detachment) from lesion itself. The preoperation diagnostic accuracy of CT was 53% compared with that of 88% of MRI. Conclusion MRI is the optimal examination method in confirming suspected uveal melanoma.

Objective To investigate the value of volume reconstruction ( VR ) in showing the ear lesions and to compare with otherpost-processed methods.Methods 15 patients with hearing loss underwent MSCT ultra-high resolution scan with collimation 0.5 mm,pitch 0.5~0.875 and increment 0.1mm.An additional enhanced CT was performed in 10 of 15 cases. The various reconstructions including VR , MPR, SSD, VE, Ray Sum and MIP/MinIP were performed for the comparison of their ability in showing the ear lesions and the invasion of the auditory ossicals and the inner ear of the adjacent tumors. Results VR was the most optimal way, in showing the ear lesions,successionally in descending order was VR, MPR, SSD, MIP/MinIP. Ray Sum and VE were only suitable for showing the lesions of the inner ear and auditory ossicles respectively.Conclusion The data acquired from the scan of MSCT using 0.5 mm collimation can be used in various reconstruction methods including VR, MPR, SSD and MIP/MinIP to display the lesions of the ear and the invasion of the ear fine structures from the plane and three-dimensional view. VR is the best way, which can display the relationship between tumor and facial/auditory nerve, labyrinth .

Objective:To investigate the value of low-energy virtual monoenergetic image (VMI) at 45 keV in visualizing the primary tumor and T staging of hypopharyngeal squamous cell carcinoma.Methods:The clinical and imaging data of 58 patients with hypopharyngeal squamous cell carcinoma from April 2018 to January 2020 at Eye & ENT Hospital, Fudan University were analyzed retrospectively. All the patients underwent a venous phase contrast-enhanced dual-source dual-energy CT scan before treatment. The VMI at 45 keV and standard linearly blended image (30% 80 kV+70% 140 kV) were acquired from dual-energy post-processing software. One senior radiologist and one junior radiologist independently assessed the visibility of the tumor on the 45 keV VMI and standard linearly blended image using a 5-point Likert rating scale. Furthermore, the senior radiologist assessed the visibility of the tumor at each subsite (piriform fossa, posterior pharyngeal wall, postcricoid region) and determined the invasion depth of the tumor (extension to esophagus, invasion to strip muscles and prevertebral muscles) and performed the T staging of the primary tumor using the two sets of images blindly. The accuracy of T staging was calculated, using pathological T staging (surgical cases) or clinical T staging (non-surgical cases) as the gold standard. The image scores of the two sets of images were compared using Wilcoxon rank sum test. McNemar-Bowker test was used to compare the accuracy of T staging using the two sets of images.Results:The overall image scores of the 45 keV VMI and standard linearly blended image from the senior radiologist were 3.5 (3, 4) and 3 (2, 3) respectively ( Z=-7.03, P<0.001), and the scores from the junior radiologist were 3 (3, 4) and 2 (2, 3) ( Z=-6.93, P<0.001). The scores of the 45 keV VMI were significantly higher than those of the standard linearly blended image in visualizing tumors in the piriform fossa, posterior pharyngeal wall, and postcricoid region, as well as in detecting invasion to the strip muscles ( P<0.05). There was no significant difference in the scores of the two sets of images in determining whether the tumor extended to esophagus or invaded prevertebral muscles ( P>0.05). Referring to pathological and clinical T stage, the accuracy of T staging determined by the 45 keV VMI and standard linearly blended image was 87.9% (51/58) and 81.0% (47/58) respectively, and the difference was not significant (χ 2=3.33, P=0.189). Conclusions:The 45 keV VMI is superior to the standard linearly blended image in visualizing tumors and detecting invasion to the strip muscles of hypopharynx squamous cell carcinoma. However, the accuracy of determining T staging using 45 keV VMI is slightly improved than that of standard linearly blended image, and the difference is not statistically significant. In determining whether the tumor extends to esophagus or invades prevertebral muscles, 45 keV VMI shows no significant advantage over standard linearly blended image.