Objective To compare the difference between linear blending and non-linear blending function for dual-energy CT, and to evaluate the effect on CT image quality. Methods The model was made of a piece of fresh pork liver inserted with 5 syringes containing various concentrations of iodine solutions ( 16.3,26.4,48.7,74.6 and 112.3 HU). Linear blending images were automatically reformatted after the model was scanned in the dual-energy mode. Non-linear blending images were reformatted using the software of optimal contrast in Syngo workstation. Images were divided into 3 groups, including linear blending group, non-linear blending group and 120 kV group. Contrast noise ratio (CNR) were measured and calculated respectively in the 3 groups and the different figure of merit (FOM) values between the groups were compared using one-way ANOVA. Twenty patients scanned in the dual-energy mode were randomly selected and the SNR of their liver, renal cortex, spleen, pancreas and abdominal aorta were measured. The independent sample t test was used to compare the difference of signal to noise ratio (SNR) between linear blending group and non linear blending group. Two readers' agreement score and single-blind method were used to investigate the conspicuity difference between linear blending group and non linear blending group.Results With models of different CT values, the FOM values in non-linear blending group were 20. 65 ±8.18,11.40±4.25, 1.60±0.82,2.40±1.13, 45.49±17. 86. In 74. 6 HU and 112. 3 HU models, the differences of the FOM values observed among the three groups were statistically significant ( P＜0.05),which were 0.30±0. 06 and 14. 43 ±4. 59 for linear blending group, and 0. 22 ±0. 05 and 15.31±5.16 for 120 kV group. And non-linear blending group had a better FOM value. The SNR of renal cortex and abdominal aorta were 19.2±5.1 and 36. 5 ± 13.9 for non-linear blending group, while they were 12.4 ±3.8 and 22.6±7.0 for linear blending group. There were statistically significant differences between the two groups. The conspicuity scores of different organs in non linear blending group were far higher than those in the linear bending group. The Kappa value was 0. 64, showing good consistency. Conclusion Compared to the linear blending method, the non-linear blending method was a better option, especially in improving SNR and conspicuity.

Objective To explore the imaing features of hemangioma in the long bone and improve the diagnostic level of this disease. Methods The X-ray(14 cases), CT(9 cases) and MRI(6 cases)findings of 18 patientswith histologically proven hemangioma in the long bone after surgery were retrospectively reviewed. Results Ten tumors occurred in medullary cavity or bone end(medullary type),6 on the surface of bone (periosteal type) and 3 in cortex (intracortical type). X-ray findings: among 8 cases of medullary type, 3 showed honeycomb appearance, 3 lytic areas with sclerotic borders, one purely osteolyticchanges, and 1 frosted glass; 3 cases of periosteal type showed sclerosis and thickening of the underlying cortex; 3 cases of intracortical type showed well-defined osteolytic foci. CT findings: among 6 cases of medullary type, 5 appeared as expansile lytic lesion with uneven selerotic rim (3 cases)orhoneycomb appearance (2 cases), 1 cribriform appearance in the cortical bone, 2 periostealnew bone formation in vertical radiation pattern; 1 ground-glass appearance; among 2 cases of periosteal typeone showed regular cortical thickening, and the other irregular periosteal proliferation with marrowing of medullarycavity; 1 case of intracorticaltype showed density similar to that of soft tissue, with cortical thickening and expansion . MRI findings: 2 apeared as well-defined lesions with low signal intensity on T1WI and high signal intensity on T2WI; 1 appeared as ill-defined lesion with low to intermediate signal intensity on T1 WI and T2 WI. One showed breakthrogh of cortex and formation of soft tissue mass with low signal intensity on T1 WI and high signal on T2WI. Two showed thickening of periosteumwith intermediate signal intensity in one of them and very low signal intensity in the other. Two showed abnormal signal intensity in surrounding muscles, which was high on T2 WI and intermediate on T1 WI. Conclusions The soap-bubble or honeycomb appearance is the typical radiographic finding of hemangioma in long bone. CT and MRI can provide useful information for the diagnosis of hemangioma in long bone.

Objective To explore the imaging characteristics of SAPHO syndrome in 11 cases.Methods Clinical features and imaging findings from 11 patients (6 male,5 female, 28 to 68 years old)with SAPHO syndrome were analyzed retrospectively Including DR in 9 cases, CT in 10 cases, MRI and radioisotope scanning in 3 cases. Results Multi-bones of anterior chest wall disorders were shown in 9cases on DR images including superior sternum , anterior first rib and clavicle hyperostosis. Bony fusion and bony bridge were also seen in these cases. Hyperostosis osteosclerosis, bone destruction and bony fusion of sternoclavicular articulation and first rib were shown on CT images in 9 cases. Osteosclerosis of the joint between manubrium and midsternum was seen in 1 case on CT image. Thc sign of flying sea gull was seen in 2 cases on axial anterior chest wall CT images. The disorders of anterior chest wall were bilateral in 8 cases and unilateral in 2 cases. Sacroiliitis and osteomyelitis of ilium were found accompanied in 1 case.Osteomyelitis of thoracic vertebrae were found in 2 cases, while sclerosing osteitis of lumbar vertebrae and ostearthritis of bilateral hands were observed respectively in 1 case. The thicken soft tissue surround clavicle head, thoracic vertebra disease with long T1 ,jumbly T2 and high fat suppression signal ,long T1 and short T2signal under sacroiliac joint were shown on MRI. Radioisotope scanning displayed higher radioactive uptake of radionuclides, with T shape in sternoclavicular area in 3 cases. Conclusions Multi-bones of anterior chest wall involvement was the common imaging characteristics in 11 patients. Sacroiliitis, osteomyelitis of vertebrae and ilium, sclerosing osteitis, ostearthritis of hand could be seen in some cases.

Objective To investigate the value and usefulness of optimized multislice CT enterography (MSCTE) with orally administered isosmotic mannitol (2. 5%) as negative contrast in demonstrating the small bowel and its abnormality. Methods Forty patients suspected of intestinal tumors were randomly divided into two groups and underwent conventional or optimized MSCTE. The expansion degree of bowel lumen and the thickness of bowel wall were evaluated for the six segments of the small intestine. The other 20 patients suspected of gastrointestinal diseases underwent gastrointestinal CT imaging.The expansion degree of bowel lumen and the wall thickness of bowel wall were statistically analyzed with Chi-Square test and t test. Results The wall thickness of the stomach, ileum and colon were (2. 56 ±0.52) ,(1.41 ±0. 15),(1.46 ±0. 13),(1.91 ±0. 25), (1.97 ±0.26),(2.01 ±0. 19), (2. 04 ±0.24)and (2. 05 ±0. 18)mm. Optimized method was superior to conventional method in the expansion degree of the second and third segments of the small intestine (P ＜ 0. 05) . There was no significant differences between two groups in the expansion degree and depiction of mucosa for the other segments of small intestine (P ＞ 0. 05). The gastrointestinal CT imaging was poor in the depiction of the duodenum and jejunum, but stomach, ileum and colon were fully illustrated. Conclusion Optimized MSCTE was superior to conventional method in demonstrating the small bowel, and gastrointestinal CT imaging can expand diagnostic scope because of good observation of whole gastrointestinal tract.

Objective　To optimize the oral contrast media in three-dimensional display of gastric lesions. Methods　41 cases were randomly divided into 3 groups according to different oral contrast media administered: No. 1 air contrast group (n=17), No. 2 fat emulsion group (n=7) and No. 3 positive contrast group (n=25). The 3D CT images were reconstructed using MPR, SSD, RaySum display and virtual endoscopic techniques, and compared with gastric endoscopy and/or conventional barium study.Results　The detectability of gastric lesions using fat emulsion and air contrast was 42.8%(3/7) and 80.0%(20/25), respectively, both were significantly lower than that using positive contrast (100%, 30/30) (χ2=19.22,P＜0.01;χ2=6.60, P＜0.05). The capability of showing the details of stomach lesions was significantly affected by the oral contrast media administered（χ2=17.04，P＜0.01）. Conclusion　It is very important to choose the appropriate oral contrast media for 3D display of gastric lesions in spiral CT, the positive contrast agent is the optimal choice.

Objective　To improve the recognition of the imaging appearances of inflammatory pseudotumors of the spleen (IPS), 3 cases with IPS were reported. Methods　The US (n=3), CT (n=3) and MRI (n=1) findings of IPS were reviewed and correlated with the pathologic findings. Results　On US, a well-defined solitary mass with heterogeneous echo texture was found in all 3 cases. A hyperechoic rim with associated acoustic shadowing was shown in 1 case. In all of the 3 cases in nonenhanced CT scanning, a well-defined hypoattenuated mass was found. One had a calcified egg-shell-like rim; On the venous/delayed phase of enhanced CT after contrast administration in 2 cases, slight/marked enhancement was shown. On nonenhanced MRI in 1 case, the mass was shown as heterogeneous hypointensity on T1- and T2- weighted images. Conclusion　IPS should be included in the differented diagnosis of solitary mass lesion of spleen. The imaging findings depend on the variable proportions of fibrous and granulomatous components within the lesion. IPS was characterized by well-defined solitary mass on sonogram, delayed enhancement on enhanced CT, and hypointensity on T2 weighted MR images.

Objective　To evaluate the value of gradient-echo chemical shift imaging in detection of lipid-containing abdominal lesions. Methods　35 patients with abdominal lesions were divided into 2 groups according to whether or not they contained lipid. Breath-hold in-phase(IP) and opposed-phase(OP) acquisitions were performed. The demonstration of these lesions on IP and OP was compared. The signal intensity of these lesions on both images was measured. The percentage of signal intensity variation on IP and OP was calculated and statistically analyzed with non-parametric method between the 2 groups. Results　For lesions containing lipid, the signal intensity dropped significantly from IP to OP, with median value of signal intensity variation attaining 37.1%. On the other hand, for lesions not containing lipid, the signal intensity did not change obviously between IP and OP, whose median value of signal intensity variation was 3.3%. The difference between the 2 groups was statistically significant (u=4.56，P<0.01). Conclusion　Gradient-echo chemical shift imaging is an easy and practical way, helpful to detection of lipid component within abdominal lesions.

Objective To evaluate the coronary microvascular dysfunction in patients with hypertrophic cardiomyopathy(HCM) by MR first-pass perfusion and late gadolinium enhancement. Methods From January 2011 to May 2015, 47 cases with HCM (HCM group) from the second affiliated hospital of Nanchang University were retrospectively analyzed. Additionally, 21 healthy volunteers were recruited as the control group. HCM group and control group underwent cardiac MR examinations at rest, including short axial cine, first-pass myocardial perfusion and late gadolinium enhancement scanning. Time to peak(tpeak), maximal upslope of time-intensity curve(Slopemax), peak signal intensity(SIpeak), myocardial thickening, and late myocardial gadolinium enhancement(LGE) were assessed for each myocardial segment. HCM group were divided into LGE segments group and non-LGE segments group. LGE segments group were divided into mild, moderate and severe LGE segments group. The SIpeak, Slopemax and tpeak in multiple groups were compared by one-way ANOVA and Kruskal-Wallis test. Spearman correlation tests were used to determine the relationships between perfusion parameter and LGE. Results The average values of tpeak in non-LGE segments group (527 segments), LGE segments group (225 segments) and control group (336 segments) were (67.0 ± 27.4), (79.4 ± 27.4), (59.7 ± 21.6)s, respectively. The average values of Slopemax in the three groups were 17.2±7.0, 16.4±7.4, 20.4±6.3, respectively. The average values of SIpeak in the three groups were 442.7 ± 143.2, 465.1 ± 138.4, 521.9 ± 146.7, respectively. Compared to the control group, tpeak increased and Slopemax, SIpeak decreased in non-LGE segments group and LGE segments group (P<0.01), while tpeak increased more significantly in LGE segments group. The Slopemax and SIpeak showed no statistically significant differences between non-LGE segments group and LGE segments group (P>0.05). There were significant differences among LGE segments groups, as the tpeak and SIpeak increased with increasing degrees of myocardial LGE (P<0.01). The Slopemax showed no statistically significant difference among them (P>0.05). The degree of LGE were positively correlated with tpeak (r=0.237, P<0.01). Conclusions 3.0 T magnetic resonance myocardial perfusion imaging can show microvascular dysfunction accurately and reliably in non-LGE segments. It may be helpful in the early diagnosis of coronary microvascular dysfunction for HCM.

Objective To investigate the influence of different scout image on the radiation dose of head and chest spiral scanning in CT using automatic tube current modulation(ATCM).Methods CT scanning was performed on the head-neck and chest phantom with ATCM.Five different scout images through the five various positions that include anteroposterior(AP),posteroanterior(PA),lateral,AP,lateral,PA and lateral.The phantom was scanned three times for each position.Then each scout image was scanned by spiral technique once more.ROI in orbital center and C5 upper edge level were selected for Head and neck phantom,ROI in the apical and tracheal bifurcation level were selected for chest phantom.The contrast to noise ratio (CNR) was measured and recorded.The organ dose of eye lens and mammary glands were measured by thermoluminescent dosimeters (TLD) in every scanning(the average of 3 measurements),The cumulative value of scout image and spiral scanning were calculated.The volume CT dose index (CTDIvol) of each scan was recorded,and the cumulative value of CTDIvol was calculated.Results In the five scout images of the head-neck phantom mode,the maximum value of accumulated radiation dose of eye lens and CTDIvol appear on PA scout image(18.354 and 26.43 mGy respectively),while the minimum value appear on the lateral scout image(11.847 and 18.20 mGy respectively).In the chest phantom mode,the maximum value of the accumulated radiation dose of the mammary gland and CTDIvol emerge from the AP scout image (6.873 and 9.42 mGy respectively),while the minimum value of the mammary accumulated radiation dose appear on lateral scout image(4.592 mGy),the minimum value of CTDIvol appear on AP plus lateral scout images(3.94 mGy).The accumulated radiation dose of eye lens and CTDIvol value in PA scout image mode were 54.9％(6.507/11.847) and 45.2％(8.23/18.20) higher than those of the lateral scout image mode.The accumulated radiation dose of mammary gland and CTDIvol value in AP scout image mode were 42.5％ (2.051/4.822) and 136.7％ (5.44/3.98) higher than that of PA plus lateral scout images mode.In the head-neck phantom mode,the CNR value of orbital center and C5 upper edge level were 102.55 to 115.89,161.01 to 204.52 respectively.In the chest phantom mode,the CNR value of the apical and tracheal bifurcation level were 82.74 to 164.00,83.12 to 121.49 respectively.Conclusion The choice of the scout image had significant effect on the radiation dose and the sensitive organ dose in CT.

Objective To investigate the feasibility of reducing spine metal artifacts with metal artifacts reduction technique (WARP) at 3.0 T MRI.Methods This study included 15 cervical and 14 lumbar spine cases.The image quality of WARP sequences and conventional sequences were compared (5 score evaluation scale) as well as the signal to noise ratio (SNR) and contrast noise ratio (CNR) of the image artifacts.The scanning time was recorded.Paired-t test and Mann-Whitney test were used respectively to compare the SNR and CNR,and qualitative scoring between the two sequences.P＜0.05 was considered to indicate a significant difference.Results The image distortion and blur of the WARP sequences were obviously reduced as compared to the conventional sequences.The SNR and CNR of artifacts of the WARP sequences were lower than that of the conventional sequences (All P＜0.05).The image quality scores of WARP sequences in cervical and lumbar spines[4(3 to 5) and 4(3 to 5)] were higher than that of conventional sequences[3(2 to 4),3(2 to 4)](P＜0.05).The scanning time of cervical spines in WARP sequence(14 min 9 s) was increased by 64 s (8.2％),and the time of lumbar spines (13 min 41 s) decreased by 9 s (1.1％).Conclusion The WARP sequences at 3.0 T could effectively reduce the artifacts of metallic prosthesis in cervical and lumbar spine without prolonging the scanning time at 3.0 T MRI.