Purpose To explore the diagnostic value of double exponential model for pelvic lesions using 3.0T MRI for the diagnosis of pelvic lesion. Materials and Methods Fifty patients with pelvic lesions (30 benign cases and 20 malignant cases) underwent MR750-diffusion weighted imaging (DWI) scans, with b values of 0, 50, 300, 600, 800 and 1200 s/mm2, Functool-MADC software was used on AW 451 workstations for data processing, Slow ADC value, Fast ADC value, Standard ADC value, Fraction of fast ADC value were recorded and compared between benign and malignant lesions, and Standard ADC images were fused with axial T2 fat-suppressed images. Results Slow ADC values [(1.83±0.86)×10-3 mm2/s] and Standard ADC values [(1.79±0.78)×10-3 mm2/s] of benign lesions were larger than those of the malignant lesions [Slow ADC values:(1.05±0.31)×10-3 mm2/s;Standard ADC values:(1.13±0.39)×10-3 mm2/s] (t=3.90, 3.51;P<0.01), and the difference of Slow ADC value was largest between benign and malignant lesions. Slow ADC values of both benign and malignant lesions were significantly less than the Fast ADC values [benign:Slow ADC value=(1.83±0.86)×10-3 mm2/s, Fast ADC value=(16.95±8.63)×10-3 mm2/s; malignant: Slow ADC value=(1.05±0.31)×10-3 mm2/s, Fast ADC value=(15.12±9.90)×10-3 mm2/s] (t=-10.40,-6.29;P<0.01). Conclusion Double exponential decay model is capable of differentiating benign and malignant pelvic tumors, thus is of great significance for clinical preoperative diagnosis.

3D reconstruction from medical images is widely used in diagnose, radiotherapy surgery programming and medical research. This paper deals with the framework and key techniques of 3D reconstruction system of medical images. Methods of automatic and manual segmentation of medical images are designed. Data structure of the system is proposed. The system data are encapsulated in object-oriented paradigm and the hierarchical structure of classes is presented for reconstructing 3D geometry models. The 3D surface geometry model is constructed by using MT algorithm from the tissues extracted by automatic or manual segmentation. An edge collapse algorithm of mesh simplification is proposed and implemented, and the surface model constructed by MT algorithm is simplified by using this algorithm, After reduction of 90 percent triangles, the simplified model still maintains the original feature. It can speed up surface rendering of the models.
Algorithms ; Humans ; Image Processing, Computer-Assisted ; Imaging, Three-Dimensional

Objective To evaluate the role of MSCT in diagnosing Cantrell syndrome.Methods Five patients with Cantrell syndrome were enrolled in this study retrospectively.All clinical data,especially imaging data were collected at enrolment.Maximum intensity projection(MIP),multi planar reconstruction(MPR)and volume rendering(VR)of the analysis method of MSCT were used to describe the characteristics of Cantrell syndrome.Results The age of 5 patients ranged from 2 days to 24 years,4(4/5)cases were Cantrell syndrome and 1(1/5)case was incomplete Cantrell syndrome,3(3/5)cases were confirmed by surgery.Five cases were all diagnosed as ectocardia,thoracocyllosis,pericardium defect and diaphragm defect with MSCT.Four had sternum dysplasia,abdominal wall defect, 3 had ventricular diverticulum,2 had umbilical hernia.Conclusion MSCT can be used in accurately diagnosing Cantrell syndrome in clinical work.