Accurate Image Recognition And Intelligent Training Of US-MRI Image Fusion In Ultrasound Ablation Of Uterine Leiomyoma

Leiomyoma is the most common benign tumor in fertile age women,the tumor will cause a series of health problems.High Intensity Focused Ultrasound(HIFU)ablation for uterine leiomyoma is a noninvasive method of treatment,by virtue of its characteristics and minimally invasive with preservation of fertility,is attracting more and more attention to the gynecologist.With the ultrasound guided HIFU technology is widely used in clinical,breakthrough Ultrasound Image(US)of poor quality,high noise and limitation of the narrow field of vision,the Ultrasound Image and Magnetic Resonance Image(US-MRI)registration,fusion technology to improve the HIFU diagnosis of uterine fibroids ablation,treatment,intelligent become the new demand of training.So for this purpose HIFU research team based on US-MRI registration,fusion technology,3D reconstruction technology of intelligent technology,developed the HIFU virtual reality system.The purpose of this study was to evaluate the performance of US-MRI registration,fusion technology's value inultrasound ablation of uterus.And the virtual reality training based on US-MRI registration technology were used in order to provide basis for the application and popularization of HIFU in the treatment of uterine fibroids.Objective:1.To evaluate the application value of US-MRI image registration and fusion technology in the treatment of ultrasound ablation of uterine fibroids.2.To verify the feasibility and accuracy of US-MRI image registration in ultrasound ablation of uterine fibroids.3.To evaluate the effect of HIFU virtual reality system on ultrasound ablation of uterine fibroids.Methods:1.The image registration and fusion effect of 78 groups of patients with uterine fibroids were evaluated prospectively.The registration markers were selected to evaluate the performance of the registration technique: registration accuracy,reliability and clinical feasibility of the 78 groups.2.According to Objective Structured Clinical Examination rules,the virtual test items are designed to evaluate the reliability and validity of the virtual test and the quality of the training project.3.From December 2015 to January 2017 in First Affiliated Hospital of Chongqing Medical University ultrasound ablation treatment center inultrasound ablation of uterine fibroids surgery training 8 students,the students assigned to the virtual group(virtual group,virtual reality training group)and traditional group(traditional group,traditional training group),each group of 4 people.As a senior doctor selected 240 patients for the positioning of the students,by double blind method(with each other and students can not choose)assigned to each student in 30 patients,according to the doctor in the sequence is divided into 6 phases,each doctor treated 5patients.Comparative evaluation of effect of HIFU virtual reality system aided ultrasound ablation of uterine fibroids surgery positioning training learning curve through the positioning time of two groups of students.Results:1.The image registration results,78 groups of images to achieve image registration operation of 76 groups of the image can image of the 2 groups before the uterus,because of changes in US and MRI on the position of the uterus is too large to achieve image registration operation in.78 group after image registration definition are improved.The uterus,uterine fibroids,pubic bone,sacrum and coccyx can according to the the boundary line,the standard points by US and MRI.With the boundary of registration as a symbol to the same plane of registration,registration clarity increased significantly,P=0.000;uterine leiomyoma with boundary as a symbol of the same registration of uterine fibroids plane,the distinctness of the registration increased significantly,P=0.000;with pubic boundary as asymbol of registration can be the same a pubic plane,registration clarity increased significantly,P=0.000;used as a sign of the sacral promontory sacrococcyx registration,registration clarity increased significantly,P=0.000.2.The feasibility of the evaluation results,the effect of registration of A image registration after the definition is well improved: there was a good correlation between US with the MRI,expand their horizons,improve the lesion boundary definition.The proportion of 82.05%.The effect of B,image registration after the resolution does not improve: there is no correlation between US and MRI: cannot be diagnosed by image registration.The proportion of 2.56%.effect of C on US has been very clear,do not need to be diagnosed by registration the proportion of 15.38%.Image.3.Registration accuracy,in the womb image,76 groups of image registration operations can be realized,the registration error registration of uterus was 3.20±0.10 mm,with uterine leiomyoma boundary markers for registration,registration registration of uterine leiomyoma was3.00±1.25 mm.45 patients with endocardial border visible on the US,the intima boundary marker registration,registration registration endometrium was 2.20±0.45mm;33 cases of sacral promontory patients seen by US,so as to the sacral promontory as registration markers,registration registration sacrococcyx is 1.82±0.48mm;the pubic boundary of 22 patients seen byUS,so the pubic boundary marker registration,registration registration of pubis was 1.64±0.32 mm.4.Image fusion effect evaluation,on uterine myoma boundary as a symbol of image registration after the implementation of the global MRI and US fusion group.70 image fusion,3 groups of C(4.29%)(fusion effect of image definition and registration of C.image fusion after the lesions and the surrounding anatomical structures compared to no improvement.),67 groups of fusion effect for D(95.71%)level(image definition and registration of D.image fusion after compared lesion borders and surrounding anatomical clarity declined.)5.The whole project positioning ability test,reliability,Cronbach's ?=0.90.Positioning ability test total score,virtual group scored 40.00±1.63 points,traditional group scored 26.75±5.74 points.The virtual group was significantly higher than that of the traditional group score(t=4.421,P=0.004).Positioning capability testing project:(1)To test the ability to operate the project positioning equipment.The credibility of Cronbach 's? =0.91,test scores,virtual group scored 16.25±2.36 points,traditional group scored 12.00±4.24 points.Scores of virtual group and traditional group compared the difference was not statistically significant(t=1.753,P=0.131);(2)Positioning in the anatomic landmarks identification test project,the credibility of Cronbach' s ? =0.87,test scores,virtual group scored 14.00±1.41 points,the traditional group scored 10.50±1.92 points.Group was significantly higher than that of the traditional virtual group score(t=2.942,P=0.026);(3)The difficult location discrimination ability test project,credible Cronbach 's ?=0.88,test scores,virtual group score9.75±2.63,the traditional group score of 4.25±1.25.virtual group score was significantly higher than the traditional group score(t=3.774,P=0.017).6.The learning curve of the Virtual Reality Technology Training Group(VRTTG)6 stages and learning curve of the Traditional Training Group(TTG)6 stages' results: in the 1 stage to 6 stage positioning operation practice,the learning curve of the VRTTG below the learning curve of TTG;the positioning operation practice 1 to 4 stage,the distance of learning curve in the VRTTG and the TTG showed oscillation asymptotic state,after the 4 stage,showing a steady trend of asymptotic in learning curve of the VRTTG and TTG.Conclusion1.US-MRI image registration technique can be implemented by doctor In the treatment of ultrasound ablation.It can improve improve the recognition effect of anatomical structure in sometime,make up for the real time monitoring MRI shortcomings and disadvantages of low resolution US,improve the clarity of the image,expand horizons,assist doctors to understand the US by blocking artifacts of image information.2.Evaluation of the results in this study show that the degree of credibility evaluation scheme of virtual reality,general adaptability is verygood.Virtual Reality Assisted Training and Traditional Training in students' learning curve compared to the initial stage has better training effect.