Prostate ultra-high b-value DWI image generation and clinical evaluation based on unsupervised learning model
Objective To explore the unique advantages of ultra-high b-value DWI image acquisition in the diagnosis of prostate cancer,and to solve the problems of long acquisition time and artifact.Methods In order to improve the contrast and sensitivity of prostate disease detection,the prostate ultra-high b-value DWI images were generated by clinical collection of high-b-value magnetic resonance images and ultra-high-b-value images,and the feature relationship between the two domains was used unsupervised.To evaluate the clinical performance of the model and assist young physicians to improve the ability to diagnose prostate cancer.Results The generated ultra-high b value DWI images of the prostate had sufficient contrast to improve the sensitivity of prostate disease detection,and the model performed well in clinical performance evaluation.Conclusion The ultra-high b-value DWI image generation technology based on unsupervised learning not only has outstanding technical advantages,but also shows certain application potential in clinical practice.This technology is expected to be an important auxiliary tool in the field of imaging diagnosis of prostate diseases,bringing important progress and improvement to medical practice.
NETL
NSTL
万方数据
