I-Doctor Clinical Decision Support System For Brain Tumour Using Multimodal Magnetic Resonance Imaging

Global statistics show that brain tumors have become one of the malignant diseases that seriously threaten human life,health and quality of prognosis,and the number of new patients has increased year by year in all ages.The conventional diagnosis methods for brain tumors are imaging examination and tissue biopsy.Although up to now,tissue biopsy or pathological examination is still the gold standard for clinical diagnosis of tumors,pathological examination based on tissue biopsy still has many limitations.Magnetic resonance imaging technology(MRI)can show the structure of the human body clearly in detail.It is one of the standard imaging method for brain tumors.It is also increasingly used in tumors in other parts of the body.Noninvasive detection methods such as magnetic resonance imaging combined with imaging play a key role in tumor diagnosis,treatment plan selection and treatment monitoring.For example,the relative concentration information of metabolites provided by magnetic resonance spectroscopy can be used as a powerful basis for early diagnosis of brain tumors.As the information obtained from MRI(magnetic resonance imaging,MRI)scans become more and more complex,it is difficult for clinicians to process large amounts of data in real time without additional decision support systems.Therefore,with the rapid development of artificial intelligence(AI)technology,integrating a multi-modal image data storage,preprocessing and analysis into an automated clinical decision support system for the diagnosis of brain tumor,and developing a user-friendly user interface for clinicians can greatly reduce the burden on doctors,promote better distribution of medical resources,and ease the current situation of shortage of medical resources.In this thesis,multimodal MRI data from Children’s Hospital of Nanjing Medical University and Nanjing Drum Tower Hospital were collected,and I-Doctor Clinical Decision Support System for Brain Tumor diagnosis is designed.The I-Doctor prototype can provide doctors with a standardized process for data analysis and real-time classification.Its main functions include the following parts:(1)Collect multimodal MRI data of brain tumors.The format of the data includes clinical information,cognitive scale,magnetic resonance imaging(MRI),magnetic resonance spectroscopy(MRS)and its post-processing files.(2)Establish a data management system.The I-Doctor prototype provides uploading management functions for data in different formats,and automatic extraction module.Combined with the WPF user interface,this I-Doctor prototype provides image display and quality control functions for MR images and MRS post-processing files.(3)Classify common brain tumors.The I-Doctor prototype classifies pediatric brain tumors and adult brain tumors through classic machine learning methods and deep learning methods,finally provides intelligent reports for user browsing and collects feedback.This thesis designed and implemented an I-Doctor Clinical Decision Support system for Brain Tumor using multimodal Magnetic Resonance Imaging.As a prototype of a decision support system for brain tumors,I-Doctor has formed a modular method of knowledge feedback structure to provide further support for data amplification and module iteration.