Huimin Hou, PhD | Beijing Hospital; Ming Liu，PhD; Min Chen, Phd; Chunmei Li, Phd; Wei Zhang, Phd; Wen Zhen, Phd; Mengxiao Peng; Phd Candidate; Menglin WU, PhD; Shiping Zhang, MS; Han LI; Longteng Liu, PhD; Wenrui Xu, Phd; Tongzheng Lu, MS; Haoran Wang, MS; Wen Liu, MS; Jiening Wang, MS
Competition Sponsor: Chinese Academy of Medical Sciences
Awardee Year: 2023
Prostate cancer is one of the most common malignant tumors in men. Early diagnosis is very important to improve the survival rate of patients. However, the existing diagnostic techniques that rely on MRI and prostate biopsy rely too much subjectively, resulting in inadequate diagnostic rate. This project aims to develop an artificial intelligence image recognition system and intelligent biopsy robot based on prostate cancer pathology and MRI, so as to improve the early diagnosis and treatment of prostate cancer.
Our system will utilize deep learning technology and computer vision algorithm to automate analysis and recognition of large pathological section images of prostate cancer. By constructing a deep convolutional neural network model and training and optimization of large-scale pathological image data of prostate cancer, accurate classification and segmentation of cancer cells, normal tissues and diseased areas were realized. The model will be able to accurately identify and label the location and extent of prostate cancer, providing diagnostic capability.
We will also develop an intelligent biopsy robot. By combining image recognition results and real-time ultrasonic guidance and matching image processing algorithm, the puncture target will be provided to the biopsy robot, and the biopsy will be completed by the robotic arm under the control of magnetic field.
The key technical challenges of the project include tagging a large number of pathological images of prostate cancer and repetition training, optimizing the accuracy and robustness of deep learning models, and achieving accurate matching of image recognition results to ultrasound images, and robotic arm guidance. We will rely on the hospital’s clinical resources and professional team to work closely together to ensure the smooth progress of the project and the effective implementation of the results.
The expected project outcomes include a stable and reliable prostate imaging AI diagnostic system that can accurately identify the location and extent of prostate cancer in a relatively short period of time, and an intelligent puncture system to achieve integrated diagnostic results and improve diagnostic efficiency.