보고서 정보
주관연구기관 |
연세대학교 Yonsei University |
연구책임자 |
최병욱
|
보고서유형 | 최종보고서 |
발행국가 | 대한민국 |
언어 |
한국어
|
발행년월 | 2021-02 |
과제시작연도 |
2020 |
주관부처 |
식품의약품안전처 Ministry of Food and Drug Safety |
등록번호 |
TRKO202100007643 |
과제고유번호 |
1475011610 |
사업명 |
의료기기등안전관리(R&D) |
DB 구축일자 |
2021-07-31
|
키워드 |
의료기기.임상의사결정지원시스템.인공지능.빅데이터.표준데이터.medical device.Clinical Decision Support System.Artificial intelligence.Big data.Standard data.
|
초록
▼
인공지능 기반 임상의사결정지원시스템 의료기기의 성능 및 안전성 검증을 위하여 폐결핵, 유방암, 폐결절, 뇌경색, 관상동맥질환, 간종양의 6개 대표질환에 대한 검증용 표준데이터를 구축하였음. 6개 질환은 의료영상을 중심으로 임상정보와 함께 데이터베이스로 구축하였고, 활용을 위해서 영상뷰어, 자료입력, 수정, 검토 기능, 인공지능 학습을 위한 레이블링 기능, 전문가 판독 입력 기능, 인공지능 소프트웨어 결과물 비교 기능을 탑재한 활용서버를 구축하였음. 구축된 질환 데이터를 사용하여 4개의 인공지능 기반영상진단 소프트웨어의 검증을 진행하
인공지능 기반 임상의사결정지원시스템 의료기기의 성능 및 안전성 검증을 위하여 폐결핵, 유방암, 폐결절, 뇌경색, 관상동맥질환, 간종양의 6개 대표질환에 대한 검증용 표준데이터를 구축하였음. 6개 질환은 의료영상을 중심으로 임상정보와 함께 데이터베이스로 구축하였고, 활용을 위해서 영상뷰어, 자료입력, 수정, 검토 기능, 인공지능 학습을 위한 레이블링 기능, 전문가 판독 입력 기능, 인공지능 소프트웨어 결과물 비교 기능을 탑재한 활용서버를 구축하였음. 구축된 질환 데이터를 사용하여 4개의 인공지능 기반영상진단 소프트웨어의 검증을 진행하였고, 수요기업에 결과와 피드백을 제공하였음. 데이터의 보완 및 타 질환으로의 확장을 통해 지속적인 검증 활용을 계획하고 있으며 이를 위해 세브란스병원에 검증 서비스 시스템을 구축하였음.
(출처 : 요약문 3p)
Abstract
▼
Purpose
· Standard data for clinical validation of performance and safety of AI based medical device and development of evaluation index for clinical safety and effectiveness
· Standard data for clinical validation of performance and safety of AI based medical device on pulmonary tuberculosis,
Purpose
· Standard data for clinical validation of performance and safety of AI based medical device and development of evaluation index for clinical safety and effectiveness
· Standard data for clinical validation of performance and safety of AI based medical device on pulmonary tuberculosis, lung cancer, breast cancer, lung cancer, liver cancer, cerebral infarction, and coronary artery disease.
· Developement of evaluation index for establishing standard data for validation of clinical safety and effectiveness of AI-based medical device on the 6 clinical diseases
Contents
■ The 1st year (2018)
· Making standard data for validation of performance and safety of AI-based medical device for diagnosis of pulmonary tuberculosis on chest X-ray
· Development of evaluation standard and method by establishing evaluation index for clinical safety and effectiveness of AI-based medical device for diagnosis of pulmonary tuberculosis on chest X-ray
· Making standard data for validation of performance and safety of AI-based medical device for diagnosis of breast cancer on mammography
· Development of evaluation standard and method by establishing evaluation index for clinical safety and effectiveness of AI-based medical device for diagnosis of breast cancer on mammography
■ The 2nd year (2019)
· Making standard data for validation of performance and safety of AI-based medical device for diagnosis of cerebral infarction on brain MRI
· Development of evaluation standard and method by establishing evaluation index for clinical safety and effectiveness of AI-based medical device for diagnosis of cerebral infarction on brain MRI
· Making standard data for validation of performance and safety of AI-based medical device for diagnosis of lung cancer on chest X-ray and CT
· Development of evaluation standard and method by establishing evaluation index for clinical safety and effectiveness of AI-based medical device for diagnosis of lung cancer on chest X-ray and CT
■ The 3rd year (2020)
· Making standard data for validation of performance and safety of AI-based medical device for diagnosis of liver cancer on liver CT
· Development of evaluation standard and method by establishing evaluation index for clinical safety and effectiveness of AI-based medical device for diagnosis of liver cancer on liver CT
· Making standard data for validation of performance and safety of AI-based medical device for diagnosis of coronary artery disease coronary artery CT angiography
· Development of evaluation standard and method by establishing evaluation index for clinical safety and effectiveness of AI-based medical device for diagnosis of coronary artery disease coronary artery CT angiography
Results
· Building ‘Pulmonary Tuberculosis’ standard dataset with 2000 chest radiographs and clinical information (active tuberculosis 800, stabilized tuberculosis 400, non-tuberculosis abnormality 400, normal 400)
· Building ‘Breast Cancer’ standard dataset with 2000 mammography and clinical information (breast cancer 1200, normal 800)
· Building ‘Pulmonary nodule’ standard dataset with 1200 chest radiographs, CT scans, and clinical information (solid lung cancer nodule 200, subsolid lung cancer nodule 600, benign lung nodule 200, normal 200)
· Building ‘Brain Stroke’ standard dataset with 800 CT, MRI scans, and clinical information (stroke 600, normal 200)
· Building ‘Liver tumor’ standard dataset with 800 Liver CT scans and clinical information (Liver cancer 300, metastatic liver cancer 300, benign cyst/hemangioma 200)
· Building ‘Coronary Artery Stenosis’ standard dataset with 800 cardiac CT scans and clinical information (coronary artery stenosis 400, mild stenosis 200, normal 200)
· Utilization of ‘Pulmonary Tuberculosis’ dataset for accuracy validation of CDSS (Lunit insight CXR - tuberculosis diagnosis model, 2019)
· Utilization of ‘Breast Cancer’ dataset for accuracy validation of CDSS (Lunit insight MMG – breast cancer diagnosis model, 2019)
· Utilization of ‘Pulmonary nodules’ dataset for accuracy validation of CDSS (Lunit insight CXR – pulmonary cancer nodule diagnosis model, 2020)
· Utilization of ‘Pulmonary nodules’ dataset for accuracy validation of CDSS (Samsung ALND – lung cancer nodule diagnosis model)
· Publishing an article in an international journal 1 (2019)
· Publishing articles in domestic journals 2 (2019, 2021)
· Presentation at international academic conference 1 (2019)
· Presentation at domestic academic conference 1 (2018)
· Seminars for technology communication 2회 (2018, 2020)
· The fine report (2021)
Expected Contribution
■ Plan for utilization
· Utilization for official approval of AI-based CDSS medical device for diagnosis of the 6 representive diseases
· Utilization for interim check and validation during development of AI-based CDSS medical device for diagnosis of the 6 representive diseases
· Utilization of guidelines for development of evaluation index for AI-based CDSS medical device.
■ Expected effects
○ Technological aspect
· Improvement of the performance of AI-based CDSS medical device
· Acceleration of development and application of AI-based CDSS medical device with training important factors which are provided by development of validation system about important factors affecting the performance
○ Economical and industrial aspect
· Promotion of commercialization of AI-based CDSS medical device
· Securing competitiveness in the market by fast development and utilization of AI-based CDSS medical device
○ Social aspect
· Realization of healthy wellfare society with AI-based CDSS medical device
· Reduction of medical error in diagnosis by utilization of AI-based CDSS medical device
· Reduction of medical cost with improvement of efficiency
(출처 : Summary 6p)
목차 Contents
- 표지 ... 1
- 제 출 문 ... 2
- 보고서 요약서 ... 3
- 국문 요약문 ... 4
- Summary ... 6
- 목차 ... 10
- 제1장 연구개발과제의 개요 ... 11
- 1. 연구개발 목적 ... 11
- 2. 연구개발의 필요성 ... 11
- 3. 연구개발의 범위 ... 12
- 제2장 국내·외 기술 개발 현황 ... 13
- 제3장 연구 수행 내용 및 성과 ... 16
- 1. 1차년도 (2018) 연구수행 내용 및 성과 ... 17
- 2. 2차년도 (2019) 연구수행 내용 ... 27
- 3. 3차년도 연구수행 내용 ... 38
- 제4장 목표 달성도 및 관련 분야 기여도 ... 48
- 1. 목표 달성도 ... 48
- 2. 관련분야 기여도 ... 49
- 제5장 연구개발성과의 활용계획 ... 54
- 제6장 연구 과정에서 수집한 해외 과학기술 정보 ... 58
- 제7장 연구개발성과의 보안등급 ... 60
- 제8장 국가과학기술종합정보시스템에 등록한 연구시설·장비 현황 ... 60
- 제9장 연구개발과제 수행에 따른 연구실 등의 안전 조치 이행 실적 ... 61
- 제10장 연구개발과제의 대표적 연구 실적 ... 62
- 제11장 기타 사항 ... 65
- 제12장 참고 문헌 ... 65
- 끝페이지 ... 68
※ AI-Helper는 부적절한 답변을 할 수 있습니다.