TAIWAN: Updates in the Product Registration Guidance for Artificial Intelligent/Machine Learning-Based Software as a Medical Device – April, 2022

2022-04-27

Artificial Intelligent/Machine Learning based Software as Medical Device (SaMD) is now a hot topic in enhancing the welfare of humanity. To ensure the safety, performance, and quality of medical device software that utilizes artificial intelligence/machine learning technology, TFDA has established a guidance with reference to the regulations of the United States, Japan, and Korea and the International Medical Device Regulatory Forum (IMDRF). This guidance was initially announced in September 2020 and was further updated in August 2021.

The latest requirements are summarized in the table below. However, depending on the different characteristics of each SaMD, additional documents may still be required during the reviewing process.

  1. Defining the software as:
  • Artificial Intelligence, AI
  • Machine Learning, ML
  • Deep Learning
  1. Description of software functions:

Functional types, including-

  • Computer Assisted Detection, CADe
  • Computer Aided Diagnosis, CADx
  • Computer Aided Triage

Algorithm framework-

  • Design
  • Training method
  • Testing principle and algorithm framework
  1. Data Limitation
  • Shall include the detail of the training method, structure, process, data relevancy, and quality maintenance.
  • Shall describe the training module database, in population, clinical meaning, output type, output method, and other additional clinical significances
  • There are 3 kinds of data applied in SaMD, data for training, validation, or testing.
    Among all, data for testing must be strictly separated from data for training or validation to avoid bias in the validated results.
  1. User Environment and Data Management Safety
  • Shall list out the limitations of operating environment and personnel.
  • If the SaMD can be connected to the internet, has wireless transmission functions, or is a kind of mobile application (APP), shall refer to “Cybersecurity Guidelines for Medical Device manufacturers” and provide related documents.
  1. Functional Verification and Validation

Shall refer to “Software Medical Device Guidelines”, including-

  • Level of concern
  • Software description
  • Device hazard analysis
  • Software requirements specifications
  • Architecture design chart
  • Software design specification
  • Traceability analysis
  • Software development environment
  • Verification and validation documentation
  • Revision level history
  • Unresolved anomalies/ bugs or defects

Besides the above V&V documents, the output results shall comply with the intended use of the SaMD, including-

  • How the output result will be presented
  • The limitation of the output result
  • Subsequent clinical suggestions regarding the output result
  1. Clinical Significance

Scientific evidence shall be provided to establish the applicability and suitability of the software specifications. It is recommended to include the below concept in the study protocol-

  • Intended use
  • Study objectives
  • Patient population, e.g. age, ethnicity, race…
  • Number of clinicians and qualifications
  • Description of the methodology used in gathering clinical information
  • Description of the statistical methods used to analyze the data
  • Study results


 

Reference:
Product Registration Guidance for Artificial Intelligent/Machine Learning-based Software Medical Device

分享:

TAG : Taiwan , TFDA , Medical Device , Artificial Intelligent , Machine Learning , Software , SaMD

返回