Regulatory Open Forum

 View Only

  • 1.  Special 510(k) and established methodology for a software change

    This message was posted by a user wishing to remain anonymous
    Posted 27-Sep-2021 10:37
    This message was posted by a user wishing to remain anonymous

    Hello All,

    I am trying to understand the special 510(k) eligibility criteria (established test methodology) for a software change that requires a submission.

    It is my understanding that if we create a new protocol to test the SW algorithm change, it no longer falls in the established test methodology category because the new protocol was not part of the predicate submission.

    Since 62304 is a process based standard, the new protocol will not be considered established test methodology. 

    Please share your experience about submitting a new protocol in a special 510(k) for a software change.

    Thanks in advance!


  • 2.  RE: Special 510(k) and established methodology for a software change

    Posted 27-Sep-2021 11:29
    Hello,

    For AI devices, the algo change may/may not trigger a new submission.  Have you reviewed the Artificial Intelligence and Machine Learning Discussion Paper?  It provides a few scenarios to help make your determination.  https://www.fda.gov/medical-devices/software-medical-device-samd/artificial-intelligence-and-machine-learning-software-medical-device

    ------------------------------
    Gretchen Upton
    RAC, CQA, CCRP
    QA/RA consultant
    San Antonio, TX
    ------------------------------

    Original Message


  • 3.  RE: Special 510(k) and established methodology for a software change

    This message was posted by a user wishing to remain anonymous
    Posted 27-Sep-2021 23:15
    This message was posted by a user wishing to remain anonymous

    Thanks, Gretchen. This change is not an AI software. This is a SW that controls the hardware and not an AI/ML.
    Original Message


  • 4.  RE: Special 510(k) and established methodology for a software change

    Posted 28-Sep-2021 03:43
    Hello Anon,

    You best source of information is the FDA guidance document: https://www.fda.gov/regulatory-information/search-fda-guidance-documents/deciding-when-submit-510k-software-change-existing-device for changes to software related to a 510(k).  While this concentrates mainly on Software as a Medical Device (SaMD) it can be applied to software contained in a medical device as well.  There is also the guidance document for submitting a change to a medical device with a 510(k) which has the flow charts, decision tree, and examples concerning different scenarios.

    Regarding a Special 510(k), there are basic three criteria where this can be made:
    • It is your own device that you are changing
    • There is no change to the indications for use or other significant labelling (like adding a significant warning)
    • There is no significant change to the technology or performance of testing which can be managed through design controls

    Without knowing the type of change you are making in the software it would be hard to comment.  However, if you are using a new protocol for testing this may not even require a new 510(k) submission, but just internal documentation.  Just because the original protocol was in the original 510(k) submission, if you develop something better or higher quality then utilise design controls, design changes, change management, etc., to document.  Again without the full information it would be hard to comment, though recommend reading through the guidance documents as there are some examples provided.

    ------------------------------
    Richard Vincins RAC
    Vice President Global Regulatory Affairs
    ------------------------------

    Original Message


  • 5.  RE: Special 510(k) and established methodology for a software change

    This message was posted by a user wishing to remain anonymous
    Posted 28-Sep-2021 17:12
    This message was posted by a user wishing to remain anonymous

    Thanks, Richard. Please see the excerpt from the FDA special 510 (k) guidance. The guidance mentions to reference the protocol from existing device and provide explanation of changes. Software lifecycle is based on 62304 (standard recognized by FDA), a process based standard. 62304 does not define any test method or acceptance criteria, so if a brand new test protocol is created to test a new feature/user interface will that be acceptable as a special submission?

    The guidance mentions using well established test method: previously cleared test protocol, FDA recognized standard that provide test method, acceptance criteria etc. or test method that are part MDDT or publicly recognized. It looks like a software change that requires a 510(k) and can't be tested without new protocol (not cleared previously), special 510(k) is not an appropriate pathway. Please share your experience.


    Original Message


Related Content