Skip to main content
QMSQMS
QMS
  • Welcome to your QMS
  • Quality Manual
  • Procedures
  • Records
  • TF_Legit.Health_Plus
    • Legit.Health Plus TF index
    • Legit.Health Plus STED
    • Legit.Health Plus description and specifications
    • R-TF-001-007 Declaration of conformity
    • GSPR
    • Clinical
    • Design and development
    • Design History File (DHF)
      • Version 1.0.0.0
        • Requirements
          • REQ_001 The user receives quantifiable data on the intensity of clinical signs
          • REQ_002 The user receives quantifiable data on the count of clinical signs
          • REQ_003 The user receives quantifiable data on the extent of clinical signs
          • REQ_004 The user receives an interpretative distribution representation of possible ICD categories represented in the pixels of the image
          • REQ_005 The user can send requests and get back the output of the device as a response in a secure, efficient and versatile manner
          • REQ_006 The data that users send and receive follows the FHIR healthcare interoperability standard
          • REQ_007 If something does not work, the API returns meaningful information about the error
          • REQ_008 Notify the user if the image does not represent a skin structure
          • REQ_009 Notify the user if the quality of the image is insufficient
          • REQ_010 The device detects if the image is of clinical or dermatoscopic modality
          • REQ_011 The user specifies the body site of the skin structure
          • REQ_012 Users can easily integrate the device into their system
          • REQ_013 The user receives the pixel coordinates of possible ICD categories
          • ignore-this
            • SWR-001- Users of the REST API can log in and receive an access token
            • SWR-002- The REST API enforces HTTPS for all communications to ensure data security
            • SWR-003- The REST API implements rate limiting to prevent abuse
            • SWR-004- The REST API verifies the access token for every request to secure endpoints
            • SWR-005- Data exchanged with clinical endpoints of the API adhere to the FHIR standard
            • SWR-006- The REST API only accepts and outputs images in Base64 format
            • SWR-007- The diagnosis support service accepts multiple images to deliver more accurate results
            • SWR-008- The user's password is stored in the database as a hashed password
            • SWR-009- New users of the device are only created by an internal user registration service
        • Test plans
        • Test runs
        • 🥣 SOUPs
    • IFU and label
    • Post-Market Surveillance
    • Quality control
    • Risk Management
  • Licenses and accreditations
  • External documentation
  • Procedures
  • Records
  • TF_Legit.Health_Plus
  • Design History File (DHF)
  • Version 1.0.0.0
  • Requirements
  • ignore-this
  • SWR-007- The diagnosis support service accepts multiple images to deliver more accurate results

SWR-007- The diagnosis support service accepts multiple images to deliver more accurate results

Internal IDSWR_007
TitleThe diagnosis support service accepts multiple images to deliver more accurate results
CategoryFUNCTIONAL REGULATORY
ImportanceHIGH
SystemBackend, AI model
Editor(s)Alejandro Carmena Magro , JD-017
SupervisorAlfonso Medela , JD-005
ApprovalPENDING
Created at24 Jun 2024

Description​

The diagnosis support service must be capable of accepting 1 to 5 clinical images per request to generate a more accurate preliminary diagnosis report. The AI model analyzes each image for visual signs of skin abnormalities or diseases, integrating findings to enhance diagnostic precision. This feature ensures that multiple aspects of the skin condition are captured, which could be overlooked if only one image is examined.

Rationale​

Examining multiple images offers a more complete perspective on a patient's skin condition, capturing different angles and stages of the disease that a single image might miss. This approach improves the precision and thoroughness of the diagnostic process, helping dermatologists make more informed decisions. Utilizing multiple images allows the AI model to cross-check findings, which decreases the chances of misdiagnosis and increases the reliability of the preliminary reports generated.

Source​

  • Alfonso Medela, JD-005

Tested by software tests​

  • PLAN_013: Diagnosis support endpoint accepts multiple images
  • PLAN_014: Improved accuracy with multiple images

Activities generated​

  • Update the API endpoint for the diagnosis support service to handle multiple image inputs.
  • Improve the AI model to analyze and integrate findings from multiple images, or develop code that uses heuristic rules to aggregate the results from these images using mathematical formulas like mean or median.

Implements user needs​

This requirement focuses on providing accurate and reliable interpretative distribution of skin conditions for healthcare providers. By enabling the software to analyze a wide range of clinical images, it enhances the precision of preliminary clinical reports.

Regulatory requirements​

7.1: The device shall be compliant with MDR 2017/745, Annex I, points 1, 17.1

Causes failure modes​

  • Getting worse results compared to working with a single image.
  • Inability of the AI model or aggregation rules to handle or integrate findings from multiple images.

Implements risk control measures​

  • Mitigate risks of inaccurate diagnosis.

Acceptance criteria​

  • The API endpoint for diagnosis support service successfully accepts and processes 1 to 5 images in a single request.
  • The AI model (or heuristic rules) integrates findings from all provided images to generate a preliminary diagnosis report.
  • The preliminary diagnosis report demonstrates improved accuracy over single-image analysis in testing.
  • The system handles increased computational load within acceptable performance parameters.

Constraints​

  • The images sent to the API should be within a maximum size limit, which will be determined based on the computational capabilities and characteristics of the deployment environment.

Dependencies​

  • The AI model for the diagnosis support service has been effectively trained and validated on a test dataset, meeting the established performance metrics.

Performance considerations​

  • Ensuring the system processes multiple images efficiently without significant delays.
  • Scaling the backend infrastructure to handle increased computational demands.

Additional notes​

No additional remarks are required.

Revision history​

Document versionDateAuthorDescription
Previous
SWR-006- The REST API only accepts and outputs images in Base64 format
Next
SWR-008- The user's password is stored in the database as a hashed password
  • Description
  • Rationale
  • Source
  • Tested by software tests
  • Activities generated
  • Implements user needs
  • Regulatory requirements
  • Causes failure modes
  • Implements risk control measures
  • Acceptance criteria
  • Constraints
  • Dependencies
  • Performance considerations
  • Additional notes
  • Revision history
All the information contained in this QMS is confidential. The recipient agrees not to transmit or reproduce the information, neither by himself nor by third parties, through whichever means, without obtaining the prior written permission of Legit.Health (AI LABS GROUP S.L.)