Retrospective study on the technical feasibility of analysis algorithms for the detection of dermatological and radiological diseases at a medical level using image data

Recruiting

• Conditions: L98.9; Disorder of skin and subcutaneous tissue, unspecified

• Registration Number: DRKS00024650
• Lead Sponsor: MU München

Brief Summary

Not available

Detailed Description

Not available

Study Timeline

• Study Start: 2023-02-15
• Last Update Posted: 19 August 2024

Recruitment & Eligibility

• Status: Recruiting
• Sex: All
• Target Recruitment: 900000

Inclusion Criteria

Use only images with a registered dermatological diagnosis
Cross-validation with the histological diagnosis

Exclusion Criteria

Insufficient image quality
Pronounced image artifacts
Complete anonymization not possible

Study & Design

• Study Type: observational
• Study Design: Not specified

Primary Outcome Measures

Name	Time	Method
The aim of the present study is to demonstrate the technical feasibility of analysis algorithms for the detection of dermatological and radiological diseases at the medical level using image data through programming, training and validation using retrospective image data.

Secondary Outcome Measures

Name	Time	Method
The results of the algorithms are compared with the reference standards specified by doctors. Key statistical parameters include: accuracy, sensitivity, specificity, F1 score and confusion matrix.