Handling of AI tools in the reporting process
Conditions
Interventions
Group 1: This prospective, single-center study evaluates the use and performance of various CE-certified AI tools in routine radiological practice. Structured assessments of the AI results are conduct
approximately 10,000 datasets will be analyzed per tool.
Sponsors
Klinikum der Universität München, Klinik und Poliklinik für Radiologie
Eligibility
Sex/Gender
All
Inclusion criteria
Inclusion criteria: - Medical staff of the Clinic and Polyclinic for Radiology - Training regarding the study process - Willingness to participate in the study
Exclusion criteria
Exclusion criteria: - Missing participant information sheet/consent form
Design outcomes
Primary
| Measure | Time frame |
|---|---|
| - The use and performance of various AI algorithms in routine clinical practice - The influence of AI results on the diagnostic workflow by recording diagnostic features or existing pathologies before and after reviewing the AI results Target parameters: - Number of examinations during the study period - Number of available AI results - Temporal sequences in the diagnostic workflow - Performance parameters of the AI tools (sensitivity, specificity, AUC/AUROC, negative predictive value, positive predictive value) - Assessment regarding the presence of a diagnostic feature or pathology before reviewing the AI results - Assessment regarding the presence of a diagnostic feature or pathology after reviewing the AI results | — |
Secondary
| Measure | Time frame |
|---|---|
| Subsequently, the handling of AI tools in the reporting process will be discussed. Further target parameters: Acceptance of AI support by radiologists Integration of AI tools into the clinical workflow | — |
Countries
Germany
Contacts
Public ContactClemens Cyran
Klinikum der Universität München, Klinik und Poliklinik für Radiologie
Outcome results
None listed