Radiomics for the differentiation of benign and (pre)malignant lesions in virtual CT colonography

• Conditions: D12.6; C18.9; Colon, unspecified

• Registration Number: DRKS00014828
• Lead Sponsor: Klinikum der Universität München, Campus Großhadern

Brief Summary

In this proof-of-concept study, deep learning enabled the differentiation of premalignant from benign colorectal polyps detected with CT colonography and the visualisation of image regions important for predictions. The approach did not require polyp segmentation and thus has the potential to facilitate the identification of high-risk polyps as an automated second reader.

Detailed Description

Not available

Study Timeline

• Results First Posted: 2021-02-23
• Study Completion: 2022-09-28
• Study Start: 2023-02-08
• Last Update Posted: 19 August 2024

Recruitment & Eligibility

• Status: Complete
• Sex: All
• Target Recruitment: 620

Inclusion Criteria

Data sets with a potentially pathological finding in the VK and histopathology present.

Exclusion Criteria

- Data sets with an inconspicuous VK
- Data sets with a potentially pathological finding in the VK, but without existing histopathology

Study & Design

• Study Type: observational
• Study Design: Not specified

Primary Outcome Measures

Name	Time	Method
Semantic/agnostic image parameters with high test-retest stability as well as high predictive value are to be determined. The reference standard is histopathology.

Secondary Outcome Measures

Name	Time	Method
Potentially pathologic lesions of the colon are recorded and subdivided with respect to their image morphology.