Automated segmentation and classification of pleural effusion subtypes in computed tomography using machine learning

Automated segmentation and classification of pleural effusion subtypes in computed tomography using machine learning - ASCPE

Status

Recruiting

Phases

Unknown

Study type

Observational

Source

DRKS

Registry ID

DRKS00025407

Enrollment

635

Registered

2021-06-10

Start date

2021-05-23

Completion date

Unknown

Last updated

2025-04-07

For informational purposes only — not medical advice. Sourced from public registries and may not reflect the latest updates. Terms

Conditions

J91 J86

Interventions

Group 1: Patients with pleural effusion (CT based diagnosis): segmentation, training and evaluation of pleural effusion and the lung. For testing 30% of the data are reserved vor evaluation. Group 2:

Eligibility

Sex/Gender

All

Age

18 Years to 105 Years

Inclusion criteria

Inclusion criteria: CT thorax 1mm WT between 01 / 2016-01 / 2021 1st cohort with pleural effusion, randomized 2nd cohort without pleural effusion, randomized 3. Cohort for classification, consecutively with pathological correlation.

Exclusion criteria

Exclusion criteria: CTs from adjacent organ areas. Other slice thicknesses.

Design outcomes

Primary

Measure	Time frame
Sensitivity and specificity for detection and for classification. Dice coefficient and volume ICC for segmentation.	—

Secondary

Measure	Time frame
AUC for detection and for classification. Jaccard index for segmentation.	—

Countries

Switzerland

Contacts

Public ContactRaphael Sexauer

Unispital Basel

raphael.sexauer@usb.ch+41 6132 86584

Outcome results

None listed

Source: DRKS (via WHO ICTRP) · Data processed: Feb 4, 2026