Sign in to view source links and access this dataset
Description
Between 2010 and 2024, approximately 200,000 bedside chest radiographs were collected from around 50,000 intensive care patients at University Hospital Aachen, Germany. Trained radiologists provided structured reports assessing key findings such as cardiomegaly, pulmonary congestion, pleural effusion, pulmonary opacities, and atelectasis on an ordinal scale. The dataset is hosted by TLAIM on Hugging Face.
Use Cases
Train models for automated detection of cardiomegaly based on radiologist assessments.
Develop algorithms to quantify pulmonary congestion severity using ordinal scale labels.
Benchmark multi-label classification models for pleural effusion, pulmonary opacities, and atelectasis.
Study temporal progression of chest pathologies in intensive care patients.
Validate AI tools for structured reporting in radiology.
Strengths
Approximately 200,000 images provide a substantial sample size for training.
Data from around 50,000 patients offers a diverse patient cohort.
Structured reports by trained radiologists provide expert annotations.
Collection spans 14 years (2010-2024), enabling longitudinal studies.
Limitations
Column-level documentation is absent; field semantics must be inferred after download.
Row count is unknown, which may limit suitability assessment.
Data may reflect geographic and institutional bias inherent to a single German hospital.
Provenance
Source
University Hospital Aachen, Germany.
Collection Method
Bedside chest radiographs collected during routine clinical care.
Time Range
2010 to 2024.
Freshness
Last updated 2026-04-20 11:13:10; freshness should be verified.
Geography
Aachen, Germany.
License is unknown; users must verify terms of use before downloading.