PET/CT Radiomics and Biomarker Data for Lymphoma Diagnosis in Fever of Unknown Origin

203 patient records (114 lymphoma, 89 benign) from a retrospective study by Fenglian Jing, last updated in April 2026. The dataset integrates PET/CT radiomics features with clinical biomarkers like procalcitonin and serum amyloid A to build diagnostic models for differentiating lymphoma from benign lymphadenopathy in patients with fever of unknown origin. A combined model demonstrated an area under the curve of 0.958 in the testing cohort.

Use Cases

• Training binary classification models to identify lymphoma based on integrated radiomics and biomarker features.
• Validating the performance of multimodal diagnostic models against a clinical testing cohort.
• Analyzing the predictive power of specific biomarkers like procalcitonin (PCT) and serum amyloid A (SAA) for lymph node disorders.
• Developing risk estimation tools, such as nomograms, for individualized patient assessment in clinical settings.

Strengths

• Includes data from 203 patients with a clear class distribution (114 lymphoma, 89 benign).
• Model performance is quantitatively reported, with the combined model achieving an AUC of 0.958 in the independent testing cohort.
• Integrates multiple data modalities: PET/CT radiomics features and specific blood-based biomarkers.

Limitations

• Row count and column-level documentation are unknown, which may limit suitability assessment and require manual inspection after download.
• The dataset is small at 12.2 KB, indicating limited scope and feature count.
• Data is from a retrospective single-center study, which may introduce selection bias.

Provenance

Source

figshare, author Fenglian Jing.

Collection Method

Retrospective study of patients with fever of unknown origin who underwent PET/CT; radiomics features were extracted using LIFEx software.

Freshness

Last updated 2026-04-21 05:35:27.