Differential Diagnosis of Testicular Cancers with Enhanced CT Features, 2012-2024

Zhixu Zhang's retrospective study compares 15 cases of testicular embryonal rhabdomyosarcoma (ERMS) and 30 cases of testicular seminoma using enhanced CT features and laboratory indicators. The dataset includes CT attenuation measurements across non-enhanced, arterial, and venous phases, as well as patient age and tumor marker levels like AFP, LDH, and β-HCG. Data was collected from December 2012 to May 2024 and analyzed to identify diagnostic features, with results showing a model achieving an AUC of 0.961.

Use Cases

• Train a binary classifier to differentiate testicular ERMS from seminoma based on CT attenuation values and patient age.
• Build a feature importance model to identify which CT signs, such as the vascular ball sign, are most predictive of cancer type.
• Analyze the correlation between laboratory indicators (AFP, LDH, β-HCG) and specific tumor characteristics visible on CT scans.
• Validate diagnostic models using the reported sensitivity (86.7%) and specificity (96.7%) metrics from the ROC curve analysis.

Strengths

• Includes data from 45 patients with pathologically confirmed diagnoses, providing a verified ground truth.
• Contains multi-phase CT measurements (non-enhanced, arterial, venous) and key laboratory indicators for each case.
• Statistical analysis, including ROC curve results with an AUC of 0.961, is documented, indicating a strong diagnostic model.

Limitations

• Row count is unknown, which may limit suitability assessment for machine learning tasks.
• Column-level documentation is absent; field semantics must be inferred after download from the PDF.
• The dataset is small in scale (70.8 KB), reflecting a limited patient cohort of 45 individuals.

Provenance

Source

Zhixu Zhang via figshare.

Collection Method

Retrospective collection of pathologically confirmed cases, with CT images retrieved, reviewed, and reanalyzed.

Time Range

December 2012 to May 2024.

Freshness

Last updated 2026-06-03 04:47:14; freshness should be verified.