Prostate Health Index Scores for Early-Stage Prostate Cancer in a South African Cohort

A clinical dataset from a prospective cohort study of 29 men with early-stage prostate cancer at a South African academic hospital. It contains quantified serum concentrations of total PSA, free PSA, and [-2]proPSA used to derive Prostate Health Index (PHI) scores. The research by Sabelo Manzini, last updated in May 2026, evaluated the PHI's diagnostic accuracy for classifying cancer aggressiveness.

Use Cases

• Validate the Prostate Health Index (PHI) as a standalone diagnostic biomarker based on the described tPSA, fPSA, and [-2]proPSA measurements.
• Build predictive models for prostate cancer aggressiveness using derived PHI scores.
• Compare the performance of biomarker-only models against models incorporating clinicopathological variables.
• Study diagnostic methodologies for prostate cancer within a South African clinical context.

Strengths

• Data is from a prospective clinical cohort study, suggesting a structured collection process.
• The PHI-only diagnostic model achieved a reported specificity of 80% and an AUC of 72.9%.
• Dataset is openly licensed under CC-BY-4.0.

Limitations

• The dataset is very small, with only 29 patient records.
• Column-level documentation is absent; field semantics must be inferred after download.
• Row count is unknown, which may limit suitability assessment.

Provenance

Source

Sabelo Manzini via figshare.

Collection Method

Prospective cohort study analyzing serum samples with a Beckman Coulter Access 2 Immunoassay Analyser.

Freshness

Last updated 2026-05-15 10:34:06; freshness should be verified.

Geography

South Africa (single academic hospital cohort).