Improving Clinical Risk Stratification at Diagnosis in Primary Prostate Cancer: A Prognostic Modelling Study
View abstract on PubMed
Summary
This summary is machine-generated.A new five-stratum risk stratification system for prostate cancer significantly improves the prediction of prostate cancer-specific mortality (PCSM) compared to the current three-stratum system. This enhanced model aids in better treatment selection for men with non-metastatic disease.
Area Of Science
- Urology
- Oncology
- Pathology
Background
- Over 80% of prostate cancer diagnoses are localized or locally advanced non-metastatic disease.
- Current risk stratification (PSA, Gleason grade, clinical stage) shows significant outcome heterogeneity.
- The International Society of Urological Pathology (ISUP) prognostic score is a recent development not yet integrated into risk stratification.
Purpose Of The Study
- To develop and test a novel five-stratum risk stratification system for non-metastatic prostate cancer.
- To incorporate the ISUP prognostic score into a new risk stratification model.
- To compare the prognostic performance of the new system against the current standard system.
Main Methods
- Utilized clinicopathological data from 10,139 men with non-metastatic prostate cancer, divided into training and testing sets.
- Developed a new system based on individual risk factors and the ISUP prognostic score.
- Compared prognostic power using Cox hazards model, log-rank test, Kaplan-Meier curves, competing-risks regression, and concordance indices.
Main Results
- The new five-stratum system identified distinct subgroups with significantly different risks of prostate cancer-specific mortality (PCSM) (p < 0.0001).
- The new system demonstrated improved prognostic performance with a higher concordance index (0.75) compared to the current system (0.69) in the training set.
- External validation showed superior performance of the new system (concordance index 0.79) versus the current system (0.66).
Conclusions
- A novel five-stratum risk stratification system effectively predicts PCSM in non-metastatic prostate cancer, outperforming the current three-stratum system.
- This new model allows for better identification of clinically relevant subgroups for tailored therapy.
- Further validation in external datasets and exploration of additional variables are warranted.