Introduction: Risk stratification is the cornerstone of management for newly diagnosed non-metastatic prostate cancer. Here we developed and tested a new risk stratification system using the number of individual risk factors and the ISUP 2014 prognostic scores.
Patients and methods: Diagnostic data was derived from 10,139 UK men divided into a training (n=6226) and testing set (n=4113). An external validation cohort (n=1706) was also used. Risk groups were first assigned as low, intermediate and high (NICE 2014) then sub-stratified by the number of risk factors and grade re-classified using the new ISUP scores. A 5 strata criteria was produced and prognostic power compared against the NICE criteria with prostate cancer specific mortality (PCSM) as the outcome.
Results: In both training and testing sets the new classification identified a very low-risk group (Group 1), a subgroup of intermediate-risk cancers with a low PCSM risk (Group 2, HR 1.62[0.97-2.75]) and a further subgroup with an increased PCSM risk (Group 3, HR 3.35 [2.04-5.59]) (p<0.0001). High-risk cancers were also sub-classified into a better and worse outcome group: Group 4 (HR 5.03 [3.25-7.80]) and Group 5 (HR 17.28 [11.2-26.67]) (p<0.0001). In comparison to NICE,the new criteria demonstrated improved prognostic performance (Concordance index 0.75 (95% CI 0.72-0.77) versus 0.67 (95% CI 0.64-0.69) (p<0.0001). This was recapitulated in an external cohort (Concordance index of 0.83 (95% CI 0.80-0.87) for predicting PCSM versus 0.67 (95% CI 0.64-0.70) for the NICE criteria.
Conclusion: A novel 5 strata risk classification out-performs the NICE criteria in predicting the risk of PCSM.
Introduction: Risk stratification is the cornerstone of management for newly diagnosed non-metastatic prostate cancer. Here we developed and tested a new risk stratification system using the number of individual risk factors and the ISUP 2014 prognostic scores.
Patients and methods: Diagnostic data was derived from 10,139 UK men divided into a training (n=6226) and testing set (n=4113). An external validation cohort (n=1706) was also used. Risk groups were first assigned as low, intermediate and high (NICE 2014) then sub-stratified by the number of risk factors and grade re-classified using the new ISUP scores. A 5 strata criteria was produced and prognostic power compared against the NICE criteria with prostate cancer specific mortality (PCSM) as the outcome.
Results: In both training and testing sets the new classification identified a very low-risk group (Group 1), a subgroup of intermediate-risk cancers with a low PCSM risk (Group 2, HR 1.62[0.97-2.75]) and a further subgroup with an increased PCSM risk (Group 3, HR 3.35 [2.04-5.59]) (p<0.0001). High-risk cancers were also sub-classified into a better and worse outcome group: Group 4 (HR 5.03 [3.25-7.80]) and Group 5 (HR 17.28 [11.2-26.67]) (p<0.0001). In comparison to NICE,the new criteria demonstrated improved prognostic performance (Concordance index 0.75 (95% CI 0.72-0.77) versus 0.67 (95% CI 0.64-0.69) (p<0.0001). This was recapitulated in an external cohort (Concordance index of 0.83 (95% CI 0.80-0.87) for predicting PCSM versus 0.67 (95% CI 0.64-0.70) for the NICE criteria.
Conclusion: A novel 5 strata risk classification out-performs the NICE criteria in predicting the risk of PCSM.