Development of the STRATified CANcer Surveillance protocol for men with favourable-risk prostate cancer
BAUS ePoster online library. Light A. 06/21/21; 319061; p2-3
Disclosure(s): Nil
Dr. Alexander Light
REGULAR CONTENT
Login now to access Regular content available to all registered users.
Abstract
Discussion Forum (0)
Rate & Comment (0)
Introduction
For men with favourable-risk prostate cancer on active surveillance (AS), there is little evidence to inform more personalised AS strategies. We previously reported a three-tier model for predicting progression to unfavourable-intermediate risk cancer (Table 1), denoted by Cambridge Prognostic Group 3 or higher (≥CPG3). We aimed to validate and improve this model.
Patients & Methods
The STRAtified CANcer Surveillance (STRATCANS) model was retested in an AS cohort from a tertiary UK centre (Table 1). We studied the association between risk group and progression to ≥CPG3 disease with Cox regression. Six further categorical variables were then fitted with STRATCANS groups using backward selection, and a new model (STRATCANS+) derived and retested. Discrimination was assessed using 5-year C-index.
Results
588 men were eligible. The 3-tier STRATCANS model had good discrimination (C-index 0.721, 95% confidence interval (CI) 0.666-0.776), and defined distinct risk groups (Group 2 hazard ratio (HR) 2.89, 95% CI 1.77-4.72, p<0.001; Group 3 HR 7.02, 95% CI 4.15-11.86, p<0.001; pairwise log rank all p<0.001). 5-year progression rates were 4.6%, 13.7%, and 30.9% for Groups 1-3, respectively. Next, PI-RADS score and core positivity were added to form a four-tier STRATCANS+ model (Table 1). Each new tier represented a distinct risk group (pairwise log rank all p<0.004; Fig.1), and STRATCANS+ improved discrimination (C-index 0.741, 95% CI 0.690-0.792). 5-year progression rates were 0.8%, 7.6%, 16.0% and 30.9% for Groups 1-4, respectively.
Conclusions
The STRATCANS model demonstrates good discrimination for predicting progression in men on AS, and performance improved with MRI and biopsy data.
For men with favourable-risk prostate cancer on active surveillance (AS), there is little evidence to inform more personalised AS strategies. We previously reported a three-tier model for predicting progression to unfavourable-intermediate risk cancer (Table 1), denoted by Cambridge Prognostic Group 3 or higher (≥CPG3). We aimed to validate and improve this model.
Patients & Methods
The STRAtified CANcer Surveillance (STRATCANS) model was retested in an AS cohort from a tertiary UK centre (Table 1). We studied the association between risk group and progression to ≥CPG3 disease with Cox regression. Six further categorical variables were then fitted with STRATCANS groups using backward selection, and a new model (STRATCANS+) derived and retested. Discrimination was assessed using 5-year C-index.
Results
588 men were eligible. The 3-tier STRATCANS model had good discrimination (C-index 0.721, 95% confidence interval (CI) 0.666-0.776), and defined distinct risk groups (Group 2 hazard ratio (HR) 2.89, 95% CI 1.77-4.72, p<0.001; Group 3 HR 7.02, 95% CI 4.15-11.86, p<0.001; pairwise log rank all p<0.001). 5-year progression rates were 4.6%, 13.7%, and 30.9% for Groups 1-3, respectively. Next, PI-RADS score and core positivity were added to form a four-tier STRATCANS+ model (Table 1). Each new tier represented a distinct risk group (pairwise log rank all p<0.004; Fig.1), and STRATCANS+ improved discrimination (C-index 0.741, 95% CI 0.690-0.792). 5-year progression rates were 0.8%, 7.6%, 16.0% and 30.9% for Groups 1-4, respectively.
Conclusions
The STRATCANS model demonstrates good discrimination for predicting progression in men on AS, and performance improved with MRI and biopsy data.
Introduction
For men with favourable-risk prostate cancer on active surveillance (AS), there is little evidence to inform more personalised AS strategies. We previously reported a three-tier model for predicting progression to unfavourable-intermediate risk cancer (Table 1), denoted by Cambridge Prognostic Group 3 or higher (≥CPG3). We aimed to validate and improve this model.
Patients & Methods
The STRAtified CANcer Surveillance (STRATCANS) model was retested in an AS cohort from a tertiary UK centre (Table 1). We studied the association between risk group and progression to ≥CPG3 disease with Cox regression. Six further categorical variables were then fitted with STRATCANS groups using backward selection, and a new model (STRATCANS+) derived and retested. Discrimination was assessed using 5-year C-index.
Results
588 men were eligible. The 3-tier STRATCANS model had good discrimination (C-index 0.721, 95% confidence interval (CI) 0.666-0.776), and defined distinct risk groups (Group 2 hazard ratio (HR) 2.89, 95% CI 1.77-4.72, p<0.001; Group 3 HR 7.02, 95% CI 4.15-11.86, p<0.001; pairwise log rank all p<0.001). 5-year progression rates were 4.6%, 13.7%, and 30.9% for Groups 1-3, respectively. Next, PI-RADS score and core positivity were added to form a four-tier STRATCANS+ model (Table 1). Each new tier represented a distinct risk group (pairwise log rank all p<0.004; Fig.1), and STRATCANS+ improved discrimination (C-index 0.741, 95% CI 0.690-0.792). 5-year progression rates were 0.8%, 7.6%, 16.0% and 30.9% for Groups 1-4, respectively.
Conclusions
The STRATCANS model demonstrates good discrimination for predicting progression in men on AS, and performance improved with MRI and biopsy data.
For men with favourable-risk prostate cancer on active surveillance (AS), there is little evidence to inform more personalised AS strategies. We previously reported a three-tier model for predicting progression to unfavourable-intermediate risk cancer (Table 1), denoted by Cambridge Prognostic Group 3 or higher (≥CPG3). We aimed to validate and improve this model.
Patients & Methods
The STRAtified CANcer Surveillance (STRATCANS) model was retested in an AS cohort from a tertiary UK centre (Table 1). We studied the association between risk group and progression to ≥CPG3 disease with Cox regression. Six further categorical variables were then fitted with STRATCANS groups using backward selection, and a new model (STRATCANS+) derived and retested. Discrimination was assessed using 5-year C-index.
Results
588 men were eligible. The 3-tier STRATCANS model had good discrimination (C-index 0.721, 95% confidence interval (CI) 0.666-0.776), and defined distinct risk groups (Group 2 hazard ratio (HR) 2.89, 95% CI 1.77-4.72, p<0.001; Group 3 HR 7.02, 95% CI 4.15-11.86, p<0.001; pairwise log rank all p<0.001). 5-year progression rates were 4.6%, 13.7%, and 30.9% for Groups 1-3, respectively. Next, PI-RADS score and core positivity were added to form a four-tier STRATCANS+ model (Table 1). Each new tier represented a distinct risk group (pairwise log rank all p<0.004; Fig.1), and STRATCANS+ improved discrimination (C-index 0.741, 95% CI 0.690-0.792). 5-year progression rates were 0.8%, 7.6%, 16.0% and 30.9% for Groups 1-4, respectively.
Conclusions
The STRATCANS model demonstrates good discrimination for predicting progression in men on AS, and performance improved with MRI and biopsy data.
Code of conduct/disclaimer available in General Terms & Conditions
{{ help_message }}
{{filter}}