Related Experiment Videos
Predicting life expectancy in men with clinically localized prostate cancer
Mark E Cowen1, Lakshmi K Halasyamani, Michael W Kattan
1Quality Institute, St. Joseph Mercy Hospital, Ann Arbor, Michigan; Department of Medicine, St. Joseph Mercy Hospital, Ann Arbor, Michigan, USA. cowenm@trinity-health.org
The Journal of Urology
|January 13, 2006
Summary
Accurately predicting life expectancy for localized prostate cancer is crucial for treatment decisions. A new prediction rule showed moderate accuracy, aiding patient-clinician discussions but not solely determining management.
Area of Science:
- Oncology
- Biostatistics
- Geriatric Medicine
Background:
- Localized prostate cancer management requires accurate life expectancy estimation.
- Existing prediction rules for life expectancy have unknown accuracy.
- Prostate cancer treatment decisions are influenced by patient prognosis.
Purpose of the Study:
- To develop a new prediction rule for life expectancy in localized prostate cancer.
- To evaluate the accuracy of existing life expectancy prediction rules.
- To improve prognostic accuracy for guiding treatment strategies.
Main Methods:
- Retrospective cohort study of 506 men with localized prostate cancer (1987-1989).
- Development of a nomogram using proportional hazards regression.
- Assessment of prediction rule accuracy using c-statistic and calibration curves.
Main Results:
- Developed a new prediction rule with a c-statistic of 0.73.
- External rules (Albertsen, Tewari et al) showed comparable discrimination (c-statistics 0.71, 0.70).
- New rule and Tewari's rule approximated actual survival; Albertsen's rule underestimated survival.
Conclusions:
- Life expectancy in localized prostate cancer can be predicted with moderate accuracy.
- Predictions are sufficient for patient-clinician discussions.
- Life expectancy prediction alone is inadequate for determining optimal management.