Misop Han, Alan W. Partin, Marianna Zahurak, Steven Piantadosi, Jonathan I. Epstein and Patrick C. Walsh

Introduction

The Han tables were developed by urologists Misop Han, M.D., Alan W. Partin, M.D., Ph.D., and Patrick C. Walsh, M.D., based on accumulated data from thousands of patients who had been treated for prostate cancer at the James Buchanan Brady Urological Institute at The Johns Hopkins Hospital. After using the Partin tables to predict the definitive pathological stage (the extent of cancer spread), men and their doctors may want to know the probability of recurrence following surgery (radical prostatectomy). The Han tables were designed to predict the probability of the first evidence of recurrence (detectable PSA level) after surgery.

Similar to the Partin tables, the Han tables correlate the three common factors known about a man’s prostate cancer, PSA level, Gleason score, and clinical stage (or pathological stage). While the Partin tables are used to predict pathological stage, the Han tables are used to predict the probability of prostate cancer recurrence up to 10 years following surgery. Based on the result of the probability of recurrence, men and their doctors can choose the best course of treatment.

Which model should I use?

If you are considering surgery, but have not had surgery yet, you want to use the preoperative model.

If you have had surgery already, you want to use the postoperative model.

1. Preoperative Model

• For men who are considering surgery for prostate cancer, but have not had surgery yet
• Prediction of recurrence probability following surgery using the available information BEFORE the surgery (PSA level, biopsy Gleason score, and clinical stage) 

2. Postoperative Model

• For men who have had surgery for prostate cancer
• Prediction of recurrence probability following surgery using the available information BEFORE and AFTER the surgery (PSA level, surgical Gleason score and pathological stage)

Knowing the probability of recurrence following surgery would help patients rationally choose appropriate treatment (either primary and/or adjuvant therapy) for prostate cancer. The tables will be updated soon to combine information from prostate cancer patients across multiple institutions.

Statistical Methods to Build Biochemical Recurrence Prediction Models

Available preoperative parameters for multivariate analysis were age, clinical TNM stage, Gleason score from biopsy specimen (categorical), PSA (categorically divided into 4 or less, 4.1 to 10, 10.1 to 20 and greater than 20 ng/ml) and acid phosphatase level. Available postoperative variables were organ confinement, focal extraprostatic extension, extensive extraprostatic extension, lymph node involvement, seminal vesicle invasion, surgical margin status and Gleason score from surgical specimen.

Actuarial analysis was performed comparing freedom from biochemical recurrence after radical retropubic prostatectomy (PSA 0.2 ng./ml. or greater). Patients were censored if they were lost to followup or there was no recurrence. Event time distributions for the time to recurrence end point were estimated with the Kaplan-Meier method and compared using the log rank statistic or the proportional hazards regression model. The simultaneous effect of two or more factors was studied using the multivariate proportional hazards model. Covariates and interactions marginally significant (p <0.19) in univariate analyses were entered into the multivariate regression model and insignificant effects were removed in a stepwise fashion. The first model was developed using preoperative variables only and the second model using all available variables.

Early detection programs encouraged in the later years of this study produced significant shifts towards early stage disease over time. Interactions of PSA, Gleason score and TNM stage with calendar time were tested to determine if the risks attributable to these variables were also functions of time. For all analyses, PSA, Gleason score and TNM stage were categorized into discrete levels and entered in regression models in unfactored form. When calendar year of surgery was entered into the proportional hazards model, a restricted cubic spline transformation was used to allow nonlinearity.

In addition to the shift toward early stage disease, the relative risk of biochemical recurrence following surgery decreased significantly over time. Most importantly, when the relative risk of biochemical recurrence was adjusted for clinical TNM stage, preoperative PSA and Gleason score, there was still a significant decrease in relative risk of biochemical recurrence over time.

Observed and predicted recurrence-free survival curves for two proportional hazards models (one stratified for year of surgery and one including year of surgery as predictor) and three parametric survival models (Weibull, log-normal and gamma) were compared to select a model for calculation of predicted recurrence-free probabilities and confidence intervals. For each model coefficients from the multivariate regression were used to generate a prognostic factor score for each individual. This score is the weighted average of prognostic factor values with weights determined by the estimated coefficients from the model. The proportional hazards regression model is the patient’s hazard relative to a patient with the most favorable level of all prognostic factors. These scores were then ranked and categorized to form 4 risk groups. Plots of the observed survival, Kaplan-Meier curves, for these risk groups were then compared to the model of predicted recurrence-free survival for each risk group (data not shown). From the chosen model (proportional hazards), the nomograms were constructed from biochemical recurrence-free survival probability with corresponding 95 percent confidence intervals at three, five, seven and 10 years following radical retropubic prostatectomy, adjusting for the latest year in which surgical data were available (1999). All p values are two-sided. All statistical analyses were performed using the Intercooled Stata 6.0 statistics and graphics data management system (Stata Corporation, College Station, Texas), SAS system (SAS Institute, Inc., Cary, N.C.), EGRET (Cytel Statistical Software, Cambridge, Mass.) or the S-plus Design package (Mathsoft Data Analysis Products Division, Seattle, Wash.).

Discussion

Our models are based on the followup information from a large group of men who underwent radical retropubic prostatectomy. All five variables integrated in the final model are commonly used and readily reproducible for men who are considering or have already undergone radical retropubic prostatectomy. We excluded from analysis men with clinical stage T1a/b or T3a disease or a Gleason score of less than 5 since the percentage of men with those diseases has decreased significantly in the contemporary patient population. Our nomograms provide biochemical recurrence-free survival probability, which is easier to explain to patients.

The most important distinction of our models is that we integrated a significant downward stage migration and an improved surgical outcome over time into the models. We have previously demonstrated the decreasing relative risk of biochemical recurrence following surgery in the modern era. That change may reflect the benefits of early detection, better preoperative selection of patients for surgery as well as lead time bias. In the current study we attempted to delineate whether downward stage migration alone could account for the improved therapeutic outcome over time. When the relative risk of biochemical recurrence was adjusted for clinical TNM stage, preoperative PSA and Gleason score, there was still a significant decrease in relative risk of biochemical recurrence over time. Since patients have a decreasing relative risk of biochemical recurrence over time, our nomograms were generated for men who underwent radical retropubic prostatectomy at the latest year of followup (1999). Therefore, these nomograms can be applied to men with clinically localized prostate cancer who underwent or plan to undergo surgery in the modern era.

A man with newly diagnosed prostate cancer has to make important decisions regarding treatment. The Partin tables enabled physicians and patients to make more informed treatment decisions based on the probability of pathological stage for clinically localized prostate cancer. When a patient experiences biochemical recurrence following radical retropubic prostatectomy, the study by Pound et al can be informative as well as comforting regarding the interval from PSA detection to evidence of metastasis. They demonstrated that disease progression from an isolated PSA increase to metastasis and cancer-specific mortality is generally a protracted process. The algorithm in their study provided the risk for developing metastatic cancer so that patients and physicians could decide on the need for and timing of the most appropriate adjuvant therapy following postoperative biochemical recurrence.

In addition to the Partin tables and the Pound algorithm, the tables in the present study can help patients rationally decide on the best treatment options depending on the probability of recurrence-free survival following radical prostatectomy according to disease characteristics in the modern era. In addition, these tables can guide physicians caring for men with prostate cancer in determining how often and what type of monitoring tests should be performed following surgery. Finally, the nomograms can help physicians determine whether adjuvant therapy may be beneficial or which patients would benefit from novel adjuvant therapy.

Conclusions

We reviewed a large series of men who underwent radical prostatectomy for clinically localized prostate cancer to identify indicators of biochemical recurrence. Using three preoperative or postoperative variables, we developed multivariate proportional hazards models to determine the three-, five-, seven- and 10-year biochemical recurrence-free survival probabilities among men who undergo radical prostatectomy for clinically localized prostate cancer. These nomograms were adjusted for the decreasing relative risk of biochemical recurrence over time. They may be helpful in guiding treatment decisions for men who are considering or have undergone radical prostatectomy for clinically localized prostate cancer in the modern era.

References

1. M. Han, A.W. Partin, M. Zahurak, S. Piantadosi, J.I. Epstein, and P.C. Walsh, Biochemical (prostate specific antigen) recurrence probability following radical prostatectomy for clinically localized prostate cancer, J Urol 169(2003), p. 157

2. C.R. Pound, A.W. Partin, M.A. Eisenberger, D.W. Chan, J.D. Pearson and P.C. Walsh, Natural history of progression after PSA elevation following radical prostatectomy, JAMA 281 (1999), p. 1591.

3. C.R. Pound, A.W. Partin, J.J. Epstein and P.C. Walsh, Prostate-specific antigen after anatomic radical retropubic prostatectomy. Patterns of recurrence and cancer control, Urol Clin North Am 24 (1997), p. 395.

4. W.J. Catalona and D.S. Smith, 5-Year tumor recurrence rates after anatomical radical retropubic prostatectomy for prostate cancer, J Urol 152(1994), p. 1837.

5. W.J. Catalona and D.S. Smith, Cancer recurrence and survival rates after anatomic radical retropubic prostatectomy for prostate cancer: intermediate-term results, J Urol 160 (1998), p. 2428.

6. M. Ohori, J.R. Goad, T.M. Wheeler, J.A. Eastham, T.C. Thompson and P.T. Scardino, Can radical prostatectomy alter the progression of poorly differentiated prostate cancer?, J Urol 152 (1994), p. 1843.

7. J.G. Trapasso, J.B. de Kernion, R.B. Smith and F. Dorey, The incidence and significance of detectable levels of serum prostate specific antigen after radical prostatectomy, J Urol 152 (1994), p. 1821.

8. H. Zincke, J.E. Oesterling, M.L. Blute, E.J. Bergstralh, R.P. Myers and D.M. Barrett, Long-term (15 years) results after radical prostatectomy for clinically localized (stage T2c or lower) prostate cancer, J Urol 152 (1994), p. 1850.

9. M. Han, A.W. Partin, C.R. Pound, J.I. Epstein and P.C. Walsh, Long term biochemical disease-free and cancer-specific survival following anatomic radical retropubic prostatectomy: the 15-year Johns Hopkins experience. In: H. Lepor, Editor, Urologic Clinics of North America, W. B. Saunders Co., Philadelphia (2001), pp. 555–565.

10. M. Han, A.W. Partin, S. Piantadosi, J.I. Epstein and P.C. Walsh, Era specific biochemical recurrence-free survival following radical prostatectomy for clinically localized prostate cancer, J Urol 166 (2001), p. 416.

11. O.H. Beahrs, D.E. Henson, R.V.P. Hutter and M.H. Myers, American Joint Committee on Cancer Staging Manual (4th ed.), J. B. Lippincott Co., Philadelphia (1992), pp. 181–186.

12. A.W. Partin, S. Piantadosi, M.G. Sanda, J.I. Epstein, F.F. Marshall and J.L. Mohler et al., Selection of men at high risk for disease recurrence for experimental adjuvant therapy following radical prostatectomy, Urology 45(1995), p. 831.

13. A.W. Partin, C.R. Pound, J.Q. Clemens, J.I. Epstein and P.C. Walsh, Serum PSA after anatomic radical prostatectomy. The Johns Hopkins experience after 10 years, Urol Clin North Am 20 (1993), p. 713.

14. E.L. Kaplan and P. Meier, Nonparametric estimation from incomplete observations, J Am Stat Assoc 53 (1958), p. 457.

15. D.R. Cox, Regression models and life-tables, J R Stat Soc Ser B 34 (1972), p. 187.

16. W.G. Reiner and P.C. Walsh, An anatomical approach to the surgical management of the dorsal vein and Santorini's plexus during radical retropubic surgery, J Urol 121 (1979), p. 198.

17. P.C. Walsh and P.J. Donker, Impotence following radical prostatectomy: insight into etiology and prevention, J Urol 128 (1982), p. 492.

18. J.L. Stanford, R.A. Stephenson, L.M. Coyle, M.A. Carol Kosary, J. Cerhan and R. Correa et al., Prostate Cancer Trends 1973–1995, SEER Program, National Institutes of Health, National Cancer Institute, Bethesda (1999).

19. W.W. Roberts, E.J. Bergstralh, M.L. Blute, J.M. Slezak, M. Carducci and M. Han et al., Contemporary identification of patients at high risk of early prostate cancer recurrence after radical retropubic prostatectomy, Urology57 (2001), p. 1033.

20. M.W. Kattan, J.A. Eastham, A.M. Stapleton, T.M. Wheeler and P.T. Scardino, A preoperative nomogram for disease recurrence following radical prostatectomy for prostate cancer, J Natl Cancer Inst 90 (1998), p. 766.

21. M.W. Kattan, T.M. Wheeler and P.T. Scardino, Postoperative nomogram for disease recurrence after radical prostatectomy for prostate cancer, J Clin Oncol 17 (1999), p. 1499.

22. C.L. Amling, M.L. Blute, E.J. Bergstralh, T.M. Seay, J. Slezak and H. Zincke, Long-term hazard of progression after radical prostatectomy for clinically localized prostate cancer: continued risk of biochemical failure after 5 years, J Urol 164 (2000), p. 101.

23. X.J. Yang, K. Lecksell, S.R. Potter and J.I. Epstein, Significance of small foci of Gleason score 7 or greater prostate cancer on needle biopsy, Urology54 (1999), p. 528.

24. A.W. Partin, M.W. Kattan, E.N. Subong, P.C. Walsh, K.J. Wojno and J.E. Oesterling et al., Combination of prostate-specific antigen, clinical stage, and Gleason score to predict pathological stage of localized prostate cancer. A multi-institutional update, JAMA 277 (1997), p. 1445.