If you have comments or features you'd like to see added to the app, please contact us at starcap.app@umich.edu.

Purpose

Our staging model is for patients diagnosed with prostate cancer who have not yet begun treatment. We predict the long-term chances of dying from prostate cancer after standard treatments including surgical removal of the prostate gland or curative radiation therapy with or without hormonal therapy. In external validation, this system outperformed similar prediction models such as NCCN, AJCC 8th edition, and CAPRA.

Outputs

In addition to providing predictions from the STAR-CAP model, we also summarize patient characteristics and provide the NCCN risk group for ease of clinical use. NOTE: we have collapsed the Very Low/Low NCCN risk group and the Very High/High NCCN risk group because our model does not use the predictors needed to distinguish between Very Low and Low risk or Very High and High risk. Patients grouped as being NCCN Low risk may be NCCN Low or Very Low risk, and similarly with High and Very High.

Methods

The models were built based on a collection of over 40,000 prostate cancer patients diagnosed and treated in North America and Europe from 1992-2013 as part of the International Staging Collaboration for Cancer of the Prostate (STAR-CAP) cohort. The variables include age, clinical T-stage, clinical N-stage, primary and secondary Gleason score based on biopsy, percent positive cores (percent of total biopsy cores positive/total biopsy cores), and pretreatment serum prostate specific antigen. The model given here provides clinical prognostic stage grouping (Stage IA-IIIC).

Patients input:

• Age

• Clinical T-stage

• Clinical N-stage

• Gleason grade

• Number of positive cores

• Number of negative cores

• PSA

Based on these inputs, we assign patients a Clinical Prognostics Staging Group between IA-IIIC, and provide mortality and survival predictions based on that stage.

Stage Assignment

For each risk factor, patients receive a number of points as follows:

These points are then summed. The final sum is grouped into stages as follows:

Model Building and Validation

The STAR-CAP data were collected from seven North American and European cancer centers, the SEARCH collaborative, and the CaPSURE registry. The final cohort included patients with prostate cancer treated with curative intent between 1992 and 2013 with follow-up through Dec. 31, 2017. The STAR-CAP data were split 50/50 into training and validation datasets, and follow-up was administratively censored at 15 years. A Fine-Gray regression model with age, T-stage, Gleason grade, percent positive cores, and PSA was fit in the training sample of 9,915 patients, with non-cancer death treated as a competing risk. Regression splines were used to develop cutpoints for continuous covariates, and each group was assigned a point value, to develop the point system given above.

The model was validated in the validation portion of the STAR-CAP data (n = 9,769), and additionally in patients in SEER (n = 125,575). In the STAR-CAP validation data, our model had a 10-year C-index of 0.796, which outperformed the AJCC 8th edition staging system and the NCCN 3- and 4-tier classification systems, as well as CAPRA. In the SEER validation data, our model had a 5-year C-index of 0.838.

References

Dess RT, Suresh K, Zelefsky MJ, Freedland SJ, Mahal BA, Cooperberg MR, Davis BJ, Horwitz EM, Terris MK, Amling CL, Aronson WJ, Kane CJ, Jackson WC, Hearn JWD, DeVille C, DeWeese TL, Greco S, McNutt TR, Song DY, Sun Y, Mehra R, Kaffenberger SD, Morgan TM, Nguyen PL, Feng FY, Sharma V, Tran PT, Stish BJ, Pisansky TM, Zaorsky NG, Moraes FY, Berlin A, Finelli A, Fossati N, Gandaglia G, Briganti A, Carroll PR, Karnes RJ, Kattan MW, Schipper MJ and Spratt DE. 2020. Development and Validation of a Clinical Prognostic Stage Group System for Non-Metastatic Prostate Cancer: Disease Specific Mortality Results from the International Staging Collaboration for Cancer of the Prostate (STAR-CAP). JAMA Oncology.

---

Disclaimer: The content provided does not provide medical advice. By proceeding, you acknowledge that viewing or use of this content does not create a medical professional-patient relationship, and does not constitute an opinion, medical advice, professional service or treatment recommendation of any condition. Content provided is for educational purposes only. The information and Content provided are not substitutes for medical or professional care, and you should not use the information in place of a visit, call, consultation or the advice of your physician or other healthcare provider. You are liable or responsible for any advice, course of treatment, or any other information, services that are based on Content through this site.