STAR-CAP as an advanced prognostic method for localized prostate cancer

31 Dec 2020

Prostate cancer was the third most common cancer among males in Hong Kong in 2018, accounting for 12.9% of total cancer cases and 5.5% of total cancer deaths.1 For decades, the anatomical extent or stage of cancer has been classified using the tumor-node-metastasis (TNM) system, but not for localized prostate cancer.2 Instead, the non-anatomical factors [Gleason grade and Prostate Specific Antigen (PSA) level] have been incorporated into the American Joint Committee on Cancer (AJCC) 8th edition cancer staging manual for the localized prostate cancer staging.2 But the current staging system is constructed from discovery data sets which may not be reflective and diversified enough, therefore, it is not an ideal model to be generalized to all kinds of population.3 Recently, a group of international researchers has published a new staging system for localized prostate cancer, the International Staging Collaboration for Cancer of the Prostate (STAR-CAP) model, that can provide a more accurate treatment decision than the current system.1,2

Combining the anatomical factors from Group I-IV and TNM grouping, the current cancer staging system approved by the AJCC 8th edition has only achieved a 10-year C index of 0.757 (95% CI: 0.719-0.792).2 This non-ideal scoring can be partly attributed to the end point used by the current staging system for prostate cancer, i.e. mortality, which is highly suspectable to errors from post-treatment recurrence and non-prostate cancer-specific mortality.3 In addition, the AJCC 8th edition mainly focuses on metastatic prostate cancer and has only approved 2 out of 15 prostate cancer prognostic models as official models.3-5

To improve the current cancer staging system, the STAR-CAP model was developed and validated by analyzing over 19,000 patients suffering from cT1-4N0-1M0 prostate adenocarcinoma that were treated from January 1, 1992 to December 31, 2013.2 The population was randomized into data training set and data validation set for analysis.2 The scoring criteria included T category, clinical N category, age, pre-treatment PSA level, percentage of positive core biopsy result, and Gleason score.2 Ethnicity was not included as a parameter for data analysis but was included for model performance examination.2 To generate a higher specificity towards prostate cancer, prostate cancer-specific mortality instead of general mortality was taken as the end point of the STAR-CAP model.3 The model was then validated using the Surveillance, Epidemiology and End Result (SEER) cohort as a second validation set.2,3

When compared to the AJCC 8th edition cancer staging system, STAR-CAP had a better performance across age, race and treatment methods with a 10-year C index of 0.796 (95% CI: 0.760-0.828).2 Instead of 8 categorial groups, the STAR-CAP model has also increased the group number to 9 for a more refined and accurate description towards each prostate cancer stage.2 Clinically, a STAR-CAP score can be easily generated using an interactive program which can help predict the 5 and 10-year prostate cancer-specific mortality corresponding to the localized prostate cancer stage.6 This information can then be shared with the patients and allows an easier mutual decision-making process between clinicians and patients.6

However, the STAR-CAP system is limited by ethnicity as only the North Americans and Europeans were included in the study.2 As a result, the efficacy of STAR-CAP in staging localized prostate cancer patients in Hong Kong remains uncertain. A similar study for Hong Kong patients should be done to validate its effectiveness for local use. By utilizing a more accurate staging system, clinicians can then compose a more individualized treatment plan for patients and improve their quality of life during the treatment journey.

Get access to our exclusive articles.
Related Articles