SUO 2017: Can Free PSA Be Used As A Biomarker in Biochemical Recurrence To Predict Castrate Resistant Prostate Cancer?

Washington, DC ( While prostate specific antigen (PSA) has long been established as a serum screening tool for prostate cancer, and has often come under fire for its utility, there has been growing interesting in PSA in its various forms as a marker for prostate cancer. With creation of the 4K score and the prostate health index (PHI), we have gained better understanding of the role of the PSA complexes and their relationship to total PSA levels. 

Most serum PSA is complexed to proteases, but 5%-45% of PSA exists as enzymatically inactive free PSA (fPSA). PSA produced from prostate cancer (PC) cells appears to escape proteolytic processing, resulting in a greater fraction of complexed PSA and a lower %fPSA. While previously evaluated in the role of prostate cancer screening, the role of fPSA in biochemical recurrence (BCR) after radical prostatectomy (RP) is not known.

The authors of this study utilize their institutional large tertiary-care center database to evaluate the role of fPSA in assessing for BCR following definitive therapy with RP. All patients in the last decade who had BCR after RP and had at least one fPSA blood test were included. Patients were stratified according the %fPSA cut-off of 0.15 (or 15%). 

81 men with BCR after RP were identified, of whom 41 patients (50.6%) had %fPSA < 0.15 (group 1) and 40 patients (49.4%) had %fPSA >0.15 (group 2). Median age and age adjusted Charlson score were similar between group 1 (61.4 [7.1], 3.8 [0.75]) and group 2 (63.2 [7.6], 4.1 [0.88]), p=0.285 and p=0.143, respectively. While mean PSA at diagnosis was similar between the two groups (10.7 and 8.6, p=0.234), %fPSA was lower in group 1 (0.121 and 0.156, p=0.042). 

Mean prostate volume at RP was also lower in group 1 (45.8 ml vs. 57.7 ml, p=0.012). Importantly, 23% of the patients with %fPSA > 0.15 (group 2) had adjuvant radiotherapy compared to only 5% of those in Group 1 (p = 0.043). Median follow-up time was 104 and 95.6 months for group 1 and 2, respectively, p=0.44. 

Interestingly, 20% (group 1) vs. 60% (group 2) become castrate resistant (CRPC), and the time to reach CRPC state was much shorter in group 2 (33.5 months) vs. group 1 (57.9 months), p=0.05. Additionally, 60% of group 2 patients vs. 32.5% of group 1 patients developed metastasis, p=0.014. Lastly, Kaplan Meier curve demonstrated median survival of 193 months for group 2 patients, while median survival was not reached for patients in group 1 (p=0.023). On multivariable logistic regression analysis, %fPSA > 0.15 predicted CRPC status (OR 7.99, CI 95% 2-31.95, p=0.003).

Based on these results, it would appear that %fPSA does have implications for the risk of progression to CRPC and time to progression. Higher %fPSA (>0.15) in the setting of BCR after RP confers a higher risk of more aggressive disease. This manifests in a faster development of CRPC, metastasis and death. 

This is interesting, because it is associated with a reversal in the significance of % fPSA values in BCR patients compared to screened patients, and should be validated in larger cohorts.

Limitations / Discussion Points:

1. Group 1 had lower %fPSA even at the time of initial diagnosis. It would have been more relevant to compare patients who had similar %fPSA prior to RP but different %fPSA after RP.

2. Current study excludes patients treated with EBRT as primary definitive therapy. It will be interesting to see if the same findings hold up for patients with BCR after EBRT.

3. 3% of the patients with %fPSA > 0.15 (group 2) had adjuvant radiotherapy compared to only 5% of those in Group 1. This may account for some of the differences – biased towards higher risk groups?

Presented by: Hanan Goldberg

Co-Authors: Ally Hoffman, Zachary Klaassen, Thenappan Chandrasekar, Douglas Cheung, Alejandro Berlin, Rashid Sayyid, Neil Fleshner

Institution: Princess Margaret Cancer Center, UHN, Toronto, Ontario, Canada

Written by: Thenappan Chandrasekar, MD, Clinical Fellow, University of Toronto, twitter: @tchandra_uromd at the 18th Annual Meeting of the Society of Urologic Oncology, November 20-December 1, 2017 – Washington, DC