The Impact of the Percent of Residual Prostate-Specific Antigen on Metastasis-Free Survival in Patients with Persistent Prostate-Specific Antigen after Radical Prostatectomy

Radical prostatectomy (RP) can provide good long-term oncological outcomes in patients with localized and locally advanced prostate cancer (PCa) [1, 2]. RP aims to fully remove the tumor burden, and the goal following surgery is for levels of prostate-specific antigen (PSA) to be undetectable [3]. Undetectable levels of PSA (<0.1 ng/mL) after RP means there is no biochemically active tumor burden. Conversely, persistent levels of PSA (≥0.1 ng/mL) mean that there is a residual biochemically active tumor burden, regardless of imaging findings [4]. Therefore, persistent PSA is a poor prognostic factor for recurrence after RP [4, 5, 6, 7, 8, 9, 10, 11].

Several investigators have reported the impact of persistent PSA levels after RP on oncological outcomes [4, 6, 7, 8, 9, 10, 11]. However, most studies have focused on persistent PSA levels (≥0.1 ng/mL) only and not on the accurate level of PSA. Within 15 years, approximately 75% and 47% of patients with persistent PSA levels after RP develop biochemical recurrence (BCR) and metastasis, respectively [7, 11]. These patients have heterogeneous progression patterns. Therefore, we should stratify the risk of progression in this population to determine the appropriate salvage treatment.

In men with persistent PSA levels after RP, presence of a measurable residual tumor on conventional imaging, such as computed tomography/magnetic resonance imaging, and bone scan, is uncommon. Thus, we need another test to measure the residual tumor burden in addition to imaging tests. Theoretically, levels of PSA after RP may indicate residual tumor burden because PCa cells express PSA [3, 12]. Therefore, measurement of postoperative PSA levels may provide more information regarding PCa tissue burden.

Here, we describe the relationship between persistent levels of PSA at 4 to 8 weeks after RP and oncological outcomes within a single tertiary referral hospital database. To stratify the risk of progression in men with persistent PSA levels, we investigated the impact of the first postoperative PSA (fPSA) level and the percentage of residual PSA (%rPSA) [(postoperative PSA/preoperative PSA)×100] on the metastasis-free survival (MFS) rate.

Levels of PSA represent the cornerstone of the follow-up for patients who have undergone RP. In particular, early PSA values after RP could help predict worse oncological outcomes. Previous studies have shown poor oncological outcomes based on the presence of PSA persistence measured 4 to 8 weeks after RP [5, 6, 7, 8, 9, 10, 11]. However, most of these studies evaluated PSA persistence itself, with no analysis based on PSA values. If early PSA values after RP indicate the burden of residual tumor, the prognosis will vary according to its value. Moreover, it would be helpful if we could measure the burden of residual tumor to determine if salvage treatment should be administered, such as salvage radiation.

PSA is secreted by epithelial cells of the prostate gland and PCa tissues [3, 12]. The half-life of PSA is 3.15 days and should reach an undetectable level within 4 weeks in patients who have undergone complete pathological resection [12]. As a result, persistently detectable PSA levels 4 to 8 weeks after RP indicate either residual PCa, residual benign tissue, recurrence in the prostatic bed, distant micro-metastasis, or a combination of these [4].

To the best of our knowledge, this is the first study to investigate the impact of %rPSA after RP on oncological outcomes. In the present study, fPSA levels and %rPSA were evaluated as indicators of residual tumors. Unfortunately, a simple fPSA level was not a predictor of distant metastasis. However, %rPSA was well established in this study as an independent predictor of MFS in patients with persistent PSA levels after RP. In particular, the MFS rate differed between patients with a %rPSA ≥6% and those with a %rPSA <6% groups (log-rank test, p<0.001).

The fPSA level has already been evaluated as a predictor of cancer-specific mortality [13]. Although the fPSA level was a predictor of cancer-specific mortality in the overall cohort with RP, it was not observed in men with persistent PSA levels after RP [13]. Therefore, the fPSA level has limited usefulness. We need to clarify the relationship between residual tumor cells in aggressive and benign cancers. The %rPSA reflects these characteristics. Therefore, it could be an ideal indicator for predicting metastatic progression.

A recent meta-analysis reported that approximately 12% of patients treated with RP experienced PSA persistence 4 to 8 weeks after RP [4]. A recent German study reported that 8.8% of patients had persistent PSA levels 6 weeks after RP [11]. The incidence of PSA persistence in our study was 14.9%. The reason for the increased incidence of these populations in our study is that our data included more aggressive cancers. The incidence of pGGG 4–5 in our study was 17.2% vs. 59.0% in patients with undetectable vs. persistent PSA levels after RP. However, those in the German data were 3.3% and 21.6%, respectively [11]. The incidence of pT3b in our study was 8.2% vs. 52.2% in patients with undetectable vs. persistent PSA levels after RP. However, the incidences in the German data were 8.1% vs. 45.2%, respectively [11]. In addition, we measured the PSA levels at 4 to 6 weeks after RP in most patients. In previous studies, PSA levels were measured 4 to 8 weeks after RP. Measuring PSA levels at an earlier time point may result in an insufficient reduction in the PSA level. Conversely, Bianchi et al [8] reported that the incidence of persistent PSA levels after RP was 26%. The reason for the high incidence of persistent PSA levels in that study was due to the high proportion of lymph node metastasis. In an earlier French study, 34.6% of their cohort showed persistently elevated PSA levels after RP [6].

Previous studies have demonstrated that the prognosis of men with PSA persistence is not invariably poor. Recently, long-term follow-up data of 11,604 patients who underwent RP were reported [11]. Fifteen years after RP, the MFS rate, overall survival, and cancer-specific survival were 53.0% vs. 93.2% (p<0.001), 64.7% vs. 81.2% (p<0.001), and 75.5% vs. 96.2% (p<0.001) for persistent vs. undetectable PSA levels, respectively [11]. In multivariable Cox regression models, a persistent level of PSA was an independent predictor of metastasis (HR, 3.59; p<0.001), death (HR, 1.86; p<0.001), and cancer-specific death (HR, 3.15; p<0.001) [11]. In a recent meta-analysis, PSA persistence was associated with BCR (HR, 4.44; 95% CI, 2.84–6.93), disease recurrence (HR, 3.43; 95% CI, 1.62–7.25), and cancer-specific mortality (HR, 2.32; 95% CI, 1.83–2.95) [4]. Our data showed concordant results that persistent PSA levels after RP was associated with MFS (HR, 3.94; p=0.010).

Additional treatment should be considered for patients with persistent levels of PSA after RP to control the residual tumor [8, 11, 13, 14]. To select appropriate candidates for adjuvant or salvage treatment, additional risk factors must be considered, such as pT3b, pGGG 3–5, positive surgical margin, or pN1 disease [8, 11, 13, 14]. Preisser et al [11] demonstrated that salvage radiation was associated with improved overall survival (HR, 0.37; p=0.02) and cancer-specific survival (HR, 0.12; p<0.01). Most studies support the benefit of salvage radiation in patients with persistent levels of PSA after RP [8, 13, 14]. Fossati et al [14] developed an accurate risk stratification tool to facilitate the individualized recommendation for eSRT based on PCa characteristics. Among patients with persistent levels of PSA after RP, eSRT administration provided better MFS only in patients with a Gleason score ≤7. Patients with a Gleason score ≥8 did not appear to benefit from eSRT, as these patients had a relatively constant high rate of metastases regardless of the PSA level when they received radiation.

The current study has several limitations. First, we included retrospective data, excluding patients from the analysis if complete information was not available. This may have created a selection bias. Second, the median follow-up duration was relatively short. Third, the proportion of patients with pelvic lymph node dissection was low in our institution. Finally, we used only computed tomography/magnetic resonance imaging and bone scan to measure the residual tumor after RP. Prostate-specific membrane antigen positron emission tomography is the most recommended test for patients with low PSA levels in these days [15].

Persistent levels of PSA were associated with worse MFS after RP. However, the fPSA level after RP was not a predictor of distant metastasis in patients with persistent levels of PSA. To our knowledge, this is the first study to demonstrate that %rPSA is an independent predictor of MFS in patients with persistent levels of PSA after RP. Further, in men with persistent levels of PSA after RP, biochemical residual tumors can be measured by %rPSA. Therefore, we hypothesize that %rPSA may improve the currently available prognostic models to help physicians decide whether to use a salvage treatments, such as radiation therapy, after RP for patients with PCa and scheduling follow-up patient counseling.