ABSTRACT
Background
Outcomes of metastatic castration‐resistant prostate cancer (mCRPC) treated with androgen receptor signaling inhibitors (ARSIs) vary widely and may not be fully reflected by PSA and standard clinicopathologic factors. We evaluated the prognostic value of serum neuron‐specific enolase (NSE) at baseline and early on‐treatment dynamics.
Methods
We retrospectively analyzed ARSI‐treated mCRPC patients with available baseline NSE (n = 120). NSE was dichotomized by the upper limit of normal (ULN, 12.5 ng/mL). Radiographic progression‐free survival (rPFS) and overall survival (OS) were estimated by Kaplan–Meier methods and compared by log‐rank test. Multivariable Cox models adjusted for age, baseline PSA (per 100 ng/mL), line of therapy (2–4 vs 1), Gleason score (> 7 vs ≤ 7), metastatic timing (synchronous vs metachronous), and visceral metastases (present vs absent). NSE dynamics were evaluated using a 28‐day landmark approach among patients alive and at risk at day 28 with available baseline and week‐4 NSE (rPFS landmark n = 47; OS landmark n = 48); follow‐up was redefined from day 28 and dynamics were modeled as log2(NSE4w/NSE0) (per doubling).
Results
Median follow‐up was 44.7 months (95% CI 35.7–48.3). Baseline NSE > ULN was associated with shorter rPFS (9.4 vs 20.8 months) and OS (23.5 vs 38.2 months). In multivariable analyses, baseline NSE > ULN remained independently associated with inferior outcomes (rPFS HR 1.81, 95% CI 1.15–2.85, p = 0.011; OS HR 1.85, 95% CI 1.12–3.05, p = 0.017). In landmark models, higher NSE dynamics were independently associated with worse outcomes (rPFS HR 4.30, 95% CI 1.06–17.39, p = 0.041; OS HR 8.78, 95% CI 2.15–35.77, p = 0.002).
Conclusions
In ARSI‐treated mCRPC, baseline NSE above ULN and early 4‐week NSE dynamics provide independent prognostic information for rPFS and OS, supporting NSE as a pragmatic adjunct to clinical risk stratification.
Keywords: androgen receptor signaling inhibitors, biomarker, castration‐resistant prostate cancer, neuron‐specific enolase, prognosis
1. Introduction
Prostate cancer represents a worldwide health concern and is one of the most commonly diagnosed cancers in men worldwide [1]. Metastatic castration‐resistant prostate cancer (mCRPC) is a clinically and therapeutically challenging form of the disease, characterized by progression due to acquired resistance to androgen deprivation therapy (ADT) at castrate testosterone levels [2, 3]. Despite considerable advancements in novel systemic therapies for mCRPC, patient outcomes remain heterogeneous and reliable prognostic stratification is still needed in routine clinical practice.
Next‐generation androgen receptor signaling inhibitors (ARSIs), including abiraterone acetate (ABI) and enzalutamide (ENZ), constitute key treatment options in mCRPC and have demonstrated clinically meaningful improvements in disease control and survival in randomized phase III trials [4, 5]. However, treatment responses vary substantially, reflecting biological diversity, and de‐novo or acquired drug resistance. In mCRPC, tumors develop reduced dependence on androgen receptor (AR) signaling and acquire features of lineage plasticity, including neuroendocrine differentiation, which has been associated with aggressive clinical behavior and poor outcomes [6, 7]. In this context, circulating biomarkers reflecting non–AR‐driven biology may provide important complementary prognostic information beyond conventional biomarkers mainly represented by prostate‐specific antigen (PSA). Notably, post‐treatment changes in PSA levels have not been shown to have a robust association with survival of mCRPC patients and have not been qualified as an endpoint to support regulatory approval [8, 9, 10]. Neuron‐specific enolase (NSE), a glycolytic enzyme expressed in neuroendocrine cells, is widely used as a serum marker in neuroendocrine malignancies and has been suggested as a blood‐based indicator of neuroendocrine features in prostate cancer, albeit with imperfect specificity [6, 7]. A growing body of evidence suggests that serum NSE is a surrogate prognostic biomarker in prostate cancer [11]. Notably, its role in mCRPC patients treated with ARSIs remains underexplored.
The aim of the present retrospective study was to evaluate the prognostic value of baseline serum NSE and its early dynamics in patients with mCRPC receiving ABI or ENZ.
2. Patients and Methods
2.1. Study Design
Data from mCRPC patients receiving ABI or ENZ were retrospectively analyzed. We assessed the association between outcome of patients and serum NSE at baseline and 4 weeks after treatment initiation. Clinical data were extracted from the hospital information system.
2.2. Patients and Treatment
Patients with histologically confirmed mCRPC were treated between 2007 and 2026 at the Department of Oncology and Radiotherapeutics, University Hospital in Pilsen, Czech Republic. ABI (Zytiga, Janssen Pharmaceuticals Co., Beerse, Belgium) was administered orally in the standard approved schedule (1000 mg daily) in combination with prednisone (10 mg daily). ENZ (Xtandi, Astellas Pharma Inc., Tokyo, Japan) was administered orally in the standard approved dose (160 mg daily). The therapy was continued until disease progression, unacceptable toxicity, or patient refusal.
2.3. Outcome Assessment
The routine clinical checks including physical examination and biochemical laboratory tests were performed every 4 weeks. Radiographic controls using computed tomography (CT) or positron emission tomography‐CT (PET/CT) or PET‐magnetic resonance (PET/MR) were performed every three to 6 months. The radiographic progression was assessed according to the Response Evaluation Criteria in Solid Tumors (RECIST) [12]. Radiographic progression‐free survival (rPFS) was defined as the time from the initiation of ARSI to the progression confirmed by the aforementioned radiographic assessment or death, whichever occurred first. Overall survival (OS) was defined as time from ARSI initiation to death from any cause. Patients without an event were censored at last follow‐up.
2.4. Assessment of Serum NSE and PSA
Peripheral blood was drawn using VACUETTE Z Serum Sep tubes (Greiner Bio‐One, Kremsmünster, Austria) and allowed to clot. Serum was separated within 3 h of blood collection and analyzed. Serum NSE was measured using immunoradiometric titration method (IRMA) (Beckman‐Immunotech, USA), and ACCESS chemiluminescent assay (Beckman Coulter, Brea, CA, US) was used for the measurement of serum PSA.
2.5. Variables and Definitions
Baseline NSE was defined as the serum NSE value obtained prior to ARSI initiation (NSE0) and dichotomized using the upper limit of normal (ULN) 12.5 ng/mL (NSE‐high > 12.5 vs NSE‐low ≤ 12.5). An on‐treatment NSE value at approximately 4 weeks (NSE4w) was recorded, and early NSE dynamics was expressed as the ratio NSE4w/NSE0 (computed only when both values were available and > 0). For regression modeling, this ratio was log2‐transformed, such that hazard ratios (HR) can be interpreted per doubling of the week‐4‐to‐baseline NSE ratio. Covariates considered for multivariable adjustment were age at treatment initiation (years), baseline PSA (per 100 ng/mL), line of therapy (2–4 vs 1), Gleason score (> 7 vs ≤ 7), synchronous vs metachronous metastases, and visceral metastases (present vs absent).
2.6. Statistical Analysis
Standard descriptive statistics and frequencies were used to characterize the dataset. Group comparisons used the Mann–Whitney U test for continuous variables and χ² or Fisher's exact test for categorical variables, as appropriate. Kaplan–Meier estimates were compared by log‐rank test; median follow‐up was calculated using reverse Kaplan–Meier. Cox proportional hazards models were used to estimate hazard ratios (HRs) with 95% confidence intervals (CIs). The multivariable Cox model for both rPFS and OS included baseline NSE (ULN‐based high vs low) and age (per 1 year), PSA (per 100 ng/mL), line of therapy (2–4 vs 1), Gleason score (> 7 vs ≤ 7), synchronous metastases, and visceral metastases. To evaluate NSE dynamics while avoiding immortal time bias, a 28‐day landmark analysis was performed among patients with available baseline and ~4‐week NSE. Time‐to‐event was redefined from day 28, and the landmark multivariable Cox model additionally included NSE dynamics as the week‐4‐to‐baseline NSE ratio (NSE4w/NSE0), modeled on the log2 scale to yield hazard ratios per doubling of the ratio. For visualization of NSE dynamics, unadjusted Kaplan–Meier curves were plotted using a median split of NSE4w/NSE0 ratio and compared by log‐rank test; and adjusted survival curves were derived from the landmark multivariable Cox model using the same median split, holding other covariates at typical values (median/mode) and displaying 95% CIs for adjusted estimates. Two‐sided p‐values < 0.05 were considered statistically significant. Analyses were performed in Python (v3.11.2; Python Software Foundation, Wilmington, DE, USA), with pandas (v2.2.3), statsmodels (v0.14.3), and Matplotlib (v3.7.5).
3. Results
3.1. Patient Cohort
A total of 120 patients with mCRPC treated with ABI or ENZ were included. Baseline characteristics for the overall cohort and stratified by baseline NSE (≤ 12.5 ng/mL vs > 12.5 ng/mL) are presented in Table 1. Only patients with ECOG PS 0 or 1 were included. Median follow‐up (reverse Kaplan–Meier) was 44.7 months (95% CI 43.9–48.3) overall and was comparable between the NSE ≤ 12.5 group (44.5 months, 95% CI 41.7–48.3) and the NSE > 12.5 group (45.8 months, 95% CI 44.4–57.4).
Table 1.
Baseline clinical and demographic characteristics at initiation of treatment with androgen receptor signaling inhibitors (ARSIs).
| Characteristic | Overall (n = 120) | NSE ≤ 12.5 ng/mL (n = 76) | NSE > 12.5 ng/mL (n = 44) |
|---|---|---|---|
| Demographics | |||
| Age at treatment initiation, years (median) | 73.5 (67.9–78.9) | 72.7 (68.5–78.0) | 74.8 (67.0–79.4) |
| Baseline biomarkers | |||
| PSA before treatment, ng/mL (median) | 17.9 (4.7–68.6) | 13.1 (4.6–55.6) | 34.1 (5.7–138.0) |
| NSE before treatment, ng/mL (median) | 11.5 (9.6–13.5) | 10.1 (9.2–11.2) | 14.3 (13.4–16.7) |
| Disease characteristics | |||
| Gleason score | |||
| ≤ 7 | 52 (43.3%) | 32 (42.1%) | 20 (45.5%) |
| > 7 | 65 (54.2%) | 42 (55.3%) | 23 (52.3%) |
| Unknown | 3 (2.5%) | 2 (2.6%) | 1 (2.3%) |
| Synchronous metastases | |||
| No | 72 (60.0%) | 46 (60.5%) | 26 (59.1%) |
| Yes | 43 (35.8%) | 26 (34.2%) | 17 (38.6%) |
| Visceral metastases | |||
| No | 92 (76.7%) | 59 (77.6%) | 33 (75.0%) |
| Yes | 27 (22.5%) | 16 (21.1%) | 11 (25.0%) |
| Bone metastases | |||
| No | 17 (14.2%) | 13 (17.1%) | 4 (9.1%) |
| Yes | 103 (85.8%) | 63 (82.9%) | 40 (90.9%) |
| Lung metastases | |||
| No | 22 (18.3%) | 12 (15.8%) | 10 (22.7%) |
| Yes | 97 (80.8%) | 63 (82.9%) | 34 (77.3%) |
| Treatment | |||
| Previous prostatectomy | |||
| No | 90 (75.0%) | 57 (75.0%) | 33 (75.0%) |
| Yes | 30 (25.0%) | 19 (25.0%) | 11 (25.0%) |
| Previous radiotherapy | |||
| No | 63 (52.5%) | 41 (53.9%) | 22 (50.0%) |
| Yes | 57 (47.5%) | 35 (46.1%) | 22 (50.0%) |
| Line of treatment with ARSI | |||
| 1 | 90 (75.0%) | 58 (76.3%) | 32 (72.7%) |
| 2 | 26 (21.7%) | 15 (19.7%) | 11 (25.0%) |
| 3 | 2 (1.7%) | 2 (2.6%) | 0 (0.0%) |
| 4 | 0 (0.0%) | 0 (0.0%) | 0 (0.0%) |
| Unknown | 2 (1.7%) | 1 (1.3%) | 1 (2.3%) |
| ARSI type | |||
| Enzalutamide | 76 (63.3%) | 45 (59.2%) | 31 (70.5%) |
| Abiraterone acetate | 44 (36.7%) | 31 (40.8%) | 13 (29.5%) |
Abbreviations: ARSI, androgen receptor signaling inhibitor; NSE, neuron‐specific enolase; PSA, prostate‐specific antigen.
3.2. Baseline Serum NSE and Survival Outcomes
Patients with baseline NSE above the ULN (12.5 ng/mL) experienced significantly worse outcomes than those with NSE ≤ 12.5 ng/mL. Median rPFS was 20.8 months (95% CI 12.7–28.4) in the NSE ≤ 12.5 group vs 9.4 months (95% CI 5.6–17.2) in the NSE > 12.5 group (log‐rank p = 0.009; Figure 1). Median OS was 38.2 months (95% CI 30.1–50.7) vs 23.5 months (95% CI 18.6–27.9), respectively (log‐rank p= 0.003; Figure 1). In the multivariable Cox regression model baseline NSE above ULN remained an independent predictor of both rPFS (HR 1.551, p = 0.041) and OS (HR 1.672, p = 0.031) (Table 2).
Figure 1.
Kaplan–Meier curves for radiographic progression‐free survival (rPFS) and overall survival (OS) according to baseline serum neuron‐specific enolase (NSE) using the upper limit of normal (ULN, 12.5 ng/mL) cut‐off. Shaded areas indicate 95% confidence intervals; p‐values are from the log‐rank test.
Table 2.
Univariable and multivariable Cox proportional hazards models for radiographic progression‐free survival (rPFS) and overall survival (OS) according to baseline serum neuron‐specific enolase (NSE; ULN‐based high vs low, ULN 12.5 ng/mL) and clinical covariates.
| Variable | rPFS | OS | ||||
|---|---|---|---|---|---|---|
| HR (95% CI) | p‐value | Events | HR (95% CI) | p‐value | Events | |
| Age at ARSI initiation (per 1 yr increase) | 0.98 (0.96–1.01) | 0.287 | 84 | 0.99 (0.96–1.02) | 0.594 | 69 |
| Baseline PSA (per 100 ng/mL increase) | 1.10 (0.99–1.23) | 0.067 | 84 | 1.13 (1.02–1.25) | 0.025 | 69 |
| Line of treatment with ARSI (2–4 vs 1) | 1.78 (1.09–2.91) | 0.021 | 84 | 1.52 (0.88–2.63) | 0.132 | 69 |
| Gleason score (> 7 vs ≤ 7) | 0.92 (0.59–1.45) | 0.726 | 84 | 1.02 (0.61–1.70) | 0.936 | 69 |
| Metastases (synchronous vs metachronous) | 1.00 (0.62–1.62) | 0.996 | 84 | 1.10 (0.64–1.89) | 0.723 | 69 |
| Visceral metastases (present vs absent) | 1.00 (0.58–1.71) | 0.996 | 84 | 0.82 (0.44–1.53) | 0.542 | 69 |
| Baseline NSE > 12.5 ng/mL vs ≤ 12.5 ng/mL (ULN) | 1.81 (1.15–2.85) | 0.011 | 84 | 1.85 (1.12–3.05) | 0.017 | 69 |
Abbreviations: ARSI, androgen receptor signaling inhibitor; CI, confidence interval; HR, hazard ratio; NSE, neuron‐specific enolase; OS, overall survival; PSA, prostate specific antigen; rPFS, radiographic progression‐free survival; ULN, upper limit of normal.
3.3. Early Serum NSE Dynamics and Survival Outcomes
In a 28‐day landmark analysis including patients with available baseline and 4‐week serum NSE values (rPFS landmark n = 47; OS landmark n = 48), a higher week‐4‐to‐baseline NSE ratio (NSE4w/NSE0) was independently associated with inferior outcomes (modeled as log2(NSE4w/NSE0), yielding hazard ratios per doubling of the ratio: rPFS HR 4.30, 95% CI 1.06–17.39, p = 0.041; OS HR 8.78, 95% CI 2.15–35.77, p = 0.002; Table 3). For visual presentation, survival from the 4‐week landmark is shown using unadjusted Kaplan–Meier curves based on a median split of the NSE ratio and compared by log‐rank test (Figure 2A,B), and adjusted survival curves derived from the landmark multivariable Cox model using the same median split, with 95% confidence intervals and the Wald p‐value for the NSE‐dynamics term (Figure 2C,D). Because follow‐up in the landmark analysis begins at week 4 and includes only patients alive and at risk at the landmark with available 4‐week NSE, effect estimates from landmark models are not directly comparable to those from baseline models.
Table 3.
Landmark (week 4) univariable and multivariable Cox proportional hazards models for rPFS and OS evaluating on‐treatment NSE dynamics (NSE4w/NSE0) while adjusting for baseline NSE (ULN‐based high vs low) and clinical covariates. Time‐to‐event was redefined from week 4 among patients alive and at risk with available week‐4 NSE.
| Variable | rPFS | OS | ||||
|---|---|---|---|---|---|---|
| HR (95% CI) | p‐value | n/events | HR (95% CI) | p‐value | n/events | |
| Age at ARSI initiation (per 1 year increase) | 0.95 (0.89–1.01) | 0.102 | 43/33 | 1.00 (0.94–1.06) | 0.971 | 44/33 |
| Baseline PSA (per 100 ng/mL increase) | 2.35 (1.17–4.71) | 0.016 | 43/33 | 2.94 (1.38–6.27) | 0.005 | 44/33 |
| Line of treatment with ARSI (2–4 vs 1) | 2.07 (0.64–6.72) | 0.228 | 43/33 | 6.64 (2.00–22.05) | 0.002 | 44/33 |
| Gleason score (> 7 vs ≤ 7) | 1.29 (0.59–2.83) | 0.520 | 43/33 | 3.04 (1.17–7.91) | 0.023 | 44/33 |
| Metastases (synchronous vs metachronous) | 0.67 (0.28–1.60) | 0.371 | 43/33 | 1.00 (0.43–2.32) | 0.994 | 44/33 |
| Visceral metastases (present vs absent) | 0.79 (0.27–2.33) | 0.666 | 43/33 | 1.72 (0.59–4.98) | 0.318 | 44/33 |
| Baseline NSE > 12.5 ng/mL vs ≤ 12.5 ng/mL (ULN) | 4.71 (1.91–11.60) | < 0.001 | 43/33 | 3.97 (1.68–9.35) | 0.002 | 44/33 |
| NSE change at 4 weeks (NSE4w/NSE0)* | 4.30 (1.06–17.39) | 0.041 | 43/33 | 8.78 (2.15–35.77) | 0.002 | 44/33 |
Abbreviations: ARSI, androgen receptor signaling inhibitor; CI, confidence interval; HR, hazard ratio; NSE, neuron‐specific enolase; OS, overall survival; PSA, prostate specific antigen; rPFS, radiographic progression‐free survival; ULN, upper limit of normal.
Effect per doubling (log2‐transformed).
Figure 2.
Week‐4 landmark analyses of NSE dynamics. (A, B) Unadjusted Kaplan–Meier curves for rPFS and OS stratified by NSE dynamics (median split of NSE4w/NSE0); shaded areas indicate 95% confidence intervals and p‐values are from the log‐rank test. (C, D) Adjusted rPFS and OS curves derived from the multivariable Cox landmark model including baseline NSE (ULN‐based high vs low), NSE dynamics, and clinical covariates; p‐values correspond to the Wald test for the NSE‐dynamics term.
4. Discussion
In this retrospective cohort of patients with mCRPC treated with ARSIs, baseline serum NSE above the ULN (12.5 ng/mL) was associated with inferior rPFS and OS and remained independently associated with outcomes after adjustment for established clinical factors, including baseline PSA. Moreover, early on‐treatment NSE dynamics provided additional prognostic information: in a 28‐day landmark analysis, a higher week‐4‐to‐baseline NSE ratio (NSE4w/NSE0) was independently associated with both shorter rPFS and OS. Collectively, these findings support the concept that serum NSE may capture clinically relevant aspects of tumor biology that are not fully reflected by PSA and traditional clinicopathological variables.
A plausible explanation is that elevated or rising NSE reflects neuroendocrine differentiation or a shift toward an AR‐indifferent phenotype that may emerge under therapeutic pressure on the AR axis [13, 14, 15]. Such states have been linked to aggressive clinical behavior and reduced sensitivity to AR‐directed treatment strategies, which would be consistent with the adverse outcomes observed in patients with high serum NSE. Neuroendocrine differentiation, as assessed in tumor tissue and suggested by circulating biomarkers, has been linked to prostate cancer prognosis. However, the present study does not directly examine neuroendocrine differentiation at the tissue level. Instead, it uses NSE as an easily accessible serum biomarker that may correlate with these biological processes. Previous studies investigating the prognostic relevance of serum NSE in metastatic prostate cancer suggested association between elevated pretreatment circulating neuroendocrine biomarkers and inferior survival outcomes in patients treated with ADT [16, 17, 18, 19, 20]. In addition, NSE has been evaluated alongside other circulating neuroendocrine biomarkers such as chromogranin A (CgA) across disease stages, including cohorts with mCRPC [20, 21, 22, 23]. In a retrospective study of 138 CRPC patients conducted by Hvamstad, elevated serum NSE and CgA were common among CRPC patients and provided prognostic information [20]. The available evidence on serum NSE specifically in mCRPC patients receiving ARSIs remains very limited and heterogeneous, with most prior studies being modest in size and often focusing on baseline measurements. A retrospective study by Heck et al focused on baseline serum NSE and CgA in a cohort of 45 mCRPC patients treated with ABI stratified in to low‐, intermediate‐ or high‐risk groups according to elevation of none, one, or both neuroendocrine serum biomarkers, respectively. They observed that the risk groups correlated with decreasing median OS (not reached vs 15.3 vs 6.6 months; p < 0.001) and decreasing median clinical or radiographic PFS (8.3 vs 4.4 vs 2.7 months; p = 0.001) [21]. Fan et al. performed a retrospective study with a similar design, including 40 mCRPC patients. They assessed baseline levels of and variations in serum NSE and CgA during treatment with ABI and found that elevated baseline CgA and NSE, as well as elevation during the first 3 months of treatment, were independent prognostic factors for OS (baseline: HR 13.71 (95% CI 1.03–183.31), p = 0.048; dynamics: HR 12.89 (95% CI 1.09–152.69), p = 0.043) and rPFS (baseline: HR 1.02 (95% CI 1.00–1.04), p = 0.041; dynamics: HR 12.89 (95% CI 1.44–24.99), p = 0.014) [22]. Notably, these studies were modest in size, used heterogeneous thresholds, and evaluated dynamics at later timepoints (e.g., 3 months), underscoring the need for additional data aligned with early routine follow‐up. Szarvas et al. reported the largest ABI/ENZ‐treated mCRPC cohort to date (n = 143) with systematic measurement of serum CgA and NSE. In their ABI/ENZ cohort, baseline NSE showed prognostic signal in univariable analyses (median split: HR 1.580, p = 0.030), and stronger associations were observed using higher thresholds (e.g., the previously published BJU cut‐off for NSE > 19 ng/mL: HR 3.457, p < 0.001) and serum biomarker combinations (≥ 1 elevated marker: HR 2.316, p = 0.001; both elevated: HR 4.216, p < 0.001). In multivariable analyses, however, the independent association was more consistent for CgA (ROC‐derived CgA > 81.29 ng/mL: HR 1.926, p = 0.047), while NSE effects attenuated (ROC‐derived NSE > 10.22 ng/mL: HR 1.239, p = 0.431; BJU cut‐off for NSE > 19 ng/mL: HR 1.984, p = 0.099). On‐treatment marker dynamics at 3 months were also more informative for CgA (> 50% increase associated with worse OS; p = 0.040), whereas NSE changes showed only a trend among patients with elevated baseline NSE (p = 0.061) [23]. However, serum CgA was not systematically available in the present cohort and therefore was not included in the current analyses. Consequently, we were unable to evaluate combined neuroendocrine biomarker models or directly compare the relative prognostic contribution of NSE and CgA. The present study was intentionally focused on NSE as a pragmatic and routinely accessible biomarker evaluated at baseline and during early on‐treatment follow‐up.
Taken together, these findings highlight both the potential and the remaining uncertainty regarding serum NSE in ARSI‐treated mCRPC, supporting the value of additional studies evaluating NSE using clinically grounded thresholds and early on‐treatment timepoints. In this context, our study adds incremental evidence by evaluating NSE both at baseline and as an early on‐treatment dynamic biomarker within a landmark framework, while accounting for standard clinical covariates including baseline PSA. Methodologically, the 28‐day landmark design used in the present study mitigates immortal time bias by restricting the dynamics analysis to patients alive and at risk at day 28 with available 4‐week NSE, thereby improving interpretability of post‐baseline biomarker changes [24, 25]. If validated, baseline and early NSE dynamics could help clinicians identify ARSI‐treated mCRPC patients at higher risk of early progression and death. This could prompt closer monitoring, earlier imaging, consideration of alternative systemic approaches, or earlier biopsy in cases suspicious for treatment‐related neuroendocrine prostate cancer (t‐NEPC) and/or aggressive variant prostate cancer (AVPC), especially when PSA kinetics appear discordant with clinical or radiographic behavior [14, 15]. However, NSE should not be interpreted in isolation: hemolysis and non‐malignant conditions can influence NSE levels, and neuroendocrine differentiation is heterogeneous and not always captured by a single circulating biomarker [11]. For clinical interpretation, NSE dynamics can be read directly as the week‐4‐to‐baseline ratio (NSE4w/NSE0), whereas the Cox model expresses this ratio on the log2 scale to yield hazard ratios per doubling. An important consideration is the evolving treatment landscape of advanced prostate cancer. In the contemporary upfront‐ARSI era, many patients receive ABI or ENZ already in the metastatic castration‐sensitive prostate cancer (mCSPC) setting, potentially influencing subsequent tumor biology, AR dependence, and emergence of neuroendocrine features at the mCRPC stage. Therefore, whether baseline NSE and early NSE dynamics retain similar prognostic value in ARSI‐treated mCSPC or in patients previously exposed to ARSIs during the mCSPC phase remains uncertain and warrants prospective validation. In this context, prior ARSI exposure may represent a clinically relevant variable beyond treatment line alone in future biomarker‐oriented studies.
The present study has several strengths and limitations. Key strengths include the clinically relevant population of ARSI‐treated mCRPC, the evaluation of both baseline neuroendocrine differentiation (NSE) and early on‐treatment change, and multivariable adjustment for major prognostic covariates. The limitations include the retrospective design, the relatively modest single‐center sample size (particularly for the landmark cohort, this might limit the generalizability of the dynamic analysis, given that NSE levels can be influenced by non‐malignant conditions), potential residual confounding factors, and the lack of systematic tissue confirmation of neuroendocrine differentiation and histological correlation at disease progression with confirmation of AVPCs. In addition, the absence of systematic CgA assessment prevented evaluation of combined neuroendocrine biomarker approaches described in previous studies.
5. Conclusions
The present retrospective study findings suggest that baseline NSE and early NSE dynamics are associated with outcomes in ARSI‐treated mCRPC, supporting NSE as a pragmatic adjunct for risk stratification pending prospective validation and integration with markers of lineage plasticity.
Author Contributions
O.F. designed the study; O.F., M.T., Z.V., M.H., J.K., P.S., V.L., K.P., J.F., and O.T. collected clinical data; O.F. wrote the manuscript with support from all authors; M.S. and G.L.B. performed critical review and editing; O.F. performed statistical analyses, tables and figures.
Ethics Statement
The study protocol and the informed consent form covering patient follow‐up, processing of clinical data, and publication of anonymized data were approved by the Ethics Committee of the Faculty of Medicine and University Hospital in Pilsen on June 7, 2021 (Approval No. 155/23), in accordance with the Declaration of Helsinki, International Ethical Guidelines for Biomedical Research, and applicable local regulations. All participants provided informed consent in accordance with institutional requirements.
Consent
All authors have approved the article for publication.
Conflicts of Interest
Ondrej Fiala received honoraria from Novartis, Janssen, Merck, Ipsen, BMS, MSD, Pierre Fabre, and Pfizer for consultations and lectures unrelated to this project.
Milan Hora received honoraria from Merck Sharp & Dohme for consultations and lectures unrelated to this project.
Giuseppe Luigi Banna reports personal fees from: advisory boards (Accord, AstraZeneca, Amgen, Merck); speaker bureaus (Astellas, AstraZeneca, Amgen, Bayer, Merck, Pfizer); patents (n. 4 patents with ST Microelectronics); travel, accommodations for scientific conferences (Accord, Merck, Janssen).
Matteo Santoni has received research support and honoraria from Janssen, Bristol Myers Squibb, Ipsen, MSD, Astellas and Bayer, all unrelated to the article.
Jindřich Fínek has received honoraria from Astra Zeneca, Roche, and Novartis for consultations and lectures unrelated to this project.
Other authors declare that they have no conflicts of interest that might be relevant to the contents of this article.
Acknowledgments
The authors would like to thank all patients who voluntarily took part in the study. This study was supported by the Charles University Research Fund (Cooperatio No. 43—Surgical Disciplines), the Institutional Research Fund of University Hospital Pilsen, I‐FNP‐10—urooncology II (669806) and by the European Union's Horizon 2020 research and innovation programme under grant agreement N856620 and by the project National Institute for Cancer Research—NICR (Programme EXCELES, ID Project No. LX22NPO5102)—Funded by the European Union—Next Generation EU. Funders played no role in the study design, data collection, analysis, interpretation of data or the writing of this article. Open access publishing facilitated by Univerzita Karlova, as part of the Wiley ‐ CzechELib agreement.
Fiala O., Tkadlecová M., Váchalová Z., et al., “Prognostic Role of Serum Neuron‐Specific Enolase at Baseline and its Early Dynamics in Metastatic Castration‐Resistant Prostate Cancer Treated With Androgen Receptor Signaling Inhibitors,” The Prostate 86 (2026): 1187‐1194, 10.1002/pros.70197.
Ondřej Topolčan and Jindřich Fínek contributed equally as senior authors.
Data Availability Statement
The data that support the findings of this study are available on request from the corresponding author. The data are not publicly available due to privacy or ethical restrictions.
The datasets generated and/or analyzed during the current study are not publicly available due to patient data security but are available from the corresponding author on reasonable request.
References
- 1. Bray F., Laversanne M., Sung H., et al., “Global Cancer Statistics 2022: Globocan Estimates of Incidence and Mortality Worldwide for 36 Cancers in 185 Countries,” CA: A Cancer Journal for Clinicians 74, no. 3 (2024): 229–263, 10.3322/caac.21834. [DOI] [PubMed] [Google Scholar]
- 2. European Association of Urology (EAU) . EAU Guidelines on Prostate Cancer: Follow‐up. EAU Guidelines. 2024 edition. Available from: EAU website, 2026.
- 3. European Association of Urology (EAU) . EAU Guidelines on Prostate Cancer: Treatment (CRPC Definition and Systemic Management). EAU Guidelines. 2024 edition. Available from: EAU website, 2026.
- 4. Ryan C. J., Smith M. R., de Bono J. S., et al., “Abiraterone in Metastatic Prostate Cancer Without Previous Chemotherapy,” New England Journal of Medicine 368, no. 2 (2013): 138–148, 10.1056/NEJMoa1209096. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5. Beer T. M., Armstrong A. J., Rathkopf D. E., et al., “Enzalutamide in Metastatic Prostate Cancer before Chemotherapy,” New England Journal of Medicine 371, no. 5 (2014): 424–433, 10.1056/NEJMoa1405095. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6. Terry S. and Beltran H., “The Many Faces of Neuroendocrine Differentiation in Prostate Cancer Progression,” Frontiers in Oncology 4 (2014): 60. 10.3389/fonc.2014.00060. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7. Parimi V., Goyal R., Poropatich K., and Yang X. J., “Neuroendocrine Differentiation of Prostate Cancer: A Review,” American Journal of Clinical and Experimental Urology 2, no. 4 (2014): 273–285. [PMC free article] [PubMed] [Google Scholar]
- 8. Collette L., Burzykowski T., and Schröder F. H., “Prostate‐Specific Antigen (PSA) Alone is not an Appropriate Surrogate Marker of Long‐Term Therapeutic Benefit in Prostate Cancer Trials,” European Journal of Cancer 42, no. 10 (2006): 1344–1350, 10.1016/j.ejca.2006.02.011. [DOI] [PubMed] [Google Scholar]
- 9. Scher H. I., Morris M. J., Basch E., and Heller G., “End Points and Outcomes in Castration‐Resistant Prostate Cancer: From Clinical Trials to Clinical Practice,” Journal of Clinical Oncology 29, no. 27 (2011): 3695–3704, 10.1200/JCO.2011.35.8648. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10. Halabi S., Armstrong A. J., Sartor O., et al., “Prostate‐Specific Antigen Changes as Surrogate for Overall Survival in Men With Metastatic Castration‐Resistant Prostate Cancer Treated With Second‐Line Chemotherapy,” Journal of Clinical Oncology 31, no. 31 (2013): 3944–3950, 10.1200/JCO.2013.50.3201. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11. Muoio B., Pascale M., and Roggero E., “The Role of Serum Neuron‐Specific Enolase in Patients With Prostate Cancer: A Systematic Review of the Recent Literature,” International Journal of Biological Markers 33, no. 1 (2018): 10–21. 10.5301/ijbm.5000286. [DOI] [PubMed] [Google Scholar]
- 12. Schwartz L. H., Litière S., de Vries E., et al., “RECIST 1.1—Update and Clarification: From the RECIST Committee,” European Journal of Cancer 62 (2016): 132–137, 10.1016/j.ejca.2016.03.081. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13. Farinea G., Calabrese M., Carfì F., et al., “Impact of Neuroendocrine Differentiation (NED) on Enzalutamide and Abiraterone Efficacy in Metastatic Castration‐Resistant Prostate Cancer (mCRPC): A Retrospective Analysis,” Cells 13, no. 16 (2024): 1396. 10.3390/cells13161396. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14. Merkens L., Sailer V., Lessel D., et al., “Aggressive Variants of Prostate Cancer: Underlying Mechanisms of Neuroendocrine Transdifferentiation,” Journal of Experimental and Clinical Cancer Research 41, no. 1 (2022): 46. 10.1186/s13046-022-02255-y. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15. Zhang Q., Han Y., Zhang Y., et al., “Treatment‐Emergent Neuroendocrine Prostate Cancer: A Clinicopathological and Immunohistochemical Analysis of 94 Cases,” Frontiers in Oncology 10 (2021): 571308. 10.3389/fonc.2020.571308. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16. Kamiya N., Akakura K., Suzuki H., et al., “Pretreatment Serum Level of Neuron Specific Enolase (NSE) as a Prognostic Factor in Metastatic Prostate Cancer Patients Treated With Endocrine Therapy,” European Urology 44, no. 3 (2003): 309–314, 10.1016/S0302-2838(03)00303-8. [DOI] [PubMed] [Google Scholar]
- 17. Berruti A., Dogliotti L., Mosca A., et al., “Circulating Neuroendocrine Markers in Patients With Prostate Carcinoma,” Cancer 88, no. 11 (2000): 2590–2597. 10.1002/1097-0142(20000601)88:11<2590::aid-cncr23>3.0.co;2-d. [DOI] [PubMed] [Google Scholar]
- 18. Isshiki S., Akakura K., Komiya A., Suzuki H., Kamiya N., and Ito H., “Chromogranin a Concentration as a Serum Marker to Predict Prognosis After Endocrine Therapy for Prostate Cancer,” Journal of Urology 167, no. 2 Pt 1 (2002): 512–515. 10.1016/S0022-5347(01)69075-X. [DOI] [PubMed] [Google Scholar]
- 19. Chuang C. K., Wu T. L., Tsao K. C., and Liao S. K., “Elevated Serum Chromogranin A Precedes Prostate‐Specific Antigen Elevation and Predicts Failure of Androgen Deprivation Therapy in Patients With Advanced Prostate Cancer,” Journal of the Formosan Medical Association = Taiwan yi zhi 102, no. 7 (2003): 480–485. [PubMed] [Google Scholar]
- 20. Hvamstad T., Jordal A., Hekmat N., Paus E., and Fosså S. D., “Neuroendocrine Serum Tumour Markers in Hormone‐Resistant Prostate Cancer,” European Urology 44, no. 2 (2003): 215–221, 10.1016/S0302-2838(03)00257-4. [DOI] [PubMed] [Google Scholar]
- 21. Heck M. M., Thaler M. A., Schmid S. C., et al., “Chromogranin A and Neurone‐Specific Enolase Serum Levels as Predictors of Treatment Outcome in Patients With Metastatic Castration‐Resistant Prostate Cancer Undergoing Abiraterone Therapy,” BJU International 119, no. 1 (2017): 30–37. 10.1111/bju.13493. [DOI] [PubMed] [Google Scholar]
- 22. Fan L., Wang Y., Chi C., et al., “Chromogranin A and Neurone‐Specific Enolase Variations During the First 3 Months of Abiraterone Therapy Predict Outcomes in Patients With Metastatic Castration‐Resistant Prostate Cancer,” BJU International 120, no. 2 (2017): 226–232. 10.1111/bju.13781. [DOI] [PubMed] [Google Scholar]
- 23. Szarvas T., Csizmarik A., Fazekas T., et al., “Comprehensive Analysis of Serum Chromogranin A and Neuron‐Specific Enolase Levels in Localized and Castration‐Resistant Prostate Cancer,” BJU International 127, no. 1 (2021): 44–55. 10.1111/bju.15086. [DOI] [PubMed] [Google Scholar]
- 24. Mi X., Hammill B. G., Curtis L. H., Lai E. C. C., and Setoguchi S., “Use of the Landmark Method to Address Immortal Person‐Time Bias in Comparative Effectiveness Research: A Simulation Study,” Statistics in Medicine 35, no. 26 (2016): 4824–4836. 10.1002/sim.7019. [DOI] [PubMed] [Google Scholar]
- 25. Gleiss A., Oberbauer R., and Heinze G., “An Unjustified Benefit: Immortal Time Bias in the Analysis of Time‐Dependent Events,” TRANSPLANT INTERNATIONAL 31, no. 2 (2018): 125–130. 10.1111/tri.13081. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Data Availability Statement
The data that support the findings of this study are available on request from the corresponding author. The data are not publicly available due to privacy or ethical restrictions.
The datasets generated and/or analyzed during the current study are not publicly available due to patient data security but are available from the corresponding author on reasonable request.