Tumor volume and prostate-specic antigen kinetics effects on outcomes of metastatic prostate cancer patients receiving androgen deprivation therapy

Background The present study aimed to analyse the effects of androgen deprivation therapy (ADT) in patients with newly diagnosed metastatic castration-naïve prostate cancer (mCNPC) and explore predictors, particularly prostate-specic antigen (PSA) kinetics, associated with poor prognosis according to tumor volume, a new sub-classication of metastatic prostate cancer established by the CHAARTED trial. Methods We reviewed 648 patients with newly diagnosed mCNPC receiving ADT at Chang Gung Memorial Hospital from January 2007 to December 2016. Basic characteristics and PSA kinetics prole were subsequently evaluated. Results Among patients with high-volume disease, those with faster to PSA nadir (TTN) (< 7 months), higher PSA nadir ( ≥ 2 ng/mL), and faster PSA doubling time (PSADT) (< 2 months) had higher risk for faster disease progression or shorter overall survival (OS) compared to those with slower TTN (> 7 months), lower PSA nadir (< 2 ng/mL), and slower PSADT (> 2 months). Multivariate analysis of those with low-volume disease showed that only PSADT (< 4 months) was tended to be associated with faster disease progression or shorter OS. Conclusions kinetics are

related factors, pretreatment parameters, including Gleason grade group, haemoglobin (Hb), and alkaline phosphatase (ALP), have also been identi ed as prognostic or predictive biomarkers in patients with mPCa under ADT. [19] Most previous studies had included populations comprising patients with metastatic castration-naïve prostate cancer (mCNPC) or metastatic castration-resistant prostate cancer (mCRPC). However, since the CHAARTED trial, the concept of tumor volume had been integrated into the management of metastatic prostate cancer. As such, identifying reliable early prognostic factors according to different tumor volumes during ADT would be helpful in the prompt formulation of better treatment strategies for patients with poor prognostic characteristics. While several studies have investigated factors in uencing disease burden, such factors still remain unclear across different tumor volumes. [19,20] Thus, the current study primarily aimed to explore the risk factors, particularly PSA kinetics, associated with poor prognosis, including shorter OS and shorter time to CRPC (the duration from initiation of ADT to biochemical CRPC status) according to tumor volume under ADT. Furthermore, we analysed the effects of ADT in patients with newly diagnosed mCNPC who had different tumor volumes.

Methods
We retrospectively evaluated patients with new metastatic prostate cancer diagnosed by histologic con rmation of at least one metastatic lesion after staging through computed tomography, magnetic resonance imaging, and/or bone scan according to the American Joint Committee on Cancer 8th edition between January 2007 to December 2016 who had received primary ADT (either surgical or medical castration, with or without anti-androgen) at Chang Gung Memorial Hospital (CGMH). Patients were grouped into HVD and low-volume disease (LVD) according to the CHAARTED trial, where HVD was de ned as presence of visceral metastases and/or four or more bone metastases with at least one outside the vertebral column and pelvis. Patients who did not satisfy the aforementioned de nition were classi ed as LVD. This study adhered to the tenets of the Declaration of Helsinki and was approved by the Institutional Review Board (IRB) of Chang Gung Medical Foundation (IRB number 201801377B0), and patient consent is not required for observational studies.
CRPC was diagnosed based on biochemical progression (three consecutive spikes in PSA 1 week apart, of which two were 50% higher than the nadir, and PSA > 2 ng/mL) according to the European Association of Urology guideline. OS was de ned as the period from diagnosis until death by any cause. Baseline patient demographics and post-treatment characteristics, including age at diagnosis, clinical M staging, iPSA (PSA level upon diagnosis), Gleason grade group (according to the classi cation of International Society of Urological Pathology[21]: Grade group 1, Gleason score ≤ 6; Grade group 2, Gleason score 3 + 4 = 7; Grade group 3, Gleason score 4 + 3 = 7; Grade group 4, Gleason score 8; and Grade group 5, Gleason score ≥ 9), initial Hb level, initial calcium (Ca) level and initial ALP level, were determined. PSA kinetics pro les were de ned based on previous related studies, [11,19,22,23] including TTN (de ned as the duration from ADT initiation to PSA nadir), nadir PSA level, PSA reduction rate (PSARR) [de ned as 100 × (iPSA − PSA nadir / iPSA) / TTN],[24] time to CRPC (the duration from ADT initiation to biochemical CRPC status), time from PSA nadir to CRPC (TFNTC) (the duration from PSA nadir to CRPC status), and PSADT [de ned as log2×(time interval)/log (PSA value) -log (nadir PSA)]. Patients with insu cient imaging reports for determining volume status or excessive missing data were excluded.
Previous studies have demonstrated that several cutoff points for PSA-related factors, including iPSA, TTN, PSA nadir, and PSADT, predicted disease progression or OS. [10,11,13,24,25] Moreover, some studies have utilised the receiver operating characteristic curve to determine cutoff points, while others use median values. The current study chose median values as the optimal cutoff points for the different parameters.
All statistical analyses were performed using SAS software version 9.4 (SAS Institute Inc, Cary, NC). Nominal variables are presented as means and standard deviations, while non-nominal variables are presented as medians and interquartile ranges. The Chi-square test was used to compare categorical variables, while the independent t-test was used to compare continuous variables. OS and CPRC-free survival were evaluated using the Kaplan-Meier method. The multivariate Cox proportional-hazards model was used to determine the association between risk factors and OS or CPRC-free survival. All p values reported were two sided with p < 0.05 indicating statistical signi cance.

Results
A total of 918 patients with newly diagnosed metastatic prostate cancer at CGMH from January 2007 to December 2016 were identi ed. After excluding those who had received either chemotherapy or radiotherapy and those with incomplete information, a total of 648 patients receiving primary ADT were ultimately analysed. In the study population (Table 1), included patients had a median age of 75 (IQR 68-80) years, with 352 (54.3%) classi ed as HVD and 296 (45.7%) as LVD according to the CHAARTED trial. A total of 371 (57%) patients died during the study period. The median OS was 34 months (Table 2), while those with HVD had a signi cant shorter median OS than those with LVD (30 vs. 43 months; p < 0.0001) (Fig. 1a).
Multivariate analysis of all included patients (mCNPC) revealed that those with TTN < 9 months, nadir PSA level ≥ 1 ng/mL, and PSADT < 3 months had increased tendency for biochemical progression (Table 3), while those with TTN < 9 months, nadir PSA level ≥ 1 ng/mL, duration of PSA nadir < 5 months, PSADT < 3 months, and time to CRPC < 17 months had increased risk for shorter OS (Table 4). Moreover, univariate and multivariate analyses of factors affecting disease progression to CRPC and OS in HVD and LVD are done. Accordingly, multivariate analysis of patients with HVD showed that those with faster TTN (< 7 months), higher PSA nadir (≥ 2 ng/mL), and faster PSADT (< 2 months) had higher risk for faster disease progression or shorter OS compared to those with slower TTN (> 7 months), lower PSA nadir (< 2 ng/mL), and slower PSADT (> 2 months), respectively (Tables 5 and 6). Multivariate analysis of patients with LVD showed that only PSADT (< 4 months) was associated with increased risk for faster disease progression to CRPC or shorter OS (Tables 7 and 8).

Discussion
Research has shown that upfront chemotherapy with rst-line ADT signi cantly improved OS in patients with high-volume metastatic hormone-sensitive prostate cancer. [ HVD accounted for 54% of our study population, 63% of that in the CHARTEED trial, and 49% of that in previous studies within Taiwan.
[20] This indicates that Asian men had a lower prevalence of HVD than the occidental population, which corresponds with our general conception that prostate cancer incidence rates are higher among occidental than among Asian men.
[28] The present study found that patients on ADT with HVD had signi cant shorter OS (30 vs. 43 months) and time to the development of CRPC (13 vs. 26 months) compared to those with LVD. This indicates that regardless of whether patients received chemohormonal therapy [29] or ADT alone, those with HVD had worse prognosis than those with LVD, which agrees with the general consensus that HVD promotes worse prognosis due to disease severity. The current study also found that 58% of the patients progressed to CRPC status, among whom 66% and 48% had HVD and LVD, respectively (p < 0.0001). The ndings obtained herein were similar to those presented in previous studies, which demonstrated CRPC status in over 50%, and even 60-70%, of patients with mPCa[16, 20, 22, 30] and showed that over half of the patients with mPCa under ADT eventually progressed to CRPC. Furthermore, the present study found signi cant differences in PSA kinetics and factors affecting OS and disease progression to CRPC between HVD and LVD. Accordingly, TTN, nadir PSA, level and PSADT were identi ed as signi cant predictors of both OS and progression to CRPC in HVD, whereas only PSADT was identi ed as a signi cant predictor in LVD. The aforementioned results indicated that among patients with HVD, more attention should be provided to those with faster TTN, higher nadir PSA level and faster PSADT, with such patients possibly becoming suitable candidates for upfront chemotherapy or new generation hormone therapy. However, among patients with LVD, more attention should be provided to those with increasing PSA levels and PSADT considering that TTN, PSA nadir and even PSA reduction rates were not major predictors.
PSA level has been the most widely used biomarker for evaluating disease progression and predicting survival in clinical practice. Furthermore, several retrospective clinical studies and even some metaanalyses have determined that PSA kinetics, including PSA response, nadir PSA level, TTN, or PSADT, predicted OS or disease progression.[9, 10, 13, 24] Among such factors, nadir PSA levels and TTN have been considered important predictors of survival and progression period. Generally, a faster decline in PSA levels has been associated with more cancer cell death, which promotes a more favourable prognosis and survival.
[31] However, reports have shown that rapidly decreasing PSA levels during initial ADT was a risk factor for early progression to CRPC.
[16] Accordingly, a recent study introduced the novel concept suggesting that tumor-regulating broblasts play an important role in the mechanisms associated with TTN after primary ADT. [32] Other studies have also demonstrated that rapidly decreasing PSA levels may be associated with transcriptional outcomes of ADT rather than cancer cell death.
Moreover, heterogeneous prostate cancer cells, including hormone-resistant prostate cancer cells and hormone-sensitive prostate cancer cells, often coexist in the same patient. The rapid decline in PSA levels may indicate downregulation of PSA expression in hormone-sensitive prostate cancer cells, which are regulated by androgen via the androgen receptor pathway,[33] though it is doubtful that whether hormone-sensitive prostate cancer cells account for more of the cancer cells in HVD than in LVD, it might be a possible explanation for patient with HVD have shorter TTN and faster time to disease progression, even shorter OS. Moreover, longer TTN and lower PSA nadir levels during ADT have been known to be associated with signi cantly longer OS and period of disease progression.[10, 11, 13, 24, 30] Despite the lack of internationally accepted PSA nadir or TTN cutoff points for predicting disease progression or survival outcomes until present, observed tendencies have suggested that higher PSA nadir levels and shorter TTN promoted a shorter disease progression period and poor prognosis and survival of patients with mPCa receiving ADT.[9] PSADT predicts outcomes has been known for nearly 30 years, and it may closely indicate changes in prostate tumor volume, an independent predictor of biochemical relapse among patients that either underwent radical prostatectomy or endocrine treatment. [ [39] The present study identi ed lower cutoff point of PSADT than before as a useful clinical predictor in patients with mPCa receiving ADT, regardless of whether they had HVD (PSADT < 2 months) or LVD (PSADT < 4 months). The possible cause of short median PSADT in our study is that the PSA nadir level is low, so that the doubled PSA level could be reached easily. In addition, PSA level is checked more frequently in our study than in other studies, so the PSADT may also be in uenced. Although PSADT cannot be evaluated during the early stages of ADT, it may still be a clinically effective predictor of disease progression and OS.
Previous reports have showed that PSA-producing CRPC cells have the ability to grow under low androgen environments and may be present in heterogeneous prostate cancer with bone metastases.[29, 40] Moreover, CRPC cells may exhibit faster PSA decline under ADT compared to androgen-dependent prostate cancer cells due to the rapid reduction in PSA, which may be affected via the downregulation of PSA expression, regulated by androgen through androgen receptors, rather than cancer cell death. Another explanation is that the rapid removal of androgen-dependent prostate cancer cells may provide a suitable environment for CRPC cells. [29,33] Given that HVD has been associated with a higher CRPC rate and more severe bone metastases and disease status compared to LVD, TTN and PSA nadir might be more effective predictors for OS and disease progression in HVD rather than in LVD.
The current study has some limitations worth noting. The most important limitation is our exclusion of patients who received chemotherapy or radiotherapy, however those with subsequent hormone therapy, including abiraterone or enzalutamide after CRPC were not excluded, which may have affected our results with regard to OS. Second, given that this was a single-centre retrospective study, our results may not be generalisable to other populations. Moreover, considering that CGMH is a medical centre in Taiwan, the disease severities of our study population may be higher than those of general population. Lastly, not all PSA parameters and factors, including Gleason score, Hb, Ca and ALP, were regularly assessed at our institution, which might have in uenced our results.

Conclusion
Among patients receiving ADT, those with high-volume metastatic prostate cancer had signi cantly shorter OS and faster disease progression compared to those with low-volume metastatic prostate cancer. Moreover, the current study identi ed TTN ≥ 7 months, nadir PSA level ≤ 2 ng/mL, and PSADT ≥ 2 months as signi cant predictors of slower disease progression and better OS in HVD. However, only PSADT ≥ 4 months had been identi ed as an effective predictor for slower disease progression and better OS in LVD. PSA kinetics can therefore be effective clinical predictors of disease progression and survival.