Systemic treatments for metastatic prostate cancer have grown exponentially over the last two decades and exhibited significant survival benefits in randomized phase 3 trials (1–8). However, trial findings may be difficult to replicate in real-world conditions. Indeed, only one report demonstrated a modest benefit in overall survival after chemotherapy in contemporary, de novo metastatic prostate cancer patients (Weiner et al., National Cancer Database 2014-2015) (9). We addressed the same endpoint within the same study period. Within a different, large-scale database (SEER), we focused on the most contemporary patients (2014-2015) diagnosed with de novo metastatic prostate cancer. We hypothesized that chemotherapy use may result in a survival benefit for de novo metastatic prostate cancer patients (9). Unlike Weiner et al., we relied on propensity score matching to maximally reduce potential differences between chemotherapy-exposed and chemotherapy-naïve patients.

We hypothesized that, in line with trial-derived findings and smaller population-based studies, chemotherapy exposed de novo metastatic prostate cancer patients exhibit better survival rates compared to their chemotherapy naïve counterparts. We tested this hypothesis within a large population-based cohort de novo metastatic prostate cancer patients diagnosed between 2014 and 2015.

First, we observed significantly worse cancer characteristics in chemotherapy-exposed patients compared to their chemotherapy naïve counterparts. Specifically, they exhibited higher proportions of high PSA, higher proportions of GGG V, higher proportions of cN1-stage and higher proportions of cM1c-stage. It is of note that despite an obvious prostate cancer phenotype disadvantage in chemotherapy-exposed prostate cancer patients, their overall survival was better, as will be outlined below. These observations are similar to NCDB patient characteristics. In consequence, it may be postulated that both databases (NCDB and SEER) indicate that chemotherapy is offered to patients with more aggressive prostate cancer phenotype than average (9). The same observations regarding prostate cancer characteristics were made in smaller scale, retrospective studies (9, 12).

Second, within the current study cohort the rate of chemotherapy was 21.1% (n=905) de novo metastatic prostate cancer patients. This proportion is disappointingly low, however it is very comparable to NCDB, where chemotherapy was also given to a minority of patients (27.6%). Similarly low rates were recorded in other, smaller scale population-based studies (12, 13). These observations indicate a relatively low confidence level in systemic therapy. Additionally, risk of chemotherapy-related adverse events, which vary in regard to the dose and type of chemotherapeutical agent administered, may result in tendencies towards more restrictive chemotherapy administration policies. Even though that recent studies have recorded an increase of chemotherapy rates in more contemporary years (14, 15), efforts are further required to encourage referrals from within the urological community for systemic therapy, when metastatic prostate cancer is diagnosed (1).

Third, we recorded more favorable survival in chemotherapy-exposed vs chemotherapy-naïve patients (76.3 vs 69.3% and 61.6 vs 54.3% at 18 and 30 months). These rates resulted in a highly protective multivariable hazard ratio of 0.82 (CI:0.72-0.96, p=0.01). Finally, even after detailed propensity score matching for differences in patient and prostate cancer characteristics, a protective hazard ratio of 0.77 (CI:0.66-0.90, p<0.001) was recorded. Additionally, to propensity score matching, we furthermore repeated the survival analyses after landmark analyses was applied to maximally reduce potential biases that might have occurred due to immortal time biases. Irrespectively of these two strict methodological approaches to maximally reduce any biases which may arise from differences between chemotherapy exposed vs naïve de novo mPCa patients, survival trends remained in its quantity and quality unchanged.

These observations are highly consistent with NCDB-derived findings on the same topic (Weiner et al.) (9). Conversely, to the best of our knowledge, no other reports identified a survival benefit in contemporary, metastatic prostate cancer patients exposed to chemotherapy compared to their chemotherapy naïve counterparts. In consequence, it may be postulated that the survival benefit only became apparent in the most contemporary population-based metastatic prostate cancer patients, in both the SEER and the NCDB. To the best of our knowledge, prior to Weiner et al. and to the current study, a formal comparison between chemotherapy-exposed vs chemotherapy-naïve patients was not reported. Instead, previous population-based analyses examined survival trends regardless of chemotherapy exposure status. These trends exhibited only marginal improvement over time (14). For example, Cattrini et al. reported only a modest improvement of median overall survival (30 vs 26 months) in contemporary (2011-2014) metastatic prostate cancer patients in comparison to historical (2000-2003) metastatic prostate cancer patients exposed to chemotherapy. Since Cattrini et al. did not furthermore account for any treatment approach and primarily focused on the cohort of metastatic prostate cancer patients from an epidemiological aspect, results cannot directly be compared to the current study (16). In consequence, the current study, as well as the study by Weiner et al., cannot be directly compared to previous population-based studies with different designs and endpoints. Similarly, our findings cannot be directly compared to phase 3 trials, that focused on specific molecules and treatment regimens in randomized designs. In the current and Weiner et al. studies chemotherapy-exposed patients may have received one or multiple chemotherapy lines. Unfortunately, their specific time and duration of exposure is unknown in the current study, as well as in the Weiner et al. study. Consequently, some chemotherapy-exposed patients may have received a single line of chemotherapy with no overall survival benefit. Conversely, others may have received multiple lines with an important overall survival benefit. It is of note that combination therapies, including chemotherapeutic agents, are likely to play an important role in the near future. Recently, results derived from the PEACE-1 trial demonstrated for example that addition of abiraterone to androgen deprivation therapy (ADT) and docetaxel significantly improved radiographic progression-free survival in de novo metastatic castration sensitive prostate cancer patients (17). Last, but not least, the current study differed from Weiner et al. in its design. We relied on propensity score matching to maximally attenuate differences between chemotherapy-exposed and chemotherapy-naïve metastatic prostate cancer patients. Despite propensity score matching use in the current study, the previously recorded overall survival benefit observed in chemotherapy-exposed patients remained in the current analyses. Similarly, its magnitude remained virtually unchanged. It is noteworthy, that the magnitude of the benefit in the current study, as well as in the study of Weiner et al., cannot be directly compared to the magnitude of survival benefit recorded in phase 3 trials for specific systemic approaches for metastatic prostate cancer (18, 19). It is of note, that the magnitude of the survival benefit in most of phase 3 studies addressing overall survival in metastatic prostate cancer was greater than the magnitude recorded in our study, as well as that recorded in the study of Weiner et al. and other small scale institutional studies (20–22).

Regardless of the very important beneficial survival rates in chemotherapy exposed metastatic prostate cancer patients in respect to chemotherapy naïve patients, several limitations need to be acknowledged. First, the rate of chemotherapy exposure is low in the current study. It is nonetheless very similar to the rate observed in the study of Weiner et al. Moreover, the nature of administered chemotherapy is unknown with respect to the number of lines, their duration, as well as their individual efficacy. Furthermore, treatment approaches such as palliative care or observational approaches, are not available in the SEER database. Therefore, potential biases which may have occurred due to different supportive care measurements cannot be ruled out and should be taken into account when data is interpreted. Similar to Weiner et al., we could not adjust or circumvent these limitations.

Second, the retrospective nature of the study introduces a number of selection biases, that distinguish chemotherapy exposed patients from others. As reported, chemotherapy-exposed patients tended to harbor more aggressive prostate cancer phenotypes. The same limitation applied to the study cohort focusing on NCDB. In the Weiner et al. study, these differences were addressed in multivariable analyses. Conversely, in the current study, multivariable analyses were complemented by propensity score matching to more completely and strictly address these differences.

Third, certain additional unmeasured variables could not be addressed. These variables, including performance status and comorbidities, were unavailable in the current study. Some of these variables, including comorbidities, were available in the Weiner et al. study (9). Despite their availability, overall survival rates virtually perfectly agreed with rates recorded in the current study. In consequence, lack of comorbidities does not appear to represent a rate limiting factor. Fourth, strict stratification according to low- and high-volume tumor burden, as performed in previously reported phase-3 trials is limited by the nature of SEER data collection (18). Finally, a number of established predictors of survival (Lactate dehydrogenase, hemoglobin) for metastatic prostate cancer patients were unavailable in both the current and NCDB analyses (23).

In the largest contemporary, North-American population-based study, chemotherapy exposure for metastatic prostate cancer patients was associated with a prolonged overall survival, however the magnitude of previous trial-based survival benefits could not be reassured in real-life population-based data.

The original contributions presented in the study are included in the article/supplementary material. Further inquiries can be directed to the corresponding author.

Ethical review and approval were not required for the study on human participants in accordance with the local legislation and institutional requirements. The patients/participants provided their written informed consent to participate in this study.

BHoe - conceptualization, methodology, formal analysis, writing original draft, writing review and editing, and visualization. FrC – writing review and editing and visualization. CW: writing review and editing and visualization. RF: writing review and editing and visualization. BHor: writing review and editing and visualization. GS: writing review and editing and visualization. ZT: methodology, software, validation, formal analysis, and resources. FS: writing review and editing and supervision. MaG - writing review and editing and supervision. MiG: writing review and editing and supervision. LK: writing review and editing and visualization. PM -writing review and editing and supervision. AB - writing review and editing and supervision. DT - writing review and editing and supervision. FeC: writing review and editing and supervision. SS: writing review and editing and supervision. CT: writing review and editing and supervision. PK: writing review and editing, supervision, project administration, and conceptualization. All authors contributed to the article and approved the submitted version.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.