Cell‐free DNA aneuploidy score as a dynamic early response marker in prostate cancer

Abstract

Cell‐free circulating tumor DNA (ctDNA) has emerged as a promising biomarker for response evaluation in metastatic castration‐resistant prostate cancer (mCRPC). The current study evaluated the modified fast aneuploidy screening test‐sequencing system (mFast‐SeqS), a quick, tumor‐agnostic and affordable ctDNA assay that requires a small input of DNA, to generate a genome‐wide aneuploidy (GWA) score in mCRPC patients, and correlated this to matched metastatic tumor biopsies. In this prospective multicenter study, GWA scores were evaluated from blood samples of 196 mCRPC patients prior to treatment (baseline) with taxanes (docetaxel and cabazitaxel) and androgen receptor signaling inhibitors (ARSI; abiraterone and enzalutamide), and from 74 mCRPC patients at an early timepoint during treatment (early timepoint; median 21 days). Z‐scores per chromosome arm were tested for their association with tumor tissue genomic alterations. We found that a high tumor load in blood (GWA^high) at baseline was associated with poor response to ARSI [HR: 2.63 (95% CI: 1.86–3.72) P < 0.001] but not to taxanes. Interestingly, GWA^high score at the early timepoint was associated with poor response to both ARSIs [HR: 6.73 (95% CI: 2.60–17.42) P < 0.001] and taxanes [2.79 (95% CI: 1.34–5.78) P = 0.006]. A significant interaction in Cox proportional hazards analyses was seen when combining GWA status and type of treatment (at baseline P = 0.008; early timepoint P = 0.018). In summary, detection of ctDNA in blood by mFast‐SeqS is cheap, fast and feasible, and could be used at different timepoints as a potential predictor for outcome to ARSI and taxane treatment in mCRPC.

Abbreviations

• ARSI
• androgen receptor signaling inhibitor
• CN
• copy number
• CS
• cellsave
• ctDNA
• cell-free circulating tumor DNA
• dPCR
• digital PCR
• FFS
• failure-free survival
• GWA
• genome-wide aneuploidy
• HMF
• Hartwig Medical Foundation
• IQR
• interquartile range
• mCRPC
• metastatic castration-resistant prostate cancer
• mFast-SeqS
• modified fast aneuploidy screening test-sequencing system
• OS
• overall survival
• REMARK
• recommendations for tumor marker prognostic studies
• SD
• standard deviation
• TTNT
• time to next treatment line
• WGS
• whole genome sequencing

Introduction

With the increasing number of available therapies for patients with metastatic castration-resistant prostate cancer (mCRPC), accurate response prediction has become increasingly important [1–5]. Currently available blood-based biomarkers, such as prostate specific antigen (PSA), have limitation in assessing response in patients with mCRPC [6,7]. Cell-free circulating tumor DNA (ctDNA) has emerged as a promising biomarker with diagnostic, predictive and prognostic applications in cancer [8–13]. The fraction of ctDNA within the total pool of cell-free DNA (cfDNA) has been associated with tumor burden and may be used to predict and monitor response to treatment [12]; however, prior prostate cancer studies were hampered by inconsistent sampling and studies to test whether baseline and on-treatment ctDNA fraction can predict response to treatment are lacking. Most techniques to measure the fraction of ctDNA rely on prior knowledge of genomic aberrations determined in a tumor biopsy. Therefore, they are dependent of upfront deep sequencing of tumor tissue, which is time consuming and costly.

The modified Fast aneuploidy screening test-sequencing system (mFast-SeqS) was originally established as a noninvasive screening method for fetal aneuploidy in maternal blood [14,15]. mFast-SeqS uses unique LINE-1 elements scattered across the human genome, the aneuploidy of a cfDNA sample can be estimated globally using z-statistics without prior knowledge of the genomic make-up of the tumor in situ [16]. If this score, termed the genome-wide aneuploidy (GWA) score, is larger than 5, the tumor fraction in cfDNA is usually larger than 10% [14]. This way, the mFast-SeqS-based GWA score serves as a fast, easily applicable, and affordable method to estimate the tumor fraction in cfDNA and only requires a minute amount of cfDNA (0.5−1 ng). Therefore, mFast-SeqS may be a promising technique for response evaluation in patients with mCRPC in daily clinical practice. This technique has been used by our group before in mCRPC in which we validated the GWA score as a prognostic biomarker in prostate cancer [17]. The current study aimed to evaluate the clinical value of aneuploidy in mCRPC patients using mFast-SeqS. Longitudinal cfDNA samples and metastatic tumor biopsies were collected in mCRPC patients treated with androgen receptor signaling inhibitors (ARSI) or taxanes, and these samples were assessed for the GWA score at baseline and during treatment. Additionally, the concordance between the copy number alterations (CNA) of tumor tissue biopsy and ctDNA was evaluated in matched ctDNA-tumor samples.

Materials and methods

Patient inclusion

The current analysis encompasses patients with mCRPC treated with abiraterone, enzalutamide, docetaxel or cabazitaxel and enrolled in the CPCT-02 study (NCT01855477) with or without simultaneous inclusion in the CIRCUS study (NTR5732) between February 2015 and December 2019. The CPCT-02 study obtained fresh-frozen tumor biopsies of a metastatic lesion for whole genome sequencing (WGS). Additionally, baseline and on-treatment blood biopsies for cfDNA were drawn. In- and exclusion criteria have been published before [18]. In short, patients were eligible if they had a locally advanced or metastatic solid tumor for which a next line of systemic treatment with a registered anti-cancer agent was indicated and a fresh tumor biopsy of a metastatic lesion could be obtained. The list of participating hospitals is available via . Patients with mCRPC enrolled in the CPCT-02 study could also enter the multicenter, prospective CIRCUS study, in which additional longitudinal blood samples for cfDNA during multiple treatment lines were obtained. The CIRCUS study was conducted at the Erasmus MC, the Franciscus Gasthuis & Vlietland and the HagaZiekenhuis. Both studies were conducted following the Declaration of Helsinki and approved by the medical ethical committee of University Medical Center Utrecht (CPCT-02) and Erasmus University Medical Center (CIRCUS; MEC-16-081). All patients provided written informed consent before any study procedure was performed. Clinical data for these patients were collected in an electronic case report form (ALEA Clinical) Abcoude, the Netherlands.

Study endpoints

The main outcomes were failure-free survival (FFS) and time to next treatment line (TTNT). For patients treated with ARSI, FFS was defined as the time from the start of treatment until the end of treatment with ARSI. For patients treated with taxanes, the outcome was slightly adapted, as a response to taxanes can continue after the end of treatment. Therefore, TTNT was defined as the time from the start of treatment until the next treatment line, best supportive care or death, whichever occurred first. The study endpoints were reported according to the Reporting Recommendations for Tumor Marker Prognostic Studies (REMARK) [19].

Tissue biopsies, blood sampling, and cfDNA extraction

Image-guided core needle biopsies of metastatic tumor lesions were obtained before the start of a new line of systemic treatment, as described previously by Priestley et al. [20].

Blood was drawn prior to starting treatment and at an early time-point (± 3 weeks) after the start of the treatment in one to three CellSave Preservative tubes (CS) (Menarini Silicon Biosystems, Castel Maggiore, Bologna, Italy) and processed within 96 h after withdrawal as described by van Dessel et al. [21].

Plasma was isolated from blood draws by two centrifugations steps of 10 min at room temperature, at 1711 g and 12 000 g, respectively. Plasma was then stored at −80 °C in 2 mL aliquots until further use for cfDNA isolation. cfDNA was extracted using the Maxwell platform (Maxwell® RSC LV ccfDNA Plasma Custom Kit, Promega, Madison, WI, USA) or QIAsymphony platform (QIAsymphony® (QS) SP Circulating DNA Kit, Qiagen, Venlo, Limburg, The Netherlands) according to the manufacturer's protocol. Afterwards, cfDNA was stored at −20 °C. cfDNA concentrations were quantified using the Quant-iT dsDNA high-sensitivity assay (Invitrogen, Life Technologies, Carlsbad, CA, USA) and the Qubit fluorometer (Invitrogen) was used for the readout according to the manufacturer's instructions.

Whole genome sequencing data analysis

Whole genome sequencing data from metastatic tumor tissue (CPCT02 cohort) were provided (data request number DR-170) by the Hartwig Medical Foundation (HMF), Amsterdam, and were described elsewhere [20]. From the received WGS-data, additional processing was performed using a workflow as previously reported [22]. For tissue, copy number alterations were estimated using purple v2.49, Hartwig Medical Foundation, Amsterdam, the Netherlands. From these data, the relative copy number deviation from the expected normal copy number was estimated per chromosomal arm and log² transformed.

mFast-SeqS, sequencing, and data analysis

To obtain copy number profiles, the mFast-SeqS protocol was performed on 1 ng cfDNA as previously described [14,23]. Briefly, LINE-1 elements scattered throughout the genome were amplified using a carefully selected primer pair [15]. Bar-coded adaptors were added, and the resulting libraries were pooled and sequenced on a MiSeq system generating 150 bp single-end reads to reach at least 90 000 reads per sample. Read counts per chromosome arm were then normalized to the total library size and then a Z-score per chromosome arm was calculated relative to healthy male controls (n = 17). Due to insufficient absolute number LINE-1 elements, the short arms of chromosomes 13, 14, 15, 21, and 22 as well as the Y chromosome were excluded from analysis. These Z-scores and the log2 tissue-based CNA per chromosomal arm were compared for correlations as described previously by Mendelaar et al. [22].

Finally, the resulting Z-scores per chromosome were squared and summed yielding a genome-wide aneuploidy (GWA) score per sample. Based on the previously defined cut-off set by Belic et al. [16], samples were dichotomized in genome-wide aneuploidy score (GWA) low (GWA < 5) or high (GWA ≥ 5) designated respectively GWA^low or GWA^high.

Evaluation of chromosome Xq and AR-amplification

As the mFast-SeqS assay is not able to detect the AR amplification directly, we verified the AR-amplification in blood by a more specific assay namely the digital PCR (dPCR). This assay quantified the number of AR genomic copies and compared these to the number of reference gene AGO1 genomic copies (on chromosome 1p34.3), resulting in an AR/AGO1 copy number (CN)-ratio similar to the AR/RNaseP CN-ratio reported by others [24]. The dPCR AR amplification assay was tested on prostate cancer cell lines LNCAP and VCAP, representing normal and amplified AR models, respectively. Afterward, the AR/AGO1 CN-ratio was determined by dPCR on available blood cfDNA from 82 mCRPC patients with mFast-SeqS GWA status determined as GWA^high (n = 35) or GWA^low (n = 47).

Statistical analysis

Baseline characteristics, clinical outcomes, and GWA scores were described as mean ± standard deviation (SD), median and interquartile range (IQR), or number of events. FFS was associated with dichotomized GWA scores at baseline and the early time point. Next, Kaplan Meier curves for FFS and OS were generated based on the dichotomized GWA scores, and log-rank tests were performed to determine the statistical significance of the differences between survival curves of subgroups. Hazard ratios and their 95% confidence intervals (95% CI) were determined by Cox regression. Depending on the format and distribution of the data, the appropriate statistical tests were used as described in the figure legends. The P-values were two-sided, and P ≤ 0.05 was considered significant. All statistical analyses were performed using either spss (v.28), SPSS inc, Chicago, IL, USA, stata (v.17), StataCorp, College Station, TX, USA or the statistical platform r (v4.3.0.), R Foundation, Vienna, Austria.

Results

Samples and baseline characteristics

Between February 2015 and December 2019, 235 patients with mCRPC were included in the CPCT-02 trial (NCT01855477) and treated directly after a fresh-frozen tumor biopsy was obtained [18]. The CPCT-02 study also collected blood before treatment (baseline) from 196 mCRPC patients. WGS was successfully performed in 155/235 (66%) patients. For 153 patients matched WGS and baseline cfDNA analysis using mFast-SeqS were available. The median age at registration of these included patients for the WGS analysis was 68 years. The median time between biopsy and baseline blood draw was 0 days (range = 0–187 days). The majority of the biopsies (51%) were taken from lymph node metastases, followed by 29.4% from bone metastases. The baseline characteristics of all cohorts are described in Table 1.

Table 1 Baseline characteristics.

Characteristics	mFast-SeqS	WGS (N = 153)
Total (n = 196)	ARSI (n = 154)	Taxanes (n = 42)
Age at registration
Median (range), year	68 (46–84)	68.5 (46–84)	67 (55–80)	68 (46–84)
Biopsy site, no. (%)
Lymph node	70 (35.7%)	59 (38.3%)	11 (26.2%)	78 (51%)
Bone	42 (21.4%)	37 (24%)	5 (11.9%)	45 (29.4%)
Liver	14 (7.1%)	12 (7.8%)	2 (4.8%)	16 (10.5%)
Lung	2 (1.0%)	2 (1.3%)		3 (2.0%)
Soft tissue	10 (5.1%)	9 (5.8%)	1 (5.8%)	9 (5.9%)
Other	3 (1.5%)	3 (1.9%)		2 (1.3%)
Missing	55 (28.1%)	32	23 (54.8%)	0
Time between biopsy and baseline blood draw
Median (range), days				0 (0–187)
Time between start treatment and baseline blood draw
Median (range), days	4.5 (0–28)	3.5 (0–28)	6 (0–27)
Prior therapy lines, no. (%)
Median (range)	1 (0–9)	1 (0–9)	2 (0–4)	2 (0–9)
Type of prior therapy, no. (%)
Abiraterone	26 (13.3%)	24 (15.4%)	7 (16.7%)	25 (16.4%)
Enzalutamide	60 (30.6%)	46 (29.5%)	26 (61.9%)	50 (32.7%)
Docetaxel	129 (65.8%)	103 (66%)	29 (69%)	99 (64.7%)
Cabazitaxel	51 (26%)	44 (28.2%)	2 (4.8%)	42 (27.5%)
Radium-223	18 (9.2%)	16 (10.3%)	1 (2.4%)	18 (11.8%)
Other ARSI	3 (1.9%)	3 (1.9%)	2 (4.8%)	3 (2.0%)
Other chemotherapy	4 (2.0%)	4 (2.6%)		3 (2.0%)
Other systemic treatment	11 (5.6%)	11 (7.1%)	1 (2.4%)	10 (6.5%)
PSA at start treatment, μg·L−1
Median (range)	60 (3.1–2375)	61.5 (3.1–2375)	55 (5–1690)
Missing	84	66	18

The mFast-SeqS was performed on blood cfDNA to generate a GWA score per patient, and this was assessed according to the study design shown in the flowchart of Fig. 1. Baseline liquid biopsies were available for analysis with clinical outcomes for 154 patients treated with ARSI and 42 patients treated with taxanes. An early on-treatment timepoint (median 21 days, IQR: 19–28 days after start of treatment) sample was available for 35 patients treated with ARSI and 39 treated with a taxane. The median age at registration of the included patients was 68 years. Patients had received a median of 1 prior treatment line (excluding ADT) for metastatic prostate cancer, mostly docetaxel (65.8%). The median PSA before the start of treatment was 60 ng·mL⁻¹ (range: 3.1–2375).

[IMAGE OMITTED. SEE PDF]

Median blood GWA scores were 4.7 (IQR: 1.4–35.25) at baseline and 1.22 (IQR:0.3–9.61) at early time point which translated after dichotomization (< 5 vs. ≥ 5) in 95 of 196 patients (48%) and 19 of 74 patients (26%) into being GWA^high at baseline and at early time point on treatment, respectively.

Whole genome sequencing of tumor tissue cohort and correlation with GWA

When comparing the CNA between blood cfDNA and tumor tissue DNA, a statistically significant positive correlation was seen (median Rho of 0.75; range: 0.24–0.93) (Fig. 2, and Figs S1 and S2). We further investigated whether GWA^high patients when compared to GWA^low had statistically significant different tumor characteristics (Oncoplot Fig. 3). Patients with blood GWA^high had a significantly higher genome ploidy (P < 0.001), a higher number of structural variants (P < 0.001), and more whole genome duplications (P < 0.001) in their tumor tissue when compared to tumor tissue of patients with blood GWA^low. Mutation analysis of tissue-based WGS revealed that somatic mutations were present in driver genes in the majority of patients in AR (65%) followed by TP53 (45%), PTEN (45%), and EDA2R (31%). The GWA^high patients had significantly more AR (P = 0.025) and EDA2R amplifications (P = 0.001) (both on chromosome Xq11-q12) in their tumor tissue when compared to GWA^low patients. As the number of LINE-1 elements on the X chromosome is low, AR amplification was evaluated in 82 cfDNA samples with the dPCR. To this end, we used dPCR to quantify AR and a reference gene AGO1 and derived the ratio between these two genes (AR/AGO ratio) as a measure of AR amplification and correlated this measure to GWA score and aneuploidy on Xq (expressed as Xq Z-scores) measured by the mFast-SeqS (Fig. S3). These analyses showed a trend towards higher AR/AGO1 ratio in GWA^high cases compared to those that are GWA^low (P = 0.08) confirming that GWA^high is in part defined by aneuploidy on Xq that coincides with AR amplification. Between the GWA score and AR/AGO1 dPCR ratio a modest correlation was observed (Pearson correlation: 0.47, R² = 0.22) (Fig. S4).

[IMAGE OMITTED. SEE PDF]

[IMAGE OMITTED. SEE PDF]

Clinical outcome and relation with GWA

We evaluated whether GWA status was associated with clinical outcome. In patients treated with taxanes (cabazitaxel or docetaxel), we found no statistically significant difference in TTNT when comparing between patients with a GWA^low and GWA^high score for the baseline while at early time point sample in patients with GWA^high, we observed a shorter TTNT [HR: 2.79 (95% CI: 1.35–5.78) P < 0.01] in the cox regression (Table 2) and Kaplan–Meier analyses (Fig. 4). In contrast, for patients that were treated with ARSI already at baseline GWA^high predicted a shorter FFS [HR: 2.63 (95% CI: 1.86–3.72) P < 0.001], difference that at early time point seemed even more pronounced [HR: 6.73 (95% CI: 2.60–17.42) P < 0.001] (Fig. 4 and Table 2). For ARSI treated patients, who were stratified on dichotomized GWA score, Kaplan–Meier analysis revealed a significant difference in FFS (P < 0.001), with a median of 280 days and of 120 days for patients with GWA^low and GWA^high, respectively (Fig. 4). Cox proportional hazards analyses for outcome at baseline samples showed a significant interaction between GWA status and treatment type (ARSI vs. taxanes; P = 0.008, Table 3) and subsequently, at early timepoint (P = 0.018, Table 3).

Table 2 Overview of GWA and hazard ratio per treatment.

Treatment	Response	Timepoint blood biopsy	N	Blood GWA	N	%	Events	Hazard ratio	95% CI	P
ARSI	Failure free survival (FFS)	At baseline (BL)	154	GWAlow	78	51%	61	1.00
GWAhigh	76	49%	75	2.63	1.86–3.72	< 0.001
At early timepoint (ET)	40	GWAlow	31	78%	27	1.00
GWAhigh	9	22%	9	6.73	2.60–17.42	< 0.001
Longitudinal (BL & ET)a	40	GWAlow	20	50%	16	1.00
GWAswitch	11	28%	11	1.36	0.63–2.96	0.43
GWAhigh	9	22%	9	7.61	2.77–20.88	< 0.001
Taxanes	Time to next line of treatment (TTNT)	At baseline (BL)	42	GWAlow	23	55%	23	1.00
GWAhigh	19	45%	19	1.01	0.54–1.88	0.98
At early timepoint (ET)	42	GWAlow	30	71%	30	1.00
GWAhigh	12	29%	12	2.79	1.34–5.78	0.006
Longitudinal (BL & ET)a	42	GWAlow	21	50%	21	1.00
GWAswitch	11	26%	11	0.72	0.33–1.54	0.39
GWAhigh	10	24%	10	2.35	1.07–5.15	0.032

[IMAGE OMITTED. SEE PDF]

Table 3 Interaction analysis of outcome.

	Hazard ratio	95% CI	P
At baseline
Treatment type (ARSI, Taxanes)	2.36	1.43–3.88	0.001
GWA status (High, Low)	1.08	0.59–1.98	0.81
Interaction (GWA status × treatment type)	2.58	1.28–5.20	0.008
At early timepoint
Treatment type (ARSI, Taxanes)	1.77	1.03–3.04	0.039
GWA status (High, Low)	2.43	1.21–4.86	0.012
Interaction (GWA status × treatment type)	3.53	1.24–10.01	0.018

To analyze the dynamics of GWA status, we combined the baseline and early time point samples. This showed that patients with GWA^low and patients with disconcordant GWA at both time points, i.e. GWA switching [from GWA^high to GWA^low (n = 9), or from GWA^low to GWA ^high (n = 2), vs. patients with GWA^high at both time points, showed significant differences in TTNT for taxanes (P < 0.04) and in FFS for ARSI treatment (P < 0.001) (Fig. 4 and Table 2)].

Discussion

Whilst treatment options in mCRPC patients have expanded, it remains hard to adequately determine treatment response. Traditional markers and current clinical tools for mCRPC have limitation in assessing response and do not accurately predict disease progression before radiographic assessments at 3 and/or 6 months [7]. Detection of ctDNA has arrived as a predictive or prognostic biomarker capable of noninvasive treatment monitoring [8–13]. Recently, we reported that the mFast-SeqS method was able to quantify ctDNA levels yielding prognostic value at baseline in mCRPC patients [17]. Whether a mFast-SeqS-derived GWA score, or dynamic changes of this measure can also serve as a treatment response marker was still up for debate. In the current study, we investigated whether a dichotomized GWA score derived by the mFast-SeqS method in cfDNA at the start of a new treatment and/or early time point could serve as a treatment response marker and whether the GWA score correlated with WGS from metastatic tumor biopsies. Our findings showed that at an early time point, dichotomized GWA scores were related with response in patients treated with ARSI and with taxanes. The GWA score at baseline was significantly associated with ARSI treatment length in patients treated with ARSI; however, GWA scores were unable to determine TTNT in patients treated with taxanes. This difference was significant in an interaction test suggesting baseline GWA score has predictive value.

Another observation we made was that by combining baseline data with data from the early time point in ARSI treated patients, it appeared discriminative power of GWA scoring increased as the hazard ratio increased from 2.63 at baseline to 7.61 in the combined measure. Although our observations will require further validation in independent cohorts, we can conclude that this minimally invasive, user-friendly, affordable and robust mFast-SeqS assay measure might have valuable clinical value. On top of that, it was reassuring that our reported GWA scores in cfDNA correlated well with corresponding tumor-tissue chromosomal profiles. Interestingly, patients with GWA^high score compared to GWA^low had a remarkably higher number of structural variants in their tumor tissue. Thus, we find a group of patients with a GWA^high score pinpointing patients having increased genomic instability.

A limitation of this study is the lack of additional clinical data such as PSA values, known prognostic markers such as ECOG, metastatic burden, neuroendocrine differentiation, International Society of Urological Pathology (ISUP) grade or imaging data. To determine whether GWA scores at early on treatment time point would be a treatment response marker, comparison with PSA values or imaging data as reference is recommended. Another potential weakness of this study was that GWA score threshold, i.e. < 5 and ≥ 5 used for the dichotomization of aneuploidy status was based on previous literature [14]. The threshold for grouping patients might be improved to better optimize the stratification; however, this remains to be shown. Finally, the mode to evaluate GWA dynamics was also exploratory and the study design (not a randomized controlled trial), limited us from truly assessing its predictive value. The low numbers of patients treated with cabazitaxel in this study could have limited us from finding prognostic value at baseline in these patients. Although our study was not randomized, the taxane cohort was small and we previously reported prognostic value of cabazitaxel treated patients [22], we now report that the aneuploidy score is an early response marker.

Conclusion

In conclusion, mFast-SeqS-based GWA scores are concordant with aneuploidy scores obtained by WGS from tumor-tissue and can predict response to ARSI treatment at baseline and, at early time point, to ARSI and taxanes. This assay can be easily performed at low cost, requiring little input of cfDNA. Ultimately, if further validated, these results can improve the cost-effectiveness of a treatment by sparing patients ineffective treatments and, at the same time, unnecessary toxicity.

Acknowledgements

This publication and the underlying study have been made possible in part by the Dutch Cancer Foundation (NKI2014-7080), Hartwig Medical Foundation is a dutch foundation that performs genomic analysis,  and the Center of Personalized Cancer Treatment (CPCT). The authors would like to thank the included patients and their families for their willingness to take part in this study. We also would like to thank all the medical oncologists, research nurses, and laboratory technicians involved in this study for their help in patient inclusion, blood sampling, processing, and data management.

Conflict of interest

JWMM declares a consultancy fee by Novartis, and institutional research grants by Pfizer and GSK (all outside the submitted work). PH has received consulting or advisory fees from Astellas, MSD, Pfizer AstraZeneca, BMS and Ipsen. MPL received institutional grants from JnJ, Astellas, MSD and Sanofi, consulting fees from Sanofi, Johnson & Johnson, Merck, Astellas, Incyte, Amgen, Janssen Cilag, Bayer, Servier and Pfizer, and is currently employed by Amgen. RW has advisory or speakers fees for Astellas, Merck, Bayer and received institutional research grants from Bayer. MSH has received institutional research funding from Bristol Myers Squibb, 4SC, Roche, Astellas Pharma Netherlands and AstraZeneca, and institutional consultancy fees from Bristol-Myers Squibb, Roche, Merck Sharp & Dohme, AstraZeneca, Pfizer, Janssen, and Seattle Genetics. The remaining authors declare no conflict of interest.

Author contributions

Study concept and design: MPL, SMW, RW, MPHMJ, JWMM. Acquisition of data: All authors. Analysis and interpretation of data: KTI, ACJ, JANG, JR, SMW, MPHMJ, JWMM, NB. Drafting of the manuscript: KTI, MPHMJ, JWMM. Critical revision of the manuscript for important intellectual content: All authors. Statistical analysis: KTI, JR, MPHMJ. Obtaining funding: MPL, RW, MPHMJ, JWMM. Administrative, technical, or material support: KTI, ACJ, LFD, VW, CB, JH, DV, SMW, MPHMJ, PH, MSH. Supervision: MPHMJ, JWMM. Other: None.

Peer review

The peer review history for this article is available at .

Data accessibility

The WGS and corresponding clinical data used in this study were made available by the Hartwig Medical Foundation (HMF; Dutch nonprofit biobank organization) and provided under data request number DR-170. Therefore, the data are available under restricted access and can be requested by contacting the HMF (). The remaining data are available upon request to the corresponding author.

References

Aldea M, Lam L, Orillard E, Llacer Perez C, Saint‐Ghislain M, Gravis G, et al. Cabazitaxel activity in men with metastatic castration‐resistant prostate cancer with and without DNA damage repair defects. Eur J Cancer. 2021;159:87–97. https://doi.org/10.1016/j.ejca.2021.09.029

Annala M, Fu S, Bacon JVW, Sipola J, Iqbal N, Ferrario C, et al. Cabazitaxel versus abiraterone or enzalutamide in poor prognosis metastatic castration‐resistant prostate cancer: a multicentre, randomised, open‐label, phase II trial. Ann Oncol. 2021;32:896–905.

Beer TM, Armstrong AJ, Rathkopf D, Loriot Y, Sternberg CN, Higano CS, et al. Enzalutamide in men with chemotherapy‐naïve metastatic castration‐resistant prostate cancer: extended analysis of the phase 3 PREVAIL study. Eur Urol. 2017;71:151–154.

Evans CP, Higano CS, Keane T, Andriole G, Saad F, Iversen P, et al. The PREVAIL study: primary outcomes by site and extent of baseline disease for enzalutamide‐treated men with chemotherapy‐naive metastatic castration‐resistant prostate cancer. Eur Urol. 2016;70:675–683. https://doi.org/10.1016/j.eururo.2016.03.017

Miller K, Carles J, Gschwend JE, Van Poppel H, Diels J, Brookman‐May SD. The phase 3 COU‐AA‐302 study of abiraterone acetate plus prednisone in men with chemotherapy‐naïve metastatic castration‐resistant prostate cancer: stratified analysis based on pain, prostate‐specific antigen, and Gleason score. Eur Urol. 2018;74:17–23.

Scher HI, Morris MJ, Basch E, Heller G. End points and outcomes in castration‐resistant prostate cancer: from clinical trials to clinical practice. J Clin Oncol. 2011;29:3695–3704. https://doi.org/10.1200/JCO.2011.35.8648

Scher HI, Morris MJ, Stadler WM, Higano C, Basch E, Fizazi K, et al. Trial design and objectives for castration‐resistant prostate cancer: updated recommendations from the prostate cancer clinical trials Working group 3. J Clin Oncol. 2016;34:1402–1418. https://doi.org/10.1200/JCO.2015.64.2702

Annala M, Vandekerkhove G, Khalaf D, Taavitsainen S, Beja K, Warner EW, et al. Circulating tumor DNA genomics correlate with resistance to abiraterone and enzalutamide in prostate cancer. Cancer Discov. 2018;8:444–457.

Dathathri E, Isebia KT, Abali F, Lolkema MP, Martens JWM, Terstappen L, et al. Liquid biopsy based circulating biomarkers in metastatic prostate cancer. Front Oncol. 2022;12:863472. https://doi.org/10.3389/fonc.2022.863472

Kohli M, Tan W, Zheng T, Wang A, Montesinos C, Wong C, et al. Clinical and genomic insights into circulating tumor DNA‐based alterations across the spectrum of metastatic hormone‐sensitive and castrate‐resistant prostate cancer. EBioMedicine. 2020;54:102728.

Kwan EM, Wyatt AW, Chi KN. Towards clinical implementation of circulating tumor DNA in metastatic prostate cancer: opportunities for integration and pitfalls to interpretation. Front Oncol. 2022;12:1054497.

Reichert ZR, Morgan TM, Li G, Castellanos E, Snow T, Dall'Olio FG, et al. Prognostic value of plasma circulating tumor DNA fraction across four common cancer types: a real‐world outcomes study. Ann Oncol. 2022;34:111–120. https://doi.org/10.1016/j.annonc.2022.09.163

Vandekerkhove G, Struss WJ, Annala M, Kallio HML, Khalaf D, Warner EW, et al. Circulating tumor DNA abundance and potential utility in de novo metastatic prostate cancer. Eur Urol. 2019;75:667–675. https://doi.org/10.1016/j.eururo.2018.12.042

Belic J, Koch M, Ulz P, Auer M, Gerhalter T, Mohan S, et al. Rapid identification of plasma DNA samples with increased ctDNA levels by a modified FAST‐SeqS approach. Clin Chem. 2015;61:838–849.

Kinde I, Papadopoulos N, Kinzler KW, Vogelstein B. FAST‐SeqS: a simple and efficient method for the detection of aneuploidy by massively parallel sequencing. PLoS One. 2012;7:e41162.

Belic J, Koch M, Ulz P, Auer M, Gerhalter T, Mohan S, et al. mFast‐SeqS as a monitoring and pre‐screening tool for tumor‐specific aneuploidy in plasma DNA. Adv Exp Med Biol. 2016;924:147–155.

Isebia KT, Mostert B, Deger T, Kraan J, de Weerd V, Oomen‐de Hoop E, et al. mFast‐SeqS‐based aneuploidy score in circulating cell‐free DNA is a prognostic biomarker in prostate cancer. Mol Oncol. 2023;17:1898–1907.

de Jong AC, Danyi A, van Riet J, de Wit R, Sjöström M, Feng F, et al. Predicting response to enzalutamide and abiraterone in metastatic prostate cancer using whole‐omics machine learning. Nat Commun. 2023;14:1968. https://doi.org/10.1038/s41467‐023‐37647‐x

Sauerbrei W, Taube SE, McShane LM, Cavenagh MM, Altman DG. Reporting recommendations for tumor marker prognostic studies (REMARK): an abridged explanation and elaboration. J Natl Cancer Inst. 2018;110:803–811. https://doi.org/10.1093/jnci/djy088

Priestley P, Baber J, Lolkema MP, Steeghs N, de Bruijn E, Shale C, et al. Pan‐cancer whole‐genome analyses of metastatic solid tumours. Nature. 2019;575:210–216. https://doi.org/10.1038/s41586‐019‐1689‐y

van Dessel LF, Beije N, Helmijr JC, Vitale SR, Kraan J, Look MP, et al. Application of circulating tumor DNA in prospective clinical oncology trials – standardization of preanalytical conditions. Mol Oncol. 2017;11:295–304.

Mendelaar PAJ, Robbrecht DGJ, Rijnders M, de Wit R, de Weerd V, Deger T, et al. Genome‐wide aneuploidy detected by mFast‐SeqS in circulating cell‐free DNA is associated with poor response to pembrolizumab in patients with advanced urothelial cancer. Mol Oncol. 2022;16:2086–2097. https://doi.org/10.1002/1878‐0261.13196

Angus L, Deger T, Jager A, Martens JWM, de Weerd V, van Heuvel I, et al. Detection of aneuploidy in cerebrospinal fluid from patients with breast cancer can improve diagnosis of leptomeningeal metastases. Clin Cancer Res. 2021;27:2798–2806.

Sumiyoshi T, Mizuno K, Yamasaki T, Miyazaki Y, Makino Y, Okasho K, et al. Clinical utility of androgen receptor gene aberrations in circulating cell‐free DNA as a biomarker for treatment of castration‐resistant prostate cancer. Sci Rep. 2019;9:4030. https://doi.org/10.1038/s41598‐019‐40719‐y

Nguyen L, Martens JWM, Van Hoeck A, Cuppen E. Pan‐cancer landscape of homologous recombination deficiency. Nat Commun. 2020;11:5584. https://doi.org/10.1038/s41467‐020‐19406‐4

Martincorena I, Raine KM, Gerstung M, Dawson KJ, Haase K, Van Loo P, et al. Universal patterns of selection in cancer and somatic tissues. Cell. 2017;171:1029–1041.e1021. https://doi.org/10.1016/j.cell.2017.09.042

Mermel CH, Schumacher SE, Hill B, Meyerson ML, Beroukhim R, Getz G. GISTIC2.0 facilitates sensitive and confident localization of the targets of focal somatic copy‐number alteration in human cancers. Genome Biol. 2011;12:R41. https://doi.org/10.1186/gb‐2011‐12‐4‐r41