Introduction

Over the past decade, the treatment landscape for de novo metastatic hormone-sensitive prostate cancer (mHSPC) has undergone a remarkable transformation. Treatment strategies combining upfront palliative radiotherapy to the prostate (RTx), docetaxel (Doc) and/ or an androgen receptor targeting agent (ARTA; e.g., abiraterone, apalutamide or enzalutamide) with androgen deprivation therapy (ADT), have demonstrated remarkable efficacy in delaying disease progression and increasing overall survival whilst preserving quality of life (QoL)1,2,3,4,5,6. The introduction of these doublet and triplet therapies has caused a paradigm shift in the treatment of de novo mHSPC, supplanting ADT monotherapy as standard-of-care. Nonetheless, uncertainty remains regarding which treatment strategy fits the individual patient best.

In parallel, prostate-specific membrane antigen (PSMA) positron emission tomography (PET)/computed tomography (CT) has revolutionised the ability to detect prostate cancer (PCa) metastases, with a 27% higher accuracy rate compared to conventional imaging (CI; i.e., CT and bone scans)7. However, in landmark trials, metastatic volume as defined by CHAARTED on CI (i.e. high volume (HV) and low volume (LV)) was considered an important prognostic and predictive parameter8,9,10. This raised the question of how we should extrapolate the results of published studies and construe them when patients are staged with PSMA-PET/CT. Therefore, the integration of PSMA-PET/CT has not only facilitated accurate diagnosis, but has also sparked a debate on the interpretation of tumour burden and therapeutic consequences of detected metastases11.

With the heterogeneous use and interpretation of next-generation molecular imaging and the approval of new treatment strategies, treating de novo mHSPC patients presents a challenging conundrum12. It is conceivable that patients are managed differently when treated in another country, hospital or even by another clinician. If we are to prevent unwanted (i.e., deviating from guidelines) practice variations, it is necessary to identify factors that may influence treatment decisions, potential reasons for practice variation, and unmet needs of treating physicians. To date, studies exploring relevant factors for the decisional process in de novo mHSPC are lacking. We aim to identify, compare, and understand how these patients are managed - and why - by exploring physicians’ attitudes, motives, priorities, experiences and unmet needs.

Materials and methods

Research design

This exploratory mixed-method study utilised quantitative methods to identify trends and patterns in the initial imaging modality, definitions, important factors for treatment selection and patient cases. Subsequently, quantitative findings were then explored qualitatively within focus group discussions. This explanatory sequential design was employed to enhance, elaborate, and clarify the findings from one method with those from another method13.

In compliance with Dutch legislation, studies of this nature are exempt from the review of a research ethics committee. After providing the participants (i.e., physicians) with adequate information, we considered their consent to complete the survey and participate in the focus groups as sufficient informed consent for the study. The study was conducted in accordance with the guidelines of good clinical practice and in accordance with the ethical principles outlined in the Declaration of Helsinki.

Data collection and participants

Quantitative study

Survey

A self-administered digital survey was conducted from November 2021 until January 2022 to identify, compare, and understand how de novo mHSPC patients are managed across the Netherlands. The TripleAiM1 network facilitated the distribution of the survey through Qualtricsxm software (Qualtrics, Provo, UT) to urologists and oncologists specialising in and/ or treating (metastatic) PCa in academic, top clinical, and general hospitals14. A representation consisting of 91 physicians from 13 (TripleAiM1) hospitals was considered sufficient for this study. Two email reminders were sent to non-responders at one and three weeks after the initial survey invitation.

The survey included five patient cases and 19 questions regarding diagnostic imaging modalities (3 items), definitions (7 items), important factors for treatment selection (3 items), treatment strategies (1 item) and follow-up of de novo mHSPC (5 items). Response formats ranged from multiple-choice questions, open-ended text responses and four-point Likert scales, to numerical sliders (Additional file 1). All five patient cases were ECOG 0 with no comorbidities, symptoms or contraindications to treatment strategies. Case 1 involved a 75-year-old patient with de novo M1a PCa and > 10 lymph node metastases. Case 2 is that of a 65-year-old patient with de novo M1b PCa and two bone metastases, while Case 3 concerns a 72-year-old patient with de novo M1b PCa and 11 bone metastases, all of which are located within the axial skeleton. High-volume cases according to CHAARTED were Case 4, which included a 70-year-old patient with de novo M1b PCa and five bone metastases within the axial skeleton and one in the peripheral skeleton, and Case 5, which described a 70-year-old patient with de novo M1c PCa with five bone metastases and one liver metastasis.

Multidisciplinary team consultations.

Addressing the same objective, but then at a multidisciplinary team (MDT) level, four de novo mHSPC patient cases were developed in collaboration with experts from Mirrors of Medicine™ (MoM) (Additional file 2). These fictitious patients were presented during MDT consultations at four hospitals in the Netherlands.

Qualitative study

Urologists and oncologists specialised in (metastatic) PCa, who were affiliated with participating hospitals in the TripleAiM1 registry, were recruited by email. Participants were then selected using the quota sampling technique, based on pre-defined criteria. As the ideal focus group consists of an expert panel of equal urologists3 and oncologists3 from different hospitals throughout the Netherlands, we selected participants based on those criteria. Participants unwilling/ unable to attend all focus group sessions were excluded. Because more than six physicians wished to partake in the focus group, further selection was based on the diversity distribution (specialty, gender, age, hospital type, and geographical ___location) of the group. Participation was voluntary and participants did not receive any compensation.

The checklists used for the semi-structured discussions in the second and third focus groups were based on quantitative findings and key themes identified during the first session, which focused on the ideal situation and framework for managing de novo mHSPC. The second session focused on definitions, imaging modalities, and follow-up imaging, while the third session explored patient cases and the treatment decision-making process in more detail (Additional file 3). All focus group sessions were conducted digitally via Microsoft TEAMS, an online communication platform, between June 2022 and February 2023. The sessions were videotaped and transcribed (Additional file 4).

Data-analysis

Quantitative study

Quantitative data was analysed using descriptive statistics: frequencies, medians, minimum and maximums. The Chi-squared test was employed to ascertain significance among categorical variables, while the Mann-Whitney U test was used when analysing continuous variables. Significance was held at the standard value of P < 0.05 in two-sided tests. Statistical analyses were performed in IBM SPSS version 27.

Qualitative study

Focus group transcripts were reviewed and coded to identify major themes in Atlas.ti23 (Additional file 5). In addition to narrative thematic analysis, themes were then analysed using the matrix methodology to examine patterns by specialty (urologist versus oncologist) and type of hospital (academic, top clinical and general)15,16. By assessing major themes in matrices with both the frequency of the responses and the detailed content of responses, the assessment of both patterns of association and the nature of the association was possible16.

Results

Of the 91 surveys distributed, 27 urologists and 19 oncologists responded, representing a total of 13 hospitals. Table 1 shows the characteristics of the survey respondents.

Table 1 Characteristics of survey respondents.
Full size table

The focus group consisted of six experts: three urologists and three oncologists, with one oncologist and one urologist from academic hospitals. The male-to-female ratio was 2:1. Additionally, one expert of nuclear medicine and one radiation oncologist participated in the session relevant to their specialisation, focus group 2 and 3, respectively.

A total of 201 codes were extracted from the focus group transcripts and grouped into 45 axial codes, 13 categories and six themes: (1) Imaging modality for staging mHSPC; (2) Factors influencing treatment decisions; (3) Treatment landscape of de novo mHSPC; (4) Potential reasons for practice variation; (5) Multidisciplinary team consultations (Additional file 6); and (6) Shared decision-making (Additional file 7).

Imaging: which modality should be used for initial staging?

Most survey respondents (36/46) used PSMA-PET/CT for initial staging of de novo mHSPC. Three used CI and seven answered ‘other’, as the modality depends on patient characteristics such as initial serum PSA and age. The majority of survey respondents agreed (38/40) that treatment decisions can be made based on PSMA-PET/CT findings and that CI can be replaced by PSMA-PET/CT for initial staging of de novo mHSPC (35/40). However, most respondents disagreed with the statement that treatment policy can only be established when patients are staged with PSMA-PET/CT (15/40). This was confirmed by our expert panel, which reached consensus on PSMA-PET/CT, as the modality of choice for initial staging of de novo mHSPC.

“Conventional imaging is really too outdated. You simply miss too many lesions. We observe (on PSMA-PET) so many lesions of three, four millimetres that you do not see on CT” – SP7.

“We have superior imaging in the form of PSMA-PET/CT. We also no longer revert to abdominal X-ray or ultrasound. We now have excellent imaging and we should simply use that. This has always been the trend in many aspects of medicine.” – SP1.

The reasons for preferring PSMA-PET/CT include its scientifically proven superior accuracy to CI, presumed better patient selection and the avoidance of unnecessary treatments (i.e., radical curative treatments). On the other hand, PSMA-PET/CT is more expensive, has longer waiting times and its effect on overall survival is currently unknown (Table 2 and Table 3). Therefore, the expert panel argued that the therapeutic goal should determine the type of imaging. They concluded that PSMA-PET/CT is the modality of choice for routine staging patients of mHSPC, unless there are valid reasons to deviate from this based on patient characteristics such as age, PSA and frailty.

“You also have the frail older patient where you have to ask yourself: ‘What is the goal of the treatment?’ Is it to go all the way and aim for survival, or do you primarily want to prevent the occurrence of symptoms? The latter can be very well observed on conventional imaging.” – SP1.

“Especially for an older patient with limited life-expectancy for who you already opt for watchful waiting in advance, we are not eager to perform a PSMA-PET/CT to thoroughly investigate the extent of metastases ” – SP2.

“ (…) So, I think it is important to say that out loud here, that it (PSMA-PET/CT) is not necessary for everyone” – SP7.

Table 2 Expert opinions on PSMA-PET/CT by specialty.
Full size table
Table 3 Expert opinions on PSMA-PET/CT by hospital setting.
Full size table

Factors influencing treatment decisions

In the context of treatment decision-making, survey respondents identified metastatic volume (8.5, range 3–10), performance (8.6, range 5–10) and patient preference (9.0, range 2.8–10) as the most influential factors. The medians did not differ significantly between specialties, except for initial PSA (p = 0.007), but showed large dispersion within specialties (Fig. 1). Similarly, no significant differences were found between physicians from academic, general and top clinical hospitals, except for initial PSA (p = 0.02). However, the top three factors for academic physicians differed to those from the overall group, as they considered metastasis ___location (8.7) to be more important than performance score (8.1) and volume (8.1), which both ranked third (Additional file 8).

Fig. 1
figure 1

Importance of factors for treatment decision-making by specialty. Whiskers represent minimum and maximum and the box is IQR range with medians. On the, X-axis factors influencing treatment decision-making are displayed, while the Y-axis represents the importance score raging from the lowest (least important) to the highest (most important). LUTS lower urinary tract symptoms, iPSA initial PSA.

Full size image

The expert panel reached consensus on metastatic volume (HV/LV), comorbidities, and patient preference, as the most important factors for treatment decision-making. In contrast to the survey respondents, the panel prioritised comorbidities over performance status, but argued that both factors are intrinsically linked. According to them, performance status plays an important role in determining the need for treatment intensification, while comorbidities are especially important in identifying risks of interactions with current medications and reducing the likelihood of side effects.

In addition to the factors mentioned above, the expert panel also discussed the importance of considering healthcare costs, on which they were divided. While some experts emphasised their responsibility to reduce healthcare burden by incorporating cost differences of agents in their treatment decisions, others expressed concern that health insurance companies may play a too prominent role. This raised the question of how to balance physician preferences based on study outcomes with cost.

“I think it is dangerous that we have to do this [consider finances] more and more. Because then we are letting a political agenda into our consultation room: we are now forced to make a certain choice because the cost of national healthcare is rising. And I think it will get worse and worse, and the next thing you know, the health insurers will be sitting next to us in the consultation room.” – SP6.

“If there is an equivalent drug with similar side effects, the choice based on cost is inconsequential, but not when decisions are made based on monetary considerations.” – SP5.

Treatment landscape of de novo mHSPC

Discrepancies in the choice of treatment were found at the level of the individual (physician), through patient cases in surveys, and at the level of the MDT. Especially LV cases showed most practice variation, ranging from treatment with curative intent to palliative triplet therapies consisting of ADT, prostate radiotherapy and docetaxel/ ARTA (Fig. 2). The expert panel argued that metastatic disease generally has a palliative approach, but with next-generation imaging techniques, therapy with curative intent may be considered in young, fit patients with limited number (up to 4) of metastases.

Fig. 2
figure 2

Practice variation (physician level) for low volume mHSPC. All patient cases were ECOG 0, without comorbidities, symptoms and contra-indications for treatment strategies.Case 1 75 year old patient with newly diagnosed M1a prostate carcinoma, > 10 lymph node metastases, low volume (per CHAARTED); Case 2 65 year old patient with newly diagnosed M1b prostate carcinoma, 2 bone metastases, low volume (per CHAARTED); Case 3 72 year old patient with newly diagnosed M1b prostate carcinoma, 11 bone metastases all within the axial skeleton, low volume (per CHAARTED). Source: survey. Number of respondents: 23 urologists, 16 oncologists. mHSPC metastatic hormone-sensitive prostate cancer, ADT androgen deprivation therapy, RTx radiotherapy to the prostate, MDT metastasis directed therapy, Doc docetaxel, Apa apalutamide, Abi abiraterone, Enza enzalutamide, ECOG Eastern cooperative oncology group performance scale.

Full size image

In HV, the treatment variation was less pronounced than in LV; the majority of the Dutch oncologists and urologists would treat HV patients with ADT + Doc (Fig. 3). The expert panel acknowledged that docetaxel traditionally was the first choice for HV patients, as the major drawback of ARTAs was the higher cost. With abiraterone (Abi) going off patent, the panel foresees a shift from docetaxel towards use of Abi. One expert expressed a strong preference for an ARTA, believing it to be more effective and less toxic than docetaxel based on indirect meta-analyses.

“But my preference nowadays - is for an ARTA based on indirect meta-analyses between all those studies. I am really convinced that an ARTA is more effective than docetaxel based on the updates of all studies, including the CHAARTED study. Also for the quality of life, there has been an indirect analysis in STAMPEDE that shows that- yes, you do have quite a dip with docetaxel, and it does recover. However, even after a year, you are still just below the level of an ARTA. Plus, I think that you do affect a patient quite a bit with docetaxel.” – SP5.

Another expert indicated a weaker preference for ARTAs, allowing the choice to be patient-driven, but agreed that patients are more affected by docetaxel in terms of QoL and side effects, highlighting the lack of equivalence between these treatment strategies. Others did not express a preference for one over the other. The expert panel agreed that equivalence between ARTAs is expected.

“I don’t think there is an oncologist who believes that Enza/ Apa is better than Abi.” – SP4.

Fig. 3
figure 3

Practice variation (physician level) for high volume mHSPC. All patient cases were ECOG 0, without comorbidities, symptoms and contra-indications for treatment strategies.Case 4 70 year old patient with newly diagnosed M1b prostate carcinoma, 5 bone metastases within the axial skeleton and 1 in the peripheral skeleton, high volume (per CHAARTED); Case 5 70 year old patient with newly diagnosed M1c prostate carcinoma, 5 bone metastases and 1 liver metastasis, high volume (per CHAARTED). Source: survey. Number of respondents: 23 urologists, 16 oncologists. mHSPC metastatic hormone-sensitive prostate cancer, ADT androgen deprivation therapy, RTx radiotherapy to the prostate, MDT metastasis directed therapy, Doc docetaxel, Apa apalutamide, Abi abiraterone, Enza enzalutamide, ECOG Eastern cooperative oncology group performance scale.

Full size image

The expert panel expressed their concern about the practice variation observed. They argued that practice variation is not necessarily problematic, but unwanted variation is, which they believed is the case here. Reasons given for this variation include the heterogeneous aspect of LV disease, ambiguous definitions, different interpretation and implementation of the same studies, and current guideline gaps.

“What worries me is the practice variation. We are now discussing cases with six people and there is quite a lot of variation in treatment plans. So it matters where you go as a patient (…). That’s quite worrying, isn’t it?” – SP1.

Potential reasons for practice variation

Inadequate definitions

Metastatic volume

Although volume was considered important for therapeutic decision-making, its definition was contested by the expert panel. They argued that the definition is based on statistical tests and that most of the gain is expected to come from outliers on either side of the spectrum. In addition, HV and LV per CHAARTED is defined on CI, whereas the majority of the Dutch urologists and oncologists use PSMA-PET/CT for initial staging and the translation of the definition to PSMA-PET/CT findings remains elusive. Nonetheless, both the survey respondents (46/46) and the expert panel used the same definition for PSMA-PET/CT and CI. One expert explained that the assumption is that there is no immediate upstaging in terms of volume.

“While additional metastases may emerge, they are not expected to be extensive.” - SP1.

This expert also questioned the validity of the assumption of upstaging (“or is it merely enhanced diagnostics”) and whether clinically relevant upstaging occurs.

Furthermore, the expert panel argued that the definition itself it is not as black and white as it sounds and that it is not really suitable for use in clinical practice. One of the reasons given was the complexity of the definition, as it includes not only the number of bone metastases but also whether the bone metastases are in the axial or peripheral skeleton. This is particularly relevant for LV, as it represents a diverse clinical picture, ranging from lymph node-only disease (M1a) to multiple bone metastases (M1b) confined to the axial skeleton. As a result of this diversity, the expert panel argued that they do not always follow the guideline when selecting treatments in clinical practice and that some patients are treated as HV when they are technically LV, depending on the volume, ___location and number of metastases. An example provided was when patients are considered to have bulky or extensive M1a disease. This was also seen in the survey where the majority (24/40) responded that bulky/ extensive M1a should be considered as HV and therefore treated as such.

M1a

Beyond the discussion of whether bulky/ extensive M1a should be considered as a separate entity and how it should be defined and managed, the definition of M1a itself appeared to be unclear to survey respondents, with ambiguous definitions of M1a being considered. The majority diagnosed M1a when lymph node metastases are above the aortic bifurcation (25/40), whilst others used the iliac bifurcation (10/40) as the cut-off point between N1M0 and M1a. Four did not use strict criteria, leaving the diagnosis of M1a to the extent of the disease. Lastly, one concluded M1a when lymph nodes are outside the surgical template for lymph node dissection. In contrast, the consensus of the expert panel was that at least one metastasis had to be above the iliac bifurcation for M1a to be diagnosed.

Different interpretation and implementation of the same studies

The expert panel noted differences in the interpretation of definitions and study results, resulting in diverse approaches to treating patients. This interpretation difference was exemplified by patients diagnosed with M1a, who were considered to be understudied, as large randomised controlled trials (RCTs) only included a limited number of M1a patients commonly, without further denoting this specific patient group in relation to outcome. Consequently, one expert reasoned that the current definition of volume excludes M1a, which prevents patients with M1a disease from being categorised as HV or LV. As a result, he recommended ADT monotherapy for M1a. Others argued that although the definition is inadequate, LV encompasses everything that is not HV, including M1a. These experts recommended ADT in combination with either radiotherapy or an ARTA, for M1a patients.

“(…) data is the same, but the way we look at it is different. That is the biggest concern for me, you know, because (…) why is the interpretation different (…) Is it not clearly written or is our perspective not clear?” - SP8.

The experts also differed in implementation of newly approved therapeutic agents. Some experts only administered new agents once they are outlined in international and/or national guidelines, while others integrated new therapies into their daily practice sooner. An example cited was the implementation of triplet therapies. Figure 2 illustrates the use of triplet therapy, despite them not being included in national and international (e.g., National Comprehensive Cancer Network (NCCN) and European Association of Urology (EAU)) guidelines. For LV, only one expert has integrated triplet therapy with ADT + ARTA + RTx into clinical practice. Others argued that RCTs have not yet demonstrated its significant superiority and therefore have not yet implemented triplet therapy with ADT + ARTA + RTx in this patient population. For HV, the expert panel agreed that triple therapy with ADT + Doc + Abi can be administered to young, healthy patients. For older patients, they argued that it may be considered in the future, but only after demonstrating that there is no significant reduction in QoL. The expert panel emphasised the importance of obtaining real-world data to better understand the impact of treatment strategies, as the average patient seen in the clinic is typically older, has a lower performance score, and more comorbidities, than those selected for RCTs.

Guideline gaps

The expert panel stated that guidelines should be the foundation for determining treatment, and any deviation from it should be a deliberate decision with appropriate justification. However, they argued that guidelines should ideally include well-defined definitions and clear standards, which they find lacking in the case of de novo mHSPC.

“It starts with the definition, which often proves challenging.” - SP3

“We have these guidelines, and yes, apparently they do not provide enough clarity to consistently give the same treatment to the patient.” – SP1.

Furthermore, guidelines exist at different levels: international, national, regional, and hospital-specific, each with their own nuances, for both urologists and oncologists. The expert panel observed a lack of awareness of the differences in guidelines and subsequent practice between specialisms, hospitals and regions. A need for greater transparency between these guidelines was expressed.

“(…) but I think that now something like interaction and communication between the regions is quite useful. Then you actually know—I was a bit shocked that there is such a big difference.” – SP8.

Two inherent disadvantages of guidelines themselves were also identified: (1) guidelines tend to lag behind, as seen in the implementation of triplet therapies; and (2) guidelines are consistently based on level 1 evidence. As a result of the latter, societal and financial considerations tend to get overlooked. The expert panel highlights the need for a more assertive approach (“a kind of guideline plus”), taking into account both financial and societal interests.

Discussion

The findings presented in this study shed light on the variation in clinical practices and perspectives regarding the management of de novo mHSPC among physicians in the Netherlands. The results of both the survey and the expert panel discussions offer valuable insights into various aspects of clinical decision-making, ranging from the use of imaging modalities to shared decision-making processes. Our findings reveal notable disparities in treatment approaches, particularly evident in cases of LV disease, where strategies ranged from those with curative intent to palliative triplet therapies. Subsequent analysis suggests that practice variation may be attributed to the heterogeneous nature and prognosis of LV disease, ambiguous definitions due to the introduction of molecular imaging modalities, varying interpretations of study data, lead-time in adoption of novel treatment strategies, and guideline gaps.

Guideline adherence

At present, it is unclear whether practice variation in itself is detrimental. However, if guidelines are deemed the gold standard, deviation from them without proper justification can be considered undesirable. We found a high level of adherence to guidelines for HV disease, with an adherence of 90% for Case 4 and 97% for Case 5, when considering ADT + Doc or ARTA as the gold standard according to the 2021/2022 Dutch and EAU guidelines17,18. For LV disease, guideline adherence was highest in Case 1 at 72%, but dropped significantly to 36% in Case 2 and 33% in Case 3, when ADT + RTx or ARTA is considered the standard-of-care17,18. These results raise the question of what deviations are causing this low adherence to guidelines in LV disease.

ADT monotherapy for M1a

First, although ADT monotherapy is considered to be obsolete and guidelines recommend reserving this treatment strategy for patients with limited life expectancy, our results indicate that ADT monotherapy is still selected for patients with M1a disease (Case 1: 2/39), regardless of life expectancy. These findings are consistent with those of a Dutch consensus meeting held in 2021, where almost a quarter of the panelists voted for ADT monotherapy as an option for M1a19. This is likely due to their underrepresentation in pivotal trials and the often-limited power of subgroup analyses. However, results from the STAMPEDE trial (arm H) and retrospective studies have highlighted the significant survival benefits of adding radiotherapy to ADT in de novo M1a20,21. Therefore, ADT monotherapy in M1a patients without limited life expectancy could be considered controversial.

Addition of prostate radiation to ADT + docetaxel or ARTA

We observed that a varying number of physicians opted for triplet therapies (e.g., ADT + RTx + ARTA or ADT + RTx + Doc), ranging from two (Case 3) to three (Case 1) and six (Case 2), despite it not being recommended in the guidelines at that time17,18,22. This trend was also noted during the 2022 Advanced Prostate Cancer Consensus Conference (APCCC), where 33% of the participants voted for ADT + ARTA + RTx for patients with LV disease with 1–3 bone lesions23. Complementary but preliminary data from the PEACE-1 trial suggests that treatment of LV disease with standard-of-care (SOC) + Abi + RTx resulted in superior PFS outcomes compared to those treated with SOC + Abi (median 7.5 versus 4.4 years, p = 0.02)24. However, in our study, the most frequently selected triplet was ADT + Doc + RTx7, followed by ADT + Enza + RTx3, and ADT + Abi + RTx1. Although no trials have researched ADT + Doc + RTx, the results from the PEACE-1 trial demonstrated an improved time to serious genitourinary events when prostate radiotherapy was added, irrespective of the treatment strategy. This was observed in both SOC + Abi + RTx (p = 0.003) and SOC + RTx (p = 0.048), where SOC consisted of ADT + Doc in 50% of the population24. This prompts the question of whether prostate radiotherapy should be added to all treatment strategies, not primarily to increase survival but to prevent or delay the onset of serious genitourinary events.

Metastasis-directed radiation therapy

Despite the lack of level 1 evidence and recommendations in both national and international guidelines, 15 physicians chose to add metastasis-directed radiation therapy (MDRT) to either short-term ADT + RTx with curative intent (10/39) or ADT + RTx with palliative intent (5/39) in patients with two bone metastases (Case 2)17,18,22. Support for the addition of MDRT is evident in consensus meetings23,25. During the 2022 APCCC, 61% of the panellists voted for systemic therapy plus local treatment of the primary and MDRT in oligometastatic (1–3 bone lesions) synchronous mHSPC23. Accordingly, 76% of the panelists recommended MDRT for all metastatic sites, mostly in combination with ADT + ARTA (40%), as shown by Zilli et al.25. However, data on the survival benefit of MDRT in mHSPC are currently limited to metachronous disease, which makes implementation of MDRT for synchronous mHSPC in clinical practice questionable26.

Use of docetaxel

ADT + Doc was identified as the preferred treatment strategy (Case 3: 21/39) for patients with multiple bone metastases (> 10) confined to the axial skeleton, although not recommended in the Dutch guideline17. The expert panel argued that they do not always adhere to the guideline when selecting treatments due to the heterogeneity of LV disease. This results in some patients being treated (and regarded) as HV when they are technically LV, depending on the volume, ___location, and number of metastases. In contrast to the CHAARTED trial, post-hoc analyses of the STAMPEDE trial (arm C) showed significant survival benefits, irrespective of volume when adding docetaxel to ADT27. However, it is important to exercise caution as these analyses were unplanned and performed posthoc, and these benefits are only marginal in this population, while the toxicity is similar10. At the 2022 APCCC meeting, the majority of the panellists (74%) voted against recommending ADT + Doc for LV disease, instead preferring ADT + ARTA. It is interesting to note that only 11 physicians in our study chose ADT + ARTA in Case 3, despite it being a treatment strategy for both HV and LV disease. This preference for docetaxel was also seen in cases with HV disease, as the majority chose ADT + Doc (Case 4: 28/39; Case 5: 32/39) instead ADT + ARTA (Case 4: 5; Case 5: 4), in line with the Dutch guideline which favored docetaxel over ARTA due to lower costs17,28. This preference for docetaxel contrasts with the results from the 2022 APCCC, where 49% voted against ADT + Doc, assuming that ARTA is available. However, our expert panel predicts a shift towards ADT + Abi in Dutch clinical practice with the expiration of abiraterone’s patent. Especially, because recent network meta-analyses have shown that all ADT + ARTA combinations have superior survival outcomes compared to ADT + Doc29,30,31. Additionally, Rush et al. demonstrated significantly better QoL outcomes in the first year and continued improvement, although not statistically significant, after two years for ADT + Abi32. These direct and indirect comparisons highlight the lack of equivalence between these treatment strategies. As a result, the current EAU and NCCN guidelines recommend ADT + ARTA and advise docetaxel only in a triplet regimen, either ADT + Doc + Abi or ADT + Doc + darolutamide (Daro), for those patients fit enough for docetaxel22,33,34.

PSMA-PET/CT and interpretation

A potential factor contributing to these guideline deviations and practice variation is the widespread implementation of PSMA-PET/CT for the initial staging of mHSPC in the Netherlands, as its results may lead to definitional ambiguity due to uncertainty about the interpretation and clinical significance of detected metastases. The proPSMA trial demonstrated that PSMA-PET/CT has a 27% higher accuracy than CI (92% versus 65%), resulting in a change in management in 28% of cases7. Because of this higher accuracy compared to CI, the NCCN guidelines do not advocate the use of CI as a prerequisite for PSMA-PET/CT, as they consider it to be a similarly effective, if not more effective, frontline imaging tool for staging34. In contrast, the EAU guideline refrains from recommending the routine use of PSMA-PET/CT for staging mHSPC due to the lack of robust data on long-term oncological outcomes following treatment modifications35. However, Bauckneht et al. found that patients upstaged by PSMA-PET/CT but treated based on CI staging had a shorter time to biochemical recurrence, suggesting potential benefits in oncological outcomes when treatment decisions align with PSMA-PET/CT stage36. This supports the expert panel’s argument that PSMA-PET/CT can enhance patient selection and mitigate unnecessary treatments.

However, challenges persist in interpreting PSMA-PET/CT findings, particularly in distinguishing between HV and LV disease. The 2022 APCCC raised concerns about potential stage migration from LV to HV disease when using PSMA-PET/CT, highlighting these uncertainties23. In Dutch clinical practice, the same definition is used for the assessment of metastatic volume on PSMA-PET/CT as for CI (46/46). Nevertheless, the relevance of differentiating HV from LV is becoming more nebulous and blurred with each successive update of landmark trials and publication of new trials4,5,6,27. The expert panel noted that there is a growing tendency to abandon the volume definition as a strict criterion, which is reflected in the EAU guidelines. However, Dutch physicians continue to adhere to it in practice and even consider it as one of the top three factors that influence treatment decision-making.

Factors influencing treatment decisions

Our results show no significant differences between specialties (i.e., urologists and oncologists) and hospital types (i.e., academic, top clinical and general) when quantifying factors influencing treatment decisions, except for initial PSA, which was considered significantly more important by urologists and physicians in academic hospitals (p = 0.007 and p = 0.02, respectively). The clinical significance of this observed between-group difference in clinical practice is expected to be limited, as impact of initial PSA on outcomes in de novo mHSPC is not demonstrated in current literature. We did observe large within-group heterogeneity, suggesting variation at the individual physician level rather than between specialties or hospital types, which could potentially influence the shared decision-making process. The expert panel discussions highlighted this influence, revealing that some treatment options were not mentioned during shared decision-making because of their higher cost. And while the patient preference was considered paramount, challenges exist in ascertaining patient preferences effectively and incorporating them into the decision-making process (Additional file 7). Variation at the individual physician level might be diminished by more internal, regional and national discussion forums, resulting in more education and consensus in areas of uncertainty. In addition, a decision aid might assist in ascertaining patient preferences while reducing the impact of individual physician preferences.

Nevertheless, tailoring the decision-making process to the individual patient is challenging due to limited evidence on the impact of patient and disease characteristics on outcomes. Recent studies have shed light on potential relevant characteristics other than metastatic volume. Vale et al. demonstrated that docetaxel had the most significant impact on survival in patients with HV disease and advanced T4 tumours, suggesting T stage as a potentially relevant characteristic8. Morgans et al. showed that older patients (≥ 70 years) treated with ADT + Doc or ADT + Abi had minimal survival benefit, suggesting an interaction between age and these treatment doublets37. However, additional trials are necessary to predict which patient responds better to which treatment strategy, to mitigate treatment side effects while improving survival outcomes and preserving QoL.

Limitations

While this study is unique and provides a comprehensive analysis using a mixed-method approach, it is not without limitations. Firstly, the survey was conducted during Q4 2021 and Q1 2022, which may yield clinical case outcomes inconsistent with current practice due to the rapidly evolving treatment landscape of mHSPC. However, the publication of the PEACE-1 and ARASENS data is expected to result in even greater variation in clinical practice, so we deem this study to remain highly relevant38,39. Additionally, this study sheds light on Dutch clinical practices that may differ from international standards, although the issue of practice variation and guideline deviation is relevant worldwide40. Secondly, the selection of participants from the TripleAiM1 network may introduce selection bias, as a subset of urologists and oncologists in the Netherlands were approached from top clinical and academic institutes and/or interested in research. As they are likely to make adoption decisions relatively early in the process of implementing novel standards of care, they are of particular interest in this study38. Finally, it is important to note that practice variation is based on survey responses, which may not accurately represent prescribing behaviours.

Conclusion

Clinical practice for de novo mHSPC differs considerably among physicians in the Netherlands, especially in cases of LV disease. Our findings highlight the complexity of treatment decision-making, which is influenced by physicians’ interpretation and implementation of study results, as well as ambiguous definitions and gaps in the guidelines. Furthermore, PSMA-PET/CT is the preferred modality for initial staging of prostate cancer, but its implementation contributes to ambiguity in definitions and subsequent variation in treatments. In conclusion, there is a need for consensus on definitions, imaging modalities for staging (and follow-up), and treatment selection given the altered diagnostic and therapeutic landscape.