ESUR: Opportunities for PSMA-PET/CT and whole-body MRI in advanced prostate cancer.

Introduction
Advanced prostate cancer (APC) refers to locally advanced prostate cancer (PCa), metastatic hormone-sensitive prostate cancer (mHSPC), and metastatic castration-resistant prostate cancer (mCRPC) [1, 2]. APC has a range of outcomes, from aggressive pelvic-confined disease to rapid progression and death from metastases. In the coming decades, we can expect an increasing reliance on image-based diagnostic, prognostic, predictive, and response biomarkers to advance patient care. Imaging biomarkers will be integrated with molecular and clinical parameters to enhance risk-based diagnoses and inform therapy selection [3].
Key imaging methods include prostate-specific membrane antigen (PSMA) positron emission tomography (PET) computed tomography (CT) and whole-body magnetic resonance imaging (WB-MRI), which demonstrate superior diagnostic capabilities to detect metastatic disease compared with conventional imaging (bone scans (BS), CT, and regional MRI). Comparative studies show that PSMA-PET/CT outperforms WB-MRI in the detection of nodal and bone metastases [4–6]. As a result, PSMA-PET/CT has become an integral part of the standard of care in several clinical settings, with WB-MRI playing a supportive role. A thorough understanding of the relationship between the treatment landscape and disease volume on conventional imaging, as well as the prognostic significance of prostate-specific antigen (PSA) response, is crucial for determining how higher-accuracy imaging can be effectively incorporated.
In this multidisciplinary collaborative effort by radiologists, nuclear medicine physicians, oncologists, and urologists with expertise in APC, we explore the specific opportunities for high-accuracy imaging in cancer detection and staging, disease characterisation, and aggressiveness assessments, as well as biopsy target selection. The impacts on treatment planning, evaluation of therapeutic response, and theranostics are assessed, and major research questions that need to be addressed in future clinical trials are posed.

Management landscape of mHSPC
mHSPC can present in men at the time of diagnosis (synchronous) or develop in those who have previously undergone definitive treatment (metachronous) [7]. Generally, patients are classified into low- and high-volume disease states using CHAARTED criteria on conventional imaging or LATITUDE criteria that additionally incorporate the Gleason score [8–10]. Both emphasise the presence of bone and visceral disease as adverse prognostic factors [11, 12], with patients with nodal disease classified as having low volume/risk [4, 5]. High-volume synchronous metastatic disease has the poorest prognosis [7]. Endpoints, including radiographic progression-free survival (rPFS), time to development of castration resistance, and overall survival (OS), are poorer for higher disease volumes [13].
The management of mHSPC primarily stems from clinical trials conducted in synchronous settings. Typically, patients are treated with androgen deprivation therapy (ADT) and androgen receptor pathway inhibitors (ARPI) as a doublet. Chemotherapy-fit patients may benefit from treatment intensification with upfront use of docetaxel (triplet therapy) when the disease burden is high. The therapeutic goal of pelvic radiotherapy (curative or palliative) is also influenced by the volume of metastatic disease, aiming to improve OS for patients with low-volume disease and achieve local tumour control for those with high-volume disease, respectively [14, 15].

Power of PSA for response assessment and discordance with imaging
Achieving a profound decline in PSA is a critical surrogate marker of treatment efficacy and long-term prognosis in mHSPC. Across multiple randomised controlled trials and real-world data from the International Registry for Men with Advanced Prostate Cancer (IRONMAN), a PSA nadir of < 0.2 ng/mL within the first 6–12 months of systemic therapy has substantial prognostic and predictive value for OS and rPFS [16, 17]. However, correlative imaging studies suggest that biochemical responses alone may not fully capture disease dynamics in men with suboptimal PSA responses in both mHSPC and mCRPC.

PSMA-PET/CT
A systematic review of 268 mCRPC patients from 10 studies observed discordance between PSA and PSMA-PET/CT responses in approximately one-quarter of cases [18]. Both discordant patterns (imaging progression/biochemical response and biochemical progression/imaging response) were documented across various therapeutic settings and imaging criteria. More recent studies have corroborated these findings [19–21]. For instance, in a prospective study involving 69 mCRPC patients treated with enzalutamide, PSA kinetics and PSMA-PET/CT were only concordant in 48% [20].

WB-MRI
Similar discordances have been described between WB-MRI and PSA responses, especially in mCRPC. For instance, in a retrospective study evaluating WB-MRI in patients with mHSPC and mCRPC undergoing doublet treatment, all mHSPC patients who achieved a PSA level of ≤ 0.2 ng/mL exhibited no imaging progression [22]. However, non-response or progression was detected in 33.3% of patients who did not achieve a PSA level of ≤ 0.2 ng/mL for mHSPC and in 54.5% of those who did not have a > 50% decrease in PSA levels for mCRPC. WB-MRI depicted disease progression with a suboptimal PSA response had a higher risk of death (hazard ratio, 8.6). Likewise, in a prospective study of mCRPC patients on Olaparib, changes in circulating tumour cell dynamics more strongly correlated with WB-MRI tumour burden than serum PSA [23].
While the clinical implications of discordances between PSA and PSMA-PET/CT or WB-MRI require further investigation, earlier imaging identification of non-responders may assist in avoiding unnecessary toxicity and costs associated with ineffective therapies, allowing for timely transition to alternative treatment strategies when available.

Biases and the need for clinical trials to assess the clinical impact of high-accuracy imaging
In comparison to conventional imaging, PSMA-PET/CT and WB-MRI enhance both sensitivity and specificity for detecting metastatic disease and assessing suboptimal treatment responses. However, potential biases can create dilemmas in treatment initiations and selections, which include [3, 24–26]:Stage and risk migration. High-accuracy imaging can detect cancer at earlier stages than conventional imaging, leading to earlier treatment exposure for some patients. The reclassification of patients into different groups (e.g., from non-metastatic to metastatic or low-volume to high-volume disease) can inflate group survival rates without an actual improvement of individuals in each group, a phenomenon known as the “Will Rogers effect”.

Lead-time bias. High-accuracy imaging can identify the presence of disease or progression earlier than conventional imaging. If earlier therapy initiation for a lower volume of detected disease yields no additional therapeutic benefit, the apparent extended period of survival is spurious, leading to a false impression of improved effectiveness.

Length-time bias. High-accuracy imaging can detect slower-growing cancers more effectively than conventional imaging. This can result in apparent longer group survival durations due to the over-detection and over-treatment of lower-risk disease.

To mitigate biases, rigorous methodological approaches are necessary in clinical settings for specific clinical scenarios.

Opportunities for PSMA-PET/CT and WB-MRI
The opportunities of PSMA-PET/CT and WB-MRI as biomarkers for metastasis detection and staging, disease characterisation, and aggressiveness assessment, as well as biopsy target selection, are highlighted below. While there is no question regarding improved disease detection and staging, it is unclear whether the prognosis of patients whose bone metastases are detected solely through high-accuracy imaging is the same as that of those identified using conventional imaging (i.e., in the absence of tumour engagement with the bone matrix) [25]. We also need to remember that because nodal disease is considered low-volume in the CHAARTED definition [8], the clinical impacts of improved nodal disease detection on systemic therapy and pelvic radiotherapy are uncertain. The clinical benefits of therapy changes due to oligoprogession, shown by high-accuracy imaging, also remain unknown [27]. This highlights the need for prospective clinical studies with surrogate endpoints that demonstrate clinical benefit. In addition, while preliminary studies on cost-effectiveness have demonstrated feasibility, in-depth economic analysis and feasibility of implementing PSMA-PET/CT and WB-MRI with consideration of resource availability and allocation, need further investigation [28–30]. Comparative features of the imaging modalities and central research questions in these clinical domains are highlighted in Tables 1 and 2. Example methodological clinical trials are also highlighted in Table 2.

Disease detection and staging

PSMA-PET/CT
PSMA-PET/CT has greater accuracy in detecting metastatic disease than BS-CT. In patients with high-risk PCa, the ProPSMA trial demonstrated that PSMA-PET/CT had higher sensitivity (85% versus 38%) and specificity (98% versus 91%) for detecting pelvic nodal and distant metastases [31]. The improved accuracy of PSMA-PET/CT not only leads to upstaging/up-risking (due to higher sensitivity) but also to downstaging/down-risking (due to higher specificity) (Fig. 1). For example, 57% of positive bone scans were false positives when compared with PSMA-PET at initial staging [32]. Unterrainer et al [33] evaluated 67 patients with mHSPC who had at least M1a disease on conventional imaging. They noted that 24% of patients with M1 disease had only local (N0) or pelvis-confined (N1) disease on PSMA-PET/CT, with risk assessment changes occurring in 42% (24% down-risked and 18% up-risked).

WB-MRI
Lecouvet et al [34] demonstrated in patients with high-risk PCa that WB-MRI has higher sensitivity than BS-CT (98–100% versus 86%) for detecting bone metastases and similar sensitivity to CT (77–82% for both) for detecting metastatic nodes. A comparative study between PSMA-PET and WB-MRI showed that while the capability for detecting distant metastasis was similar, WB-MRI had a slightly inferior ability to detect nodal metastases [35]. WB-MRI also alters the risk burden of men with mHSPC, primarily related to the higher detection of bone-only metastases. In more than 200 age-matched patients, Hassan et al [36] noted that WB-MRI risk classification was more effective than BS-CT for predicting overall survival in men with mHSPC. The tumour burden depicted on WB-MRI was also found to be prognostic in mCRPC [37, 38].

Disease characterisation

Identifying biopsy targets for molecular analysis
Rebiopsy and genomic analysis are recommended for patients exhibiting intrinsic or acquired resistance to guide treatment and inform potential trial involvement. For patients with metastatic PCa, bone biopsies are often the only source for molecular analysis of actionable mutations. Successful tissue sampling can aid in identifying mutations that inform personalised treatments [39].

PSMA-PET/CT
Several studies have demonstrated that PSMA-PET/CT-informed bone biopsy results in high success rates for molecular analysis, ranging from 66% to 70% [40, 41]. High maximum standardised uptake values (SUVmax) at PSMA-PET and low Hounsfield Units (HUs) at CT are strong predictors of success. In a study of 69 patients who underwent PSMA-PET/CT, samples suitable for whole-genome sequencing had a median SUVmax of 20.9 and a HU of 786 [41]. Donners et al [42] also noted that HU affected histological yields, with a 610 HU threshold having a positive predictive value of 89% for tumour-positive biopsies and a 370 HU threshold for successful next-generation sequencing.

WB-MRI
Similarly, a prospective study of 20 patients using multiparametric WB-MRI assessments found that 85% of samples were positive for bone metastasis; 72% were suitable for genomic sequencing [43]. Using biopsy yields in 43 patients evaluated on WB-MRI, the combination of hyperintensity on high b-value diffusion-weighted images (DWI), apparent diffusion coefficient (ADC) values < 1100 μm2/s, and a relative fat fraction (rFF) of < 20% based on the T1-weighted Dixon technique had a PPV of 93% [44]. Another report on 10 patients with mCRPC showed that combining multiparametric PSMA-PET/CT and WB-DWI had success rates of 90% for positive biopsy and 80% for successful molecular analysis [45].

Characterising aggressive disease variants
While ARPIs are the backbone of mHSPC treatment, 20–25% of patients do not experience durable responses beyond 2 years [46–48]. Aggressive histologic and clinical phenotypes can emerge, which are less dependent on androgen receptor signalling, including neuroendocrine, small cell, and aggressive variant adenocarcinomas (Fig. 2). Multiple molecular events (e.g., PTEN deletion, RB1 loss, and p53 deficiency; intrinsic or as a result of effective androgen receptor blockade) can contribute to lineage plasticity, resulting in the appearance of treatment-emergent aggressive variants [49]. Conventional imaging features can suggest the presence of aggressive variants, such as the presence of bulky metastatic disease, predominant lytic bone metastases, and very low PSA levels when associated with a high tumour volume [50]. The role of WB-MRI or PSMA-PET/CT for identifying and assessing the therapy response of aggressive disease variants is not well established. WB-MRI can be effective because it evaluates tissue cellularity and bone marrow replacement. On the other hand, early studies indicate that aggressive variants may downregulate PSMA expression [51] and upregulate glycolytic metabolic activity, enabling [18 F]-fluorodeoxyglucose (FDG)-PET to serve as a prognostic biomarker [52]. Furthermore, an integrated assessment using multiple PET radioligands (e.g., upfront FDG-/PSMA-PET followed by DOTATATE-PET) may provide prognostic information to assist in decision-making [53].

Impacting treatment management
The higher accuracy of PSMA-PET/CT is refining radiotherapy (RT) planning at initial staging, for salvage and metastasis-directed radiotherapy [54].

Metastasis-directed therapy for oligometastatic disease
The ability to accurately identify small lesions can guide highly focused therapies, such as stereotactic body radiation therapy, thereby minimising collateral damage to healthy tissues. Oligometastatic disease represents an intermediate state in which targeted local therapy can be beneficial [55]. Both PSMA-PET/CT and WB-MRI are used for patient selection and to guide metastasis-directed treatments (MDT) in patients who typically have five or fewer metastatic sites of disease [55]. MDT has demonstrated favourable disease-free survival and OS, particularly for patients with low-volume bone disease in the metachronous mHSPC setting, especially when informed by PSMA-PET/CT [56–58].
The PEACE V-STORM study showed that the MDT approach may not be equally applicable to oligo-metachronous PET-detected nodal recurrences. In this Phase 2 randomised study comparing MDT and regional radiotherapy, metastasis-free survival was better with regional radiotherapy, presumably because microscopic metastatic disease is not seen [59]. This result is consistent with the findings of the POP-RT study, which also noted that pelvic node radiotherapy cannot be omitted based solely on negative PSMA-PET/CT results, given the low node detection sensitivity of 58% (95% confidence interval, 50–66%) [60, 61]. The role of MDT for oligoresidual disease and oligoprogressive disease is under investigation (Table 2).

Optimising radiotherapy fields and surgical approaches
MRI-guided focal boost radiotherapy involves delivering a higher dose of radiation to specific areas of the prostate, identified as tumours on MRI. This treatment improves treatment outcomes, such as biochemical disease-free survival, and potentially reduces the risk of recurrence, without significantly increasing side effects [62].
PSMA-PET/CT can play a significant role in refining radiotherapy by enhancing target identification and dose escalation to the disease that may not have been visible with conventional imaging [63]. The long-term treatment outcomes of PSMA-PET/CT-guided radiotherapy plans are currently under investigation (Table 2). Although WB-MRI using diffusion-weighted sequences is excellent for depicting the location of all lymph nodes, its use for radiotherapy planning has not been explored due to its poor sensitivity for detecting involved nodes.
PSMA-radioguided surgery may also assist surgeons in identifying involved lymph nodes in patients undergoing extended pelvic lymph node dissection and for sentinel lymph node sampling [64, 65]. However, further follow-up data are needed to assess the impact of PSMA-radioguided surgery on long-term oncological outcomes.

PSMA-theranostics
PSMA-targeted radioligand therapy (RLT) specifically targets PSMA expressed on the surface of malignant cells. PSMA-PET imaging plays a pivotal role in identifying suitable candidates for treatment by enabling the visualisation of PSMA expression in disease sites. The prognostic and predictive value of baseline PSMA-PET imaging with PSMA-targeted RLT is being established [66–68]. Patients with metastases demonstrating a higher degree of PSMA expression are considered to have a greater likelihood of benefiting from PSMA-targeted RLT [69]. The predictive ability of PSMA uptake was not shown for the combined use of ARPI and RLT [70]. The European Association of Nuclear Medicine (EANM) and Society of Nuclear Medicine and Molecular Imaging (SNMMI) have published procedure guidelines for the use of PSMA-targeted RLT, encompassing eligibility criteria, patient selection, the treatment process, and follow-up [71]. While PSMA-PET is valuable in its own right, evidence suggests that combining FDG-PET with PSMA-PET, although not mandatory, may enhance patient selection and better predict treatment outcomes compared to using PSMA-PET alone by excluding patients who have discordant PSMA-negative and FDG-positive disease [71, 72]. Readers should note that PSMA-targeted RLT is currently approved based on PSMA-PET positivity alone [73–76].

Disease monitoring and response assessments
Evaluating the treatment response of PCa patients with metastatic bone disease is challenging. BS-CT report on the interactions of marrow disease with mineralised bone, primarily through osteolytic and osteosclerotic mechanisms, which may not reliably indicate treatment efficacy [77]. Apparent worsening of BS-CT (“flare reactions”) is frequently observed when patients respond clinically [78]. On the contrary, clinical deterioration not observed with BS-CT is also common [79]. Furthermore, BS/CT progression without PSA progression has also been repeatedly demonstrated in mHSPC with ADT alone and combined with enzalutamide and apalutamide treatments [46, 47, 80]. These findings support the NCCN v1.2025 guideline for periodic imaging to monitor treatment of mHSPC, which currently does not endorse the use of PET imaging in this context. However, both PSMA-PET/CT and WB-MRI can overcome these limitations because the tumour response within the marrow space is directly depicted [23, 81].

PSMA-PET
PSMA-PET response assessment criteria, such as PSMA PET Progression (PPP) and Response Evaluation Criteria in PSMA PET/CT (RECIP), were developed to standardise interpretations of responses to RLT [82, 83]. Progression using the PSMA PET-specific criteria after PSMA-targeted RLT has been significantly associated with shorter overall survival (OS) [84, 85]. The RECIP framework employed an evidence-based approach that considers changes in total tumour volume (TTV). RECIP 1.0 categorises scans into response assessment categories based on changes in PSMA-positive TTV and the appearance of new lesions. RECIP 1.0 has prognostic value for OS and exhibits excellent interreader agreement for both visual and quantitative assessments [86]. It was initially developed for late-stage mCRPC but has been successfully used in mHSPC for monitoring ARPI use [87]. The Prostate Cancer molecular imaging standardised evaluation (PROMISE) v2 framework proposes standardised parameters for longitudinal reporting of PSMA-PET using PPP, RECIP, and tumour volume assessments [88].
Tumour response assessments after ADT or ARPI may pose a specific challenge for PSMA-PET. Suppression of androgen receptor signalling can increase PSMA expression (“flare phenomenon”) even as the tumour responds, especially in mCPRC [89]. On the other hand, this can also cause a decrease in PSMA expression [90]. Such modulations in PSMA expression can make it difficult to interpret changes in PSMA uptake reliably, and further studies are needed to clarify the role of PSMA-PET in this context. There is emerging data on PSMA-PET for monitoring chemotherapy response [91]. The optimal timing for end-of-treatment versus interim imaging remains under investigation, and clear guidelines for timing have yet to be established. Data on PSMA-PET as a response biomarker for other systemic therapies (e.g., PARP inhibitors, Radium-223) are limited, with early studies suggesting a potential role [92], but further investigations are required.

WB-MRI
Multiparametric tumour response assessments using WB-MRI potentially offer more precise differentiation of bone metastasis response with a lesser susceptibility to flare responses compared to those using PSMA-PET. Multiple WB-MRI studies have reported that changes in tumour volume, ADC values, and rFF% are associated with clinical response [44, 45]. These parameters are incorporated into the METastasis Reporting and Data System for Prostate Cancer (MET-RADS-P) [93]. Garcia-Ruiz et al [94] evaluated quantitative WB-MRI biomarkers for their ability to predict bone disease progression in patients with metastatic PCa treated mainly with ARPI. An increase in tumour fat was the most powerful prognostic factor for a more extended response. Interestingly, changes in ADC values were not predictors of survival benefits. In addition, response assessment categories (RACs) - which include multiparametric assessments of morphological findings, ADC, and rFF% were predictive of treatment benefits in a secondary analysis [22].
A prospective study of 109 patients with mHSPC receiving enzalutamide reported a high bone response rate, with 80% achieving complete/partial responses (RAC 1–2) at 6-12 months [95]. PSA responses were consistent with MRI in 78.5% of cases (Cohen’s k of 0.324). Critically, lower RAC scores correlated with a lower risk of death, with a hazard ratio of 0.15. The interrater agreement for RAC scoring for bone disease is substantial to excellent [96].
RECIP and MET-RADS-P represent distinct frameworks for assessing treatment response, distinguished by their underlying principles and practical requirements. RECIP quantitatively assesses changes in TTV against a baseline scan and is sensitive for tracking only PSMA-positive disease; however, as noted above, its reliance on measuring the therapeutic target itself makes it vulnerable to therapy-induced biomarker modulation and potentially blind to PSMA-negative resistance. In contrast, the therapy-agnostic MET-RADS-P framework assesses the downstream biological effects of treatments, such as changes in tumour cellularity (ADC) and marrow composition, comparing against either baseline or nadir scans to robustly detect actual cell death and repair mechanisms regardless of the treatment mechanism. This fundamental difference extends to their implementation: RECIP’s volumetric analysis may need specialist software for clinical trials (although visual assessments may be equally effective for clinical practice [86]. MET-RADS-P uses standard radiological tools, making it highly accessible and practical for routine clinical practice. While both assessment methods are superior to conventional imaging, there is a lack of comparative analysis between PSMA-PET and WB-MRI [84, 97].

Conclusion
The increasing availability of PSMA-PET/CT and WB-MRI has undeniably transformed the detection and characterisation of APC. While these tools offer detailed biomarker information, their potential to fundamentally alter the disease course and improve long-term patient outcomes via treatment adaptations remains unproven. While their strong rule-in ability for identifying new disease sites makes therapy escalations generally safer, radiologists and clinicians must exercise caution with therapy de-escalations, given their moderate rule-out capability [98]. The availability of treatments for micrometastatic disease detected by PSMA-PET/CT and WB-MRI may not inherently translate into an altered clinical risk-to-benefit ratio for all patients, underscoring the urgent need for robust studies that integrate imaging with therapeutic interventions. Furthermore, the common discordance between clinical assessments, PSA measurements, and BS-CT in accurately depicting bone disease progression highlights a critical need for vigilance and regular, protocol-based response assessments. Both PSMA-PET/CT and WB-MRI offer a more accurate reflection of treatment-induced changes, encompassing both response and progression, with their respective response criteria currently undergoing validation and demonstrating prognostic value (Table 3). Prospective clinical trials will be required to evaluate whether higher accuracy imaging can genuinely improve clinically relevant endpoints for patients through precise treatment adaptations.