These trials established [¹⁷⁷Lu]Lu-PSMA-617 as an effective therapy in the post-taxane setting, demonstrating improvements in overall survival, radiographic progression-free survival, and PSA50 response compared to standard care. Subsequent data has expanded its use into earlier lines of therapy. Despite regulatory approval, the clinical challenge of how we optimally select and monitor patients for ¹⁷⁷Lu-PSMA therapy remains. We reviewed the efficacy data of various imaging approaches and explored the role of imaging biomarkers, adjunct modalities, and emerging AI-driven approaches in refining patient selection and response assessment.
PSMA-PET/CT: The Foundation of Patient Selection
PSMA-PET/CT has become the mainstay for determining eligibility for ¹⁷⁷Lu-PSMA therapy. Compared to conventional CT and bone scintigraphy, PSMA-PET/CT imaging has higher sensitivity and specificity for bone, nodal, and soft-tissue metastases, often reclassifying disease burden. Beyond confirming the presence of disease, PSMA-PET can provide quantitative biomarkers such as SUVmean, SUVmax, PSMA-avid tumour volume, total lesion PSMA, and tumour heterogeneity metrics to predict treatment response. However, these metrics are not yet universally standardised. Whole-body segmentation is resource-intensive, and inter-centre variability in acquisition protocols limits reproducibility. This highlights a need for harmonisation frameworks such as PROMISE V2 and emerging consensus initiatives to standardise acquisition, reporting, and quantitative analysis. Despite PSMA-based screening, approximately one-third of patients in VISION did not achieve a PSA50 response, highlighting how not all prostate cancer lesions express PSMA uniformly - patients often have heterogeneous uptake, borderline SUV values, or mixed PSMA-avid and low-avid lesions. Translating standardised trial criteria into patient-centred decisions without compromising efficacy remains a challenge.
Role of ¹⁸F-FDG PET
Approximately one-third of patients failed to achieve a meaningful biochemical response, and understanding why is key to refining selection strategies. Primary resistance is likely due to low or heterogeneous PSMA-expression, neuroendocrine differentiation, or aggressive metabolic phenotypes (discordant FDG-positive/PSMA-negative disease). In the TheraP trial, one-third of screened patients were excluded pre-therapeutically due to discordant FDG-positive/PSMA-negative lesions. These patients had poorer outcomes, suggesting that dual-tracer imaging may better identify those unlikely to benefit from PSMA-RLT.
Resistance following therapy could be due to many factors, such as clonal selection or insufficient tumour absorbed dose failing to induce lethal DNA damage. Integration of FDG imaging, quantitative PSMA metrics, molecular profiling through ctDNA, and individualised dosimetry may be complementary.
Post-Therapy ¹⁷⁷Lu-SPECT/CT: Role in Personalisation
While PET/CT is the mainstay of patient selection, SPECT/CT can have a crucial role after therapy administration to confirm radiotracer delivery, enable tumour and organ dosimetry, detect early new lesions, and quantify tumour volume change. The landmark trials used fixed activity dosing; emerging dosimetry data reflect significant interpatient variability in tumour absorbed dose, leading to some patients being potentially underdosed relative to tumour burden and characteristics.
Early changes in SPECT-derived tumour volume after two cycles strongly correlate with PSA progression-free survival and overall survival, and tumour absorbed dose has been shown to correlate with biochemical response. As there is now increased evidence for interpatient variability in tumour absorbed dose, the current use of fixed-activity dosing is imprecise - adaptive dosimetry guided therapy may be required to preserve the survival benefits demonstrated in landmark trials. This approach aligns with current trends in oncology towards precision medicine. Importantly for patients, SPECT can guide early treatment modification, avoiding unnecessary toxicity in non-responders. However, response criteria remain non-standardised across centres.
Rethinking Response Assessment
Traditional response frameworks rely on PSA kinetics, RECIST 1.1, and PCWG3 criteria. However, these have well-recognised limitations such as PSA flare phenomena, bone scan flare, non-PSA-secreting progression, and limited sensitivity for small volume disease. There can also be discordance between biochemical and imaging responses. As PSMA-PET increasingly informs both eligibility and response evaluation, the role of PSA in supporting continuation or discontinuation decisions should be reconsidered.
Interim PSMA-PET/CT offers earlier and more sensitive detection of progression. Emerging PCWG4 criteria are now incorporating PSMA-PET into formal progression definitions, and could replace bone scintigraphy as the primary imaging tool in advanced disease.
The optimal timing of treatment cessation also remains undefined. If early post-therapy SPECT demonstrates subtherapeutic tumour absorbed dose, should therapy be intensified, modified, or stopped? If ctDNA burden declines despite equivocal imaging, which should be used to guide management? These scenarios highlight the need for dynamic and multimodal monitoring strategies.
Implementation Challenges
The expansion of ¹⁷⁷Lu-PSMA therapy introduces considerations such as the requirement of specialised nuclear medicine infrastructure, radiopharmaceutical production capacity, and multidisciplinary coordination. Isotope availability remains constrained in many regions, and dual-tracer imaging strategies, while biologically informative, increase cost and logistical complexity. Furthermore, quantitative imaging analysis requires expertise that may not be available across many institutions. Ensuring that expansion of treatment due to broadened eligibility does not compromise quality of care will require standardised imaging protocols, reproducible quantitative metrics, and clear selection frameworks.
Key Clinical Takeaways
PSMA-PET/CT remains the foundation of patient selection for ¹⁷⁷Lu-PSMA therapy, but can be complemented by other strategies in a multimodal approach, particularly as its use expands. Quantitative imaging biomarkers improve prognostication, FDG PET identifies aggressive PSMA-low phenotypes, and post-therapy SPECT/CT supports personalised dosimetry and early add-on of treatment. PSMA-PET-based response frameworks outperform PSA alone, and PCWG4 may formalise this shift. ctDNA and genomic profiling add imaging-independent predictive value. AI and radiomics are adjunctive tools with promise, but require validation.
The field is moving toward multimodal, individualised treatment strategies. The future of ¹⁷⁷Lu-PSMA-therapy selection likely lies in composite models integrating quantitative PSMA-PET parameters, FDG mismatch patterns, ctDNA burden and genomic alterations, clinical variables, and dosimetry metrics. Overall, ¹⁷⁷Lu-PSMA-RLT should be an adaptive monitored intervention rather than a fixed protocol.
Written by: Aditi Ranjan,1 Minal Padden-Modi,2 Hoda Abdel-Aty,2 Joao Galante,3 Simon Wan,4 Azzra Maricar,5 Adetokunbo Adesina,6 Brent Drake,7 Siraj Yusuf,7 Gary Cook,8 Nicholas James,2 Sola Adeleke,9
- Oxford University Hospitals NHS Foundation Trust, Oxford, UK.
- Institute of Cancer Research and Royal Marsden Hospital NHS Foundation Trust, London, UK.
- Oncology, Guy's and St Thomas' NHS Foundation Trust, London, UK.
- Institute of Nuclear Medicine, University College London Hospitals NHS Foundation Trust, London, UK.
- King's College London, London, UK.
- Buckinghamshire Hospitals NHS Trust, Amersham, UK.
- Radiology and Nuclear Medicine Department, Royal Marsden NHS Foundation Trust, London, UK.
- King's College London, Cancer Imaging PET Centre, St Thomas' Hospital, London, UK.
- KCL, School of Biomedical Engineering and Imaging Sciences, Becket House, London, UK.