Endocrine Abstracts

Peptide receptor radionuclide therapy (PRRT) with ¹⁷⁷Lu-DOTATATE has significantly advanced the management of patients with progressive neuroendocrine tumours (NETs). The pivotal NETTER-1 trial demonstrated improved progression-free survival in midgut NETs, establishing PRRT as a standard of care (Strosberg et al., 2017). Despite this success, marked heterogeneity in treatment response and toxicity underscores the need for refined stratification strategies. This review synthesises current evidence on prognostic and predictive markers relevant to PRRT and evaluates the potential of artificial intelligence (AI) to enhance candidate selection. A systematic literature search was conducted in Pubmed, Emase and Web of Science for studies published beytween January/2025 and June/2025. Search terms included PRRT, neuroendocrine tumours, biomarkers, somatostatin receptor imaging, Ki-67, chromogranin A and artificial intelligence. After duplicate removals, 124 records were screened, of which 62 full-text articles were assess for eligibility. Following inclusion and exclusion criteria, 41 studies were retained for qualitative synthesis. Findings demonstrate the treatment outcomes may be influenced by somatostatin receptor (SSTR) PET/CT characteristics, circulating biomarkers such as chromogranin A and neuron-specific enolase, tumour proliferation index (Ki-67), and clinical staging (Oberg et al., 2015). However, current stratification methods lack validated frameworks capable of integrating these multimodal datasets. Emerging literature highlights the promise of AI and machine learning to address this limitation. AI-driven approaches have shown potential in integrating imaging and biomarker data to predict treatment response, toxicity, and survival outcomes across oncology, with implications for PRRT candidate selection (Topol, 2019). This review concludes that AI-guided stratification represents a promising step towards precision medicine in neuroendocrine oncology. Validation in retrospective cohorts, followed by prospective trials, will be essential to confirm reproducibility and clinical utility. The integration of multimodal biomarkers with predictive modelling could ultimately refine patient selection, optimise therapeutic benefit, and support efficient resource allocation in Nuclear Medicine.