Emerging advances in castration-resistant prostate cancer: from expression heterogeneity to the convergence of PSMA-PET and AI
Year: 2026
Type: article
Abstract: The management of castration-resistant prostate cancer (CRPC) remains a significant clinical challenge. Prostate-specific membrane antigen (PSMA)-PET/CT imaging has changed diagnostic assessment and treatment selection in CRPC because of its high lesion-detection capability in selected settings, and has spurred novel therapies such as PSMA-targeted radioligand therapy (RLT). However, the marked heterogeneity in PSMA expression poses a central obstacle to precision medicine. It can contribute to false-negative imaging findings, leading to underestimation of tumor burden, and a key mechanism driving resistance to PSMA-targeted treatments. To address heterogeneity at the diagnostic level, dual-tracer PET/CT can identify lesions with low or absent PSMA-ligand uptake, including discordant FDG-avid disease. Therapeutically, combining inhibitors targeting downstream signaling pathways, such as PI3K/AKT and MAPK, with existing treatments remains an investigational strategy for addressing resistance. Concurrently, the application of artificial intelligence (AI) in PSMA-PET/CT analysis is opening a promising new frontier. AI models enable automated lesion segmentation and detection. By mining deep radiomic features, these models have been evaluated for lesion detection, tumor-volume segmentation, biochemical recurrence, distant metastasis, molecular subtyping, and RLT response; however, most studies are retrospective and incompletely externally validated. Looking ahead, developing AI models that integrate multimodal data—including radiomics, genomics, and proteomics—may support personalized CRPC care, but prospective multicenter validation is required before they can guide routine therapeutic selection.
Source: Frontiers in Oncology
Authors: Surilige, Jiacheng Liang, Ruimin Ren
Institutions: Shanxi Medical University, Shanxi Academy of Medical Sciences
Language: English
Subfield: Cancer Research
Domain: Life Sciences
Open Access status: gold
Estimated APC paid: $3,801
OpenAlex ID: W7225789103