Neuroendocrine differentiation can emerge during the course of advanced prostate cancer and is often associated with aggressive disease behavior, resistance to androgen receptor pathway inhibition, and changes in molecular imaging characteristics. Although reduced PSMA expression and increased somatostatin receptor expression are often assumed in treatment-emergent neuroendocrine prostate cancer, the actual imaging phenotype may be more heterogeneous.
The study, titled “Dual-Tracer 68Ga-PSMA and 68Ga-DOTATATE PET/CT in Advanced Prostate Cancer: Persistent PSMA Expression Despite Clinically Suspected Treatment-Emergent Neuroendocrine Differentiation,” was published in Clinical Genitourinary Cancer on September 5, 2026.
Authors: Hiroki Sai, Hidetoshi Kuruma, Atsuya Okada, Yuki Enei, Takafumi Yanagisawa, Chieko Okazaki, Shusuke Akamatsu, Nat Lenzo, Kenta Miki, and Takahiro Kimura.
Evaluating Dual-Tracer Imaging
The investigators retrospectively analyzed eight patients with advanced prostate cancer who underwent both 68Ga-PSMA PET/CT and 68Ga-DOTATATE PET/CT because of clinically suspected treatment-emergent neuroendocrine differentiation.
The two imaging studies were performed on consecutive days in all eight patients, minimizing the possibility that interval disease progression could explain differences in tracer uptake. The analysis was based primarily on original nuclear medicine reports rather than centralized blinded image review.
Clinical suspicion of neuroendocrine differentiation was based on features such as aggressive progression despite prior systemic therapy, visceral metastases, extensive disease with relatively modest PSA levels, low-PSA/high-burden discordance, previous pathological mention of neuroendocrine features when available, and referral for dual-tracer characterization.
Importantly, histopathological evidence of neuroendocrine differentiation was available in only one patient. The remaining seven patients were included based on clinical suspicion rather than pathological confirmation.
Patient Characteristics
All eight patients had metastatic castration-resistant prostate cancer at the time of dual-tracer imaging. The median age was 60.5 years, with a range of 51 to 72 years, while the median PSA level at imaging was 14.1 ng/mL, ranging from 1.2 to 460 ng/mL. The median interval from initial diagnosis to dual-tracer imaging was 33.5 months.
Most patients had received substantial prior therapy. Seven had previously received androgen receptor signaling inhibitors, five had received taxane chemotherapy, one had received radium-223, and one had received olaparib. On conventional imaging before PSMA-targeted radioligand therapy, seven patients had bone metastases, six had nodal metastases, and two had visceral disease. ECOG performance status was 0–1 in all patients.
PSMA Uptake Remained Predominant
The principal observation across the cohort was persistence of PSMA expression. Based on the original clinical reports, PSMA uptake was equal to or greater than DOTATATE uptake in all eight patients. No DOTATATE-positive/PSMA-negative disease sites were documented.
In contrast, PSMA-positive/DOTATATE-negative sites were identified in several patients. These discordant sites were particularly evident in visceral locations such as the lung and liver, although they were also reported in selected nodal and skeletal sites. For each patient, the investigators selected one anatomically matched index lesion with SUVmax values available for both tracers. The median SUVmax was 16.9 for 68Ga-PSMA, ranging from 5.8 to 170.4, compared with 5.75 for 68Ga-DOTATATE, ranging from 2.8 to 12.0.
PSMA SUVmax was higher than DOTATATE SUVmax in all eight matched index lesions. The exploratory paired analysis using a Wilcoxon signed-rank test resulted in a P value of 0.0078.
The investigators also observed substantial heterogeneity within individual patients. Some patients showed broad concordance between PSMA and DOTATATE uptake but with stronger PSMA expression, while others had DOTATATE-positive sites accompanied by additional PSMA-positive/DOTATATE-negative disease. Across the cohort, all disease sites showing DOTATATE uptake were also visualized on PSMA PET/CT.
Pathologically Confirmed Neuroendocrine Differentiation
One patient was analyzed in greater detail because complete imaging datasets were available for both tracers. The 52-year-old man had undergone radical prostatectomy and received multiple systemic therapies and chemotherapy. A biopsy of a pulmonary lesion demonstrated adenocarcinoma with neuroendocrine differentiation, making this the only patient in the cohort with histopathological evidence of neuroendocrine differentiation.
At the time of dual-tracer imaging, his PSA level was 1.2 ng/mL and neuron-specific enolase was 105.0 ng/mL. 68Ga-DOTATATE PET/CT showed uptake in lymph nodes, bone, and pulmonary lesions. These DOTATATE-positive sites were also detected on 68Ga-PSMA PET/CT and demonstrated higher PSMA uptake. Additional liver, spleen, and adrenal lesions were PSMA-positive but DOTATATE-negative.
Quantitative imaging further demonstrated a greater disease burden with PSMA imaging. 68Ga-PSMA PET/CT identified 128 lesions, compared with 67 lesions on 68Ga-DOTATATE PET/CT. SUVmax was 42.6 versus 11.0, respectively, while total tumor volume was 1,438.5 mL versus 346.9 mL.
Because of extensive PSMA-avid disease and limited therapeutic alternatives, the patient subsequently received PSMA-targeted radioligand therapy in routine clinical care. Following treatment, PSA decreased from 1.2 to 0.3 ng/mL and neuron-specific enolase decreased from 105.0 to 12.4 ng/mL, accompanied by reduced tracer uptake and resolution of pleural effusion. However, the study was not designed to evaluate the efficacy of PSMA-targeted radioligand therapy.
A Heterogeneous Imaging Phenotype
The findings suggest that clinically suspected treatment-emergent neuroendocrine differentiation does not necessarily correspond to uniform loss of PSMA expression or uniform acquisition of somatostatin receptor expression. Instead, the imaging patterns observed in this cohort were heterogeneous, both between patients and across different metastatic sites within the same patient.
The authors emphasized that neither PSMA nor DOTATATE uptake alone can establish a specific histological lineage. Lesion-level pathological assessment would be required to determine whether individual metastatic sites represent adenocarcinoma, neuroendocrine differentiation, or another treatment-induced lineage state.
The study also discussed double-negative prostate cancer as another potential lineage-plastic state. However, this phenotype could not be formally assessed because androgen receptor activity and canonical neuroendocrine markers were not systematically evaluated across metastatic lesions.
Study Limitations
Several limitations should be considered when interpreting the findings. The study was retrospective and included only eight patients from a clinically selected cohort. Only one patient had histopathologically confirmed neuroendocrine differentiation, while the remaining seven had clinically suspected disease. Patients were identified from a PSMA-targeted radioligand therapy program, introducing potential referral and selection bias.
In addition, imaging findings were derived from routine clinical reports rather than centralized blinded image review. Lesion matching and the paired SUVmax analysis were therefore pragmatic, and the quantitative comparison should be regarded as exploratory.
Androgen receptor activity and canonical neuroendocrine markers were not systematically assessed across metastatic sites, preventing formal classification of specific lineage-plastic phenotypes. 18F-FDG PET was also not routinely performed, meaning potentially aggressive FDG-positive/PSMA-negative disease could not be evaluated.
Conclusion
In this small cohort of patients with metastatic castration-resistant prostate cancer and clinically suspected treatment-emergent neuroendocrine differentiation, PSMA expression remained prominent. PSMA uptake exceeded DOTATATE uptake in all eight paired index lesions, while no DOTATATE-positive/PSMA-negative disease sites were documented in the available clinical reports. Additional PSMA-positive/DOTATATE-negative sites were observed in several patients.
These findings suggest that suspected neuroendocrine differentiation should not automatically be interpreted as loss of PSMA expression. Assessment of the actual PSMA imaging phenotype may therefore be important before excluding PSMA-targeted approaches, although larger prospective studies with centralized imaging review and pathological correlation are needed to determine the clinical significance of these findings.
The full article is available in the Clinical Genitourinary Cancer.
