Introduction:

Accurate detection of local recurrence in men with biochemical recurrence (BCR) post radical prostatectomy (RP) is crucial for guiding salvage therapy. PSMA PET/CT is widely adopted for restaging, though its utility in detecting local recurrence is limited by urinary tracer excretion. Multiparametric MRI (mpMRI) remains the most sensitive modality for local recurrence and is likely underutilized. Though less frequently used nowadays, 11C-Choline PET/CT offers lower urinary excretion and may offer advantages in detecting local recurrence. Our institution’s unique access to high volume 11C-Choline PET/CT and PSMA with routine mpMRI provides a rare opportunity to conduct a large-scale comparison of their diagnostic performance for local recurrence.  

Methods:

We retrospectively identified 1525 men with BCR (PSA ≥ 0.2 ng/mL post-RP) who underwent mpMRI with either PSMA PET/CT or 11C-Choline PET/CT within 3 months between July 2009 to May 2025. We excluded patients who received salvage or systemic therapy prior to imaging. Imaging reports were parsed with a natural language processing pipeline using foundation model prompting and Gemini Flash 2.0 to extract standardized categorical variables. PSMA PET/CT (n=464) and 11C-Choline PET/CT (n=655) were assessed for detection of local recurrence, defined as any suspicious finding in the prostate bed. mpMRI served as the reference standard. Sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), were calculated overall and stratified by PSA level (<1.0 vs >1.0ng/mL).  

Results:

Among patients with mpMRI-confirmed local recurrence, the median age at RP was 63 years. Most had Gleason 7 disease (68%), and pT2 stage (61%). 33% had positive surgical margins, and 8% had nodal involvement. At time of imaging, mean PSA was 0.7 (SD 1.5) ng/mL, with a median of 0.3 (IQR 0.2, 0.6) ng/mL. 

Overall: PSMA PET/CT detected local recurrence in 53% (95% CI: 45–62%) of patients with a reference-positive mpMRI. It correctly identified the absence of recurrence in 88% (95% CI: 83–93%) of reference-negative cases. 11C-Choline PET/CT had a slightly better detection rate of 56% (95% CI: 50–63%) for local recurrence with a reference-positive mpMRI and showed the same specificity of 88% (95% CI: 84–93%).  

PSA <1.0ng/mL: PSMA PET/CT detected local recurrence in 55% (95% CI: 47–64%). While 11C-Choline PET/CT detected recurrence in 53% (95% CI: 45–65%). 

PSA >1.0ng/mL: PSMA PET/CT detected local recurrence in 53% (95% CI: 29–77%) of cases, while 11C-Choline PET/CT demonstrated a higher detection rate of 71% (95% CI, 58-84%) in this subgroup.  

Conclusion:

To our knowledge, this study represents the largest direct comparison of PSMA and 11C-Choline PET/CT for local recurrence detection after RP. Both PET modalities show limited detection rates for true positives, particularly at low PSA levels. 11C-Choline PET/CT outperformed PSMA PET/CT at PSA >1.0ng/mL, likely due to its reduced urinary tracer interference. These findings challenge assumptions about PSMA PET/CT superiority in the post-RP setting. mpMRI remains an essential imaging modality in the post-RP setting, and its routine use should be preserved when evaluating BCR. Combined mpMRI and PET imaging may offer the most accurate localization to guide salvage therapy and optimize patient outcomes and is the target of future work.

Funding: N/A