Predicting side-specific extraprostatic extension in prostate cancer using an 18F-DCFPyL PSMA-PET/CT-based nomogram.

BACKGROUND

Extraprostatic extension (EPE) in prostate cancer (PCa) has implications for nerve-sparing approaches. mpMRI-based nomograms show modest accuracy, highlighting the need for improved predictive models. This study evaluates 18F-DCFPyL prostate-specific membrane antigen (PSMA) positron emission tomography/ computed tomography (PET/CT) for predicting side-specific EPE using maximum standardized uptake value (SUVmax).

METHODS

This single-center cohort study included patients undergoing RALP by a single surgeon (AKT) from January 2022 to September 2024. Baseline variables included demographics, PSA, biopsy, MRI, and PSMA parameters (SUVmax, EPE, SVI). The primary endpoint was side-specific EPE on final pathology. Univariable and multivariable logistic regression identified significant predictors. A nomogram was built based on this. To evaluate model performance, a 1000-iteration bootstrap approach was used to compare (1) the institutional MRI-only 2018 model, (2) an MRI + PSMA Fixed Model, and (3) a retrained MRI + PSMA Model built on each bootstrap sample.

RESULTS

Three hundred fifty-five patients were analyzed. EPE was detected in 18.9% by MRI and 5.4% by PSMA PET. Median intraprostatic SUVmax was 11.30. EPE-positive sides were more likely to have MRI/PSMA-detected EPE (p < 0.001), PIRADS 5 lesions (p < 0.001), aggressive biopsy GGG (p < 0.001), higher positive cores (p < 0.001), and greater percent tumor involvement (p < 0.001). Median SUVmax was significantly higher in the EPE group (9.1 vs. 5.4; p < 0.001). Multivariable analysis identified PSA, MRI-detected EPE, GGG, tumor involvement percentage, and SUVmax ≥13 as significant predictors. The PSMA + MRI Fixed Model outperformed the MRI-only model (median AUC: 0.7542 vs. 0.7350) with p < 0.001. Calibration plots showed strong agreement between predicted and observed probabilities, and decision curve analysis demonstrated greater net clinical benefit across relevant thresholds.

CONCLUSION

We developed a nomogram integrating PSMA PET with MRI and clinicopathological variables, outperforming our institutional model. PSMA uptake strongly predicts side-specific EPE, which can enhance preoperative assessment and improve postoperative functional outcomes.