Integrating 68Ga-PSMA-11 PET/CT with Clinical Risk Factors for Enhanced Prostate Cancer Progression Prediction

Abstract

Background/objectives: This study evaluates whether combining ⁶⁸Ga-PSMA-11-PET/CT derived imaging biomarkers with clinical risk factors improves the prediction of early biochemical recurrence (eBCR) or clinical progress in patients with high-risk prostate cancer (PCa) after primary treatment, using machine learning (ML) models.

Methods: We analyzed data from 93 high-risk PCa patients who underwent ⁶⁸Ga-PSMA-11 PET/CT and received primary treatment at a single center. Two predictive models were developed: a logistic regression (LR) model and an ML derived probabilistic graphical model (PGM) based on a naïve Bayes framework. Both models were compared against each other and against the CAPRA risk score. The models' input variables were selected based on statistical analysis and domain expertise including a literature review and expert input. A decision tree was derived from the PGM to translate its probabilistic reasoning into a transparent classifier.

Results: The five key input variables were as follows: binarized CAPRA score, maximal intraprostatic PSMA uptake intensity (SUVmax), presence of bone metastases, nodal involvement at common iliac bifurcation, and seminal vesicle infiltration. The PGM achieved superior predictive performance with a balanced accuracy of 0.73, sensitivity of 0.60, and specificity of 0.86, substantially outperforming both the LR (balanced accuracy: 0.50, sensitivity: 0.00, specificity: 1.00) and CAPRA (balanced accuracy: 0.59, sensitivity: 0.20, specificity: 0.99). The decision tree provided an explainable classifier with CAPRA as a primary branch node, followed by SUVmax and specific PET-detected tumor sites.

Conclusions: Integrating ⁶⁸Ga-PSMA-11 imaging biomarkers with clinical parameters, such as CAPRA, significantly improves models to predict progression in patients with high-risk PCa undergoing primary treatment. The PGM offers superior balanced accuracy and enables risk stratification that may guide personalized treatment decisions.

Keywords: 68Ga-PSMA-11 PET/CT; CAPRA score; SUVmax; early biochemical recurrence; outcome prediction; prostate cancer.

Figure 1

Study flow chart illustrating the development of two predictive models (LR and PGM) using harmonized clinical and ⁶⁸Ga-PSMA-11 PET/CT data (same input variables), alongside a CAPRA-based comparator. Input variable selection integrated data-driven methods and domain knowledge. All models were evaluated using five-fold cross-validation. Finally, a decision tree was derived from the PGM to improve interpretability.

Figure 2

⁶⁸Ga-PSMA-11 PET/CT staging and multiparametric MRI of the prostate in a 68-year-old patient with high-risk prostate cancer (Gleason score 8, initial PSA 20.3 ng/mL) prior to therapy planning. (A,B) Maximum intensity projection (MIP) PET images (frontal and lateral) demonstrating multiple bone metastases and a clearly visualized locally advanced primary prostate tumor. (C) Axial PET/CT demonstrating PSMA ligand uptake in the intraprostatic tumor. (D) Axial T2-weighted MRI showing a hypointense tumor in the left prostate lobe with evidence of extraprostatic extension (arrows). (E) Sagittal PET/CT reconstruction showing PSMA ligand uptake in the intraprostatic tumor and left seminal vesicle. (F) Sagittal T2-weighted MRI illustrating infiltration of the left seminal vesicle (arrows).

Figure 3

Logistic regression coefficients are presented in absolute values. The SVI, CAPRA, PSA, Gleason score, and SUVmax showed the strongest relationship with PCa progression.

Figure 4

Correlation matrix of input variables. The analysis confirmed low inter-feature correlation overall, supporting their joint inclusion in the model, with only SVI and extracapsular infiltration showing moderate overlap (r = 0.64).

Figure 5

Comparative performance of the LR model and PGM with the CAPRA score as standalone comparator with the respective key validation metrics. The PGM shows the best balance between sensitivity and specificity, while the LR model and the CAPRA score demonstrate perfect specificity at the cost of sensitivity.

Figure 6

The decision tree uses binary input variables, where a value of 0 indicates a negative or low-risk finding. Specifically, SUVmax_binned = 0 denotes a primary tumor SUVmax below 12; capra_pred = 0 indicates a CAPRA-based progression probability below 0.15. For the remaining variables, a value of 0 corresponds to the absence of the respective pathological feature: no bone metastases (OS_Mx), no seminal vesicle infiltration (SVI), and no lymph node involvement at the common iliac bifurcation (CIB).