68Ga-PSMA PET/CT SUVmax and its derivative in predicting clinical significance of prostate cancer or prostate cancer

Introduction: 68Gallium-labelled prostate-specific membrane antigen positron emission tomography/computed tomography (68Ga-PSMA PET/CT) is a mature staging tool, and shows tremendous potential for diagnosing primary prostate cancer. We aim to seek the threshold of SUVmax and its derivative to indicate the prostate cancer (PCa) or non-PCa, clinically significant PCa (csPCa) or non-csPCa in the patient-based analysis.

Methods: We retrospectively enrolled 254 men with benign or malignant pathology of prostate lesions and without treatment before 68Ga-PSMA PET/CT. 68Ga-PSMA PET/CT images were reviewed by two experienced nuclear medicine physicians who were blinded to pathological results. The focal lesions excluding physiological uptake was reported. The lesion with the highest SUVmax was defined as the index lesion. Additionally, the SUVmax was also measured in the background (SUVBGp), liver (SUVL), blood pool (SUVBP), gluteal muscle (SUVGm). Clinical-biological data such as age, TPSA, FPSA, Gleason score were collected from medical records. Mann-Whitney U test was used to compare SUVmax, ratio of SUV max to SUVBGp (SUVT/BGp), ratio of SUV max to SUVL (SUVT/L), ratio of SUV max to SUVBL (SUVT/BL), ratio of SUVmax to SUVGm (SUVT/Gm) between csPCa and no-csPCa, PCa and no-PCa. Receiver operating characteristic (ROC) curve analysis was used to evaluate the csPCa and PCa predictive value of each risk factor. Then we calculated sensitivity and specificity for PCa and csPCa detection via the optimal cutoff value calculated by Youden’s index. Multivariable logistic regression analyses were used to identify predictors of csPCa. In addition, association between SUVmax and pathologic upgrading (non-csPCa to csPCa from biopsy to surgery) was calculated.

Results: Totally, 172/254 (67.5%) men had PCa and 156/254 (61.4%) men had csPCa pathology on radical prostatectomy (RP) or targeted biopsy. Men with csPCa had higher SUVmax (13.50 vs. 5.0; P<0.001), higher SUVT/BGp (3.7 vs. 1.0; P<0.001), higher SUVT/L (2.6 vs. 0.9; P<0.001), higher SUVT/BP (3.4 vs. 1.3; P<0.001), higher SUVT/Gm (13.0 vs. 4.8; P<0.001) compared to those without. With the area under the curve (AUC) of 0.931, the ROC analysis of SUVT/BGp significantly improved csPCa diagnosis compared to SUVmax (ΔAUC, 0.016; p=0.030), SUVT/L (ΔAUC, 0.024; p=0.033), SUVT/BP (ΔAUC, 0.030; p=0.004), and showed insignificantly higher performance for csPCa diagnosis compared to SUVT/Gm (ΔAUC, 0.010; p=0.321). With the cutoff value of 1.8, SUVT/BGp had a sensitivity, specificity, PPV, NPV of 87.18%, 89.80%, 93.15%, 81.48% for csPCa. In addition, ROC analysis of SUVT/BGp (AUC=0.949) significantly improved PCa diagnosis compared to SUVmax (ΔAUC, 0.024; p=0.003), SUVT/L (ΔAUC, 0.040; p=0.001), SUVT/BP (ΔAUC, 0.028; p=0.006), and showed insignificantly higher performance for PCa diagnosis compared to SUVT/Gm (ΔAUC, 0.021; p=0.077). With cutoff value of SUVT/BGp 1.8, the sensitivity, specificity, PPV, NPV were 81.40%, 96.34%, 97.90%, 71.17% for PCa. SUVT/BGp was an independent predictor of csPCa (odds ratio [OR] 6.97 per unit increase, 95% confidence interval [CI] 3.73–13.00) and PCa (OR 42.19 per unit increase, 95%CI 12.70–140.13). 15 of 31 (48.39%) men with non-csPCa in biopsy underwent upgrading into csPCa after RP. There was significant difference of SUVmax (7.2 VS. 5.3; P=0.045) between men with pathological upgrading and men without. Threshold 5.5 predicted pathologic upgrading with 80% sensitivity and 56.2% specificity.

Conclusions: Greater SUVT/BGp on 68Ga-PSMA PET/CT is associated with detection of csPCa and PCa. SUVmax was potentially capable to detect csPCa with a negative biopsy.

• Download figure
• Open in new tab
• Download powerpoint