Objective  To explore the diagnostic value of the molecular imaging prostate specific membrane antigen (miPSMA) score based on ¹⁸F-PSMA-1007 PET/CT and improved miPSMA score for benign and malignant prostate lesions.

Methods  A cross-sectional study design was employed and a retrospective analysis was conducted on 125 patients with suspected prostate cancer (PCa) who underwent ¹⁸F-PSMA-1007 PET/CT imaging at Cancer Hospital Affiliated to Shanxi Medical University from July 2018 to April 2023. The patients were aged (67.1±9.4) years, with an age range of 39–87 years. An improved miPSMA score was established as follows: histopathological examination results were adopted as the "gold standard", a receiver operating characteristic (ROC) curve was drawn to diagnose PCa on the basis of the maximum standardized uptake value (SUV_max) of prostate lesions, the optimal diagnostic threshold was determined, and the ratio of the optimal diagnostic threshold to the mean standardized uptake value (SUV_mean) of the spleen (denoted as Δ%) was calculated. With Δ% of spleen SUV_mean as the boundary, point 1 (benign) in the miPSMA score was refined into 1a points (benign) and 1b points (suspected malignant) in the improved miPSMA score. Improved miPSMA scores≤1a, =1b, and ≥2 were diagnosed as benign prostate disease, suspected malignant prostate disease, and PCa, respectively. Two independent sample t-tests or the Mann-Whitney U test was used to compare the quantitative data, while the four-grid chi square test was used to analyze the counting data. The Kappa test was employed to analyze the consistency between the diagnostic results of the two scoring methods and the histopathological examination results. ROC curves were utilized to compare the diagnostic efficacy of the two scoring methods for benign and malignant prostate lesions.

Results  The prostate-specific antigen levels of the 125 patients were 24.50 (10.60, 75.50) ng/ml. Among the 125 patients, 81 were confirmed to have PCa via histopathological examination, and 44 were confirmed to have benign prostate lesions via puncture and follow-up for more than six months. The differences in age, prostate-specific antigen, and SUV_max between the patients with PCa and those with benign prostate lesions were statistically significant (t=1.20, Z=4.57, 8.08; all P<0.05). Compared with the false negative rate of the miPSMA score, that of the improved miPSMA score was largely reduced (24.69%(20/81) vs. 4.94%(4/81)), and the difference was statistically significant (χ²=26.17, P<0.001). Sensitivity (75.31%(61/81) vs. 95.06%(77/81)), accuracy (84.00%(105/125) vs. 94.40%(118/125)), and the negative predictive value (68.75%(44/64) vs. 91.11%(41/45)) substantially increased, and the differences were statistically significant (χ²=28.02, 18.25, 32.11; all P<0.05). Specificity (100%(44/44) vs. 93.18%(41/44)) and the positive predictive value (100%(61/61) vs. 96.25%(77/80)) slightly decreased, but the differences were not statistically significant (χ²=9.61, 8.05; both P>0.05). The false positive rate (0%(0/44) vs. 6.82%(3/44)) increased, but the difference was not statistically significant (χ²=13.40, P=0.053). The Kappa test showed that the improved miPSMA score had high consistency with the histopathological examination results (Kappa=0.878, P<0.001), and the consistency between the miPSMA score and histopathological examination results was moderate (Kappa=0.682, P<0.001). ROC curve analysis revealed that the area under the curve (AUC) of the miPSMA score and the improved miPSMA score for diagnosing benign and malignant prostate lesions was 0.877 (95%CI: 0.806–0.929) and 0.941 (95%CI: 0.884–0.975), respectively. The difference in AUC between the two scoring methods was statistically significant (Z=2.19, P=0.028).

Conclusions  The diagnostic efficacy of the improved miPSMA score for benign and malignant prostate lesions is better than that of the miPSMA score. The former substantially reduces the false negative rate.