Search

The aim of this study was to investigate whether the favorable characteristics of novel digital PET/CT (dPET) scanners compared to analog systems (aPET) could translate into an improved disease localization in prostate cancer (PCa) patients with early biochemical recurrence/persistence (BCR/BCP). A retrospective analysis was conducted on 440 consecutive analog ( n = 311) or digital ( n = 129) 68 Ga-PSMA-11 PET/CT scans performed in hormone-sensitive ADT-free PCa patients with early-BCR/BCP (PSA at PET ≤ 2.0 ng/mL), previously treated with radical intent (radical-prostatectomy/radiotherapy). dPET showed a higher positivity rate compared to aPET (48.8% [63/129] vs. 37.3% [116/311], p = 0.03), despite the slightly lower median PSA value of the dPET cohort (0.33 [IQR: 0.26-0.61] vs. 0.55 [IQR: 0.40-0.85] ng/mL, p < 0.01). dPET detection rate was higher in both PSA ranges 0.2-0.5 ng/mL (39.0% [32/82] vs. 25.2% [34/135], p = 0.03) and 0.5-1.0 ng/mL (63.2% [24/38] vs. 40.8% [53/130], p = 0.02), but not for PSA ≥ 1.0 ng/mL. dPET detected a higher per patient median number of pathologic findings (PSMA-RADS ≥ 3) and multi-metastatic cases (>3 lesions) among N1/M1-positive scans (21.7% [10/46] vs. 8.6% [9/105], p = 0.03). Moreover, the proportion of uncertain findings among pathological lesions was significantly lower for dPET than aPET (24.4% [39/160] vs. 38.5% [60/156], p = 0.008). Overall, 68 Ga-PSMA-11 dPET showed a better performance compared to aPET, resulting in a higher scan-positivity rate, a higher number of detected pathological lesions, and a lower rate of uncertain findings.

High-resolution intraoperative PET/CT specimen imaging, coupled with prostate-specific membrane antigen (PSMA) molecular targeting, holds great potential for the rapid ex vivo identification of disease localizations in high-risk prostate cancer patients undergoing surgery. However, the accurate analysis of radiotracer uptake would require time-consuming manual volumetric segmentation of 3D images. The aim of this study was to test the feasibility of using machine learning to perform automatic nodal segmentation of intraoperative 68 Ga-PSMA-11 PET/CT specimen images. Six ( n = 6) lymph-nodal specimens were imaged in the operating room after an e.v. injection of 2.1 MBq/kg of 68 Ga-PSMA-11. A machine learning-based approach for automatic lymph-nodal segmentation was developed using only open-source Python libraries (Scikit-learn, SciPy, Scikit-image). The implementation of a k-means clustering algorithm ( n = 3 clusters) allowed to identify lymph-nodal structures by leveraging differences in tissue density. Refinement of the segmentation masks was performed using morphological operations and 2D/3D-features filtering. Compared to manual segmentation (ITK-SNAP v4.0.1), the automatic segmentation model showed promising results in terms of weighted average precision (97-99%), recall (68-81%), Dice coefficient (80-88%) and Jaccard index (67-79%). Finally, the ML-based segmentation masks allowed to automatically compute semi-quantitative PET metrics (i.e., SUVmax), thus holding promise for facilitating the semi-quantitative analysis of PET/CT images in the operating room.