BJU International最新文献

Objectives: To develop a novel transparent and lightweight machine learning model, the Green Learning (GL), for automated prostate segmentation (PS) and clinically significant prostate cancer (csPCa) detection on magnetic resonance imaging (MRI).

Patients and methods: Men who underwent 3-T MRI and prostate biopsy (PBx) were identified. MRI was acquired and interpreted according to the Prostate Imaging-Reporting and Data System (PI-RADS), version 2 or 2.1. The GL was created to automate PS and csPCa detection on biparametric MRI. The performance was compared to the standard-of-care radiologists using PI-RADS, and a conventional deep learning (DL) U-Net model as benchmarking. The PS performance was evaluated by the Dice similarity coefficient (DSC). The area under the curve (AUC) for patient-level csPCa detection was assessed. Model size and computational workload, measured by floating point operations (FLOPs), were reported.

Results: A total of 602 MRIs were randomly divided for training (N = 483) and testing (N = 119). Overall, 224 patients had csPCa on PBx. The median DSC for PS was higher for GL than U-Net (0.91 vs 0.88, P < 0.001). The AUC for csPCa detection of GL was similar to PI-RADS (0.75 vs 0.76, P = 0.8) and U-Net (vs 0.74, P = 0.3). A combination of GL and PI-RADS showed a higher AUC of 0.81 than PI-RADS alone (P = 0.02). Compared with U-Net, the GL had smaller magnitude parameters (1.21× 10⁶ vs 177× 10⁶) and less computational workload (9.8× 10⁹ vs 1027× 10⁹ FLOPs).

Conclusion: A novel GL model fully automatically detects csPCa on prostate biparametric MRI with comparable performance to PI-RADS and DL. Combined with PI-RADS, GL significantly improves csPCa detection.

Objective: To provide a histological catalogue and defined criteria for the diagnosis of prostatic adenocarcinoma and its variants using the fluorescent confocal microscope (FCM) platform.

Methods: A total of 22 radical prostatectomy specimens were serially sectioned and scanned with the FCM. The prostatic tissue then underwent routine histological preparation. The FCM images were assessed against traditional haematoxylin and eosin images to document FCM prostatic histology.

Results: We first comprehensively catalogued FCM images of a wide range of benign and malignant prostate tissue. We then analysed and identified FCM features characteristic of cancer glands. They include fused or connected solid glands, lack of lumens, heterogeneous nuclear signals and lack of three-dimensional depth on greyscale imaging. Gleason patterns 3 and 4 cancer glands were morphologically different on FCM. The smallest focus of cancer detected on FCM was 0.7 × 1.2 mm Gleason score 3 + 3 = 6 cancer.

Conclusion: The image atlas and the diagnostic criteria presented in this study will help train pathologists to proficiently use FCM for rapid intraoperative evaluation of surgical margins during radical prostatectomy.

Objective: To assess prostate volume (PV) changes with age in symptomatic and asymptomatic men with and without clinically significant prostate cancer (csPCa). In symptomatic patients, we additionally analysed the effect of age and csPCa on PSA and PSA-density (PSA-D) and compared these to current National Institute for Health and Care Excellence (NICE) recommended PSA age-range thresholds. Patients and Methods This single-centre retrospective cross-sectional study included 2512 men: 760 asymptomatic, disease-free men and 1752 patients referred on a PCa diagnostic pathway. Magnetic resonance imaging-derived whole-gland PV was recorded for all patients. A machine-learning pipeline with k-fold cross validation modelled relationships between PV and age.

Results: In asymptomatic men (median PV 25.4 mL), the mean PV per age-group increased non-linearly with age, from 18.7 mL at an increase of 0.10 mL/year aged 18 years, to 41.3 mL at 0.68 mL/year aged 89 years, with increased rate of change from the age of 48.9 years. Significant positive relationships were shown between PSA and age in patients with and without csPCa (r² = 0.09 vs 0.13, respectively), with PSA increasing by mean 0.17 ng/mL/year across groups. Patients with csPCa had consistently higher PSA levels. PSA-D showed significant age-related linear increases in patients with csPCa but remained consistently lower in those without csPCa at all ages (0.10-0.11 ng/mL²), allowing differentiation at a threshold of >0.15 ng/mL².

Conclusion: In asymptomatic men, PV changed non-linearly with age. Age-related PSA thresholds are supported; however, a static PSA-D threshold of 0.15 ng/mL² can be applied across all age ranges.