This paper presents the implementation of two automated text classification systems for prostate cancer findings based on the PI-RADS criteria. Specifically, a traditional machine learning model using XGBoost and a language model-based approach using RoBERTa were employed. The study focused on Spanish-language radiological MRI prostate reports, which has not been explored before. The results demonstrate that the RoBERTa model outperforms the XGBoost model, although both achieve promising results. Furthermore, the best-performing system was integrated into the radiological company\'s information systems as an API, operating in a real-world environment.

MRI is the primary imaging approach for diagnosing prostate cancer. Prostate Imaging Reporting and Data System (PI-RADS) on multiparametric MRI (mpMRI) provides fundamental MRI interpretation guidelines but suffers from inter-reader variability. Deep learning networks show great promise in automatic lesion segmentation and classification, which help to ease the burden on radiologists and reduce inter-reader variability. In this study, we proposed a novel multi-branch network, MiniSegCaps, for prostate cancer segmentation and PI-RADS classification on mpMRI. MiniSeg branch outputted the segmentation in conjunction with PI-RADS prediction, guided by the attention map from the CapsuleNet. CapsuleNet branch exploited the relative spatial information of prostate cancer to anatomical structures, such as the zonal location of the lesion, which also reduced the sample size requirement in training due to its equivariance properties. In addition, a gated recurrent unit (GRU) is adopted to exploit spatial knowledge across slices, improving through-plane consistency. Based on the clinical reports, we established a prostate mpMRI database from 462 patients paired with radiologically estimated annotations. MiniSegCaps was trained and evaluated with fivefold cross-validation. On 93 testing cases, our model achieved a 0.712 dice coefficient on lesion segmentation, 89.18% accuracy, and 92.52% sensitivity on PI-RADS classification (PI-RADS ≥ 4) in patient-level evaluation, significantly outperforming existing methods. In addition, a graphical user interface (GUI) integrated into the clinical workflow can automatically produce diagnosis reports based on the results from MiniSegCaps.

BACKGROUND: Prostate cancer (PCa) is the most common malignancy in men. The multiparametric MRI (mpMRI) significantly improved the diagnostic approach of PCa. Although PCa is highly likely to be present in prostate imaging-reporting and data system (PI-RADS) 5 lesions, there are up to 18% of PI-RADS 5 lesions with benign histopathology after targeted biopsy.
METHODS: We present the case of a 66-year-old man who was referred to our hospital for MRI/ultrasound fusion-based targeted biopsy due to an elevated PSA and a PI-RADS 5 lesion described in the mpMRI. After 2 consecutive biopsies, the mpMRI target showed no malignancy. The lesion was described as PI-RADS 2 two years later.
CONCLUSIONS: This case demonstrates the risk of false-positive classified PI-RADS 5 lesions in the mpMRI and the challenge in some cases to distinguish between BPH nodules and cancer. Until today, a limited amount of studies exists concerning this issue. However, further studies are required to evaluate further characteristics associated with a higher possibility of histopathologically benign findings in PI-RADS 5 lesions.

Multiparametric magnetic resonance imaging (mpMRI) of the prostate and mpMRI-guided biopsy have proved to be a valuable part of the diagnostic pathway for prostate cancer. This review reports on the current results in terms of clinical performance of these diagnostic tools and their role in clinical decision-making.
UNASSIGNED: Die multiparametrische Magnetresonanztomographie (mpMRT) der Prostata und die mpMRT-gesteuerte Biopsie sind ein wichtiger Bestandteil der Diagnostik des Prostatakarzinoms. In dieser Übersichtsarbeit berichten wir über die aktuelle Studienlage zur klinischen Anwendung dieser diagnostischen Mittel und bewerten deren Stellenwert in der klinischen Entscheidungsfindung.

BACKGROUND: Multiparametric-Magnetic Resonance Imaging (mpMRI)-Ultrasound fusion guided biopsy (Fbx) has emerged as the new standard of risk stratification for prostate cancer (PCa) with superior detection rates of clinically significant PCa than randomized biopsy. In the present study, we evaluated patients with suspicion of clinically significant PCa on mpMRI, but histopathologically proven Gleason 6 PCa in Fbx.
METHODS: Between 2015 and 2019, 849 patients underwent Fbx and concurrent systematic 12-core biopsy at our department. 234 patients were diagnosed with Gleason 6 PCa in either mpMRI-targeted and/or concurrent systematic biopsy. Patients were analyzed regarding PSA, mpMRI findings according to PI-RADS classification, histopathological results of Fbx and systematic 12-core biopsy. 99/234 patients were also analyzed in regards of histopathology of the whole-mount specimen of subsequent radical prostatectomy (RP).
RESULTS: In 131/234 patients (56%), Gleason 6 PCa was detected in the mpMRI target. In 103/234 patients (44%), Gleason 6 PCa was detected in the concurrent systematic 12-core biopsy with negative mpMRI-targeted biopsy. Men with evidence of Gleason 6 in the mpMRI target had significantly higher amounts of overall positive biopsies (median 4 vs. 2, p < 0.001) and higher maximum tumor infiltration per biopsy core (30% vs. 20%, p < 0.001) compared to men with negative mpMRI-targeted biopsy. Detection of Gleason 6 in mpMRI Target lesions correlated significantly with the PI-RADS score (p < 0.001). Patients with positive mpMRI-target had significantly higher tumor infiltration in whole-mount specimen after prostatectomy (20% vs. 15%, p = 0.0026) compared to men without detection of Gleason 6 in mpMRI-targeted biopsy but in additional systematic biopsy.
CONCLUSIONS: Detection of Gleason 6 PCa in mpMRI-targeted biopsy indicates higher tumor burden compared to detection of Gleason 6 PCa in concurrent systematic biopsy and negative mpMRI-targeted biopsy.