Classification Of Prostate Cancer Research Articles

Novel Treatment Strategies for Low-Risk Metastatic Castration-Sensitive Prostate Cancer

Background: The treatment strategy for metastatic castration-sensitive prostate cancer (mCSPC) has changed significantly in recent years. Based on various guidelines, an upfront androgen receptor signaling inhibitor (ARSI) is the first choice, but in patients of Asian descent, including Japanese patients, there are a certain number of cases in which androgen deprivation therapy (ADT) and CAB are more effective. If patients can be identified who show a marked response to ADT within 12 weeks after the initiation of ADT, which is the inclusion criterion for ARSI clinical trials targeting mCSPC, it would be valuable from an economic standpoint. Methods: A total of 218 patients with pure prostate adenocarcinoma and treated with ADT at the Kanazawa University Hospital between January 2000 and December 2020 were included in this study. As a risk classification for mCSPC, in addition to the LATITUDE and CHAARTED criteria, we used the castration-sensitive prostate cancer classification proposed by Kanazawa University (Canazawa), developed by the Department of Urology of Kanazawa University. The Canazawa classification was based on three factors: Gleason pattern 5, bone scan index (BSI) ≥ 1.5, and lactate dehydrogenase (LDH) ≥ 300 IU/L. It defined patients with one factor or less as low-risk and patients with two or three factors as high-risk. The overall survival (OS) and time to castration resistance (TTCR) were estimated retrospectively using the Kaplan–Meier method, and factors associated with TTCR were identified using univariate and multivariate analyses. Results: The median follow-up period was 40.4 months, the median OS period was 85.2 months, and the median TTCR period was 16.4 months. The Canazawa risk classification provided the clearest distinction between the OS and TTCR in mCSPC patients. Multivariate analysis revealed a decrease in PSA levels of <95% at 12 weeks after ADT initiation and was a predictor of short TTCR in low-risk, low-volume patients across all risk classifications. Conclusion: The Canazawa classification differentiated the prognosis of mCSPC patients more clearly. A PSA reduction rate of <95% at 12 w after starting ADT in low-risk, low-volume patients of all risk classifications was significantly shorter than the TTCR. We propose a new treatment strategy, in which patients with low-risk mCSPC are treated with ADT and switched to ARSIs based on the rate of PSA reduction at 12 w.

Texture graph transformer for prostate cancer classification

Prostate cancer classification plays a pivotal role in the diagnosis and treatment of this disease. In this paper, we present a novel approach called the texture graph transformer, which combines texture analysis techniques, graph-based representations, and multi-head attention to enhance the accuracy of prostate cancer classification. Our texture graph transformer adeptly captures intricate texture details and patterns within prostate cancer lesion regions, efficiently improving the model’s ability to discern between different cancer types. By leveraging graph-based representations, our model constructs texture graphs that effectively capture the spatial and textural relationships among various regions in MRI images, enabling comprehensive feature extraction. The incorporation of a texture graph multi-head attention further enhances the model’s capacity to capture long-range dependencies and global context, with a specific focus on informative regions. Through extensive experiments conducted on a dataset comprising 226 MRI images, we demonstrate the efficacy of our proposed approach in significantly improving the performance of prostate cancer classification.

Diffusion MRI in prostate cancer with ultra-strong whole-body gradients.

Diffusion-weighted MRI (dMRI) is universally recommended for the detection and classification of prostate cancer (PCa), with PI-RADS recommendations to acquire b-values of ≥1.4ms/μm2. However, clinical dMRI suffers from a low signal-to-noise ratio (SNR) as the consequence of prolonged echo times (TEs) attributable to the limited gradient power in the range of 40-80 mT/m. To overcome this, MRI systems with strong gradients have been designed but so far have mainly been applied in the brain. The aim of this work was to assess the feasibility, data quality, SNR and contrast-to-noise ratio (CNR) of measurements in PCa with a 300 mT/m whole-body system. A cohort of men without and with diagnosed PCa were imaged on a research-only 3T Connectom Siemens MRI system equipped with a gradient amplitude of 300 mT/m. dMRI at high b-values were acquired using high gradient amplitudes and compared with gradient capabilities mimicking clinical systems. Data artefacts typically amplified with stronger gradients were assessed and their correction evaluated. The SNR gains and lesion-to-healthy tissue CNR were statistically tested investigating the effect of protocol and b-value. The diagnostic quality of the images for different dMRI protocols was assessed by an experienced radiologist using a 5-point Likert scale and an adapted PI-QUAL scoring system. The strong gradients for prostate dMRI allowed a significant gain in SNR per unit time compared with clinical gradients. Furthermore, a 1.6-2.1-fold increase in CNR was observed. Despite the more pronounced artefacts typically associated with strong gradients, a satisfactory correction could be achieved. Smoother and less biased parameter maps were obtained with protocols at shorter TEs. The results of this study show that dMRI in PCa with a whole-body 300-mT/m scanner is feasible without a report of physiological effects, SNR and CNR can be improved compared with lower gradient strengths, and artefacts do not negate the benefits of strong gradients and can be ameliorated. This assessment provides the first essential step towards unveiling the full potential of cutting-edge scanners, now increasingly becoming available, to advance early detection and diagnostic precision.

Mixed Supervision of Histopathology Improves Prostate Cancer Classification From MRI.

Non-invasive prostate cancer classification from MRI has the potential to revolutionize patient care by providing early detection of clinically significant disease, but has thus far shown limited positive predictive value. To address this, we present a image-based deep learning method to predict clinically significant prostate cancer from screening MRI in patients that subsequently underwent biopsy with results ranging from benign pathology to the highest grade tumors. Specifically, we demonstrate that mixed supervision via diverse histopathological ground truth improves classification performance despite the cost of reduced concordance with image-based segmentation. Where prior approaches have utilized pathology results as ground truth derived from targeted biopsies and whole-mount prostatectomy to strongly supervise the localization of clinically significant cancer, our approach also utilizes weak supervision signals extracted from nontargeted systematic biopsies with regional localization to improve overall performance. Our key innovation is performing regression by distribution rather than simply by value, enabling use of additional pathology findings traditionally ignored by deep learning strategies. We evaluated our model on a dataset of 973 (testing n=198 ) multi-parametric prostate MRI exams collected at UCSF from 2016-2019 followed by MRI/ultrasound fusion (targeted) biopsy and systematic (nontargeted) biopsy of the prostate gland, demonstrating that deep networks trained with mixed supervision of histopathology can feasibly exceed the performance of the Prostate Imaging-Reporting and Data System (PI-RADS) clinical standard for prostate MRI interpretation (71.6% vs 66.7% balanced accuracy and 0.724 vs 0.716 AUC).

Protecting Prostate Cancer Classification From Rectal Artifacts via Targeted Adversarial Training.

Magnetic resonance imaging (MRI)-based deep neural networks (DNN) have been widely developed to perform prostate cancer (PCa) classification. However, in real-world clinical situations, prostate MRIs can be easily impacted by rectal artifacts, which have been found to lead to incorrect PCa classification. Existing DNN-based methods typically do not consider the interference of rectal artifacts on PCa classification, and do not design specific strategy to address this problem. In this study, we proposed a novel Targeted adversarial training with Proprietary Adversarial Samples (TPAS) strategy to defend the PCa classification model against the influence of rectal artifacts. Specifically, based on clinical prior knowledge, we generated proprietary adversarial samples with rectal artifact-pattern adversarial noise, which can severely mislead PCa classification models optimized by the ordinary training strategy. We then jointly exploited the generated proprietary adversarial samples and original samples to train the models. To demonstrate the effectiveness of our strategy, we conducted analytical experiments on multiple PCa classification models. Compared with ordinary training strategy, TPAS can effectively improve the single- and multi-parametric PCa classification at patient, slice and lesion level, and bring substantial gains to recent advanced models. In conclusion, TPAS strategy can be identified as a valuable way to mitigate the influence of rectal artifacts on deep learning models for PCa classification.

Dilated TransUNet + +-Based Segmentation with Multi-Scale Adaptive DenseNet with Bi-LSTM Layer-Aided Prostate Cancer Classification Model

Dilated TransUNet + +-Based Segmentation with Multi-Scale Adaptive DenseNet with Bi-LSTM Layer-Aided Prostate Cancer Classification Model

Prostate Cancer: MRI Image Detection Based on Deep Learning: A Review

This comprehensive study delves into the transformative role of artificial intelligence (AI) and deep learning (DL) in the realm of prostate cancer care, an issue of paramount importance in men’s health worldwide. Prostate cancer, marked by the unchecked growth of cells in the prostate gland, poses risks of tumor formation and eventual metastasis. The crux of combating this disease lies in its early detection and precise diagnosis, for which traditional screening methodologies like Prostate-Specific Antigen (PSA) tests and multiparametric Magnetic Resonance Imaging (mp-MRI) are fundamental. The introduction of AI and DL into these diagnostic avenues has been nothing short of revolutionary, enhancing the precision of medical imaging and significantly reducing the rates of unnecessary biopsies. The advancements in DL, particularly through the use of convolutional neural networks (CNNs) and the application of multiparametric MRI, have been instrumental in improving the accuracy of diagnoses, foreseeing the progression of the disease, and tailoring individualized treatment regimens. This paper meticulously examines various DL models and their successful application in the detection, classification, and segmentation of prostate cancer, establishing their superiority over conventional diagnostic techniques. Despite the promising horizon these technologies offer, their implementation is not without challenges. The requisite for specialized expertise to handle these advanced tools and the ethical dilemmas they present, such as data privacy and potential biases, are significant hurdles. Nevertheless, the potential of AI and DL to inaugurate a new chapter in prostate cancer management is undeniable. The emphasis on interdisciplinary collaboration among scientists, clinicians, and technologists is crucial for pushing the boundaries of current research and clinical practice, ensuring the ethical deployment of AI and DL technologies. This collaborative effort is vital for realizing the full potential of these innovations in providing more accurate, efficient, and patient-centric care in the fight against prostate cancer, heralding a future where the burden of this disease is significantly mitigated.

When surface-enhanced Raman spectroscopy meets complex biofluids: A new representation strategy for reliable and comprehensive characterization

BackgroundSurface-enhanced Raman spectroscopy (SERS) has gained increasing importance in molecular detection due to its high specificity and sensitivity. Complex biofluids (e.g., cell lysates and serums) typically contain large numbers of different bio-molecules with various concentrations, making it extremely challenging to be reliably and comprehensively characterized via conventional single SERS spectra due to uncontrollable electromagnetic hot spots and irregular molecular motions. The traditional approach of directly reading out the single SERS spectra or calculating the average of multiple spectra is less likely to take advantage of the full information of complex biofluid systems. ResultsHerein, we propose to construct a spectral set with unordered multiple SERS spectra as a novel representation strategy to characterize full molecular information of complex biofluids. This new SERS representation not only contains details from each single spectra but captures the temporal/spatial distribution characteristics. To address the ordering-independent property of traditional chemometric methods (e.g., the Euclidean distance and the Pearson correlation coefficient), we introduce Wasserstein distance (WD) to quantitatively and comprehensively assess the quality of spectral sets on biofluids. WD performs its superiority for the quantitative assessment of the spectral sets. Additionally, WD benefits from its independence of the ordering of spectra in a spectral set, which is undesirable for traditional chemometric methods. With experiments on cell lysates and human serums, we successfully achieve the verification for the reproducibility between parallel samples, the uniformity at different positions in the same sample, the repeatability from multiple tests at one location of the same sample, and the cardinality effect of the spectral set. SERS spectral sets also manage to distinguish different classes of human serums and achieve higher accuracy than the traditional prostate-specific antigen in prostate cancer classification. SignificanceThe proposed SERS spectral set is a robust representation approach in accessing full information of biological samples compared to relying on a single or averaged spectra in terms of reproducibility, uniformity, repeatability, and cardinality effect. The application of WD further demonstrates the effectiveness and robustness of spectral sets in characterizing complex biofluid samples, which extends and consolidates the role of SERS.

Prostate Cancer Detection from MRI Using Efficient Feature Extraction with Transfer Learning.

Prostate cancer is a common cancer with significant implications for global health. Prompt and precise identification is crucial for efficient treatment strategizing and enhanced patient results. This research study investigates the utilization of machine learning techniques to diagnose prostate cancer. It emphasizes utilizing deep learning models, namely VGG16, VGG19, ResNet50, and ResNet50V2, to extract relevant features. The random forest approach then uses these features for classification. The study begins by doing a thorough comparison examination of the deep learning architectures outlined above to evaluate their effectiveness in extracting significant characteristics from prostate cancer imaging data. Key metrics such as sensitivity, specificity, and accuracy are used to assess the models' efficacy. With an accuracy of 99.64%, ResNet50 outperformed other tested models when it came to identifying important features in images of prostate cancer. Furthermore, the analysis of understanding factors aims to offer valuable insights into the decision-making process, thereby addressing a critical problem for clinical practice acceptance. The random forest classifier, a powerful ensemble learning method renowned for its adaptability and ability to handle intricate datasets, then uses the collected characteristics as input. The random forest model seeks to identify patterns in the feature space and produce precise predictions on the presence or absence of prostate cancer. In addition, the study tackles the restricted availability of datasets by utilizing transfer learning methods to refine the deep learning models using a small amount of annotated prostate cancer data. The objective of this method is to improve the ability of the models to generalize across different patient populations and clinical situations. This study's results are useful because they show how well VGG16, VGG19, ResNet50, and ResNet50V2 work for extracting features in the field of diagnosing prostate cancer, when used with random forest's classification abilities. The results of this work provide a basis for creating reliable and easily understandable machine learning-based diagnostic tools for detecting prostate cancer. This will enhance the possibility of an early and precise diagnosis in clinical settings such as index terms deep learning, machine learning, prostate cancer, cancer identification, and cancer classification.

Prostate classification network (PC-Net) for automated classification of Prostate cancer in Magnetic resonance imaging

Prostate classification network (PC-Net) for automated classification of Prostate cancer in Magnetic resonance imaging

Focused active learning for histopathological image classification

Active Learning (AL) has the potential to solve a major problem of digital pathology: the efficient acquisition of labeled data for machine learning algorithms. However, existing AL methods often struggle in realistic settings with artifacts, ambiguities, and class imbalances, as commonly seen in the medical field. The lack of precise uncertainty estimations leads to the acquisition of images with a low informative value. To address these challenges, we propose Focused Active Learning (FocAL), which combines a Bayesian Neural Network with Out-of-Distribution detection to estimate different uncertainties for the acquisition function. Specifically, the weighted epistemic uncertainty accounts for the class imbalance, aleatoric uncertainty for ambiguous images, and an OoD score for artifacts. We perform extensive experiments to validate our method on MNIST and the real-world Panda dataset for the classification of prostate cancer. The results confirm that other AL methods are ‘distracted’ by ambiguities and artifacts which harm the performance. FocAL effectively focuses on the most informative images, avoiding ambiguities and artifacts during acquisition. For both experiments, FocAL outperforms existing AL approaches, reaching a Cohen’s kappa of 0.764 with only 0.69% of the labeled Panda data.

Efficient Classification of Prostate Cancer Using Artificial Intelligence Techniques

Efficient Classification of Prostate Cancer Using Artificial Intelligence Techniques

Long Short-Term Memory (LSTM) Network for Efficient Classification of Prostate Cancer and Prostate Enlargement

Long Short-Term Memory (LSTM) Network for Efficient Classification of Prostate Cancer and Prostate Enlargement

The Role of PSMA PET Imaging in the Classification of the Risk of Prostate Cancer Patients: A Systematic Review on the Insights to Guide an Active Surveillance Approach.

active surveillance (AS) is a suitable strategy for patients with prostate cancer (PCa). Prostate-specific membrane antigen (PSMA) positron emission tomography (PET) imaging is an established tool used to assess PCa. The aim of this review was to evaluate the role of PSMA imaging to guide correct risk-based classification and the AS approach in PCa patients. The Scopus, Embase, Web of Science, Cochrane Library, and PubMed/MEDLINE databases were screened to find relevant published articles. 1774 articles were revealed with the literature search. A total of 1764 articles were excluded after applying exclusion criteria (data not within the field of interest, preclinical papers, conference proceedings, reviews, or editorials). Ten studies were finally included in the review, revealing that PSMA PET could have the ability to guide risk-based classification of PCa and the choice of AS, and to guide the execution of biopsies for the research of high-grade PCa, therefore precluding AS. this systematic review underlined a possible role of PSMA PET imaging in patients with PCa by correctly re-classifying them on the basis of their risk and guiding AS.

Prostate Cancer Gleason Score Classification Using Transfer Learning Models

This research discusses the application of transfer learning models in the classification of prostate cancer based on Gleason scores. Gleason scoring is crucial in determining the aggressiveness of prostate cancer, guiding treatment decisions. Transfer learning, a technique where knowledge from one task is applied to another, has gained traction in medical image analysis. This study explores the efficacy of transfer learning models, such as convolutional neural networks (CNNs), in accurately classifying Gleason scores from histopathological images. Leveraging pre-trained CNNs like ResNet and VGG, the research demonstrates significant improvements in classification accuracy compared to traditional machine learning approaches. The methodology involves fine-tuning these pre-trained models on a dataset of prostate cancer histopathological images annotated with Gleason scores. Experimental results showcase promising performance metrics, including high accuracy, precision, recall, and F1-score, highlighting the potential of transfer learning in enhancing prostate cancer diagnosis and prognostication. This work contributes to the growing body of research utilizing deep learning techniques for improving cancer classification and personalized treatment strategies.

Imaging classification of prostate cancer with extracapsular extension and its impact on positive surgical margins after laparoscopic radical prostatectomy.

To explore the impact of different imaging classifications of prostate cancer (PCa) with extracapsular extension (EPE) on positive surgical margins (PSM) after laparoscopic radical prostatectomy. Clinical data were collected for 114 patients with stage PT3a PCa admitted to Ningbo Yinzhou No. 2 Hospital from September 2019 to August 2023. Radiologists classified the EPE imaging of PCa into Type I, Type II, and Type III. A chi-square test or t-test was employed to analyze the factors related to PSM. Multivariate regression analysis was conducted to determine the factors associated with PSM. Receiver operating characteristic curve analysis was used to calculate the area under the curve and evaluate the diagnostic performance of our model. Clinical decision curve analysis was performed to assess the clinical net benefit of EPE imaging classification, biopsy grade group (GG), and combined model. Among the 114 patients, 58 had PSM, and 56 had negative surgical margins. Multivariate analysis showed that EPE imaging classification and biopsy GG were risk factors for PSM after laparoscopic radical prostatectomy. The areas under the curve for EPE imaging classification and biopsy GG were 0.677 and 0.712, respectively. The difference in predicting PSM between EPE imaging classification and biopsy GG was not statistically significant (P>0.05). However, when used in combination, the diagnostic efficiency significantly improved, with an increase in the area under the curve to 0.795 (P<0.05). The clinical decision curve analysis revealed that the clinical net benefit of the combined model was significantly higher than that of EPE imaging classification and biopsy GG. EPE imaging classification and biopsy GG were associated with PSM after laparoscopic radical prostatectomy, and their combination can significantly improve the accuracy of predicting PSM.

A systematic review of deep learning methods for the classification and segmentation of prostate cancer on magnetic resonance images

AbstractProstate Cancer (PCa) is a prevalent global threat to male health, contributing significantly to male cancer‐related mortality. Timely detection and management are pivotal for improved outcomes, as successful cure rates are highest in the early stages. Deep learning (DL) methodologies offer a promising avenue to enhance the precision of detection, potentially reducing mortality rates. Magnetic resonance imaging (MRI) stands out as an effective tool for PCa diagnosis, providing comprehensive visuals of the prostate and adjacent tissues. This technology aids in identifying cancerous developments early on, crucial for treatment planning. Utilizing MRI for PCa detection has demonstrated increased accuracy, minimizing unnecessary biopsies, and facilitating personalized treatment decisions. Recent studies showcase the potential of DL methods in identifying and segmenting the prostate in MRI scans. These techniques not only improve diagnostic precision but also assist in treatment planning and monitoring. Incorporating DL in MRI‐based PCa diagnosis holds immense potential for enhancing efficiency and accuracy, promising better treatment outcomes. However, further research is imperative to explore these methods comprehensively, especially in larger and more diverse patient populations. This review evaluates the progress in employing DL techniques, rooted in artificial intelligence, for categorizing and outlining the prostate and lesions in MR images, underscoring the need for continued investigation and validation in varied clinical settings.

An Automated Prostate-cancer Prediction System (APPS) Based on Advanced DFO-ConGA2L Model using MRI Imaging Technique

The prostate cancer is a deadly form of cancer that assassinates a significant number of men due of its mediocre identification process. Images from people with cancer include important and intricate details that are difficult for conventional diagnostic methods to extract. This work establishes a novel Automated Prostate-cancer Prediction System (APPS) model for the goal of detecting and classifying prostate cancer utilizing MRI imaging sequences. The supplied medical image is normalized using a Coherence Diffusion Filtering (CDFilter) approach for improved quality and contrast. The appropriate properties are also extracted from the normalized image using the morphological and texture feature extraction approach, which helps to increase the classifier's accuracy. In order to train the classifier, the most important properties are also selected utilizing the cutting-edge Dragon Fly Optimized Feature Selection (DFO-FS) algorithm. Using this method greatly improves the classifier's overall disease diagnosis performance in less time and with faster processing. More specifically, the provided MRI input data are used to categorize the prostate cancer-affected and healthy tissues using the new Convoluted Gated Axial Attention Learning Model (ConGA2L) based on the selected features. This study compares and validates the performance of the APPS model by looking at several aspects using publicly available prostate cancer data.

Prostate cancer grading framework based on deep transfer learning and Aquila optimizer

Prostate cancer is the one of the most dominant cancer among males. It represents one of the leading cancer death causes worldwide. Due to the current evolution of artificial intelligence in medical imaging, deep learning has been successfully applied in diseases diagnosis. However, most of the recent studies in prostate cancer classification suffers from either low accuracy or lack of data. Therefore, the present work introduces a hybrid framework for early and accurate classification and segmentation of prostate cancer using deep learning. The proposed framework consists of two stages, namely classification stage and segmentation stage. In the classification stage, 8 pretrained convolutional neural networks were fine-tuned using Aquila optimizer and used to classify patients of prostate cancer from normal ones. If the patient is diagnosed with prostate cancer, segmenting the cancerous spot from the overall image using U-Net can help in accurate diagnosis, and here comes the importance of the segmentation stage. The proposed framework is trained on 3 different datasets in order to generalize the framework. The best reported classification accuracies of the proposed framework are 88.91% using MobileNet for the “ISUP Grade-wise Prostate Cancer” dataset and 100% using MobileNet and ResNet152 for the “Transverse Plane Prostate Dataset” dataset with precisions 89.22% and 100%, respectively. U-Net model gives an average segmentation accuracy and AUC of 98.46% and 0.9778, respectively, using the “PANDA: Resized Train Data (512 × 512)” dataset. The results give an indicator of the acceptable performance of the proposed framework.

MRI-based prostate cancer classification using 3D efficient capsule network.

Prostate cancer (PCa) is the most common cancer in men and the second leading cause of male cancer-related death. Gleason score (GS) is the primary driver of PCa risk-stratification and medical decision-making, but can only be assessed at present via biopsy under anesthesia. Magnetic resonance imaging (MRI) is a promising non-invasive method to further characterize PCa, providing additional anatomical and functional information. Meanwhile, the diagnostic power of MRI is limited by qualitative or, at best, semi-quantitative interpretation criteria, leading to inter-reader variability. Computer-aided diagnosis employing quantitative MRI analysis has yielded promising results in non-invasive prediction of GS. However, convolutional neural networks (CNNs) do not implicitly impose a frame of reference to the objects. Thus, CNNs do not encode the positional information properly, limiting method robustness against simple image variations such as flipping, scaling, or rotation. Capsule network (CapsNet) has been proposed to address this limitation and achieves promising results in this domain. In this study, we develop a 3D Efficient CapsNet to stratify GS-derived PCa risk using T2-weighted (T2W) MRI images. In our method, we used 3D CNN modules to extract spatial features and primary capsule layers to encode vector features. We then propose to integrate fully-connected capsule layers (FC Caps) to create a deeper hierarchy for PCa grading prediction. FC Caps comprises a secondary capsule layer which routes active primary capsules and a final capsule layer which outputs PCa risk. To account for data imbalance, we propose a novel dynamic weighted margin loss. We evaluate our method on a public PCa T2W MRI dataset from the Cancer Imaging Archive containing data from 976 patients. Two groups of experiments were performed: (1) we first identified high-risk disease by classifying low + medium risk versus high risk; (2) we then stratified disease in one-versus-one fashion: low versus high risk, medium versus high risk, and low versus medium risk. Five-fold cross validation was performed. Our model achieved an area under receiver operating characteristic curve (AUC) of 0.83 and 0.64 F1-score for low versus high grade, 0.79 AUC and 0.75 F1-score for low + medium versus high grade, 0.75 AUC and 0.69 F1-score for medium versus high grade and 0.59 AUC and 0.57 F1-score for low versus medium grade. Our method outperformed state-of-the-art radiomics-based classification and deep learning methods with the highest metrics for each experiment. Our divide-and-conquer strategy achieved weighted Cohen's Kappa score of 0.41, suggesting moderate agreement with ground truth PCa risks. In this study, we proposed a novel 3D Efficient CapsNet for PCa risk stratification and demonstrated its feasibility. This developed tool provided a non-invasive approach to assess PCa risk from T2W MR images, which might have potential to personalize the treatment of PCa and reduce the number of unnecessary biopsies.