Background Micro-ultrasound (micro-US) is a clinically available novel high-resolution imaging technology for guiding prostate biopsies. However, clinical image interpretation during live biopsies remains a challenge. Purpose To develop a convolutional neural network (CNN) to classify prostate tissues as benign versus clinically significant prostate cancer (csPCa) from the power spectrums (PS) derived from raw micro-US, with the eventual goal of developing a tool for automating interpretation during image-guided biopsies. Methods Retrospective Micro-US data were obtained from 491 men (mean age 62 y, SD 8) undergoing prostate biopsy across 5 sites between 2013 and 2016; associated raw data and prostate-specific antigen (PSA) were obtained for each targeted biopsy location and used to obtain spatially mapped PSs. The dataset was split at a patient-level into a train/validation (80%) and a set-aside test set (20%). This includes up to 12 image-frames at distinct prostate locations (total of 6530 single image-frames), each with a corresponding biopsy. No specific prostate tissue segmentation was carried out. A custom CNN named PSNet was developed to classify benign from csPCa in non-segmented regions of micro-US data, and its performance was compared to traditional CNNs trained on associated conventional B-Mode images. Biopsy histopathology served as the clinical standard labels. The area under the receiver operator curve (ROC-AUC) was used to evaluate all models; sensitivity, specificity, precision and the F1 score were also computed; 95% confidence interval is shown in parenthesis. Results For frame-level performance, PSNet without PSA achieved an ROC-AUC of 82% (0.77, 0.85), a sensitivity of 0.73 (0.66, 0.80) and a specificity of 0.74 (0.71, 0.77) for classifying benign versus csPCa. After inclusion of PSA, the ROC-AUC increased to 85% (0.83, 0.88), with a sensitivity of 0.72 (0.65, 0.79) and a specificity of 0.82 (0.80, 0.84). For patient-level performance, which was obtained by aggregating image-level predictions, the models without and with PSA achieved patient-level ROC-AUCs of 85% (0.77, 0.92) and 91% (0.85, 0.97), sensitivities of 0.74 (0.70, 0.79) and 0.70 (0.65, 0.75) and specificities of 0.88 (0.76, 0.84) and 0.99 (0.98, 1.00), respectively. Conclusion In this pilot development study, we suggest that deep learning can capture unique tissue acoustic properties in raw micro-US data to help identify prostate cancer, without the need for segmentation of the prostate gland, and that the diagnostic value of these tissue properties can be augmented by PSA measurements to increase specificity. Our approach may be further leveraged to guide targeted prostate biopsy.

Reinforcement learning (RL) has emerged as a promising approach to improve large language model (LLM) reasoning, yet most open efforts focus narrowly on math and code, limiting our understanding of its broader applicability to general reasoning. A key challenge lies in the lack of reliable, scalable RL reward signals across diverse reasoning domains. We introduce Guru, a curated RL reasoning corpus of 92K verifiable examples spanning six reasoning domains–Math, Code, Science, Logic, Simulation, and Tabular–each built through domain-specific reward design, deduplication, and filtering to ensure reliability and effectiveness for RL training. Based on Guru, we systematically revisit established findings in RL for LLM reasoning and observe significant variation across domains. For example, while prior work suggests that RL primarily elicits existing knowledge from pretrained models, our results reveal a more nuanced pattern: domains frequently seen during pretraining (Math, Code, Science) easily benefit from cross-domain RL training, while domains with limited pretraining exposure (Logic, Simulation, and Tabular) require in-domain training to achieve meaningful performance gains, suggesting that RL is likely to facilitate genuine skill acquisition. Finally, we present Guru-7B and Guru-32B, two models that achieve state-of-the-art performance among open models RL-trained with publicly available data, outperforming best baselines by 7.9

The positive predictive value of prostate multiparametric MRI for detecting clinically significant prostate cancer improves significantly when lower apparent diffusion coefficient values and higher prostate-specific antigen density are integrated.

BACKGROUND AND OBJECTIVE:To assess whether conventional brightness-mode (B-mode) transrectal ultrasound images of the prostate reveal clinically significant cancers with the help of artificial intelligence methods. METHODS:This study included 2986 men who underwent biopsies at two institutions. We trained the PROstate Cancer detection on B-mode transrectal UltraSound images NETwork (ProCUSNet) to determine whether ultrasound can reliably detect cancer. Specifically, ProCUSNet is based on the well-established nnUNet frameworks, and seeks to detect and outline clinically significant cancer on three-dimensional (3D) examinations reconstructed from 2D screen captures. We compared ProCUSNet against (1) reference labels (n = 515 patients), (2) eight readers that interpreted B-mode ultrasound (n = 20-80 patients), and (3) radiologists interpreting magnetic resonance imaging (MRI) for clinical care (n = 110 radical prostatectomy patients). KEY FINDINGS AND LIMITATIONS:ProCUSNet found 82% clinically significant cancer cases with a lesion boundary error of up to 2.67 mm and detected 42% more lesions than ultrasound readers (sensitivity: 0.86 vs 0.44, p < 0.05, Wilcoxon test, Bonferroni correction). Furthermore, ProCUSNet has similar performance to radiologists interpreting MRI when accounting for registration errors (sensitivity: 0.79 vs 0.78, p > 0.05, Wilcoxon test, Bonferroni correction), while having the same targeting utility as a supplement to systematic biopsies. CONCLUSIONS AND CLINICAL IMPLICATIONS:ProCUSNet can localize clinically significant cancer on screen capture B-mode ultrasound, a task that is particularly challenging for clinicians reading these examinations. As a supplement to systematic biopsies, ProCUSNet appears comparable with MRI, suggesting its utility for targeting suspicious lesions during the biopsy and possibly for screening using ultrasound alone, in the absence of MRI.

Magnetic Resonance Imaging (MRI) is increasingly being used to detect prostate cancer, yet its interpretation can be challenging due to subtle differences between benign and cancerous tissue. Recently, Denoising Diffusion Probabilistic Models (DDPMs) have shown great utility for medical image segmentation, modeling the process as noise removal in standard Gaussian distributions. In this study, we further enhance DDPMs by introducing the knowledge that the occurrence of cancer varies across the prostate (e.g., ∼70% of prostate cancers occur in the peripheral zone). We quantify such heterogeneity with a registration pipeline to calculate voxel-level cancer distribution mean and variances. Our proposed approach, ProstAtlasDiff, relies on DDPMs that use the cancer atlas to model noise removal and segment cancer on MRI. We trained and evaluated the performance of ProstAtlasDiff in detecting clinically significant cancer in a multi-institution multi-scanner dataset, and compared it with alternative models. In a lesion-level evaluation, ProstAtlasDiff achieved statistically significantly higher accuracy (0.91 vs. 0.85, p<0.001), specificity (0.91 vs. 0.84, p<0.001), positive predictive value (PPV, 0.50 vs. 0.35, p<0.001), compared to alternative models. ProstAtlasDiff also offers more accurate cancer outlines, achieving a higher Dice Coefficient (0.33 vs. 0.31, p<0.01). Furthermore, we evaluated ProstAtlasDiff in an independent cohort of 91 patients who underwent radical prostatectomy to compare its performance to that of radiologists, relative to whole-mount histopathology ground truth. ProstAtlasDiff detected 16% (15 lesions out of 93) more clinically significant cancers compared to radiologists (sensitivity: 0.90 vs. 0.75, p<0.01), and was comparable in terms of ROC-AUC, PR-AUC, PPV, accuracy, and Dice coefficient (p≥0.05). Furthermore, we evaluated ProstAtlasDiff in a second independent cohort of 537 subjects and observed that ProsAtlasDiff outperformed alternative approaches. These results suggest that ProstAltasDiff has the potential to assist in localizing cancer for biopsy guidance and treatment planning.

Pre-biopsy magnetic resonance imaging (MRI) is increasingly used to target suspicious prostate lesions. This has led to artificial intelligence (AI) applications improving MRI-based detection of clinically significant prostate cancer (CsPCa). However, MRI-detected lesions must still be mapped to transrectal ultrasound (TRUS) images during biopsy, which results in missing CsPCa. This study systematically evaluates a multimodal AI framework integrating MRI and TRUS image sequences to enhance CsPCa identification. The study included 3110 patients from three cohorts across two institutions who underwent prostate biopsy. The proposed framework, based on the 3D UNet architecture, was evaluated on 1700 test cases, comparing performance to unimodal AI models that use either MRI or TRUS alone. Additionally, the proposed model was compared to radiologists in a cohort of 110 patients. The multimodal AI approach achieved superior sensitivity (80 (42 to radiologists, the multimodal model showed higher specificity (88 and Lesion Dice (38 demonstrate the potential of multimodal AI to improve CsPCa lesion targeting during biopsy and treatment planning, surpassing current unimodal models and radiologists; ultimately improving outcomes for prostate cancer patients.

PURPOSE:Evaluation of artificial intelligence (AI) algorithms for prostate segmentation is challenging because ground truth is lacking. We aimed to: (1) create a reference standard data set with precise prostate contours by expert consensus, and (2) evaluate various AI tools against this standard. METHODS AND MATERIALS:We obtained prostate magnetic resonance imaging cases from six institutions from the Qualitative Prostate Imaging Consortium. A panel of 4 experts (2 genitourinary radiologists and 2 prostate radiation oncologists) meticulously developed consensus prostate segmentations on axial T2-weighted series. We evaluated the performance of 6 AI tools (3 commercially available and 3 academic) using Dice scores, distance from reference contour, and volume error. RESULTS:The panel achieved consensus prostate segmentation on each slice of all 68 patient cases included in the reference data set. We present 2 patient examples to serve as contouring guides. Depending on the AI tool, median Dice scores (across patients) ranged from 0.80 to 0.94 for whole prostate segmentation. For a typical (median) patient, AI tools had a mean error over the prostate surface ranging from 1.3 to 2.4 mm. They maximally deviated 3.0 to 9.4 mm outside the prostate and 3.0 to 8.5 mm inside the prostate for a typical patient. Error in prostate volume measurement for a typical patient ranged from 4.3% to 31.4%. CONCLUSIONS:We established an expert consensus benchmark for prostate segmentation. The best-performing AI tools have typical accuracy greater than that reported for radiation oncologists using computed tomography scans (the most common clinical approach for radiation therapy planning). Physician review remains essential to detect occasional major errors.

OBJECTIVES:To improve sensitivity and inter-reader consistency of prostate cancer localisation on micro-ultrasonography (MUS) by developing a deep learning model for automatic cancer segmentation, and to compare model performance with that of expert urologists. PATIENTS AND METHODS:We performed an institutional review board-approved prospective collection of MUS images from patients undergoing magnetic resonance imaging (MRI)-ultrasonography fusion guided biopsy at a single institution. Patients underwent 14-core systematic biopsy and additional targeted sampling of suspicious MRI lesions. Biopsy pathology and MRI information were cross-referenced to annotate the locations of International Society of Urological Pathology Grade Group (GG) ≥2 clinically significant cancer on MUS images. We trained a no-new U-Net model - the Prostate Micro-Ultrasound Network (ProMUS-NET) - to localise GG ≥2 cancer on these image stacks in a fivefold cross-validation. Performance was compared vs that of six expert urologists in a matched sub-cohort. RESULTS:The artificial intelligence (AI) model achieved an area under the receiver-operating characteristic curve of 0.92 and detected more cancers than urologists (lesion-level sensitivity 73% vs 58%; patient-level sensitivity 77% vs 66%). AI lesion-level sensitivity for peripheral zone lesions was 86.2%. CONCLUSIONS:Our AI model identified prostate cancer lesions on MUS with high sensitivity and specificity. Further work is ongoing to improve margin overlap, to reduce false positives, and to perform external validation. AI-assisted prostate cancer detection on MUS has great potential to improve biopsy diagnosis by urologists.

Vision-based tactile sensors have recently become popular due to their combination of low cost, very high spatial resolution, and ease of integration using widely available miniature cameras. The associated field of view and focal length, however, are difficult to package in a human-sized finger. In this paper we employ optical fiber bundles to achieve a form factor that, at 15 mm diameter, is smaller than an average human fingertip. The electronics and camera are also located remotely, further reducing package size. The sensor achieves a spatial resolution of 0.22 mm and a minimum force resolution 5 mN for normal and shear contact forces. With these attributes, the DIGIT Pinki sensor is suitable for applications such as robotic and teleoperated digital palpation. We demonstrate its utility for palpation of the prostate gland and show that it can achieve clinically relevant discrimination of prostate stiffness for phantom and ex vivo tissue.

We detail the training of the LLM360 K2-65B model, scaling up our 360-degree OPEN SOURCE approach to the largest and most powerful models under project LLM360. While open-source LLMs continue to advance, the answer to "How are the largest LLMs trained?" remains unclear within the community. The implementation details for such high-capacity models are often protected due to business considerations associated with their high cost. This lack of transparency prevents LLM researchers from leveraging valuable insights from prior experience, e.g., "What are the best practices for addressing loss spikes?" The LLM360 K2 project addresses this gap by providing full transparency and access to resources accumulated during the training of LLMs at the largest scale. This report highlights key elements of the K2 project, including our first model, K2 DIAMOND, a 65 billion-parameter LLM that surpasses LLaMA-65B and rivals LLaMA2-70B, while requiring fewer FLOPs and tokens. We detail the implementation steps and present a longitudinal analysis of K2 DIAMOND's capabilities throughout its training process. We also outline ongoing projects such as TXT360, setting the stage for future models in the series. By offering previously unavailable resources, the K2 project also resonates with the 360-degree OPEN SOURCE principles of transparency, reproducibility, and accessibility, which we believe are vital in the era of resource-intensive AI research.