Citations

Image-derived artificial intelligence (AI) risk models have shown promise in short-term risk assessment for improving breast cancer (BC) screening. No image-derived long-term AI risk model for primary prevention has been developed and externally validated. Individuals aged 31 to 94 years were recruited between 2010 and 2020 in two population-based case cohorts (Olmsted County, Minnesota, US, and KARMA, Sweden) and one hospital-based case-control study (EMBED, Atlanta, US). Median follow-up in the case cohorts was 10 years, with BCs diagnosed through June 2022. Additional validation was performed in EMBED with 3-year follow-up. An AI risk model was developed in Sweden, and we report independent validation in Olmsted/KARMA/EMBED. Absolute 10-year risks were estimated at study entry. Time-dependent discriminatory performance [AUC( t )] and expected-to-observed events (E/O) were estimated. Comparisons were performed with clinical risk tools and the image-based Mirai tool. Across Olmsted/KARMA, 8696 individuals (mean age, 54.4±10.6) and 1633 individuals with incident BC (mean age, 57.0±10.6) were included. Average 10-year risks were 3.83 and 3.14%, and E/O ratios were 0.99 and 0.99. Ten-year AUC( t ) for invasive BCs were 0.72 in Olmsted and 0.72 in KARMA. Similar discriminatory results were observed in EMBED. Our AI risk tool performed significantly better than Mirai in Olmsted/KARMA/EMBED. In KARMA, in the top 10% of high-risk individuals, the AI risk tool predicted 33% of BCs compared with 23%/20%/24% predicted by Tyrer-Cuzick-v8/BCSC-v3/Mirai (all P < 0.01). The 10-year image-derived AI risk model showed significantly higher performance than clinical risk models and Mirai in diverse populations, supporting its clinical potential for primary prevention.

Patients with advanced liver disease challenge anesthesiologists, especially those providing regional anesthesia for postoperative pain management. Coagulopathy, encephalopathy, sepsis and other comorbidities complicate care. Liver failure lowers albumin and α1-acid glycoprotein and slows hepatic clearance, increasing the unbound fraction and effective exposure to amide local anesthetics. Risks are magnified by acidosis, low cardiac output, low body mass index, sepsis, and concurrent renal dysfunction. Neuraxial options are often limited by coagulopathy and the risk of catastrophic neurologic complications while superficial, compressible peripheral blocks are generally considered safer than deep plexus targets. Safe practice includes the use of ultrasound guidance, incremental injection with frequent aspirations, avoidance of unnecessary high-dose multiple-site injections, and conservative dosing and infusion strategies. Regional anesthesia is feasible in patients with hepatic failure when dose and technique are adjusted to disease biology and when local anesthetic systemic toxicity (LAST) preparedness is routine. Catheters should run at lower concentrations and rates with scheduled reassessment and clearly communicated stop times. Teams should keep 20% lipid emulsion immediately available for treatment of neurologic or cardiovascular signs of LAST. We provide a practical pathway to select the block, choose the agent, and set doses that keep a reasonable safety margin for local anesthetics.

Foundation models are reshaping computational pathology by enabling scalable task-agnostic representations of histopathological whole-slide images (WSIs). Unlike earlier task-specific deep learning systems, pathology foundation models (PFMs) leverage massive whole-slide image repositories and self-supervised Vision Transformer architectures to achieve broad generalization and few-shot adaptability. Their evolution reflects a shift from weakly supervised approaches such as Clustering-Constrained Attention Multiple Instance Learning (CLAM) and hierarchical architectures such as Hierarchical Image Pyramid Transformer (HIPT) to large-scale efforts including foundation models, UNI, Virchow, Phikon, CONtrastive learning from Captions for Histopathology (CONCH), GigaPath, H-Optimus, Transformer-Based Pathology Image and Text Alignment Network (TITAN), and the Mayo Clinic Atlas. These models demonstrate impressive performance across diagnostic and prognostic benchmarks while also opening pathways for multimodal integration with genomics and clinical data. Yet significant barriers remain including inconsistent generalization across institutions, interpretability lagging behind clinical needs, and slow integration into routine laboratory workflows. Certain domains of anatomic pathology such as cytopathology, transplant pathology, frozen sections, and rare tumor subtypes remain particularly resistant to current models. Here, we review the development of PFMs, critically evaluate their strengths and limitations, and outline priorities for their safe and effective clinical translation. We argue that the next phase of PFM development will depend on rigorous benchmarking, pathologist-in-the-loop deployment, and multimodal fusion ensuring these models evolve from research tools into clinically robust systems.

OBJECTIVE The objective was to evaluate the trajectories of hematoma resolution and functional improvement after middle meningeal artery embolization (MMAE) for chronic subdural hematoma (cSDH), model the temporal pattern of cSDH resolution, and identify factors associated with favorable outcomes. METHODS This real-world multicenter retrospective study included cSDH patients treated with MMAE at 24 centers between 2019 and 2024. Hematoma thickness was measured at baseline and at follow-up intervals (1–4 weeks, 1–3 months, 3–6 months, 6–12 months, and > 12 months after embolization). Resolution patterns were modeled using exponential decay functions. Modified Rankin Scale (mRS) scores assessed functional outcomes. Good functional outcome was defined as mRS score ≤ 2. Resolution patterns were modeled using exponential decay functions to estimate time to 50% and 80% reduction. Patients were categorized as complete resolution (≥ 99%) or by quartile of the remaining distribution (substantial [73%–98%], moderate [53%–73%], partial [27%–53%], minimal [< 27%]). The primary outcome was good functional status (mRS score ≤ 2). RESULTS The authors analyzed a total of 1781 patients with 2295 cSDHs who underwent MMAE. The mean ± SD age was 72.8 ± 12.4 years and 68.1% of patients were male. The initial mean hematoma thickness was 15.31 ± 6.53 mm, decreasing to 5.24 ± 5.91 mm at final follow-up (mean reduction 64.3% ± 42.1%). Resolution followed an exponential decay pattern, with an estimated time to 50% reduction of 1.8 months and to 80% reduction of 8.9 months. Complete resolution occurred in 1031 of 2224 patients (46.4%) with complete follow-up. The median (IQR) mRS score improved from 1 (0–3) at baseline to 0 (0–2) at > 12 months. Good functional outcomes were more common in patients with complete versus minimal resolution (68.9% vs 35.0%, p < 0.001). Achieving ≥ 73% resolution within 90 days was associated with better outcomes (good outcome in 76.9% of those with ≥ 73% resolution vs 67.3% in those without, p < 0.001). Neurological deterioration was the strongest predictor of lack of good outcome (23.7% in patients with neurological deterioration vs 82.4% without, p < 0.001). CONCLUSIONS After MMAE for cSDH, reduction in hematoma thickness follows a predictable exponential decay pattern. Greater extent (≥ 80%) and faster timing (within 90 days) of resolution are valuable prognostic indicators. Functional outcomes improve progressively through 6–12 months after the procedure. The relationship between resolution extent and functional outcomes provides quantitative benchmarks for evaluating treatment response.