Frontiers in Medicine最新文献

Anesthesiology and intensive care medicine are among the most data-rich fields of medicine, where accurate and timely outcome prediction or risk stratification is important. During patient care, heterogeneous data streams, including structured electronic health records, free-text documentation, and high-frequency physiologic time series are recorded. This provides a fertile ground for machine learning (ML) models to make individualized risk predictions. Yet, secondary use of routine data remains difficult due to heterogeneity, missingness, variable granularity, ambiguously defined outcomes, or poor representation of clinical concepts in routine data. Reproducibility and transparency are difficult to achieve with hospital-specific complex data pipelines. New complexities arise when combining different data modalities. This perspective article discusses three common modalities-tabular data, clinical text, and time series-and outlines data modality-specific challenges, data preprocessing strategies, and ML modeling approaches. We examine multimodal fusion strategies through the common taxonomy of early, intermediate, and late fusion. In early fusion, generated features are aggregated into a unified tabular representation, offering simplicity and often serve as first baseline prediction models. Intermediate fusion uses modality-specific encoders with shared layers to learn cross-modal dependencies. This strategy yields the most complex and powerful models. Late decision-level fusion combines outputs from modality-optimized models, providing modularity and robustness to missing modalities, leading to advantages for real-time deployment where data arrive asynchronously. The growth of multi-centric datasets and federated infrastructures may enable intermediate-fusion architectures and multimodal foundation models to better capture patient trajectories, supporting risk stratification and personalized therapy in perioperative and intensive care settings.

Heart failure (HF) is a prevalent and debilitating condition that significantly affects patients' quality of life and places a substantial burden on healthcare systems. In recent years, digital technologies have been increasingly explored in cardiac rehabilitation (CR), particularly through their integration within Internet of Things (IoT) ecosystems to support remote monitoring and personalized care. This review aimed to provide a focused overview of emerging digital technologies applicable to the rehabilitation of patients with HF, with emphasis on solutions compatible with IoT-based systems. A targeted literature search was conducted in PubMed, Scopus, Cochrane and Web of Science, including studies published between 2019 and 2024. Studies addressing digital technologies in HF rehabilitation. Following a structured selection process, 59 articles were included in the narrative synthesis. The findings indicate a growing body of literature investigating wearable physiological monitoring devices, telehealth-based CR programs, digital platforms, and smart sensors, many of which have been explored for integration within IoT infrastructures. These technologies have been associated with improved remote follow-up, patient engagement, and real-time physiological data collection outside traditional clinical settings. Emerging applications of artificial intelligence within IoT-enabled systems have also been examined to support clinical workflows and adaptive rehabilitation strategies. Despite increasing interest, challenges remain, including heterogeneity in study designs, usability concerns, data privacy and security issues, economic barriers, and limited large-scale clinical validation. Overall, this review suggests that IoT-enabled technologies represent a promising area of research in CR, warranting further investigation to support their sustainable integration into routine HF care.

Chronic constipation (CC) is a prevalent functional gastrointestinal disorder involving complex interactions among the brain-gut-microbiome axis, with 5-hydroxytryptamine (5-HT) as a key signaling node. Aberrations in 5-HT synthesis, release, receptor expression, or reuptake disrupt gastrointestinal motility, contributing to CC pathogenesis. Traditional Chinese Medicine (TCM), including herbal compounds, monomers, acupuncture, and tuina, exerts therapeutic effects by modulating the 5-HT signaling pathway. Animal studies demonstrate that TCM interventions regulate gut microbiota, promote 5-HT production via metabolites like short-chain fatty acids (SCFAs) and bile acids, and target receptors (e.g., 5-HT₃R, 5-HT₄R) to enhance intestinal motility. Clinical trials validate TCM's efficacy in normalizing 5-HT levels and improving symptoms, with advantages in safety and holistic regulation. However, important gaps remain, including incomplete understanding of upstream and downstream 5-HT signaling mechanisms, paradoxical 5-HT expression, and limited investigation of comorbid emotional disorders. Future studies should explore how TCM interventions modulate the gut microbiota-5-HT axis and inflammation-related pathways to provide novel insights into CC management.

Background: Cutaneous squamous cell carcinoma (cSCC) is a common non-melanoma skin cancer with potential for local invasion and metastasis. Accurate preoperative assessment is essential for optimal treatment planning.

Materials and methods: We report a case of an 83-year-old female patient who presented with a progressively enlarging scalp mass over 3 months. HR-MRI revealed a mixed-signal lesion (22 × 15 × 26 mm) in the right scalp. On T1-weighted imaging (T1WI) and T2-weighted imaging (T2WI), the lesion exhibited heterogeneous signals with an irregular, crater-like surface. Post-contrast imaging demonstrated significant heterogeneous enhancement. The tumor was located within the epidermis, dermis, and subcutaneous fat, closely adhering to the galea aponeurotica with well-defined margins. Surgical resection and histopathological examination confirmed a (exophytic type) moderately to well-differentiated cSCC (2.8 × 2.3 × 2.0 cm) infiltrating the subcutaneous tissue but without perineural invasion or deeper tissue involvement.

Results: HR-MRI provided clear visualization of tumor morphology, infiltration depth, and relationship with surrounding structures. Compared to conventional MRI, HR-MRI improved the accuracy of tumor boundary delineation, offering valuable information for preoperative planning.

Conclusion: HR-MRI plays a significant role in the evaluation of cSCC, particularly in assessing tumor infiltration depth and differentiating it from other cutaneous malignancies. Its high-resolution imaging facilitates early detection, precise surgical planning, and improved patient outcomes.

Background: Relugolix, an oral gonadotropin-releasing hormone (GnRH) antagonist, represents a potentially effective and less invasive therapeutic approach for retained products of conception (RPOC). However, its impact on subsequent pregnancy outcomes remains unclear. This study aimed to evaluate these outcomes following GnRH antagonist treatment for RPOC.

Methods: This single-center cohort study encompassed 20 patients diagnosed with RPOC following miscarriage or abortion before 22 gestational weeks who were treated with oral relugolix, a GnRH antagonist, from January 2022 to July 2024. Following treatment completion and hysteroscopic confirmation of complete resolution, 12 patients subsequently conceived and were prospectively followed for pregnancy and neonatal outcomes. To contextualize outcomes, results were compared with the non-GnRH antagonist group managed at the same institution from 2014 to 2021 without GnRH antagonist exposure.

Results: The GnRH antagonist group had fewer patients requiring surgical intervention than the non-GnRH antagonist group [30.0% (6/20) vs. 70.1% (54/77), p = 0.002], underscoring the reduced invasiveness of medical therapy. Of the 12 pregnancies following relugolix treatment, 4 led to early miscarriage and 8 progressed beyond 22 gestational weeks. The GnRH antagonist and non-GnRH antagonist groups exhibited comparable clinical pregnancy and live birth rates (53.3% vs. 73.3 and 33.3% vs. 48.9% per embryo transfer, respectively; not significant). Similarly, the interval from RPOC diagnosis in the previous pregnancy to gestational sac confirmation in the current pregnancy (328.9 ± 153.4 vs. 486.9 ± 658.8 days) and the interval from RPOC treatment completion to gestational sac confirmation (295.7 ± 166.6 vs. 445.8 ± 628.2 days) did not significantly differ between both groups.

Conclusion: Relugolix therapy for RPOC was associated with preserved fertility and favorable pregnancy outcomes comparable to conventional management.

Background: Early risk-stratification in hypertriglyceridemia-induced acute pancreatitis (HTG-AP) is challenging within the first 24 h of admission. We assessed whether pleural effusion (PE) detected on early imaging serves as an early predictor for the development of severe HTG-AP (HTG-SAP) and organ failure (OF), compared with BISAP, MCTSI, and C-reactive protein (CRP).

Methods: A retrospective study included 296 HTG-AP patients at Xuanwu Hospital from August 2013 to February 2024. PE presence within 24 h was abstracted from chest CT; outcomes were HTG-SAP and OF. Diagnostic performance was evaluated using ROC analysis and DeLong tests, comparing with BISAP, MCTSI, and CRP. The prognostic impact of PE laterality (unilateral vs. bilateral) was also analyzed.

Results: PE demonstrated the strongest positive correlation with hypertriglyceridemia-induced severe acute pancreatitis (HTG-SAP) (OR: 6.82; 95% CI: 3.13-14.88; p < 0.001). PE achieved the area under the curve (AUC) of 0.792 (95% CI: 0.742-0.837) for HTG-SAP prediction and 0.718 (95% CI: 0.663-0.768) for OF anticipation. The predictive accuracy of PE was significantly higher than that of CRP for predicting HTG-SAP (ΔAUC = 0.106; p = 0.004) and comparable for predicting OF (ΔAUC = 0.055; p = 0.112). There was no significant difference between PE and MCTSI scores in predicting either HTG-SAP or OF. However, PE had a lower predictive accuracy compared with the BISAP for both HTG-SAP (ΔAUC = 0.057) and OF (ΔAUC = 0.083; p < 0.001). Despite this, PE still demonstrated substantial predictive value for HTG-SAP and OF. There was no significant difference between unilateral and bilateral pleural effusions in predicting the risk of developing HTG-SAP and OF.

Conclusion: PE detected within 24 h is a simple, readily available early radio-graphic marker that identifies patients at high risk for progressing to severe course and OF in HTG-AP. While not superior to BISAP, PE offers pragmatic utility when composite scores or complete laboratories are unavailable and may trigger early monitoring and escalation.

Bullous pemphigoid (BP) is the most common autoimmune blistering disease affecting the skin. It is characterized by a type 2 immune-mediated pathogenesis. BP predominantly affects elderly patients who have multiple comorbidities, and treatment can be particularly challenging in refractory cases or when standard immunosuppressive therapies are limited due to toxicity or contraindications. Here, we present a case series of five patients with confirmed BP who were treated with off-label dupilumab. All five patients achieved sustained clinical improvement with dupilumab treatment, including reductions in disease activity and pruritus. No severe adverse events were reported, except for one case of manageable conjunctivitis. Dupilumab was well tolerated and resulted in marked clinical benefit in all five BP patients included in this case series. These findings support the growing evidence for dupilumab's potential role as a safe and effective treatment option for BP. Further long-term clinical trials are needed to assess its disease-modifying potential.

Aim: To assess the effect of bedside comprehensive ability training on teaching and training in operating room.

Methods: One hundred and thirty female clinical nurses who joined our hospital from January 2019 to December 2021 were randomly assigned to a control group and an observation group. The control group adopted problem-based learning (PBL), while the observation group adopted bedside comprehensive ability training. The training effect, core competence evaluation results, self-comprehensive ability as well as teaching satisfaction of 2 groups of nurses were compared.

Results: After training, nurses in the observation group demonstrated significantly higher training scores (p < 0.05), higher core competence (p < 0.05), higher self-comprehensive ability (p < 0.05), higher scores of nursing ability (p < 0.05) and better satisfaction rate (p < 0.001).

Conclusion: Bedside comprehensive ability training can effectively promote the core ability and training results of new clinical nurses and further improve their own comprehensive ability and nursing ability.

Background: Alimentary tract hemorrhage (ATH) after intracerebral hemorrhage (ICH) surgery is a common complication that can increase morbidity and mortality. Prevention of this complication is important for recovery of ICH patients, and early identification of high-risk patients would facilitate targeted prevention. Machine learning (ML) is a data-driven tool that can potentially be used to predict postoperative ATH in ICH surgical patients. However, there are currently no validated ML models for this purpose.

Methods: A retrospective cohort study was performed with 658 ICH surgical patients from a single center. Five predictors were selected with the Boruta algorithm, and a total of 12 ML models were developed. The models were validated on a 70/30 train-test split, and further performance validation was performed with 10-fold cross-validation. The primary endpoint was postoperative ATH, and long-term functional outcome at 180 days was assessed with Modified Rankin Scale (MRS).

Results: The Light Gradient-Boosting Machine (LGBM) model showed the best performance with an AUROC of 0.918 in the test set and an average AUROC of 0.949 on cross-validation. The five confirmed predictors were hemorrhage volume, Glasgow Coma Scale (GCS) score, surgery time, albumin, and glucose. In addition, ATH was significantly associated with lower odds of good functional outcome (MRS 0-2) at 180 days (log-rank p = 0.0012).

Conclusion: The present study developed an accurate and easy-to-use ML model for early prediction of ATH in ICH surgical patients. Postoperative ATH was associated with worse long-term neurological recovery, further highlighting the importance of its prevention. The developed model should be externally validated and further used to guide the development of personalized prophylactic ATH strategies.