Eauchai, Lattawat

Background/Objectives: Novel vision-language models (VLMs) can integrate patient textual data with image data to support medical diagnosis. Recent studies reported conflicting results regarding the performance of multimodal VLMs compared to other models and physician performance. This systematic review aims to assess the diagnostic performance of multimodal VLMs integrating both patient textual and image data across diverse real-world hospital settings. Methods: We performed comprehensive searches of eight resources, including Embase, MEDLINE, and SCOPUS, on 17 December 2025. Eligible studies reporting diagnostic performance of VLMs integrating both image and patient history textual data from real-world adult patients compared to that of other models and physicians were included. The review adhered to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines. The Prediction model study Risk Of Bias Assessment Tool + AI (PROBAST + AI) was used to assess the quality and risk of bias. The study protocol was registered in the PROSPERO database (CRD420251244054). This review received no external funding. Results: We screened 11,026 records, of which 18 studies met the inclusion criteria. Six studies comparing multimodal and unimodal models demonstrated the consistent superiority of the multimodal models. Four studies evaluating VLM accuracy as standalone agents compared with physician performance reported conflicting evidence. One study assessing VLMs as a clinical copilot demonstrated higher accuracy from the group of physicians using VLM assistance. A meta-analysis could not be performed due to the heterogeneity across study populations and outcomes. The majority of the studies were assessed as having a high risk of bias due to dataset quality. Primary limitations identified across studies include small sample size, a lack of external validation, and the need for prospective clinical deployment studies. No study provided documented considerations regarding model safety or data security. Conclusions: This systematic review suggests that multimodal VLMs consistently outperform unimodal models with access to only image or text. While model performance as standalone agents compared to humans remains inconclusive, a copilot model has demonstrated high diagnostic accuracy. Given substantial methodological concerns across studies, cautious interpretation is required, No firm clinical recommendation can be made regarding the use of standalone VLMs. Further research employing high-quality datasets is needed to ensure the reliability and clinical applicability of future VLMs.

Background/Objectives: With the growing population of kidney transplant recipients (KTRs) in intensive care units (ICUs), understanding their prognostic outcomes is critical. As conflicting findings exist, we aim to systematically evaluate and meta-analyze ICU outcomes in kidney transplant recipients compared to non-recipients. Methods: We conducted a comprehensive search of the PubMed, Embase, and Cochrane databases, from inception through 23 December 2024, to identify relevant studies comparing the outcomes of KTRs and non-transplant ICU patients. Odds ratios (ORs) with 95% confidence intervals (CIs) were calculated for dichotomous outcomes, and weighted mean differences (WMDs) were calculated for continuous outcomes. The risk of bias was assessed using the ROBINS-I V2 tool. The study protocol was registered in the International Prospective Register of Systematic Reviews (CRD42024595104). Results: Seven studies, including 12,062 patients, were analyzed. Demographics, including age and sex, were comparable across groups. No statistically significant associations were found for overall mortality (OR: 1.82, 95% CI: 0.79 to 4.16), ICU mortality (OR: 1.06, 95% CI: 0.45 to 2.48), or 28/30-day mortality (OR: 2.06, 95% CI: 0.30 to 14.10) in KTRs, though there was a trend suggesting a potential increase in the odds of overall mortality. KTRs tended to have longer ICU stays (WMD: +1.96 days, 95% CI: 0.81–3.11) and higher Sequential Organ Failure Assessment (SOFA) scores (WMD: +0.79, 95% CI: −0.78–2.36), but these findings did not reach statistical significance. One study reported higher 1-year and 5-year mortality for KTRs. Sensitivity analyses revealed one influential study. Begg’s test for overall mortality suggested non-significant publication bias (p = 1.0). Conclusions: KTRs in ICUs are at significantly higher risk for long-term mortality, emphasizing the need for tailored critical care strategies and long-term management. Future research should focus on standardizing methodologies, reducing heterogeneity, and addressing gaps in data to improve evidence-based care for this vulnerable population.

Abstract Background and purpose Elevation of serum creatine kinase (CK) or hyperCKemia is considered a biological marker of myopathies. However, selective elevation of serum aldolase with normal CK has been reported in a few myopathies, including dermatomyositis, immune‐mediated myopathy with perimysial pathology and fasciitis with associated myopathy. The aim was to investigate the disease spectrum of myopathies with isolated aldolase elevation. Methods Medical records were reviewed to identify patients >18 years old seen between December 1994 and June 2020 who had pathologically proven myopathies with elevated aldolase and normal CK level. Patients with alternative causes of aldolase elevation were excluded. Results Thirty‐four patients with various types of myopathies were identified. Myopathies were treatable in 27 patients. The three most common etiologies were dermatomyositis ( n = 8), overlap myositis ( n = 4) and nonspecific myopathy ( n = 4). Perimysial pathology comprising inflammation, fragmentation, vasculitis, calcified perimysial vessels or extracellular amyloid deposition was found in 17/34 patients (50%). Eight dermatomyositis patients with selective elevated aldolase were compared to 24 sex‐ and age‐matched patients with dermatomyositis and hyperCKemia. Dermatomyositis patients with normal CK significantly ( p < 0.05) had less frequent cutaneous involvement (50.0% vs. 100.0%) and fibrillation potentials (50.0% vs. 90.5%) but higher median erythrocyte sedimentation rate (33.5 vs. 13.5 mm/h) and more common perifascicular mitochondrial pathology (37.5% vs. 4.2%). Conclusion Isolated aldolase elevation can be found in a greater variety of myopathies than initially thought and most were treatable. Dermatomyositis is the most common myopathy with selective elevation of aldolase in our cohort, which features some unique characteristics compared to dermatomyositis with hyperCKemia.