Julide Gurbuz

Rapid risk stratification is crucial in managing upper gastrointestinal bleeding (UGIB). This study evaluated the prognostic value of the age‐adjusted shock index (ASI), blood urea nitrogen (BUN)/haemoglobin ratio and BUN/platelet ratio for predicting adverse clinical outcomes. This single‐centre retrospective observational cohort study included 204 adult patients (age ≥ 18 years) presenting to the emergency department with UGIB between 1 January 2022 and 31 December 2022. Patients with oesophageal variceal bleeding, lower gastrointestinal sources, trauma, pregnancy, malignancy, prior diagnosis at another institution or incomplete data were excluded. Receiver operating characteristic (ROC) curve analysis was the primary statistical method used to determine prognostic performance. Among the 204 included patients, the intensive care unit (ICU) admission rate was 14%, and the in‐hospital mortality rate was 5.4%. ROC curve analyses for in‐hospital mortality revealed statistically significant predictive capabilities across all evaluated parameters. The BUN/haemoglobin ratio demonstrated an area under the curve (AUC) of 0.779 at an optimal cut‐off of > 8.45, yielding a high negative predictive value (NPV) of 97.26%. The BUN/platelet ratio produced an AUC of 0.747 at a cut‐off of > 0.17, with an NPV of 98.32%. Additionally, the ASI achieved an AUC of 0.780 at an optimal cut‐off of > 61.72 and an NPV of 98.03%. The ASI, BUN/haemoglobin and BUN/platelet ratios are rapid, inexpensive and easily calculable parameters upon admission. Their notably high NPVs make them effective complementary tools to established scoring systems, aiding clinicians in safely identifying low‐risk patients who can be managed in the general ward rather than the ICU.

This study aims to compare emergency physicians’ performance with that of general-purpose large language models (LLMs), such as ChatGPT and Gemini, for pneumothorax (PTX) detection on chest radiographs (CXRs). This single-center, retrospective study of adults was conducted between January 2015 and February 2025 and included 265 PTX cases and 267 non-PTX controls. Exclusions included diagnoses made only by computed tomography, absence of CXR, initial treatment at another center, or incomplete data. Thirteen emergency physicians independently and blindly reviewed CXRs and recorded a binary decision. ChatGPT and Gemini evaluated the same images with a standardized yes/no prompt, with memory cleared between cases to prevent carryover. The primary outcome was LLM diagnostic performance for PTX, while the secondary outcome compared LLMs with physicians. ChatGPT and Gemini exhibited distinct diagnostic performance profiles for PTX detection on CXRs. Gemini demonstrated a sensitivity of 52.5%, whereas ChatGPT demonstrated a sensitivity of 44.5%. Conversely, ChatGPT achieved a specificity of 95.5% and an overall accuracy of 70.1%, while Gemini demonstrated a specificity of 79.0% and an accuracy of 65.8%. Agreement with the reference standard was moderate for ChatGPT, with a kappa value of 0.401, and fair for Gemini, with a kappa value of 0.315. Increasing case difficulty was associated with a reduction in diagnostic accuracy for both models, with correlation coefficients of − 0.438 for ChatGPT and − 0.274 for Gemini. For contextual clinical comparison, emergency physicians demonstrated a sensitivity of 64.5%, a specificity of 99.6%, and an overall accuracy of 82.1%. This study demonstrates model-specific differences in PTX detection by general-purpose AI systems, with Gemini showing higher sensitivity and ChatGPT showing superior specificity and accuracy, both declining with increasing case difficulty. Physician performance remained higher, but was secondary for context. Despite their accessibility and low cost, these models should be considered only adjunctive tools until task-specific optimization and clinical validation are achieved.

Objectives: This study primarily aimed to evaluate the association of EDACS and GRACE scores with angiographically significant coronary lesions in patients presenting with NSTEMI. The secondary objective was to assess their relationship with 30-day mortality. Methods: This retrospective single-center study included adults diagnosed with NSTEMI between January 2022 and December 2024. Demographic and clinical characteristics, GRACE and EDACS scores, angiographic findings, and 30-day mortality were obtained from electronic hospital records. Significant coronary lesions were defined as ≥70 % stenosis or the need for stent placement. Statistical analyses included Mann–Whitney U, Chi-square, logistic regression, and ROC analyses. Results: Among 332 patients, both GRACE and EDACS scores were significantly higher in non-survivors. GRACE high-risk classification was strongly associated with mortality, whilst EDACS (≥16) showed a non-significant trend. EDACS demonstrated superior discriminatory ability for significant coronary lesions, whereas GRACE was more closely associated with mortality risk. An EDACS cut-off >12.5 yielded 87.7 % sensitivity and 40.0 % specificity. Logistic regression identified male sex, hypercholesterolaemia, EDACS ≥16, and systolic blood pressure as independent predictors of significant lesions. Conclusions: GRACE predicted mortality, whereas EDACS better identified significant coronary lesions; combined use showed marginal improvement in model fit but did not significantly enhance discriminatory performance.

Central pontine myelinolysis (CPM) is a non-inflammatory neurological condition with various clinical features and onset patterns and also with a very wide prognostic spectrum ranging from asymptomatic cases to encephalopathy and mortality. In this case report we, aim to present the first case of central pontine myelinolysis as a result of an optimally corrected hyponatremia in a post-COVID-19 patient our knowledge while assessing the clinical and radiological findings.