Nevin Manimala

Eur Radiol. 2025 Aug 20. doi: 10.1007/s00330-025-11933-2. Online ahead of print.

ABSTRACT

OBJECTIVES: To create and test a locally adapted large language model (LLM) for automated scoring of radiology requisitions based on the reason for exam imaging reporting and data system (RI-RADS), and to evaluate its performance based on reference standards.

MATERIALS AND METHODS: This retrospective, double-center study included 131,683 radiology requisitions from two institutions. A bidirectional encoder representation from a transformer (BERT)-based model was trained using 101,563 requisitions from Center 1 (including 1500 synthetic examples) and externally tested on 18,887 requisitions from Center 2. The model’s performance for two different classification strategies was evaluated by the reference standard created by three different radiologists. Model performance was assessed using Cohen’s Kappa, accuracy, F1-score, sensitivity, and specificity with 95% confidence intervals.

RESULTS: A total of 18,887 requisitions were evaluated for the external test set. External testing yielded a performance with an F1-score of 0.93 (95% CI: 0.912-0.943); κ = 0.88 (95% CI: 0.871-0.884). Performance was highest in common categories RI-RADS D and X (F1 ≥ 0.96) and lowest for rare categories RI-RADS A and B (F1 ≤ 0.49). When grouped into three categories (adequate, inadequate, and unacceptable), overall model performance improved [F1-score = 0.97; (95% CI: 0.96-0.97)].

CONCLUSION: The locally adapted BERT-based model demonstrated high performance and almost perfect agreement with radiologists in automated RI-RADS scoring, showing promise for integration into radiology workflows to improve requisition completeness and communication.

KEY POINTS: Question Can an LLM accurately and automatically score radiology requisitions based on standardized criteria to address the challenges of incomplete information in radiological practice? Findings A locally adapted BERT-based model demonstrated high performance (F1-score 0.93) and almost perfect agreement with radiologists in automated RI-RADS scoring across a large, multi-institutional dataset. Clinical relevance LLMs offer a scalable solution for automated scoring of radiology requisitions, with the potential to improve workflow in radiology. Further improvement and integration into clinical practice could enhance communication, contributing to better diagnoses and patient care.

PMID:40836020 | DOI:10.1007/s00330-025-11933-2

Eur Radiol. 2025 Aug 20. doi: 10.1007/s00330-025-11961-y. Online ahead of print.

ABSTRACT

OBJECTIVE: Non-sustained ventricular tachycardia (NSVT) is an independent predictor of sudden cardiac death (SCD) in hypertrophic cardiomyopathy (HCM). This meta-analysis evaluates the prognostic value of late gadolinium enhancement (LGE) on cardiac magnetic resonance (CMR) for predicting NSVT and its association with SCD in HCM.

MATERIALS AND METHODS: We screened electronic databases for studies evaluating the prognostic value of LGE in predicting NSVT and SCD in HCM patients. A random-effects model estimated pooled sensitivity, specificity, accuracy, predictive values, and likelihood ratios for NSVT prediction. Association between LGE extent and NSVT incidence was analyzed using weighted mean differences (WMDs), while pooled odds ratios (ORs) with 95% CIs were calculated to assess LGE’s association with SCD.

RESULTS: Among 20 studies, LGE showed a pooled sensitivity, specificity, and accuracy of 91.33% (95% CI: 88.81-93.86), 37.45% (95% CI: 31.60-43.31), and 52.86% (95% CI: 45.73-59.98), respectively, for NSVT prediction. Positive and negative likelihood ratios and predictive values were, 1.40 and 0.23, and 36.35% and 92.03%, respectively. Patients with NSVT had a significantly greater LGE extent than those without (WMD: 5.95%, 95% CI: 3.08-8.81, p < 0.0001). NSVT prevalence was 28.73% (95% CI: 20.91-36.54). Additionally, LGE presence and SCD were significantly associated (OR 3.64, 95% CI: 2.36-5.61, p < 0.00001).

CONCLUSION: LGE on CMR shows high sensitivity but limited specificity and accuracy for NSVT prediction. Moreover, LGE presence was significantly associated with SCD, and NSVT patients had greater LGE extent. Nonetheless, variability in predictive values and likelihood ratios underscores the need to combine LGE with other imaging biomarkers.

KEY POINTS: Question Can LGE on CMR predict NSVT and SCD in HCM patients? Findings LGE demonstrated high sensitivity but limited specificity for NSVT prediction. Moreover, LGE presence was significantly associated with SCD, and NSVT patients had greater LGE extent. Clinical relevance LGE on CMR is a valuable marker for NSVT prediction and SCD in HCM patients, but it is not widely integrated into clinical practice. Our study highlights the need to integrate LGE with other imaging biomarkers for improved risk stratification.

PMID:40836019 | DOI:10.1007/s00330-025-11961-y

Eur Radiol. 2025 Aug 20. doi: 10.1007/s00330-025-11958-7. Online ahead of print.

ABSTRACT

OBJECTIVES: Preoperative assessment of World Health Organization (WHO) meningioma grading and Ki-67 expression is crucial for treatment strategies. We aimed to develop a fully automated attention-based deep learning network to predict WHO meningioma grading and Ki-67 expression.

MATERIALS AND METHODS: This retrospective study included 952 meningioma patients, divided into training (n = 542), internal validation (n = 96), and external test sets (n = 314). For each task, clinical, radiomics, and deep learning models were compared. We used no-new-Unet (nn-Unet) models to construct the segmentation network, followed by four classification models using ResNet50 or Swin Transformer architectures with 2D or 2.5D input strategies. All deep learning models incorporated attention mechanisms.

RESULTS: Both the segmentation and 2.5D classification models demonstrated robust performance on the external test set. The segmentation network achieved Dice coefficients of 0.98 (0.97-0.99) and 0.87 (0.83-0.91) for brain parenchyma and tumour segmentation. For predicting meningioma grade, the 2.5D ResNet50 achieved the highest area under the curve (AUC) of 0.90 (0.85-0.93), significantly outperforming the clinical (AUC = 0.77 [0.70-0.83], p < 0.001) and radiomics models (AUC = 0.80 [0.75-0.85], p < 0.001). For Ki-67 expression prediction, the 2.5D Swin Transformer achieved the highest AUC of 0.89 (0.85-0.93), outperforming both the clinical (AUC = 0.76 [0.71-0.81], p < 0.001) and radiomics models (AUC = 0.82 [0.77-0.86], p = 0.002).

CONCLUSION: Our automated deep learning network demonstrated superior performance. This novel network could support more precise treatment planning for meningioma patients.

KEY POINTS: Question Can artificial intelligence accurately assess meningioma WHO grade and Ki-67 expression from preoperative MRI to guide personalised treatment and follow-up strategies? Findings The attention-enhanced nn-Unet segmentation achieved high accuracy, while 2.5D deep learning models with attention mechanisms achieved accurate prediction of grades and Ki-67. Clinical relevance Our fully automated 2.5D deep learning model, enhanced with attention mechanisms, accurately predicts WHO grades and Ki-67 expression levels in meningiomas, offering a robust, objective, and non-invasive solution to support clinical diagnosis and optimise treatment planning.

PMID:40836018 | DOI:10.1007/s00330-025-11958-7

Eur Radiol. 2025 Aug 20. doi: 10.1007/s00330-025-11921-6. Online ahead of print.

ABSTRACT

OBJECTIVE: To analyze the differences in contrast-enhanced ultrasound (CEUS) patterns among synovial pathotypes in rheumatoid arthritis (RA), aiming to provide a non-invasive basis for differentiating RA subtypes.

MATERIALS AND METHODS: A total of 151 patients with RA (involving 151 joints) were included and underwent gray-scale ultrasound, Power Doppler ultrasound (PDUS), CEUS, and ultrasound-guided synovial biopsy sequentially. The synovial tissues were classified into three pathological types by immunohistochemistry, including lympho-myeloid, diffuse-myeloid, and pauci-immune. This study analyzed the differences in ultrasound features among different synovial pathotypes and proposed five distinct synovial vascular enhancement patterns. Variables with correlations > 0.3 were selected to construct three classification models, namely Model 1 (Clinical variables), Model 2 (Clinical-PDUS), and Model 3 (Clinical-PDUS-CEUS).

RESULTS: CEUS synovial vascular enhancement patterns: the pauci-immune predominantly exhibited pattern II (35.7%), the diffuse-myeloid predominantly exhibited pattern IV (42.5%), and the lympho-myeloid predominantly exhibited pattern V (72.0%). Qualitative parameters (enhancement pattern, degree, and range) and quantitative parameters (peak intensity [PI], maximum gradient [MGrad], and wash-out area under the curve [WoAUC]) of CEUS were significantly associated with pathological types (r = 0.426-0.602, all p < 0.001). In differentiating pathological subtypes, Model 3 achieved an AUC of 0.902, which was significantly higher than Model 1 (AUC = 0.728) and Model 2 (AUC = 0.751). However, the diagnostic efficacy for the diffuse-myeloid still needs optimization (AUC = 0.818).

CONCLUSIONS: CEUS features of synovium can help differentiate RA synovial pathotypes and may provide important guidance for the early diagnosis and treatment selection of RA.

KEY POINTS: Question Are there differences between RA synovial pathotype and their corresponding contrast-enhanced (CEUS) findings? Findings Five patterns of synovial vascular enhancement corresponding to pathology types were identified, and a model was constructed. Clinical relevance CEUS synovial qualitative analysis and quantitative parameters can help differentiate RA synovial pathotypes.

PMID:40836017 | DOI:10.1007/s00330-025-11921-6

Sci Rep. 2025 Aug 20;15(1):30502. doi: 10.1038/s41598-025-16364-z.

ABSTRACT

Accurate estimation of the evaporation is of great significance for the management of limited agricultural water resources. However, developing highly accurate and universal data- driven models using time-series analysis methods to achieve precise evaporation estimation remains a challenging. Specifically, integrating meta-heuristic algorithms, ensemble deep learning models, and data preprocessing techniques for evaporation prediction is notably scarce. The aim of this paper was to employ time series analysis methods to develop data-driven model with high accuracy and universality to realize accurate estimation of evaporation. To achieve this purpose, the Convolutional neural network (CNN) was integrated with Bidirectional long short-term memory network (BiLSTM) as main estimating module, and the Sparrow search algorithm (SSA) was employed to search the optimal hyperparameters of CNN-BiLSTM. To overcome the drawback that directly using measured evaporation time series to predict evaporation may lead to large error, the Variational mode decomposition (VMD) was used to extract multiscale traits of evaporation time series, and Whale optimization algorithm (WOA) was adopted to find the optimal parameters of VMD, and a novel hybrid deep learning model WOA-VMD-CNN-SSA-BiLSTM was proposed to estimate the evaporation in the Linze County, China. The estimating performance was evaluated by using the statistical accuracy metrics, including R², the mean squared error (MSE), the mean absolute error (MAE), the root mean squared error (RMSE), and the mean absolute percentage error (MAPE). The results show that the Sample entropy (SEn) remains 0.0832 when the optimal values of [Formula: see text] and [Formula: see text] of VMD are 6 and 0.1773, suggesting that VMD optimized by using WOA effectively overcomes the subjectivity in traditional VMD parameter setting and realizes amplitude-dependent feature extraction of evaporation time series in the study area. In addition, the model performance of CNN-SSA-BiLSTM can be significantly improved by coupling CNN-SSA-BiLSTM with WOA-VMD, and the hybrid model WOA-VMD-SSA-CNN-BiLSTM with MSE = 0.1258, RMSE = 0.3547, MAE = 0.2833, and MAPE = 6.17% in testing stage is superior than other hybrid models and ensemble models, which could be highly recommended for estimating evaporation in study area.

PMID:40836003 | DOI:10.1038/s41598-025-16364-z

Burns. 2025 Aug 6;51(8):107644. doi: 10.1016/j.burns.2025.107644. Online ahead of print.

ABSTRACT

BACKGROUND: For many burn survivors, the symptomatic burn scarring can limit the functionality and compromise the quality of life. Various modalities have been in use over the years to minimize the scar-related distress of the patients but none effective in totality. Due to availability of sparse data on efficacy of a most potent combination therapy, we used split-scar technique to compare Er:YAG and triamcinolone versus triamcinolone alone.

METHODS: The study was a prospective, interventional, double blind, parallel, two-arm, randomized controlled trial including 34 patients from November 2022 to April 2024. Scars were divided into two halves using lottery system; one half was given Er:YAG laser along with intralesional triamcinolone (group A) while the other half was given Er:YAG laser alone (group B). Results were compared objectively by measuring the decrease in thickness and subjectively on UNC4P scar scale and patient satisfaction index.

RESULTS: In group A, 52.9 % (n = 18) achieved > 75 % reduction in thickness as compared to 26.5 % (n = 9) in group B (p < .05). UNC4P scar scale revealed better symptomatic relief in the scars of group A and the difference was found statistically significant for pruritus, paraesthesia, and pliability; and comparable for pain. Patient satisfaction index was also higher in group A.

CONCLUSION: A combination treatment of Er:YAG laser along with intralesional corticosteroid is more efficacious in improving the scar thickness and providing symptomatic relief, better safety profile and improved patient satisfaction–thereby establishing it as a better modality of treatment compared to laser monotherapy of scars.

PMID:40834485 | DOI:10.1016/j.burns.2025.107644

Burns. 2025 Aug 5;51(8):107650. doi: 10.1016/j.burns.2025.107650. Online ahead of print.

ABSTRACT

BACKGROUND: Burn injuries significantly impact quality of life and physical functionality. Early, accurate evaluation of burn wounds is essential, yet assessing burns remains a challenge, especially for non-specialists. This pilot study examines the efficacy of an AI-powered diagnostic tool using multispectral imaging (MSI) to help medical teams determine whether conservative or surgical management is required for burn wounds.

METHODS: Thirty-one acute burn wounds in adult patients (within 7 days of injury) were assessed at a super-regional burn center. Clinical examinations were performed by an experienced burn doctor, with two adjunct devices used: the AI-driven MSI DeepView SnapShot Imaging (Version 1.0.1) and the Moor Laser Doppler Imaging (LDI) device, as per NICE recommendations. Wounds on the face, hands, feet, and genitals were excluded. Predictive outcomes from MSI and LDI were compared to final clinical management decisions.

RESULTS: MSI predicted clinical outcomes in 58 % of cases, while LDI achieved 90 % accuracy. Concordance between MSI and LDI was observed in 52 % of cases, with a statistically significant difference between their outcomes (McNemar’s test p = 0.012).

CONCLUSION: This study highlights the potential of AI in burn wound management. However, the binary classification of current AI models may not fully address the complexities of burn healing. The observed accuracy suggests limitations in AI’s ability to capture the multifactorial nature of burn wounds, indicating the need for further refinement and collaboration with clinical expertise.

PMID:40834474 | DOI:10.1016/j.burns.2025.107650

Burns. 2025 Jul 21;51(8):107626. doi: 10.1016/j.burns.2025.107626. Online ahead of print.

ABSTRACT

INTRODUCTION: Lower urinary tract symptoms (LUTS) are frequently observed following perineal burn injuries, but their long-term clinical implications and predictors of prognosis remain poorly understood. This study aimed to investigate the clinical characteristics of newly developed LUTS in patients with perineal burns and to identify factors associated with persistent symptoms.

METHODS: We retrospectively reviewed 172 patients who were admitted to a burn center with perineal burns and subsequently referred to the urology department for evaluation of new-onset LUTS between August 2010 and December 2023. LUTS were evaluated at the time of urologic referral using available clinical data. Patients were categorized into transient and persistent symptom groups based on whether pharmacologic treatment was required beyond 3 months. Multivariate logistic regression was used to identify independent predictors.

RESULTS: The most common LUTS were urinary frequency (44.2 %), nocturia (28.5 %), and incomplete bladder emptying (25.6 %). Storage symptoms were observed in 64.5 % of patients, while voiding symptoms were noted in 58.1 %. Most patients showed symptom resolution with short-term standard treatment. However, 30 % experienced persistent LUTS beyond 3 months, requiring continued pharmacologic therapy. Eleven patients (6.4 %) ultimately needed catheter-based management. Among burn-related factors, electrical burns were independently associated with persistent LUTS (OR 7.7, 95 % CI 2.3-26.0, p = 0.001), whereas other variables were not statistically significant.

CONCLUSION: LUTS following perineal burns generally resolve with short-term treatment, but a substantial proportion of patients experience persistent symptoms requiring long-term care. Early identification and management are particularly important in patients with electrical burns.

PMID:40834473 | DOI:10.1016/j.burns.2025.107626

Geriatr Nurs. 2025 Aug 19;66(Pt A):103581. doi: 10.1016/j.gerinurse.2025.103581. Online ahead of print.

ABSTRACT

Fatigue is one of the most common and distressing symptoms in patients with chronic obstructive pulmonary disease (COPD). This study applied MICT, EFT and their combination (MICT and EFT) to COPD patients with fatigue through factorial design analysis method, aiming to explore non-drug management strategies that are more conducive to relieving fatigue in COPD patients. This study had a total of 112 participants. Patients were randomly divided into Control group, MICT group, EFT group and Combination group (MICT-plus-EFT group) by random number table method. The intervention duration was 8 weeks. The results showed that there were statistically significant differences in fatigue scores, anxiety and depression scores, sleep quality scores, and 6MWT scores among the four groups at T₁, T₂, and T₃(P<0.05). The results of factorial analysis showed that there was no significant interaction between MICT and EFT on fatigue score, HADS score, PSQI score and 6WMT.The overall effect of the combined intervention was better than that of the single intervention, but the effect of the combined intervention may simply be the superposition of the effects of the two measures.

PMID:40834442 | DOI:10.1016/j.gerinurse.2025.103581

Geriatr Nurs. 2025 Aug 19;66(Pt A):103556. doi: 10.1016/j.gerinurse.2025.103556. Online ahead of print.

ABSTRACT

OBJECTIVE: Evidence-based development and empirical application of a discharge preparation service program for cataract day surgery patients.

METHODS: Evidence-based construction of a discharge preparation service programme for cataract day surgery patients. Using the convenience sampling method, 84 patients undergoing cataract day surgery in July-August 2024 were divided into a control group and an intervention group; Comparing the level of readiness for hospital discharge, quality of discharge teaching, caregiver preparedness, patients’ ability to self-care, complication rate and unplanned readmission rate 1 month post-discharge between the two groups.

RESULTS: The intervention group scored higher than the control group in terms of readiness for hospital discharge, quality of discharge teaching, caregiver readiness, and patient self-care ability (P < 0.05), but the differences in post-discharge complications and unplanned readmissions were not statistically significant.

CONCLUSION: Evidence-based construction of a discharge preparation service programme for cataract day surgery patients has good practical value.

PMID:40834438 | DOI:10.1016/j.gerinurse.2025.103556