Limits Of Agreement Research Articles

Clinical utility of non-gated 4-minute delayed dual-energy computed tomography for myocardial extracellular volume quantification.

To develop and validate the myocardial extracellular volume (ECV) obtained from non-electrocardiography (ECG)-gated delayed CT images acquired 4 minutes post-contrast infusion (4-min-nonECG-ECV) compared with the ECV obtained from ECG-gated delayed CT images acquired 10 min post-contrast infusion (Conv-ECV). We retrospectively analysed 29 patients (males: 21) after a comprehensive CT protocol of both 4-min-nonECG-ECV and Conv-ECV on a dual-layer CT scanner. The mean volume of contrast medium administered: 90 ± 11.8 mL, the average heart rate during the CT examinations: 74.2 ± 18.2 bpm. Two independent observers calculated the respective 4-min-nonECG-ECV and Conv-ECV. We determined the correlation between the ECV obtained by the two methods and conducted a Bland-Altman analysis to identify systematic errors and determine the limits of agreement (LOA) between the 4-min-nonECG-ECV and Conv-ECV values. The respective median ECV values for observer 1 were 27.3 for 4-min-nonECG-ECV and 26.5 for Conv-ECV; for observer 2, they were 27.8 and 27.1. The correlation between the methods was 0.97 for both observers (p < 0.01). The Bland-Altman plots for observers 1 and 2 demonstrated a minor bias (-0.2% and -0.5%, respectively), with the 95% LOA ranges at - 4.4%-4.0% and -5.0%-4.0%, respectively. The 4-min-nonECG-ECV provided ECV values comparable to those obtained by Conv-ECV. Myocardial ECV quantification is feasible using a non-gated, 4-minute delayed dual-energy CT scan with an already established CT acquisition method. This approach achieves ECV accuracy comparable to that of the conventional CT-ECV calculation method (gated 10-minute delayed imaging) while enhancing clinical efficacy and diagnostic throughput.

Opportunistic osteoporosis screening in intraoperative CT can accurately identify patients with low volumetric bone mineral density and osteoporosis during spine surgery.

To evaluate the accuracy of opportunistic measurements of volumetric bone mineral density (vBMD) in intraoperative multi-detector CT (MDCT) scans, using preoperative MDCT as the reference. This retrospective, single-center study included 105 patients (mean age: 73 ± 12.6 years, 53 women) who underwent spine surgery for various indications. All patients had preoperative MDCT with/without intravenous contrast and unenhanced intraoperative scans. VBMD of thoracolumbar vertebrae was automatically extracted using a convolutional neural network (CNN)-based framework with asynchronous calibration and contrast-phase correction. Vertebrae affected by artifacts, fractures, or severe degenerations were excluded. Root-mean-square errors (RMSEs) for associations between pair-wise vertebrae from preoperative and intraoperative vBMD values were calculated in linear regression models. Mean bias and 95%-limits of agreement (LOA) were calculated in Bland-Altman plots. Strong associations between preoperative and intraoperative vBMD values were observed in the thoracic (R2 = 0.94) and lumbar spine (R2 = 0.96). Intraoperative vBMD values showed high accuracy in reference to preoperative measurements with a mean bias of -1.3mg/cm3 for the thoracic spine (LOA: -18.7 to 16.1mg/cm3) and - 3.0mg/cm3 for the lumbar spine (LOA: -17.4 to 11.3mg/cm3). RMSEs between preoperative and intraoperative vBMD values slightly increased for contrast-enhanced scans (RMSEthoracic: 8.42 vs. 10.1mg/cm3; RMSElumbar: 7.75 vs. 8.87mg/cm3). Opportunistic osteoporosis screening with the presented approach is feasible and demonstrates high accuracy in reference to preoperative MDCT scans. This could enable the identification of patients with low bone mass during surgery, allowing surgeons to take measures (e.g., adapted techniques) that prevent postoperative complications and improve patient outcomes.

Multi-centre comparison between device-independent web-browser perimetry (Melbourne Rapid Fields-web) and SITA-Faster for glaucoma

PurposeVisual field testing is important for glaucoma diagnosis and management, but access to standard automated perimetry can be limited in some areas due to cost or access. Melbourne Rapid Fields-web (MRF-web) perimeter is designed to address these limitations by allowing perimetry testing on the flat screen of your personal computer.MethodsThis study is a retrospective, cross-sectional study involving two locations in Australia, one in metropolitan Melbourne and one in rural Dubbo NSW. 232 patients with stable glaucoma, glaucoma suspect or normal eyes were tested with MRF-web and outcomes were compared to the most recent Humphrey Field Analyzer (HFA) 24-2 SITA Faster test. Outcomes were compared by Deming regressions, Intraclass Correlation Coefficients (ICC) and Bland-Altman methods.ResultsPatient age ranged from 21 to 92 (average 66.3, SD 16.1). Bland-Altman found a bias of -0.50dB for Mean Deviation (MD) between the two tests, with 95% Limits of Agreement (LoA) of -6.80dB to 5.80dB. Pattern Deviation (PD) had a bias of -0.58dB with 95% LoA of -5.60dB to 4.40dB. High concordance was found for MD and PD, with ICCs of 0.87 and 0.73. No significant differences were found in false positive and fixation loss rates. Test time was approximately one minute longer for MRF-web compared to SITA-Faster. Area Under the Curve of MRF and HFA are similar indicating comparable diagnostic capacity.ConclusionMRF-web produces outcomes comparable to HFA SITA-Faster. Its portability and cost-effectiveness suggest suitability as an alternative method for visual field testing where a standard perimeter is not easily accessible.

Point-of-care ultrasound in the diagnosis and management of community-acquired Pneumonia in out-of-hospital care: a narrative review

The article reviews the use of Point-of-Care Ultrasound (POCUS) in the diagnosis and management of community-acquired pneumonia (CAP) in extra-hospital settings, such as primary healthcare, outpatient clinics, and home care. CAP is one of the main causes of morbidity and mortality across different age groups, including children and the elderly, requiring quick and accurate diagnoses to improve clinical outcomes. Traditionally, methods such as chest X-rays (CXR) and computed tomography (CT) scans are used to diagnose CAP, but they have limitations such as high cost, the need for specialized infrastructure, and radiation exposure. POCUS emerges as an effective alternative, offering high accuracy and being feasible for use by healthcare professionals who are not specialists in imaging diagnosis. Its portability and accessibility make it especially useful in remote or underserved areas. The absence of radiation makes POCUS particularly beneficial for sensitive populations, such as pregnant women. The review analyzed 27 studies published between 2015 and 2024, highlighting that POCUS not only competes with traditional methods in terms of accuracy but also provides additional advantages, such as the possibility of repeating exams for continuous clinical monitoring. In pediatrics and geriatrics, POCUS has proven particularly effective, surpassing chest X-rays in some diagnostic aspects. Moreover, the implementation of POCUS in primary care can reduce healthcare costs by decreasing the need for conventional imaging exams and hospital admissions. The tool has also proven to be valuable in public health emergencies, such as the COVID-19 pandemic, where it facilitated rapid diagnoses and monitoring of pulmonary complications. However, challenges persist, such as interobserver variability and the need for adequate training to maximize the reproducibility of the findings. The integration of artificial intelligence and the development of specific protocols for primary care are future perspectives that can further enhance the effectiveness of POCUS. It is concluded that POCUS has great potential to transform the diagnosis and management of CAP in out-of-hospital settings, promoting faster, more accurate, and accessible diagnoses, as well as contributing to greater equity in access to healthcare.

Real-world smartphone-based point-of-care diagnostics in primary health care to monitor HbA1c levels in people with diabetes

BackgroundThe lack of accurate and affordable monitoring of glycated hemoglobin (HbA1c) is a common issue among patients with diabetes in low- and middle-income countries. We aimed to test a tablet- and smartphone-based point-of-care (TSB POC) device against a local laboratory-based measure of HbA1c for monitoring diabetes under real-world conditions.MethodsFor this cross-sectional clinical method applicability study, capillary and venous blood was collected in duplicate and analyzed at local primary health care centers. For a heterogeneity test, the tests were performed by an expert, and by a team of local nurses. The study was conducted in a multicenter design in rural and urban Aceh, Indonesia in 2019, and included a total of 533 adults. We mainly used Bland-Altman plots to assess the number of readings within the 95%-limits of agreement (LoA) and Deming regressions.ResultsThe results show a mean difference between capillary HbA1c on the test device and the reference method of −0.54 [CI0.95 = −1.6933; 0.6048] with 5.21% of measurements outside the LoA and a Pearson’s r = 0.91 in the Deming Regression. There is no significant difference in test concordance between local nurses and the expert (4.23% versus 5.13% results outside the LoA [CI0.95 = −0.0331; 0.0511]).ConclusionsTSB POC for analysis of HbA1c is an acceptable alternative for accessible monitoring of diabetes patients under these conditions. This method could provide access to high-quality diagnostic decisions through regular and cost-effective HbA1c monitoring directly in healthcare facilities, thus providing better access to essential health services.

Performance of ultra-high-frequency ultrasound in the evaluation of skin atrophy in patients with long-term oral glucocorticoid therapy in a tertiary rheumatology center.

Ultra-high-frequency ultrasound (UHF-US) allows visualization of the epidermis, dermis, and subcutis and precise measurement of skin thickness. The aim of this study was to assess the performance and interobserver reliability of UHF-US for measuring skin thickness in patients with long-term systemic glucocorticoid (GC) therapy compared to patients without GC therapy or treated for a shorter period.156 patients with known or suspected inflammatory rheumatic diseases underwent US evaluation for skin thickness by 3 experts in 3 anatomical sites (hand, distal, and proximal forearm). 87 patients were classified as "frequent users" who had received continuous oral GCs for at least one year or at least 3 years with various interruptions. 69 patients without any oral GC therapy in the past or treated for a shorter period were classified as "non-frequent users".UHF-US allowed a precise measurement of skin thickness. Skin thickness at all 3 anatomical sites was significantly decreased in "frequent users" of GCs compared to "non-frequent users" (distal and proximal forearm: p < 0.001; hand: p < 0.05). At all 3 anatomical sites, skin thickness was decreased in patients with clinically assessed parchment-like skin compared to patients without parchment-like skin (distal and proximal forearm: p < 0.001; hand: p < 0.05). Interobserver variability was excellent [hand intraclass correlation coefficient (ICC) = 0.99; proximal forearm ICC = 0.85; distal forearm ICC = 0.84].These data support the idea of UHF-US as an objective and reliable imaging tool for monitoring skin atrophy as adverse effects of GC therapy.

Quantifying the tumour vasculature environment from CD-31 immunohistochemistry images of breast cancer using deep learning based semantic segmentation

BackgroundTumour vascular density assessed from CD-31 immunohistochemistry (IHC) images has previously been shown to have prognostic value in breast cancer. Current methods to measure vascular density, however, are time-consuming, suffer from high inter-observer variability and are limited in describing the complex tumour vasculature morphometry.MethodsWe propose a method for automatically measuring a range of vascular parameters from CD-31 IHC images, which together provide a detailed description of the vasculature morphology. We first used a U-Net based convolutional neural network, trained and validated using 36 partially annotated whole slide images from 27 patients, to segment vessel structures and tumour regions from which the measurements are taken. The model also segments the vascular smooth muscle, benign epithelium, adipose tissue, stroma, lymphocyte clusters, nerves and CD-31 positive leukocytes, and we applied it to an additional 21 images from 15 patients. Using these segmentations, we investigated the relationship between the various tissue types and the vasculature and studied the relationship of various vascular parameters with clinical parameters. We also performed a 3D histology analysis on a separate tumour sample as a proof of principle, providing a more comprehensive visualization of vasculature morphology compared to the standard 2D cross-section of a tissue sample.ResultsUsing two-way cross-validation, we show that vessels were accurately segmented, with Dice scores of 0.875 and 0.856, and were accurately identified, with F1 scores of 0.777 and 0.748. All vascular parameters exhibit strong (r>0.70.7$$\\end{document}]]>) and significant (p<0.001) correlations with measurements taken from the manual ground truth vessel segmentations. A significant relationship between the major/minor axis ratio, a measure of elongation, and the tumour grade was found.ConclusionOur proposed method shows promise as a tool for studying the tumour vasculature and its relationship with surrounding cells and tissue types. Furthermore, the correlation with tumour grade highlights the clinical relevance of our approach. These findings suggest that our method could have substantial implications for improving prognostic assessments and personalizing therapeutic strategies in breast cancer treatment.

Deep learning and radiomics for gastric cancer serosal invasion: automated segmentation and multi-machine learning from two centers

ObjectiveThe objective of this study is to develop an automated method for segmenting spleen computed tomography (CT) images using a deep learning model. This approach is intended to address the limitations of manual segmentation, which is known to be susceptible to inter-observer variability. Subsequently, a prediction model of gastric cancer (GC) serosal invasion was constructed in conjunction with radiomics and deep learning features, and a nomogram was generated to explore the clinical guiding significance.MethodsThis study enrolled 311 patients from two centers with pathologically confirmed of GC. we employed a deep learning model, U-Mamba, to obtain fully automatic segmentation of the spleen CT images. Subsequently, radiomics features and deep learning features were extracted from the entire spleen CT images, and significant features were identified through dimensionality reduction. The clinical features, radiomic features, and deep learning features were organized and integrated, and five machine learning methods were employed to develop 15 predictive models. Ultimately, the model exhibiting superior performance was presented in the form of a nomogram.ResultsA total of 18 radiomics features, 30 deep learning features, and 1 clinical features were deemed valuable. The DLRA model demonstrated superior discriminative capacity relative to other models. A nomogram was constructed based on the logistic clinical model to facilitate the usage and verification of the clinical model.ConclusionRadiomics and deep learning features derived from automated spleen segmentation to construct a nomogram demonstrate efficacy in predicting serosal invasion in GC. Concurrently, fully automated segmentation provides a novel and reproducible approach for radiomics research.

A novel augmented reality-based simulator for enhancing orthopedic surgical training.

Total Hip Arthroplasty (THA) is a well-established and common orthopedic surgery. Due to the complexity involved in THA, orthopedic surgeons require rigorous training. However, the current gold standard, the tutor-guided and -evaluated apprenticeship model is time-consuming, costly, and poses risks to patients. There is a pressing need for additional training resources to enhance the efficiency and safety of the training process. In this work, we present a novel Augmented Reality (AR)-based simulator designed for THA that helps enable a new self-paced training and learning paradigm without the need for instructors. The simulator reduces the need for instructors by integrating an AR guidance module and an automated performance evaluation module. Three types of AR guidance were developed: Overlay, Virtual Twin, and Sectional Views. A feasibility study was conducted with five resident surgeons and two senior surgeons to compare these guidance methods quantitatively and qualitatively. The automated performance evaluation module was assessed against manual performance evaluation using Bland-Altman analysis with limits of agreement (LoA) and Mann-Whitney U tests. The quantitative feasibility results indicate the efficacy of the developed AR guidance, characterized by mean transitional and rotational deviation errors below 3 mm and 3 degrees. Based on the qualitative results, we provide recommendations for efficient AR guidance designs. The Bland-Altman analysis results (0.22±1.32mm with LoA -2.37 to 2.81 mm for distance deviation, 0.94±2.41 degrees with LoA -3.78 to 5.66 degrees for yaw deviation, -0.34±1.30 degrees with LoA -2.90 to 2.22 degrees for pitch deviation) and p-values of Mann-Whitney U tests (0.64 for distance deviation, 0.12 for yaw deviation, 0.11 for pitch deviation) indicate no statistically significant differences between the automated and manual performance evaluation at a significance level of 0.05. This work shows the potential of AR-based simulators in introducing a novel, data-driven approach to open surgery training in orthopedics, enabling surgeons to individually assess and improve their progress.

Repeatability and reproducibility of a new spectral-domain optical coherence tomography biometer and agreement with swept-source optical coherence tomography based biometer

BackgroundTo assess the repeatability and reproducibility of the Colombo IOL biometer (Moptim, China), which utilizes spectral-domain optical coherence tomography (SD-OCT), in measuring ocular parameters of normal subjects and to compare its agreement with the swept-source optical coherence tomography (SS-OCT)-based IOLMaster 700 biometer (Carl Zeiss Meditec AG, Germany).MethodsThis prospective study included 91 eyes from 91 normal subjects. The evaluated parameters were axial length (AL), central corneal thickness (CCT), aqueous depth (AQD), anterior chamber depth (ACD), lens thickness (LT), flattest and steepest meridian keratometry (Kf and Ks), mean keratometry (Km), astigmatism (AST) magnitude, white-to-white (WTW) distance, and pupil diameter (PD). The within-subject standard deviation (Sw), test–retest repeatability (TRT), coefficient of variation (CoV), and intraclass correlation coefficient (ICC) were calculated to determine the repeatability and reproducibility. Paired t-tests and Bland–Altman plots with 95% limits of agreement (LoA) were employed to assess the agreement.ResultsWith respect to intraobserver repeatability, the Sw and TRT values of all evaluated parameters were low. Except J45 and PD, the ICCs were all higher than 0.928. The reproducibility Sw and TRT values of Colombo IOL were also low, and ICCs were not lower than 0.900. Comparing Colombo IOL and IOLMaster 700, the 95% LoA of AL, CCT, AQD, ACD, LT, Kf, Ks, Km, AST, J0, J45, WTW and PD ranged from − 0.08 to 0.03 mm, − 21.58 to 5.09 μm, 0.01 to 0.15 mm, − 0.01 to 0.14 mm, − 0.05 to 0.10 mm, − 0.14 to 0.59 D, − 0.31 to 0.40 D, − 0.13 to 0.40 D, − 0.68 to 0.32 D, − 0.09 to 0.34 D, − 0.07 to 0.25 D, 0.11 to 1.47 mm, and − 0.97 to 2.31 mm, respectively.ConclusionThe new SD-OCT-based Colombo IOL biometer demonstrates excellent repeatability and reproducibility. Moreover, it generally agrees well with the SS-OCT-based IOLMaster 700, except for the WTW and PD measurements.

Dose response relation for optic nerve atrophy at low-dose rate brachytherapy of uveal melanoma.

Dose-response relationships for optic neuropathy and optic nerve atrophy after brachytherapy for posterior uveal melanoma were poorly defined from previous studies. Here, the outcome differences were analyzed in dependence on dosimetric factors, the applicator type, and tumor dependent variables. Primary objective was to evaluate the association of applied dose and on-set of optic nerve atrophy after brachytherapy for posterior uveal melanoma in order to allow risk estimation for new patients. This retrospective study was performed at a single high volume centre for ocular oncology. Patients receiving brachytherapy with Ruthenium-106 applicators for posterior uveal melanoma with a maximum distance between optic nerve and the nearest tumor margin of 4 optical disc diameters and follow-up with fundus photographs were included. The dose distribution at the optic nerve was reconstructed from the fundus photographs at latest follow-up and the dose-distribution of the applicator using a dedicated software. A first mask with important structural elements such as optic nerve, macula, tumor and vessels was first superimposed on the fundus photograph and adapted to the real contours. In a second step, an applicator contour mask was adapted to the radiation scar in order to calculate the dose distribution in all structures. Dose-response relations were obtained by weighted logistic regression. The maximum dose at the optic disc (ODmax) in this group of 109 patients ranged from 5.8 Gy - 242.2 Gy, median 48.7 Gy. Optic nerve atrophy was observed in 29patients. Median time to optic radiation induced optic nerve atrophy was 18 months. Using weighted logistic regression, the dependence of optic nerve atrophy on ODmax was significant (p = 0.0001, chi2 test). There was a considerable interobserver variability in ODmax values (p < 0.02, signed rank test). An additional factor influencing the dose-response was the applicator type (p = 0.0315, chi2 test). The ODmax for a probability of optic nerve atrophy of 50 % (ED50) were 77.6 Gy ± 7.0 Gy for patients treated with notched COB applicators and 53.2 Gy ± 8.2 Gy for patients with other applicators. Including the applicator type, the area under ROC curve reached a value of 0.857 (95 %-CI: 0.793-0.921) for the logistic model with ODmax. The ED50 for optic nerve neuropathy, classified as grade ≥ 1 toxicity, was estimated to be 46.9 Gy ± 4.1 Gy for the maximum dose at the optic disc. Significant dose-response curves were found for optic nerve atrophy at low dose rate brachytherapy. A standard position of COB applicators was identified that allows estimation of the dose-response relation from the scleral dose of the applicator for risk estimation without fundus photographs. This larger data set enhances the knowledge of dose-response relationships for irradiation near the optic nerve.

Impact of different peak tube voltage settings on adrenal adenomas attenuation at unenhanced CT.

To assess the influence of peak tube voltage peak setting on adrenal adenomas (AA) attenuation on unenhanced abdominal CT. IRB-approved retrospective observational cohort study. We included 89 patients with imaging-defined AAs with shortest diameter > 6 mm who underwent two or more unenhanced abdominal CTs using at least two different peak tube voltage settings. Two readers independently measured adenoma attenuation on different CT acquisitions by drawing a round ROI on 3 mm thick axial MPR reconstructions encompassing at least 2/3 of the lesion's surface. The mean of the values measured by the two readers was used for further analysis. Interobserver variability was assessed (Intraclass Correlation Coefficient). Attenuation values measured on 100, 110 and 140 kVp acquisitions were compared with standard 120 kVp ones (Bland-Altman analysis). We included 275 unenhanced abdominal CTs (3.1 ± 0.9/patient) in image analysis; 131 acquired at 120 kVp, 65 at 100 kVp, 59 at 110 kVp, and 20 at 140 kVp. 107 lesions were detected in 89 patients (1-4/patient), with a mean maximum diameter of 17 ± 6 mm. Interobserver agreement in attenuation measurement was excellent (ICC: 0.95, CI (92-97)). Median adenoma attenuation was significantly lower on 100 kVp images than on 120 kVp ones (-1 HU, IQR (-5 to 3.6), vs, 2.5 HU, IQR (-1.5 to 8.5); p < 0.001) whereas we didn't find statistically significant differences in adenoma attenuation between 110 kVp or 140 kVp and 120 kVp ones. AA attenuation is significantly lower on unenhanced CT scans acquired at 100 kVp than on those acquired at "standard" 120 kVp. AA attenuation is significantly lower at 100 kVp in comparison to 120 kVp. This might be exploited to increase unenhanced CT sensitivity in adenoma characterisation, but further studies including non-adenoma lesions are mandatory to confirm this hypothesis. CT scans are often acquired using peak tube voltage settings different from the "standard" 120 kVp. AA attenuation varies if CT scans are acquired using different tube peak voltage settings. At 100 kVp AAs show a significantly lower attenuation than at 120 kVp.

Tumor-infiltrating lymphocytes in HER2-positive breast cancer: potential impact and challenges.

In this review, we evaluate the role of stromal tumor-infiltrating lymphocytes (sTILs) as a biomarker in human epidermal growth factor receptor 2 (HER2)-positive breast cancer, exploring the prognostic and predictive potential in various treatment settings. Data from multiple clinical trials in the early and metastatic settings, focusing on TILs' correlation with pathologic complete response (pCR), progression-free survival (PFS), and overall survival across early and metastatic HER2-positive breast cancer were summarized. This review also discusses TILs' assessment methods, interobserver variability, and emerging technologies to assess TILs. TILs have been identified as a highly reproducible biomarker that predicts pCR in patients receiving neoadjuvant therapy and serves as a prognostic indicator for long-term outcomes in several breast cancer subtypes, including HER2-positive. Studies indicate that higher TIL levels correlate with better recurrence-free survival rates. Despite these findings, there is no consensus on the optimal TIL threshold for clinical decision making, and further research is required on how to incorporate TILs into routine clinical practice. TILs represent a promising biomarker in HER2-positive breast cancer, particularly in early disease settings. This assessment could guide treatment de-escalation or intensification, tailoring therapies to individual patient profiles. Due to their prognostic importance, TILs can be added to pathology reports. However, further validation in clinical trials is essential for the widespread adoption of TILs in clinical practice.

Taming vision transformers for clinical laryngoscopy assessment.

Laryngoscopy, essential for diagnosing laryngeal cancer (LCA), faces challenges due to high inter-observer variability and the reliance on endoscopist expertise. Distinguishing precancerous from early-stage cancerous lesions is particularly challenging, even for experienced practitioners, given their similar appearances. This study aims to enhance laryngoscopic image analysis to improve early screening/detection of cancer or precancerous conditions. We propose MedFormer, a laryngeal cancer classification method based on the Vision Transformer (ViT). To address data scarcity, MedFormer employs a customized transfer learning approach that leverages the representational power of pre-trained transformers. This method enables robust out-of-domain generalization by fine-tuning a minimal set of additional parameters. MedFormer exhibits sensitivity-specificity values of 98%-89% for identifying precancerous lesions (leukoplakia) and 89%-97% for detecting cancer, surpassing CNN counterparts significantly. Additionally, when compared to the two selected ViT-based models, MedFormer also demonstrates superior performance. It also outperforms physician visual evaluations (PVE) in certain scenarios and matches PVE performance in all cases. Visualizations using class activation maps (CAM) and deformable patches demonstrate MedFormer's interpretability, aiding clinicians in understanding the model's predictions. We highlight the potential of visual transformers in clinical laryngoscopic assessments, presenting MedFormer as an effective method for the early detection of laryngeal cancer.

Impact of interobserver variability of echocardiography derived clinical metrics on sample size estimation in clinical trials

Abstract Introduction Clinical metrics obtained from echocardiography usually involve manual analysis, where annotation introduces variability and bias, reducing the reliability of the findings [1]. Traditionally, sample size estimations consider patient variability but often overlook interobserver variability, leading to underpowered and suboptimal studies and an increased risk of clinical study failure [2]. Purpose This study aims to assess the power reduction due to a neglected error sourced in the interobserver variability and its impact on clinical trial size. Methods Interobserver variability was evaluated in an in-house cohort (median age 72, IQR: 52-77; predominantly male, 70%) by recruiting a set of 35 raters, considerably larger than previous studies. Ventricular and atrial endocardium were delineated at end-diastole (ED) and end-systole (ES) in apical 4-chamber (A4C) and 2-chamber (A2C) views, totalling 120 independent annotations per rater. Standard clinical metrics were accordingly derived, and interobserver variability was calculated as the average of the raters’ standard deviations across patients. Population variability was estimated from this cohort and from SHaRe (hypertrophic cardiomyopathy with systolic dysfunction) and Miyazaki (myocardial infarction) registries [3,4]. Sample sizes were estimated using power analysis (two-sided alternative hypothesis assumption) for an independent two-sample t-test, considering a pooled standard deviation combining population and raters’ variability. Results The sample size required for all the evaluated metrics to maintain a given power level after adjusting for the interobserver variability was significantly higher (52.7% average increase). This increase was up to 3-fold for left ventricular ejection fraction. Likewise, the adjusted sample size doubled when considering population variability from the SHaRe and Miyazaki cohorts (Table 1). In the cases where the estimated increase was relatively small (e.g. 5% for left atrium volume), the absolute number of additional patients was still substantial, around several hundred, depending on the expected effect size. The sample size increases exponentially with the inverse of the effect size, and the proportional increase in the adjusted sample size is roughly the same for a wide range of effects (Figure 1). Therefore, reducing the interobserver variability has a major impact on the sample size for any expected effect. Only when the interobserver variability relative to the population one falls below 25% can the effect of the interobserver variability be neglected in the trial size calculations. Conclusion This study evaluates the impact of interobserver variability on trial size estimation. The findings highlight the importance of this source of variability and the elevated risk of underpowered studies when neglected. This ultimately argues for a swift towards automated echocardiography analysis to reduce interobserver variability.

Accurate estimation of interobserver variability in echocardiography segmentation measured from 35 experts and impact on derived clinical metrics

Abstract Introduction The delineation of cardiac structures is crucial to the analysis of echocardiography images and a prerequisite for the calculation of clinical metrics derived from heart chamber areas and volumes. Nevertheless, significant interobserver variability exists due to anatomical variability and echocardiography image quality, leading to discrepancies in endocardial border tracing [1]. Estimations of interobserver variability reported in the literature are based on measurements from a reduced number of annotations per sample [2]. Purpose This work aims to elaborate on previous interobserver variability studies presented in the literature to improve the accuracy and robustness of the estimations by considering (1) a large cohort of annotators and (2) a controlled environment, where the interobserver variability is isolated from other sources of variability, such as image acquisition, cardiac cycle selection or labelling method. Methods A group of 35 experienced sonographers independently delineated the heart chambers at end-systole (ES) and end-diastole (ED) in apical 4-chamber (A4C) and 2-chamber (A2C) views from a cohort of 10 patients (median age 72, IQR: 52-77; predominantly male, 70%), resulting in a total of 120 corresponding annotations per sonographer. The ES and ED frames were preselected by a clinical committee for each sequence to remove this source of variability. Segmentations of each sonographer were used to derive the area of the heart chambers, along with several clinical metrics routinely used to assess cardiac morphology and function. Variability was quantified via relative standard deviation (RSD) and intraclass correlation coefficient (ICC). Results The analysis confirmed a large interobserver variability, with a global RSD of 15.4±5.2% and ICC of 0.89 in area (Figure 1). Variability was significantly higher at ES phase (p&amp;lt;0.001) and in A2C view (p=0.02). No significant differences were found across heart chambers. The annotation variability directly affected the calculation of clinical metrics, leading to particularly large RSD values for those metrics derived from multiple annotations (Figure 2). The variability of left ventricular ejection fraction (LVEF) was 16.0±3.4%, in line with the literature [3]. However, the LVEF ICC was highly compromised (0.187), suggesting that the interobserver variability is larger than the variability of the metric itself. This raises concerns about the use of LVEF derived from manual segmentation as a robust diagnostic reference [4]. These findings argue the need for enhanced standardization protocols and a swift towards automated annotation that mitigate this variability, as evidenced in the EchoNet trial [5]. Conclusion This study accurately evaluates the echocardiography segmentation interobserver variability and its impact on clinical metrics. The findings confirm a large variability and highlight the need to mitigate it, ultimately improving diagnostic assessment.

Deep learning in echocardiography: real-time measurements of left ventricular wall thickness and chamber dimensions in the parasternal long-axis view

Abstract Background Accurate measurements of left ventricular (LV) wall thickness and chamber dimensions in the parasternal long-axis view (PLAX) are essential in the standard echocardiographic examination. Manual assessments introduce variability, and conducting the numerous measurements required for a comprehensive echocardiographic assessment is time-consuming. Therefore, there is potential for improving workflow. Purpose To develop a novel deep learning (DL)-based method for fully automated, real-time measurements of LV wall thickness and chamber dimensions while scanning the patient, and to validate its performance in an unselected prospective cohort. Methods PLAX recordings from 207 subjects were annotated to train DL networks for cardiac segmentation, timing of end-systole and end-diastole, measurements of LV wall thickness and chamber dimensions. The application presents results from three consecutive cardiac cycles in real-time to the operator for review (Figure 1). Agreement and correlation with reference measurements were evaluated, and the combined acquisition and measurement times for both the novel- and reference methods were recorded. Results Feasibility for real-time measurements in PLAX was 98%, with one patient excluded due to poor acoustic window. Agreement with reference measurements was good for interventricular septal diameter (IVSd) (bias - 0.7 mm, limits of agreement (LoA) -4.38 to 2.94 mm), posterior wall thickness (PWd) (bias 0.8 mm, LoA -2.32 to 3.84 mm) and left ventricular internal diastolic diameter (LVIDd) (bias 0.0 mm, LoA -7.0 to 6.9 mm, Figure 2A). Strong correlations were found for IVSd (R=0.84), PWd (R=0.66) and LVIDd (R=0.85), all significant (p &amp;lt;0.001). The average time for acquiring PLAX images was 53 ± 22 seconds, with an additional 42 ± 11 seconds for reference analysis, totaling 94 ± 27 seconds to complete reference PLAX measurements. In contrast, the real-time application took only 59 ± 34 seconds to obtain the same measurements (Figure 2B), resulting in a mean time reduction of 36 seconds (95% CI 24-47 seconds). Conclusion This novel DL-based method for fully automated, real-time analysis of key metrics in PLAX is time-efficient, feasible, and yields measurements in good agreement with manual reference. This method has the potential to substantially improve efficiency and precision of PLAX assessment in clinical practice.

Correlation of right heart catheterization with simultaneous echocardiography for the assessment of pulmonary hypertension in patients with severe tricuspid regurgitation

Abstract Background In severe tricuspid regurgitation (TR), it is important to accurately measure pulmonary artery pressure because pulmonary hypertension is a poor prognostic factor in severe TR. Therefore, we sought to investigate the correlation between PASP using simultaneously measured echocardiography and right heart catheterization (RHC). Method In this observational prospective study, a total of 33 patients with severe TR were enrolled between September 2021 and May 2024. Data was obtained by simultaneously measuring echocardiography and RHC. The correlation between the PASP by echocardiography and RHC was analyzed using Pearson correlation and Bland-Altman analysis. To properly evaluate right atrial pressure (RAP), additional parameters, including systolic flow reversal of the hepatic vein and the V wave cut-off sign, were incorporated. Result PASP by RHC was strongly correlated with PASP by simultaneous measured echocardiography, with a correlation coefficient (r) of 0.87. The bias was 1.5 mmHg, and the limits of agreement (LoA) ranged from -12.5 to 15.5 mmHg. Recategorizations with systolic flow reversal on hepatic vein Doppler and V wave cut-off sign as well as the size and respiratory collapsibility of inferior vena cava provide a more accurate estimate of RAP (eRAP) by echocardiography (p = 0.004). Using this approach for eRAP, a correlation between PASP by RHC and PASP by echocardiography was more prominent (r = 0.91, LoA -9.9 ~ 13.0 mmHg). Conclusion PASP simultaneously measured by echocardiography with systolic flow reversal on hepatic vein Doppler and V wave cut-off sign was highly correlated with PASP by RHC in patients with severe TR.

Epicardial adipose tissue thickness - a new predictive factor for coronary artery disease

Abstract Introduction The role of pericardial adipose tissue has been investigated in numerous studies in relation to the development of coronary artery disease (CAD) and other cardiac pathologies. Total pericardial fat (TPF) consists of epicardial adipose tissue (EAT), which is located between the myocardium and the visceral pericardium, and pericardial fat (PCF), which adheres to the parietal pericardium. The suspected mechanisms are complex and include the acceleration of inflammation and oxidative stress as well as proximity to the coronary arteries. Purpose The aim of the study was to investigate the existence of a correlation between the thickness of epicardial and pericardial adipose tissue and the presence of CAD, with the hope of using this marker in the future for risk stratification of cardiovascular disease. Methods This study enrolled 264 patients with a mean age of 59,9 +/- 10.6 years, 60,6% (160 pts) male, with cardiovascular risk factors who were investigated for suspected CAD. In all patients, the thickness of the TPF, EAT, and the PCF were measured echocardiographically during end-diastole, at the anterior atrioventricular groove. Intra- and interobserver variability was excellent (ICC 0.995-0.998 and ICC 0.992-0.995 for all three layers). Results Out of the total patients included, 54,2% (142 pts) had lesions on conventional angiography, of which 46% had hemodynamically significant lesions requiring revascularization. There were no statistically significant differences in the thickness of the three layers according to the presence of cardiovascular risk factors (diabetes mellitus, smoking, hypertension) and there was no correlation with body mass index. There were statistically significant differences in the dimensions of the pericardial adipose tissue layers between the group of patients with CAD and the group of patients without CAD (EAT 7,1 +/-2,34 vs 5,07+/-2,36 mm, PCF 9.3+/-3,01 vs 7,2+/-2,9 mm and TPF 16,7+/-3,8 vs 12.59+/-4,3 mm, p=0.001 for all cases). There were no statistically significant differences between patients who had one, two, or all three coronary arteries affected (p=0.5). There is a significant correlation between EAT and TPF (r 0.81), PCF and TPF (r 0.83), but no significant correlation between the two substrates (r 0.44). Women without CAD had a lower epicardial layer thickness than men without CAD (4,8+/-0,26 vs 5,4+/-0,29,p &amp;lt;0,001), but this difference did not exist in those with CAD (6.9+/-0.47 vs 7.2+/-0.24). Conclusions The study shows that echocardiographic measurement of the thickness of pericardial adipose tissue layers is a feasible method, easy to perform with good intra- and interobserver reproducibility. The data presented suggest that the thickness of the pericardial adipose tissue layers (EAT and TPF) may represent an independent risk factor for CAD and could be included in the pre-test risk assessment of patients with suspected CAD.

Feasibility of non-invasive echocardiographic estimates of intraventricular pressure differences by blood speckle tracking in adults

Abstract Introduction Non-invasive estimation of the intraventricular pressure difference (IVPD) in early diastole using high-frame-rate ultrasound-based blood speckle tracking (BST) has been reported in pediatric cardiology of various pathologies. This novel echocardiographic technology has recently become available in probes suitable for the adult population. Purpose We tested the feasibility by analyzing intra- and interobserver variability and investigating if IVPD in the left ventricular inflow exhibited the expected differences in controls vs a group of heart failure patients to indicate the ability of this new index to assess diastolic function. Methods Apical BST images of LV inflow were acquired using a GE Vivid E95 ultrasound scanner and the M5Sc probe in healthy volunteers (n=9) and patients with heart failure and left bundle branch block (LBBB) (n=15) implanted with cardiac resynchronization therapy (CRT) 6±2 months (mean±SD) earlier. Echocardiography was performed both with CRT turned on and off. IVPD was analyzed from the BST acquisitions using custom software (Figure 1). Blinded intra- and interobserver variability analysis were performed on the measurements from the healthy controls with the same observer and two other observers, respectively. Results All the controls and 11 of the 15 patients had BST acquisitions eligible for IVPD-analysis, giving a total feasibility of 83%. Intraobserver reliability was excellent (intraclass correlation coefficient 0.96 [0.85, 0.99], mean and 95% CI) and interobserver reliability was good (intraclass correlation coefficient 0.88 [0.59, 0.97] and 0.84 [0.48, 0.96]) respectively. IVPD was highest in controls (-1.7±0.5 mmHg), and higher with CRT on (-1.3±0.7 mmHg, p=NS vs controls) than off (-0.9±0.5 mmHg, p&amp;lt;0.05 compared to both on and controls) in the patients. Conclusions Estimating the intracavitary pressure difference that accelerates blood into the LV during early filling seems feasible and reproducible in adult hearts by BST. IVPD is reduced in patients with heart failure and LBBB, and increases in patients with LBBB when treated with CRT. IVPD based on BST is a potential new echocardiographic method to evaluate diastolic function. Future improvements of the technology are warranted for widespread use.