Limits Of Agreement Research Articles

Clinical applicability of signal heterogeneity and tumor border assessment on T2-weighted MR images to distinguish astrocytic from oligodendroglial origin of gliomas

Background and purposeRadiological features on magnetic resonance imaging (MRI) were attributed to oligodendroglioma, although the diagnostic accuracy in a real-world clinical setting remains partially elusive. This study investigated the accuracy and robustness of tumor heterogeneity and tumor border delineation on T2-weighted MRI to distinguish oligodendroglioma from astrocytoma. Materials and methodsEight readers from three different specialties (radiology, neurology, neurosurgery) with varying levels of experience blindly rated 79 T2-weighted MR images of patients with either oligodendroglioma or astrocytoma. After the first reading session, all readers were re-invited for a second reading session within three weeks. Diagnostic accuracy, including area under the receiver operator characteristics curve (AUC), and intra-observer variability and inter-observer variability were used as outcome measures. ResultsPooled sensitivity and specificity to distinguish oligodendroglioma from astrocytoma for the use of tumor heterogeneity were 59.9 % respectively 74.5 %, and 85.7 % respectively 40.1 % for tumor border. A second reading session did not result in a significant change in sensitivity or specificity for tumor heterogeneity (P = 0.752 and P = 0.733, respectively) or tumor border (P = 0.309 and P = 0.271, respectively). An AUC of 0.825 was achieved with regard to predicting oligodendroglial origin of gliomas. Intra-observer agreement ranged from moderate to very good for tumor heterogeneity (kappa-value 0.43–0.87) and tumor border (0.40–0.84). A moderate inter-oberserver agreement was achieved for tumor heterogeneity and tumor border (kappa-value of 0.50 and 0.45, respectively). ConclusionThis study demonstrates that tumor heterogeneity and tumor borders on T2-weighted MRI could be used with moderate Finter-observer agreement to non-invasively distinguish oligodendroglioma from astrocytoma.

Construction of a Multi-Label Classifier for Extracting Multiple Incident Factors From Medication Incident Reports in Residential Care Facilities: Natural Language Processing Approach.

Medication safety in residential care facilities is a critical concern, particularly when nonmedical staff provide medication assistance. The complex nature of medication-related incidents in these settings, coupled with the psychological impact on health care providers, underscores the need for effective incident analysis and preventive strategies. A thorough understanding of the root causes, typically through incident-report analysis, is essential for mitigating medication-related incidents. We aimed to develop and evaluate a multilabel classifier using natural language processing to identify factors contributing to medication-related incidents using incident report descriptions from residential care facilities, with a focus on incidents involving nonmedical staff. We analyzed 2143 incident reports, comprising 7121 sentences, from residential care facilities in Japan between April 1, 2015, and March 31, 2016. The incident factors were annotated using sentences based on an established organizational factor model and previous research findings. The following 9 factors were defined: procedure adherence, medicine, resident, resident family, nonmedical staff, medical staff, team, environment, and organizational management. To assess the label criteria, 2 researchers with relevant medical knowledge annotated a subset of 50 reports; the interannotator agreement was measured using Cohen κ. The entire data set was subsequently annotated by 1 researcher. Multiple labels were assigned to each sentence. A multilabel classifier was developed using deep learning models, including 2 Bidirectional Encoder Representations From Transformers (BERT)-type models (Tohoku-BERT and a University of Tokyo Hospital BERT pretrained with Japanese clinical text: UTH-BERT) and an Efficiently Learning Encoder That Classifies Token Replacements Accurately (ELECTRA), pretrained on Japanese text. Both sentence- and report-level training were performed; the performance was evaluated by the F1-score and exact match accuracy through 5-fold cross-validation. Among all 7121 sentences, 1167, 694, 2455, 23, 1905, 46, 195, 1104, and 195 included "procedure adherence," "medicine," "resident," "resident family," "nonmedical staff," "medical staff," "team," "environment," and "organizational management," respectively. Owing to limited labels, "resident family" and "medical staff" were omitted from the model development process. The interannotator agreement values were higher than 0.6 for each label. A total of 10, 278, and 1855 reports contained no, 1, and multiple labels, respectively. The models trained using the report data outperformed those trained using sentences, with macro F1-scores of 0.744, 0.675, and 0.735 for Tohoku-BERT, UTH-BERT, and ELECTRA, respectively. The report-trained models also demonstrated better exact match accuracy, with 0.411, 0.389, and 0.399 for Tohoku-BERT, UTH-BERT, and ELECTRA, respectively. Notably, the accuracy was consistent even when the analysis was confined to reports containing multiple labels. The multilabel classifier developed in our study demonstrated potential for identifying various factors associated with medication-related incidents using incident reports from residential care facilities. Thus, this classifier can facilitate prompt analysis of incident factors, thereby contributing to risk management and the development of preventive strategies.

Assessment for acceleration and transformation of chronic lymphocytic leukemia/small lymphocytic lymphoma using histologic and immunohistochemical features: a case series

Morphologic features of aggressive/ “accelerated” chronic lymphocytic leukemia/small lymphocytic lymphoma (aCLL/SLL) have been described. Richter transformation (RT) also occurs in a subset of CLL/SLL cases. This case series examined inter-observer variability when assessing for aCLL/SLL and RT, with attention to how immunohistochemical (IHC) markers may assist in this evaluation. Twelve cases of CLL/SLL with available FFPE tissue were identified. H&E staining and IHC (CD3, CD20, CD5, CD23, LEF1, LAG3, C-MYC, PD-1, MUM1, Cyclin D1, BCL-6, p53, and Ki-67) were performed. Three hematopathologists reviewed each case. The pathologists provided a final interpretation of (1) CLL/SLL, (2) CLL/SLL with expanded and/or confluent proliferation centers or increased Ki-67 (aCLL/SLL), or (3) large cell transformation/DLBCL. The pathologists lacked consensus in the diagnosis in 6/12 cases (50%). The reviewers disagreed on the presence of expanded/confluent proliferation centers in 8/12 cases (67%). With the exception of Ki-67, no IHC marker showed a difference in the staining profile in aCLL/SLL or RT compared to low-grade cases. This series showed inter-observer variability in the evaluation for aCLL/SLL and RT. A study that serially examines genetic alterations in FFPE tissue and correlates the features with histology and IHC, at diagnosis and throughout the disease course, may help refine indicators of aggressive disease.

Multi-rater Prism: Learning self-calibrated medical image segmentation from multiple raters

In medical image segmentation, it is often necessary to collect opinions from multiple experts to make the final decision. This clinical routine helps to mitigate individual bias. However, when data is annotated by multiple experts, standard deep learning models are often not applicable. In this paper, we propose a novel neural network framework called Multi-rater Prism (MrPrism) to learn medical image segmentation from multiple labels. Inspired by iterative half-quadratic optimization, MrPrism combines the task of assigning multi-rater confidences and calibrated segmentation in a recurrent manner. During this process, MrPrism learns inter-observer variability while taking into account the image’s semantic properties and finally converges to a self-calibrated segmentation result reflecting inter-observer agreement. Specifically, we propose Converging Prism (ConP) and Diverging Prism (DivP) to iteratively process the two tasks. ConP learns calibrated segmentation based on multi-rater confidence maps estimated by DivP, and DivP generates multi-rater confidence maps based on segmentation masks estimated by ConP. Experimental results show that the two tasks can mutually improve each other through this recurrent process. The final converged segmentation result of MrPrism outperforms state-of-the-art (SOTA) methods for a wide range of medical image segmentation tasks. The code is available at https://github.com/WuJunde/MrPrism.

Deep Learning-based Modeling for Preclinical Drug Safety Assessment.

In drug development, assessing the toxicity of candidate compounds is crucial for successfully transitioning from preclinical research to early-stage clinical trials. Drug safety is typically assessed using animal models with a manual histopathological examination of tissue sections to characterize the dose-response relationship of the compound - a time-intensive process prone to inter-observer variability and predominantly involving tedious review of cases without abnormalities. Artificial intelligence (AI) methods in pathology hold promise to accelerate this assessment and enhance reproducibility and objectivity. Here, we introduce TRACE, a model designed for toxicologic liver histopathology assessment capable of tackling a range of diagnostic tasks across multiple scales, including situations where labeled data is limited. TRACE was trained on 15 million histopathology images extracted from 46,734 digitized tissue sections from 157 preclinical studies conducted on Rattus norvegicus. We show that TRACE can perform various downstream toxicology tasks spanning histopathological response assessment, lesion severity scoring, morphological retrieval, and automatic dose-response characterization. In an independent reader study, TRACE was evaluated alongside ten board-certified veterinary pathologists and achieved higher concordance with the consensus opinion than the average of the pathologists. Our study represents a substantial leap over existing computational models in toxicology by offering the first framework for accelerating and automating toxicological pathology assessment, promoting significant progress with faster, more consistent, and reliable diagnostic processes.

Understanding interobserver variability of pathologists to improve oral epithelial dysplasia grading.

This study aimed to understand reasons for interobserver variability in the grading of oral epithelial dysplasia (OED) through a survey of pathologists to provide insight for improvements in the reliability and reproducibility of OED diagnoses. The study design included quantitative and qualitative methodology. A pre-validated 31-item questionnaire was distributed to general, head and neck, and oral and maxillofacial histopathology specialists worldwide. A total of 132 pathologists participated and completed the questionnaire. Over two-thirds used the three-tier grading system for OED, while about a third used both binary and three-tier systems. Regular reporters of OED preferred the three-tier system and grading architectural features. Continuing education significantly aided recognition of architectural and cytological changes. Irregular epithelial stratification and drop-shaped rete ridges had the lowest prognostic value and recognition scores, while loss of epithelial cell cohesion had the highest. Most participants used clinical information and often sought a second opinion when grading OED. Our study has found that frequency of OED reporting and attendance of CME/CPD can play an important role in grading OED. Variations in the prognostic value of individual histological features and the use of clinical information may further contribute to interobserver variability.

Continuous non-invasive vs. invasive arterial blood pressure monitoring during neuroradiological procedure: a comparative, prospective, monocentric, observational study

BackgroundIn the perioperative setting, the most accurate way to continuously measure arterial blood pressure (ABP) is using an arterial catheter. Surrogate methods such as finger cuff have been developed to allow non-invasive measurements and are increasingly used, but need further evaluation. The aim of this study is to evaluate the accuracy and clinical concordance between two devices for the measurement of ABP during neuroradiological procedure.MethodsThis is a prospective, monocentric, observational study. All consecutive patients undergoing a neuroradiological procedure were eligible. Patients who needed arterial catheter for blood pressure measurement were included. During neuroradiological procedure, ABP (systolic, mean and diatolic blood pressure) was measured with two different technologies: radial artery catheter and Nexfin. Bland-Altman and error grid analyses were performed to evaluate the accuracy and clinical concordance between devices.ResultsFrom March 2022 to November 2022, we included 50 patients, mostly ASA 3 (60%) and required a cerebral embolization (94%) under general anaesthesia (96%). Error grid analysis showed that 99% of non-invasive ABP measures obtained with the Nexfin were located in the risk zone A or B. However, 65.7% of hypertension events and 41% of hypotensive events were respectively not detected by Nexfin. Compared to the artery catheter, a significant relationship was found for SAP (r2 = 0.78) and MAP (r2 = 0.80) with the Nexfin (p < 0.001). Bias and limits of agreement (LOA) were respectively 9.6 mmHg (− 15.6 to 34.8 mmHg) and − 0.8 mmHg (− 17.2 to 15.6 mmHg), for SAP and MAP.ConclusionsNexfin is not strictly interchangeable with artery catheter for ABP measuring. Further studies are needed to define its clinical use during neuroradiological procedure.Trial registrationClinicaltrials.gov, registration number: NCT05283824.

Optimized target delineation procedure for the radiosurgery treatment of ventricular tachycardia: observer-independent accuracy.

Part of the current stereotactic arrythmia radioablation (STAR) workflow is transfer of findings from the electroanatomic mapping (EAM) to computed tomography (CT). Here, we analyzed inter- and intraobserver variation in a modified EAM-CT registration using automatic registration algorithms designed to yield higher robustness. This work is based on data of 10 patients who had previously undergone STAR. Two observers participated in this study: (1) an electrophysiologist technician (cardiology) with substatial experience in EAM-CT merge, and (2) a clinical engineer (radiotherapy) with minimum experience with EAM-CT merge. EAM-CT merge consists of 3 main steps: segmentation of left ventricle from CT (CT LV), registration of the CT LV and EAM, clinical target volume (CTV) delineation from EAM specific points. Mean Hausdorff distance (MHD), Dice Similarity Coefficient (DSC) and absolute difference in Center of Gravity (CoG) were used to assess intra/interobserver variability. Intraobserver variability: The mean DSC and MHD for 3 CT LVs altogether was 0.92 ± 0.01 and 1.49 ± 0.23 mm. The mean DSC and MHD for 3 CTVs altogether was 0,82 ± 0,06 and 0,71 ± 0,22 mm. Interobserver variability: Segmented CT LVs showed great similarity (mean DSC of 0,91 ± 0,01, MHD of 1,86 ± 0,47 mm). The mean DSC comparing CTVs from both observers was 0,81 ± 0,11 and MHD was 0,87 ± 0,45 mm. The high interobserver similarity of segmented LVs and delineated CTVs confirmed the robustness of the proposed method. Even an inexperienced user can perform a precise EAM-CT merge following workflow instructions.

Comparison of sonographic score and Bishop score in the prediction of successful labor induction in term patients: A prospective observational study

BackgroundThe cervical favorability for IOL is assessed by scoring systems. There are various scoring systems, including the Field system, Lange pelvic scoring system, and Bishop scoring system, which was then modified and termed ‘The modified Bishop scoring system, and it is the most widely used one. This scoring is done based on the assessment of the station of the fetal presenting part, consistency, position, dilatation, and length of the maternal cervix. Owing to its subjective assessment, there is a high chance of inter-observer variability. Therefore, radiological methods such as Trans-vaginal ultrasonogram (TVS) and Trans-abdominal ultrasonogram (TAS) are preferred over manual methods. This study aims to determine the role of assessment of the cervix by Modified Bishop Score and assessment of the cervical parameters by Transvaginal ultrasonogram (TVS) and compare them in predicting the outcome of labour. Material and methodsThis prospective observational study was conducted by the Department of Obstetrics and Gynecology in a tertiary care hospital located at Chengalpattu, Tamil Nadu, in a term singleton pregnancy woman, and the Study period was from January 2023 to July 2023 for six months. In total, one hundred fifty pregnant women were recruited. The study inclusion criteria were singleton pregnant women with intact amniotic membranes and a period of gestation >37 weeks with cephalic presentation. The findings of TVS were blinded to the person who measured cervical parameters for the bishop score. IOL done for both favorable and unfavorable cervix according to the guidelines given by FOGSI-ICOG (2018) for IOL. ResultsReceiver operating characteristic (ROC) curve analysis showed the area under the curve (AUC) for the TVS score was 0.705(95 % CI: 0.616–0.794) against the standard bishop score. TVS Score≥6.5 cm is the best cut-off value to ascertain successful IOL with a sensitivity of 99 % and specificity of 94 % among singleton pregnancy women. ConclusionThis prospective observational study deduced that TVS scoring is a valid alternative tool against the standard bishop score in predicting cervical favorability for IOL among singleton pregnant women with plausible sensitivity and specificity. TVS scoring for assessing cervical favorability acts as an outstanding marker in predicting the IOL with a cut-off value of ≥6.5 cm in singleton pregnant women.

Diagnostic Efficacy of Dynamic Maneuver in Contrast-Enhanced Computed Tomography Compared With Conventional Contrast-Enhanced Computed Tomography in Imaging the Neck Region.

Introduction Dynamic contrast-enhanced computed tomography (DCE-CT) and conventional contrast-enhanced computed tomography (CE-CT) are widely used to evaluate neck lesions, including lymph node metastases, thyroid nodules, salivary gland tumors, and other soft tissue masses. DCE-CT, which captures multiple phases of contrast enhancement over time, is hypothesized to provide superior diagnostic accuracy compared to the single-phase images obtained by CE-CT due to its ability to offer dynamic information about tissue perfusion, blood volume, and vascular permeability. Methods This retrospective observational diagnostic study included 100 patients who underwent neck imaging, divided equally into DCE-CT and CE-CT groups. Patient demographics (age, gender, body mass index) and lesion characteristics (type, location, size, enhancement pattern, margins) were recorded. Diagnostic performance metrics (sensitivity, specificity, accuracy, positive predictive value, negative predictive value) were evaluated alongside inter-observer variability using the kappa statistic. Clinical impact was assessed based on changes in treatment plans and improvements in patient outcomes. The radiation dose for each modality was documented. Statistical analysis was performed using SPSS software (IBM SPSS Statistics for Windows, IBM Corp., Armonk, NY) with chi-square tests for categorical variables and t-tests for continuous variables. Results The study included 58 males and 42 females with a mean age of 55.5 years. A total of 145 lesions were detected: 75 by DCE-CT and 70 by CE-CT. DCE-CT demonstrated higher sensitivity (93.33%) and specificity (96.00%) compared to CE-CT (sensitivity 86.67%, specificity 92.00%). The accuracy of DCE-CT was 94.00% versus 88.00% for CE-CT. Inter-observer agreement was higher for DCE-CT (kappa = 0.85) compared to CE-CT (kappa = 0.80). DCE-CT led to treatment plan changes in 40% of cases and resulted in a 75% improvement in outcomes compared to 25% and 60%, respectively, for CE-CT. The mean radiation dose was slightly higher for DCE-CT (8.5 mSv) compared to CE-CT (7.0 mSv). Conclusion DCE-CT offers superior diagnostic efficacy compared to CE-CT for imaging neck lesions with enhanced sensitivity, specificity, and accuracy. Its ability to capture multiple phases of contrast enhancement allows for detailed lesion characterization and provides crucial quantitative data on tissue perfusion and blood volume. These benefits lead to more frequent improvements in patient outcomes and changes in treatment plans. Despite the slightly higher radiation dose, the diagnostic advantages of DCE-CT outweigh the disadvantages, particularly in complex cases requiring detailed lesion analysis. Further prospective studies are recommended to validate these findings and explore the broader clinical benefits of DCE-CT.

Multicenter Study on Tumor Budding in Lung Squamous Cell Carcinoma: Comparison Between Biopsy and Resection With Interobserver Variability Assessment

Grading lung squamous cell carcinoma (LUSC) is controversial and not universally accepted. The histomorphologic feature of tumor budding (TB) is an established independent prognostic factor in colorectal cancer, and its importance is growing in other solid cancers, making it a candidate for inclusion in tumor grading schemes. We aimed to compare TB between preoperative biopsies and resection specimens in pulmonary squamous cell carcinoma and assess interobserver variability. A retrospective cohort of 249 consecutive patients primarily resected with LUSC in Bern (2000-2013, n = 136) and Lausanne (2005-2020, n = 113) with available preoperative biopsies was analyzed for TB and additional histomorphologic parameters, such as spread through airspaces and desmoplasia, by 2 expert pathologists (M.M., C.N.). Results were correlated with clinicopathologic parameters and survival. In resection specimens, peritumoral budding (PTB) score was low (0-4 buds/0.785 mm2) in 47.6%, intermediate (5-9 buds/0.785 mm2) in 27.4%, and high (≥10 buds/0.785 mm2) in 25% of cases (median bud count, 5; IQR, 0-26). Both the absolute number of buds and TB score were similar when comparing tumor edge and intratumoral zone (P = .192) but significantly different from the score obtained in the biopsy (P < .001). Interobserver variability was moderate, regardless of score location (Cohen kappa, 0.59). The discrepant cases were reassessed, and consensus was reached in all cases with identification of causes of discordance. TB score was significantly associated with stage (P = .002), presence of lymph node (P = .033), and distant metastases (P = .020), without significant correlation with overall survival, tumor size, or pleural invasion. Desmoplasia was significantly associated with higher PTB (P < .001). Spread through airspaces was present in 34% and associated with lower PTB (P < .001). To conclude, despite confirming TB as a reproducible factor in LUSC, we disclose areas of scoring ambiguity. Preoperative biopsy evaluation was insufficient in establishing the final TB score of the resected tumor.

Evaluation of artificial intelligence-assisted morphological analysis for platelet count estimation.

This study aims to assess the performance of the platelet count estimation using artificial intelligence technology on the MC-80 digital morphology analyzer. Digital morphology analyzer uses two different computational principles for platelet count estimation: based on PLT/RBC ratio (PLT-M1) and estimate factor (PLT-M2). 977 samples with various platelet counts (low, median, and high) were collected. Out of these, 271 samples were immunoassayed using CD61 and CD41 antibodies. The platelet counts obtained from the hematology analyzer (PLT-I and PLT-O), digital morphology analyzer (PLT-M1 and PLT-M2), and flow cytometry (PLT-IRM) were compared. There was no significant deviation observed before and after verification for both PLT-M1 and PLT-M2 across the analysis range (average bias: -0.845/-0.682, 95% limit of agreement (LOA): -28.675-26.985/-29.420-28.056). When platelet alarms appeared, PLT-M1/PLT-M2 showed the strongest correlation with PLT-IRM than PLT-I with PLT-IRM (r: 0.9814/0.9796 > 0.9601). The correlation between PLT-M1/PLT-M2 and PLT-IRM was strong for samples with interference, such as large platelets or RBC fragments, but relatively weak in small RBCs. The deviation between PLT-M1 and PLT-M2 is related to the number of RBCs. Compared with PLT-I, PLT-M1/PLT-M2 showed higher accuracy for platelet transfusion decisions, especially for samples with low-value PLT. The novel platelet count estimation on the MC-80 digital morphology analyzer provides high accuracy, especially the reviewed result, which can effectively confirm suspicious platelet count.

Comparison of Left Ventricular Function Derived from Subject-Specific Inverse Finite Element Modeling Based on 3D ECHO and Magnetic Resonance Images.

Three-dimensional echocardiography (3D ECHO) and magnetic resonance (MR) imaging are frequently used in patients and animals to evaluate heart functions. Inverse finite element (FE) modeling is increasingly applied to MR images to quantify left ventricular (LV) function and estimate myocardial contractility and other cardiac biomarkers. It remains unclear, however, as to whether myocardial contractility derived from the inverse FE model based on 3D ECHO images is comparable to that derived from MR images. To address this issue, we developed a subject-specific inverse FE model based on 3D ECHO and MR images acquired from seven healthy swine models to investigate if there are differences in myocardial contractility and LV geometrical features derived using these two imaging modalities. We showed that end-systolic and end-diastolic volumes derived from 3D ECHO images are comparable to those derived from MR images (R2=0.805 and 0.969, respectively). As a result, ejection fraction from 3D ECHO and MR images are linearly correlated (R2=0.977) with the limit of agreement (LOA) ranging from -17.95% to 45.89%. Using an inverse FE modeling to fit pressure and volume waveforms in subject-specific LV geometry reconstructed from 3D ECHO and MR images, we found that myocardial contractility derived from these two imaging modalities are linearly correlated with an R2 value of 0.989, a gradient of 0.895, and LOA ranging from -6.11% to 36.66%. This finding supports using 3D ECHO images in image-based inverse FE modeling to estimate myocardial contractility.

Machine Learning Analysis in Diffusion Kurtosis Imaging for Discriminating Pediatric Posterior Fossa Tumors: A Repeatability and Accuracy Pilot Study.

Background and purpose: Differentiating pediatric posterior fossa (PF) tumors such as medulloblastoma (MB), ependymoma (EP), and pilocytic astrocytoma (PA) remains relevant, because of important treatment and prognostic implications. Diffusion kurtosis imaging (DKI) has not yet been investigated for discrimination of pediatric PF tumors. Estimating diffusion values from whole-tumor-based (VOI) segmentations may improve diffusion measurement repeatability compared to conventional region-of-interest (ROI) approaches. Our purpose was to compare repeatability between ROI and VOI DKI-derived diffusion measurements and assess DKI accuracy in discriminating among pediatric PF tumors. Materials and methods: We retrospectively analyzed 34 children (M, F, mean age 7.48 years) with PF tumors who underwent preoperative examination on a 3 Tesla magnet, including DKI. For each patient, two neuroradiologists independently segmented the whole solid tumor, the ROI of the area of maximum tumor diameter, and a small 5 mm ROI. The automated analysis pipeline included inter-observer variability, statistical, and machine learning (ML) analyses. We evaluated inter-observer variability with coefficient of variation (COV) and Bland-Altman plots. We estimated DKI metrics accuracy in discriminating among tumor histology with MANOVA analysis. In order to account for class imbalances, we applied SMOTE to balance the dataset. Finally, we performed a Random Forest (RF) machine learning classification analysis based on all DKI metrics from the SMOTE dataset by partitioning 70/30 the training and testing cohort. Results: Tumor histology included medulloblastoma (15), pilocytic astrocytoma (14), and ependymoma (5). VOI-based measurements presented lower variability than ROI-based measurements across all DKI metrics and were used for the analysis. DKI-derived metrics could accurately discriminate between tumor subtypes (Pillai's trace: p < 0.001). SMOTE generated 11 synthetic observations (10 EP and 1 PA), resulting in a balanced dataset with 45 instances (34 original and 11 synthetic). ML analysis yielded an accuracy of 0.928, which correctly predicted all but one lesion in the testing set. Conclusions: VOI-based measurements presented improved repeatability compared to ROI-based measurements across all diffusion metrics. An ML classification algorithm resulted accurate in discriminating PF tumors on a SMOTE-generated dataset. ML techniques based on DKI-derived metrics are useful for the discrimination of pediatric PF tumors.

Deep learning-based measurement of split glomerular filtration rate with 99mTc-diethylenetriamine pentaacetic acid renal scan

PurposeTo develop a deep learning (DL) model for generating automated regions of interest (ROIs) on 99mTc-diethylenetriamine pentaacetic acid (DTPA) renal scans for glomerular filtration rate (GFR) measurement.MethodsManually-drawn ROIs retrieved from a Picture Archiving and Communications System were used as ground-truth (GT) labels. A two-dimensional U-Net convolutional neural network architecture with multichannel input was trained to generate DL ROIs. The agreement between GFR values from GT and DL ROIs was evaluated using Lin’s concordance correlation coefficient (CCC) and slope coefficients for linear regression analyses. Bias and 95% limits of agreement (LOA) were assessed using Bland-Altman plots.ResultsA total of 24,364 scans (12,822 patients) were included. Excellent concordance between GT and DL GFR was found for left (CCC 0.982, 95% confidence interval [CI] 0.981–0.982; slope 1.004, 95% CI 1.003–1.004), right (CCC 0.969, 95% CI 0.968–0.969; slope 0.954, 95% CI 0.953–0.955) and both kidneys (CCC 0.978, 95% CI 0.978–0.979; slope 0.979, 95% CI 0.978–0.979). Bland-Altman analysis revealed minimal bias between GT and DL GFR, with mean differences of − 0.2 (95% LOA − 4.4–4.0), 1.4 (95% LOA − 3.5–6.3) and 1.2 (95% LOA − 6.5–8.8) mL/min/1.73 m² for left, right and both kidneys, respectively. Notably, 19,960 scans (81.9%) showed an absolute difference in GFR of less than 5 mL/min/1.73 m².ConclusionOur DL model exhibited excellent performance in the generation of ROIs on 99mTc-DTPA renal scans. This automated approach could potentially reduce manual effort and enhance the precision of GFR measurement in clinical practice.

Real-time quantification of laser speckle contrast imaging during intestinal laparoscopic surgery: successful demonstration in a porcine intestinal ischemia model

BackgroundAnastomotic leakage (AL) is a dreaded complication following colorectal cancer surgery, impacting patient outcome and leads to increasing healthcare consumption as well as economic burden. Bowel perfusion is a significant modifiable factor for anastomotic healing and thus crucial for reducing AL.AimsThe study aimed to calculate a cut-off value for quantified laser speckle perfusion units (LSPUs) in order to differentiate between ischemic and well-perfused tissue and to assess inter-observer reliability.MethodsLSCI was performed using a porcine ischemic small bowel loop model with the PerfusiX-Imaging® system. An ischemic area, a well-perfused area, and watershed areas, were selected based on the LSCI colormap. Subsequently, local capillary lactate (LCL) levels were measured. A logarithmic curve estimation tested the correlation between LSPU and LCL levels. A cut-off value for LSPU and lactate was calculated, based on anatomically ischemic and well-perfused tissue. Inter-observer variability analysis was performed with 10 observers.ResultsDirectly after ligation of the mesenteric arteries, differences in LSPU values between ischemic and well-perfused tissue were significant (p < 0.001) and increased significantly throughout all following measurements. LCL levels were significantly different (p < 0.001) at both 60 and 120 min. Logarithmic curve estimation showed an R2 value of 0.56 between LSPU and LCL values. A LSPU cut-off value was determined at 69, with a sensitivity of 0.94 and specificity of 0.87. A LCL cut-off value of 3.8 mmol/L was found, with a sensitivity and specificity of 0.97 and 1.0, respectively. There was no difference in assessment between experienced and unexperienced observers. Cohen’s Kappa values were moderate to good (0.52–0.66).ConclusionReal-time quantification of LSPUs may be a feasible intraoperative method to assess tissue perfusion and a cut-off value could be determined with high sensitivity and specificity. Inter-observer variability was moderate to good, irrespective of prior experience with the technique.

Comparison of axial length and anterior segment parameters of patients with myopia measured using 2 fourier-domain optical coherent biometry devices

BackgroundThis study assessed the agreement of ocular parameters of patients with myopia measured using Colombo intraocular lens (IOL) 2 and IOLMaster 700.MethodsEighty patients (male, 22; average age, 29.14 ± 7.36 years) with myopia (159 eyes) were included in this study in May 2023. The participants’ axial length (AXL), central corneal thickness (CCT), lens thickness (LT), white-to-white distance (WTW), front flat (K1), steep (K2), mean (Km) corneal keratometry, astigmatism (Astig), J0 vector, and J45 vector were measured using the IOLMaster 700 and Colombo IOL 2. The measurements from both devices were compared using the generalized estimating equation, correlation analysis, and Bland-Altman plots.ResultsWith the Colombo IOL 2, lower values for K2 and J0 (odds ratio [OR] = 0.587, p = 0.033; OR = 0.779, p < 0.0001, respectively), and larger values for WTW, Astig, and J45 (OR = 1.277, OR = 1.482, OR = 1.1, all p < 0.0001) were obtained. All ocular measurements by both instruments showed positive correlations, with AXL demonstrating the strongest correlation (r = 0.9996, p < 0.0001). The intraclass correlation coefficients for AXL and CCT measured by both instruments was 0.999 and 0.988 (both p < 0.0001), and Bland-Altman plot showed 95% limits of agreement (LoA) of -0.078 to 0.11 mm and − 9.989 to 13.486 μm, respectively. The maximum absolute 95% LoA for LT, WTW, K1, K2, and J0 were relatively high, achieving 0.829 mm, 0.717 mm, 0.983 D, 0.948 D, and 0.632 D, respectively.ConclusionsIn young patients with myopia, CCT and AXL measurements obtained with the Colombo IOL 2 and IOLMaster 700 were comparable. However, WTW, LT, corneal refractive power, and astigmatism values could not be used interchangeably in clinical practice.

Subglottic dilatation in extremely preterm infants on prolonged mechanical ventilation

BackgroundAirway injuries are reported among preterm infants with bronchopulmonary dysplasia. We hypothesized that prolonged ventilation in preterm infants is associated with subglottic dilatation that can be reliably evaluated by point of care ultrasonography (POCUS). MethodsAll preterm infants (<29-weeks) admitted to the neonatal ICU at the Advent-Health from January-2020 to June-2022 were eligible if they required invasive ventilation for ≤7 days in the first 28 days of life (control) or remained intubated for ≥28 days (prolonged ventilation). Sonography was performed by one technician and all images were reviewed by the pediatric radiologist. The trachea size was measured 3 times by randomly selecting three images. The first 20 scans were also independently reported by a different pediatric radiologist. Intra and inter-observer variability was estimated. Mean trachea size and weight at the time of imaging were compared. ResultsOut of 417 eligible infants; 11 died before 28 days and 163 required ventilation for 8–27 days. Consent missed for 80 infants during COVID-19 pandemic. We enrolled 23 and 28 infants in the control & prolonged ventilation groups, respectively. Inter and intra-observer correlations were 0.83 and 0.97 respectively. Infants in the control group had higher gestation and birth weight. Infants on prolonged ventilation were at higher risk for infections, BPD, longer hospital stay and significant subglottic dilation (4.51 ± 0.04 vs 4.17 ± 0.02 mm, p < 0.01) despite smaller body weight at the time of imaging (884 ± 102 vs 1059 ± 123g, p < 0.01). ConclusionExtremely preterm infants on prolonged ventilation are at risk for sub-glottic dilatation that can be reliably measured by POCUS.

Impact of the adjunction of a short video to an original article for the recognition of newly described tumor entities in pathology: An interventional prospective study.

Merkel cell carcinoma diagnosis is often based on microscopic examination by pathologists. While histopathologic diagnosis primarily hinges on conscious and analytical cognition, the pathologist's decision-making process is also influenced by a rapid "gist" or "gestalt" approach. In this study, using cases of Merkel cell carcinoma as a model, we aim to assess how pathologists' viewing short videos containing conceptual clues and visual aids, in conjunction with reading an original article as a reference, may enhance their diagnostic performance. Sixteen pathologists were included in the present work. After participants had read the original article, their ability to distinguish Merkel cell polyomavirus (MCPyV)+ and MCPyV- Merkel cell carcinoma cases was evaluated on a first preliminary series of 20 cases. Following this test, the participants watched the video and then evaluated a second "experimental" series of 20 independent cases. After reading the original article, for each case, a median number of 12 participants (75%, Q1-Q3: 10-13) classified the specimen in the correct category (92 incorrect answers in the whole series). An important interobserver variability was observed in this setting (Kappa coefficient = 0.465). By contrast, following the video, all cases were correctly classified by most of the participants, with only 12 incorrect answers on the whole series and excellent interobserver reproducibility (Kappa coefficient = 0.846). Our study demonstrated that providing a short video together with an original article may enhance pathologists' performance in diagnosing Merkel cell carcinoma.

Commercially available activity monitors such as the fitbit charge and apple watch show poor validity in patients with gait aids after total knee arthroplasty

PurposeThe aim of this study is to determine the validity of consumer grade step counter devices during the early recovery period after knee replacement surgery.MethodsTwenty-three participants wore a Fitbit Charge or Apple Watch Series 4 smart watch and performed a walking test along a 50-metre hallway. There were 9 males and 14 females included in the study with an average age of 68.5 years and BMI of 32. Each patient wore both the Fitbit Charge and Apple Watch while completing the walking test and an observer counted the ground truth value using a thumb-push tally counter. This test was repeated pre-operatively with no gait aid, immediately post operatively with a walker, at 6 weeks follow up with a cane and at 6 months with no gait aid. Bland-Altman plots were performed for all walking tests to compare the agreement between measurement techniques.ResultsMean overall agreement of step count for pre-operative and at 6 months for subjects walking without gait aids was excellent for both the Apple Watch vs. actual and Fitbit vs. actual with bias values ranging from − 0.87 to 1.36 with limits of agreement (LOA) ranging between − 10.82 and 15.91. While using a walker both devices showed extremely little agreement with the actual step count with bias values between 22.5 and 24.37 with LOA between 11.7 and 33.3. At 6 weeks post-op while using a cane, both the Apple Watch and Fitbit devices had a range of bias values between − 2.8 and 5.73 with LOA between − 13.51 and 24.97.ConclusionsThese devices show poor validity in the early post operative setting, especially with the use of gait aids, and therefore results should be interpreted with caution.