Cohen's Kappa Research Articles

CNN-Based Integrated Framework for Enhanced Diabetic Retinopathy Detection

Background: Diabetic Retinopathy (DR), a significant cause of vision loss globally, is characterized by retinal damage caused by diabetes. Early detection is vital to prevent irreversible blindness, yet challenges remain in accurately identifying DR stages and enhancing blood vessel visibility in fundus images. This paper aims to develop an early detection methodology for DR, addressing the need for early diagnosis and the difficulties in distinguishing DR severity through fundus imaging. The challenges in the early detection of Diabetic Retinopathy (DR) and enhancing blood vessel visibility in fundus images are multifaceted, including issues such as data imbalance, image noise, and complex patterns. By addressing these challenges through advanced ML techniques and image processing methodologies, the proposed methodology in the paper aims to overcome the limitations in early detection and severity assessment of DR, contributing to improved patient outcomes and vision preservation. Methods: This study utilizes Machine Learning (ML) to analyze complex patterns in fundus color images of DR, employing spatial domain filtering to reduce image noise and address data imbalances across DR severity levels through data augmentation. A Convolutional Neural Network (CNN), enhanced with a Gabor filter, is applied for stage-specific DR detection and to pinpoint infected areas. The dataset includes 1000 color fundus images, with a 70:30 split for training and testing, respectively. The adoption of a Gabor filtering technique aims to refine the model’s performance further. Results: The incorporation of a CNN with a Gabor filter has shown outstanding efficacy in detecting DR from fundus color images, achieving a training accuracy of 97.5%, validation accuracy of 96.5%, Cohen kappa score of 89.76%, and testing accuracy of 95.87%. This method effectively illustrates the disease-affected areas in the fundus images provided. Conclusion: Through a comparative analysis of image processing techniques, this research highlights the advantages of using advanced DR analysis for image preprocessing. The proposed CNN-Gabor filter approach demonstrates significant success in identifying diabetic retinopathy in fundus color photographs and accurately delineating the affected regions, contributing valuable insights to the field of medical image processing.

Reliability of a Tailored Tele-Medicine Model for Otorhinolaryngology-Head and Neck Surgery (ORL-HNS) Clinical Practice in Remote Areas of Nepal: A Pilot Study from Tertiary Referral Center.

Introduction: Despite the ever-expanding scope of telemedicine, studies regarding its application in the field of otorhinolaryngology-head and neck surgery (ORL-HNS) in remote parts of the world are lacking. This pilot study aims to bridge the research gap and to assess the efficacy of our tailored telemedicine clinic model, operating in an "asynchronous" mode, in diagnosing various ORL-HNS diseases. Methods: This is a cross-sectional study conducted at a tertiary care center of Nepal. An asynchronous telemedicine portal was used with a store and share approach to telecast the pen endoscopic camera findings to the consultant in order to formulate a diagnosis and a management plan. Interobserver reliability was assessed using SPSS version 26 based on Cohen's kappa. The association between the results of the Mimi Hearing App (v. 5.0) and pure tone audiometry were performed using chi-square test. Results: The overall diagnostic agreement in our study was substantial to almost perfect with kappa values of .843, .784, .737, and .764 for rhinology, head and neck, otology, and pediatric cases. Chronic otitis media was the predominant diagnosis across all age groups. Various head and neck swellings, symptomatic deviated nasal septum and chronic rhinosinusitis with polyposis followed closely. Discussion: The future of telemedicine in otorhinolaryngology can be gauged to be effective in reducing cost and human resource use in the limited setting of Nepal with the model allowing for remote but effective communication between the concerned patients and the expert doctor in clarifying doubts and making diagnosis, early recognition of diseases and preventing or decreasing complications of those diseases.

Evaluation of the reliability and accuracy of MRI for the diagnosis of meniscal ramp lesions.

Meniscal ramp lesions are a special type of meniscal injury that affects the periphery of the posterior horn of the medial meniscus and/or its meniscocapsular attachments, strongly associated with anterior cruciate ligament (ACL) tears. Due to their location, these lesions can be missed arthroscopically so it is essential to diagnose them on preoperative magnetic resonance imaging (MRI). To evaluate the accuracy of MRI in detecting meniscal ramp lesions in patients with ACL tears using arthroscopy as the reference standard. Two musculoskeletal radiologists, blinded to the surgical findings, retrospectively and independently evaluated 106 knee MRI scans for the presence of meniscal ramp lesions in non-consecutive patients who underwent arthroscopic ACL reconstruction between January 2019 and July 2022 by a single surgeon at one institution. Having arthroscopy as reference, the diagnostic sensitivity and specificity as well as the positive and negative predictive values (PPV/NPV) of the MRI scans were calculated. Cohen's kappa coefficient was used to test inter-observer reliability. A P value <0.05 was considered statistically significant. In the study group of 106 patients (72 men, 34 women; mean age = 33.84 ±13.12 years), 76 had an arthroscopy-confirmed meniscal ramp lesion, while 30 did not. The sensitivity and specificity of MRI for the detection of meniscal ramp lesion were 88% and 87%, respectively. The PPV and NPV were 94% and 74%, respectively. Inter-rater reliability was excellent (k = 0915). This study demonstrates that MRI can accurately detect meniscal ramp lesions.

Designing an artificial intelligence-powered video assistant referee system for team sports using computer vision

This paper investigates the efficacy of an AI-powered Video Assistant Referee (VAR) system in enhancing officiating accuracy, efficiency, and consistency in team sports. Employing a combination of artificial intelligence and computer vision, the system was tested in a local championship in Almaty, involving eight football teams. Through the analysis of decision-making accuracy, time efficiency, and consistency across officiating scenarios, the study employed chi-squared tests, paired t-tests, and Cohen's Kappa statistics to quantitatively assess improvements over traditional VAR systems. Results indicated that the AI-powered VAR system significantly increased the accuracy of decisions and reduced the decision-making time, thereby maintaining the fluidity of gameplay. Although the system also demonstrated enhanced consistency in officiating decisions, it highlighted areas needing further refinement to handle complex game situations effectively. The findings suggest that AI integration into sports officiating can substantially benefit the fairness and dynamics of team sports, provided that ongoing technological advancements continue to address current limitations. This study contributes to the growing body of knowledge on the intersection of technology and sports, offering a framework for future enhancements in digital officiating systems.

Agreement between the EWGSOP2 and SDOC consensuses for sarcopenia in patients receiving hemodialysis: Findings of a cross sectional analysis from the SARC-HD study.

Differences in definitions and operational diagnoses for sarcopenia create difficulties in understanding the epidemiology of the disease. We examined the prevalences of sarcopenia using the revised European Working Group on Sarcopenia in Older People (EWGSOP2) and the Sarcopenia Definitions and Outcomes Consortium (SDOC) consensuses and analyzed their level of agreement in patients receiving hemodialysis. Data from the SARCopenia trajectories and associations with clinical outcomes in patients receiving hemodialysis (SARC-HD) multicenter study in Brazil were analyzed. Muscle strength was assessed using handgrip strength, muscle mass by calf circumference, and physical performance by the 4-m gait speed test. Sarcopenia was diagnosed according to both the EWGSOP2 (low muscle strength plus low muscle mass) and the SDOC (low muscle strength plus low physical performance). The Cohen kappa statistic was used to determine the level of agreement between the consensuses. 838 patients (57.8 ± 15.0 years; 61% men) from 19 dialysis units were included. We found similar prevalences of sarcopenia between the consensuses (EWGSOP2, n = 128, 15.3%; SDOC, n = 105, 12.5%) but with weak agreement (50 of 233 patients, 21.5%; κ = 0.34, 95% CI 0.25-0.43). Agreement was also weak within age categories (≥60 years, κ = 0.34; <60 years, κ = 0.15; both P < 0.001). Of the 51 patients diagnosed by the EWGSOP2 criterion as having severe sarcopenia, all but 1 (98.0%) met the SDOC criterion for sarcopenia (κ = 0.61, 95% CI 0.52-0.70). Low muscle strength was more frequently diagnosed using the SDOC than with the EWGSOP2 (52.3% vs 25.9%). We found a weak agreement between the EWGSOP2 and SDOC consensuses for the diagnosis of sarcopenia in patients receiving hemodialysis. Although still weak, agreement was marginally better for older patients. These findings highlight the importance of a global and standardized conceptual diagnosis of sarcopenia.

Diagnostic serology test comparison for Q fever and Rift Valley fever in humans and livestock from pastoral communities.

Q fever (QF) and Rift Valley fever (RVF) are endemic zoonotic diseases in African countries, causing significant health and economic burdens. Accurate prevalence estimates, crucial for disease control, rely on robust diagnostic tests. While enzyme-linked immunosorbent assays (ELISA) are not the gold standard, they offer rapid, cost-effective, and practical alternatives. However, varying results from different tests and laboratories can complicate comparing epidemiological studies. This study aimed to assess the agreement of test results for QF and RVF in humans and livestock across different laboratory conditions and, for humans, different types of diagnostic tests. We measured inter-laboratory agreement using concordance, Cohen's kappa, and prevalence and bias-adjusted kappa (PABAK) on 91 human and 102 livestock samples collected from rural regions in Chad. The serum aliquots were tested using ELISA in Chad, and indirect immunofluorescence assay (IFA) (for human QF and RVF) and ELISA (for livestock QF and RVF) in Switzerland and Germany. Additionally, we examined demographic factors influencing test agreement, including district, setting (village vs. camp), sex, age, and livestock species of the sampled individuals. The inter-laboratory agreement ranged from fair to moderate. For humans, QF concordance was 62.5%, Cohen's kappa was 0.31, RVF concordance was 81.1%, and Cohen's kappa was 0.52. For livestock, QF concordance was 92.3%, Cohen's kappa was 0.59, RVF concordance was 94.0%, and Cohen's kappa was 0.59. Multivariable analysis revealed that QF test agreement is significantly higher in younger humans and people living in villages compared to camps and tends to be higher in livestock from Danamadji compared to Yao, and in small ruminants compared to cattle. Additionally, RVF agreement was found to be higher in younger humans. Our findings emphasize the need to consider sample conditions, test performance, and influencing factors when conducting and interpreting epidemiological seroprevalence studies.

Refining the role of presurgical PET/4D-CT in a large series of patients with primary hyperparathyroidism undergoing [18F]Fluorocholine PET/CT.

4D-CT has garnered attention as complementary imaging for patients with primary hyperparathyroidism (pHPT). Herein we evaluated a diagnostic strategy using [18F]Fluorocholine Positron Emission Tomography/Computed Tomography (PET/CT), followed by 4D-CT integrated into PET/4D-CT after negative/inconclusive PET/CT results in a single-center retrospective cohort of 166 pHPT patients who underwent parathyroidectomy after [18F]Fluorocholine PET/4D-CT. PET/CT and 4D-CT images were interpreted by three nuclear medicine physicians and one expert radiologist. Pathological findings were documented, and concordance rates were assessed. PET/CT results were categorized as positive/negative, with positive cases rated on a 3-level certitude scale: low, moderate, high. Inconclusive cases included low/moderate positivity. The added value of PET/4D-CT was assessed for negative/inconclusive cases through joint reading. PET/CT lesion-based analysis showed almost perfect interobserver concordance (Cohen's kappa >.8). Across the cohort, PET/CT had a sensitivity of 83%, specificity of 97%, PPV of 90% and NPV of 94%. For 4D-CT, these values were sensitivity: 53%, specificity: 84%, PPV: 56% and NPV: 82%. PET/CT was significantly more accurate than 4D-CT. Among 44 patients with negative/inconclusive results, PET/CT had sensitivity: 60%, specificity: 91%, PPV: 71% and NPV: 86%. In the same patients, sensitivity and specificity of the sequential diagnostic algorithm increased to 80% and 97%, showing significantly better global accuracy (92% vs. 83%) than standard PET/CT. We support a personalized imaging algorithm for pHPT, placing [18F]Fluorocholine PET/CT at the forefront, followed by 4D-CT integrated into PET/4D-CT in the same imaging session for negative/inconclusive results. When PET/CT results are clearly positive, the additional sensitivity benefit of 4D-CT is minimal.

DR-GPT: A large language model for medical report analysis of diabetic retinopathy patients.

Diabetic retinopathy (DR) is a sight-threatening condition caused by diabetes. Screening programmes for DR include eye examinations, where the patient's fundi are photographed, and the findings, including DR severity, are recorded in the medical report. However, statistical analyses based on DR severity require structured labels that calls for laborious manual annotation process if the report format is unstructured. In this work, we propose a large language model DR-GPT for classification of the DR severity from unstructured medical reports. On a clinical set of medical reports, DR-GPT reaches 0.975 quadratic weighted Cohen's kappa using truncated Early Treatment Diabetic Retinopathy Study scale. When DR-GPT annotations for unlabeled data are paired with corresponding fundus images, the additional data improves image classifier performance with statistical significance. Our analysis shows that large language models can be applied for unstructured medical report databases to classify diabetic retinopathy with a variety of applications.

Impact of intracranial atherosclerosis burden on vasospasm risk and outcomes in aneurysmal subarachnoid hemorrhage

BackgroundCerebral vasospasm is a well-known complication after aneurysmal subarachnoid hemorrhage (aSAH) and occurs more commonly in younger patients. We hypothesized that intracranial atherosclerosis, which is seen predominantly in older patients, affects vasospasm risk. We sought to determine association between intracranial atherosclerosis burden with vasospasm and outcomes in aSAH. MethodsWe retrospectively reviewed a cohort of consecutive patients with aSAH admitted to a Comprehensive Stroke Center between 2016 and 2023. Intracranial atherosclerosis burden was quantified by using modified Woodcock (MW) score on CT angiograms. Vasospasm was defined based on transcranial Doppler (TCD) criteria. Poor outcome was defined as 3-month modified Rankin Scale 3–6. ResultsWe reviewed 392 patients and included 302 (mean age 56.8 years [SD 13.3], 65 % female and 70 % white) in the final analysis. MW scores were measured with excellent intra-rater and inter-rater reliability (Cohen's kappa coefficient 0.9 and 0.83 respectively) ranging from 0 to 3 (mean 0.59, SD 0.83) with higher scores in older patients (beta coefficient 0.019, 95 % CI 0.009–0.028; p < 0.001). Higher MW calcification score was associated with lower risk of vasospasm (OR 0.52 per point increase, 95 % CI 0.36–0.78; p = 0.001). There was an inverse correlation between MW scores and severity of vasospasm (beta coefficient −0.29, 95 % CI −0.48, −0.1; p = 0.003). However, MW score was not independently associated with poor functional outcome (p = 0.62). ConclusionsIntracranial atherosclerosis is a potential mechanism for lower TCD-based vasospasm in older patients with aSAH; however, it may not impact functional outcomes. Larger prospective studies are needed to confirm our findings.

Evaluating diagnostic accuracy and agreement of TI-RADS scoring in thyroid nodules: A comparative analysis between sonographers and radiologists.

The Thyroid Imaging Reporting and Data System (TI-RADS) is an essential tool for assessing thyroid nodules, primarily used by radiologists. This study aimed to compare the agreement of TI-RADS scores between sonographers and radiologists and to assess the diagnostic performance of these scores against histological findings in suspicious thyroid nodules. In a retrospective analysis, 168 patients with suspicious thyroid nodules classified as TR3 and above by the radiologists were included. Both sonographers and radiologists independently assigned the American College of Radiologists (ACR) TI-RADS scores, which were then compared for inter-reader agreement using Cohen's Kappa statistic. The scores were also evaluated for diagnostic performance against histological results based on the Bethesda system. The study revealed a moderate overall agreement between sonographers and radiologists in TI-RADS scoring (κ = 0.504; 95% CI: 0.409-0.599), with poor agreement noted specifically for nodule margin scores (κ = 0.102; 95% CI: -1.430-0.301). In terms of diagnostic performance against histological outcomes, sonographers' TI-RADS scores showed a sensitivity of 100% and a specificity of 44.6%, while radiologists' scores showed a sensitivity of 100% but a lower specificity of 29.3%. The findings indicate moderate agreement in TI-RADS scoring between sonographers and radiologists, with reproducibility challenges especially in scoring nodule margins. The marginally superior diagnostic performance of sonographers' scores suggests potential efficiency benefits in involving sonographers in preliminary assessments. Future research should aim to encompass a wider range of TI-RADS categories and focus on minimizing scoring variability to enhance the system's clinical utility.

Assessing the feasibility of ChatGPT-4o and Claude 3-Opus in thyroid nodule classification based on ultrasound images.

Large language models (LLMs) are pivotal in artificial intelligence, demonstrating advanced capabilities in natural language understanding and multimodal interactions, with significant potential in medical applications. This study explores the feasibility and efficacy of LLMs, specifically ChatGPT-4o and Claude 3-Opus, in classifying thyroid nodules using ultrasound images. This study included 112 patients with a total of 116 thyroid nodules, comprising 75 benign and 41 malignant cases. Ultrasound images of these nodules were analyzed using ChatGPT-4o and Claude 3-Opus to diagnose the benign or malignant nature of the nodules. An independent evaluation by a junior radiologist was also conducted. Diagnostic performance was assessed using Cohen's Kappa and receiver operating characteristic (ROC) curve analysis, referencing pathological diagnoses. ChatGPT-4o demonstrated poor agreement with pathological results (Kappa = 0.116), while Claude 3-Opus showed even lower agreement (Kappa = 0.034). The junior radiologist exhibited moderate agreement (Kappa = 0.450). ChatGPT-4o achieved an area under the ROC curve (AUC) of 57.0% (95% CI: 48.6-65.5%), slightly outperforming Claude 3-Opus (AUC of 52.0%, 95% CI: 43.2-60.9%). In contrast, the junior radiologist achieved a significantly higher AUC of 72.4% (95% CI: 63.7-81.1%). The unnecessary biopsy rates were 41.4% for ChatGPT-4o, 43.1% for Claude 3-Opus, and 12.1% for the junior radiologist. While LLMs such as ChatGPT-4o and Claude 3-Opus show promise for future applications in medical imaging, their current use in clinical diagnostics should be approached cautiously due to their limited accuracy.

Las placas intrapélvicas suprapectíneas interfieren con la evaluación postoperatoria de la calidad de reducción por radiografía simple. Hallazgos sobre una serie retrospectiva de fracturas de acetábulo

IntroductionIntrapelvic suprapectineal plates play an important role in acetabular fracture fixation. However, the shape of these implants may interfere with the quality of reduction evaluations using plain X-rays. We sought to evaluate this artifact and its relationship with CT findings. Materials and methodsIn a retrospective, single-center series of 22 acetabular fractures, postoperative AP, alar and obturator X-ray views and CT images were evaluated by three independent observers. Cohen's kappa was used to examine interrater reliability. ResultsSuprapectineal plates interfered with the evaluation of the weight-bearing surface in 75.3%, and with all three oblique views in 43.9% of cases. The central segment was most consistently interfered with, corresponding to the area where the greatest malreduction was in 46.9% coronal and 42.4% of sagittal CT views. ConclusionsSince the quality of reduction has prognostic value and is a necessary guide for the surgical team, that CT may be considered for the postoperative examination of the most challenging acetabular fracture cases.

Pilot study for the development of an automatically generated and wearable-based early warning system for the detection of deterioration of hospitalized patients of an acute care hospital.

Acute deteriorations of health status are common in hospitalized patients and are often preceded by changes in their vital signs. Events such as heart attacks, death or admission to the intensive care unit can be averted by early detection, therefore so-called Early Warning Scores (EWS) such as the National Early Warning Score 2 (NEWS2), including basic vital parameters such as heart rate, blood pressure, respiratory rate, temperature and level of consciousness, have been developed for a systematic approach. Although studies have shown that EWS have a positive impact on patient outcomes, they are often limited by issues such as calculation errors, time constraints, and a shortage of human resources. Therefore, development of tools for automatic calculation of EWS could help improve quality of EWS calculation and may improve patient outcomes. The aim of this study is to analyze the feasibility of wearable devices for the automatic calculation of NEWS2 compared to conventional calculation using vital signs measured by health care professionals. We conducted a prospective trial at a large tertiary hospital in Switzerland. Patients were given a wristband with a photoplethysmogram (PPG) sensor that continuously recorded their heart rate and respiratory rate for 3 consecutive days. Combined with data from the electronic health record (EHR), NEWS2-score was calculated and compared to NEWS2 score calculated from vital parameters in the EHR measured by medical staff. The main objective of our study was to assess the agreement between NEWS2 scores calculated using both methods. This analysis was conducted using Cohen's Kappa and Bland-Altman analysis. Secondary endpoints were compliance concerning the medical device, patient acceptance, data quality analysis and data availability and signal quality for all time stamps needed for accurate calculation. Of 210 patients enrolled in our study, NEWS2 was calculated in 904 cases, with 191 cases being directly compared to conventional measurements. Thirty-three of these measurements resulted in a NEWS2 ≥ 5, 158 in a NEWS2 < 5. Comparing all 191 measurements, accordance was substantial (K = 0.76) between conventional and automated NEWS2. No adverse effects due to the device were recorded. Patient acceptance was high. In conclusion, the study found strong agreement between automated and conventional NEWS2 calculations using wearable devices, with high patient acceptance despite some data quality challenges. To maximize the potential of continuous monitoring, further research into fully automated EWS calculations without relying on spot measurements is suggested, as this could provide a reliable alternative to traditional methods. January 26, 2023, NCT05699967.

Automated Transformation of Unstructured Cardiovascular Diagnostic Reports into Structured Datasets Using Sequentially Deployed Large Language Models.

Rich data in cardiovascular diagnostic testing are often sequestered in unstructured reports, with the necessity of manual abstraction limiting their use in real-time applications in patient care and research. We developed a two-step process that sequentially deploys generative and interpretative large language models (LLMs; Llama2 70b and Llama2 13b). Using a Llama2 70b model, we generated varying formats of transthoracic echocardiogram (TTE) reports from 3,000 real-world echo reports with paired structured elements, leveraging temporal changes in reporting formats to define the variations. Subsequently, we fine-tuned Llama2 13b using sequentially larger batches of generated echo reports as inputs, to extract data from free-text narratives across 18 clinically relevant echocardiographic fields. This was set up as a prompt-based supervised training task. We evaluated the fine-tuned Llama2 13b model, HeartDx-LM, on several distinct echocardiographic datasets: (i) reports across the different time periods and formats at Yale New Haven Health System (YNHHS), (ii) the Medical Information Mart for Intensive Care (MIMIC) III dataset, and (iii) the MIMIC IV dataset. We used the accuracy of extracted fields and Cohen's Kappa as the metrics and have publicly released the HeartDX-LM model. The HeartDX-LM model was trained on randomly selected 2,000 synthetic echo reports with varying formats and paired structured labels, with a wide range of clinical findings. We identified a lower threshold of 500 annotated reports required for fine-tuning Llama2 13b to achieve stable and consistent performance. At YNHHS, the HeartDx-LM model accurately extracted 69,144 out of 70,032 values (98.7%) across 18 clinical fields from unstructured reports in the test set from contemporary records where paired structured data were also available. In older echo reports where only unstructured reports were available, the model achieved 87.1% accuracy against expert annotations for the same 18 fields for a random sample of 100 reports. Similarly, in expert-annotated external validation sets from MIMIC-IV and MIMIC-III, HeartDx-LM correctly extracted 201 out of 220 available values (91.3%) and 615 out of 707 available values (87.9%), respectively, from 100 randomly chosen and expert annotated echo reports from each set. We developed a novel method using paired large and moderate-sized LLMs to automate the extraction of unstructured echocardiographic reports into tabular datasets. Our approach represents a scalable strategy that transforms unstructured reports into computable elements that can be leveraged to improve cardiovascular care quality and enable research.

Does the Combined Arteritis Damage Score (CARDS) in the Diagnosis of Takayasu Arteritis Predict Prognosis and Need for Biologic Therapy?

We aimed to investigate the prognostic value of the combined arteritis damage score (CARDS) in Takayasu arteritis (TAK) patients to predict the need for biologic treatment at diagnosis and the possible contribution of wall thickness (WT). Blind evaluation of MRA/CTA at the time of diagnosis was performed by a reader rheumatologist (RR) and an interventional radiologist (RIR). The CARDS damage score for 21 arterial regions was assessed as normal, mild or moderate/severe stenosis, occclusion or aneursym/dilatation. Additionally, WT was scored for all regions as present or absent. A modified CARDS (mCARDS) was calculated as the sum of CARDS and the number of WT areas. According to follow-up treatment, 10 patients with non-biologic treatment (non-BT) (F/M:8/2, median age 37.5years) and 15 patients with biologic treatment (BT) (F/M:13/2, median age 30years) were included. Indian Takatasu Arteritis Score (ITAS), CRP, and ESR levels were similar in both groups. CARDS (1.4 (0-7.2) vs 4.5 (.6-19), P: .003), WT (1.5 (0-8) vs 7 (1-21), P < .001), and mCARDS (4 (0-14.2) vs 11.4 (1.6-40), P < .001) scores were significantly higher in the BT group compared to nonBT group. Cohen's kappa coefficient between RR and RIR for WT was .99 with 99.6% aggrement, and CARDS was .98 with 99.6% agreement. The AUC values for CARDS, WT, and mCARDS scores were .748 (.605-.892), .837 (.723-.950), and .847 (.735-.958), respectively, and P value was <.0001. The prediction of prognosis and biologic treatment need at TAK diagnosis using non-invasive angiographic images can improve outcomes and prompt closer follow-up. The combination of CARDS and WT as mCARDS achieved the highest sensitivity and specificity, and all scores appear useful for predicting prognosis.

Micro-enema immediately prior to prostate MRI: effects on rectal gas, image quality and PI-QUAL score.

Our aim was to determine whether the administration of a micro-enema immediately prior to prostate MRI is associated with a reduction in rectal gas, gas related artifacts and an improvement in image quality and PI-QUAL score. This retrospective analysis enrolled 171 patients who underwent multiparametric 3T prostate MRI at our institution between January 2021 and September 2022. 86 patients received a micro-enema, and a further 85 patients did not. Two fellowship trained abdominal radiologists were blinded and independently reviewed each prostate MRI, recording scores on a dedicated scoring sheet. The quality of T2 weighted (T2W), diffusion weighted (DWI), and dynamic contrast enhancement (DCE) images were assessed according to standardised scales supported in the literature. In addition, gas related artifacts and rectal gas level were examined. An independent-samples Mann-Whitney U and t-test were performed, comparing both the median and mean score between micro-enema and no micro-enema groups for each reader. Spearman's correlation was used to determine the strength of relationship between variables. A quadratic weighted Cohen's Kappa and percent agreement were used to assess inter-observer agreement. Image quality was improved in those who received the micro-enema compared to those who did not according to the visual grading scale on the DWI sequence for both readers (reader 1: median 4 vs. 4, p < 0.001, mean 4.27 vs. 3.92, p < 0.001; reader 2: median 5 vs. 4, p < 0.001, mean 4.74 vs. 4.14, p < 0.001). PI-QUAL score was significantly improved in the micro-enema group for reader 2 only (reader 1: median 4 vs. 4, p = 0.25, mean 3.99 vs. 4.08, p = 0.21; reader 2: median 5 vs. 5, p = 0.01, mean 4.95 vs. 4.78, p = 0.01). Visual grading score for both the T2W and DCE images was not significantly different. Rectal gas level was lower in patients who received the micro-enema for both readers (reader 1: median 2 vs. 4, p < 0.001, mean 2.12 vs. 3.60, p < 0.001; reader 2: median 1 vs. 2, p < 0.001, mean 1.37 vs. 2.48, p < 0.001), correlating with a lower score for gas-related artifacts (reader 1: median 1 vs. 2, p < 0.001, mean 1.50 vs. 1.92, p < 0.001; reader 2: median 1 vs. 1, p < 0.001, mean 1.16 vs. 1.71, p < 0.001) in this group. Correlation between rectal gas level and gas-related artifacts on DWI regardless of whether a micro-enema was given was strong (rs = 0.71, p < 0.001). Correlation was moderate to strong between rectal gas level and image quality on DWI (rs = -0.63, p < 0.001). There was only 1 (1.2%) borderline diagnostic or non-diagnostic DWI sequence in those who received the micro-enema, compared to 9 (10.6%) in those who did not (p = 0.009). Interobserver agreement was moderate for image quality on DWI, gas related artifacts and rectal gas level (weighted kappa values of 0.52, 0.49 and 0.53 respectively). The administration of a micro-enema immediately prior to prostate MRI is associated with a significant improvement in image quality on the DWI sequence compared to no bowel preparation. This is mediated through a reduction in rectal gas and gas related artifacts. Improvements in PI-QUAL score was mixed between readers.

ISETECH-Statistics and Machine Learning Based Decision-Making for Diet Beverage Choice and Recommendation through Software Application

Positioned within the realm of ubiquitous beverage chains acclaimed for their diverse coffee and tea offerings, this investigation deeply explores the intricate nutritional dynamics and caloric compositions of these widely consumed beverages. At its core, the research aims to streamline decision-making for individuals actively seeking health-conscious beverage alternatives. This comprehensive endeavour unfolds through an integrated methodology that combines hypothesis testing and other statistical techniques with the establishment of a robust decision support system, complemented by powerful machine learning techniques. Distinctively, the paper integrates an intuitive React application (ReactApp) into the robust Fast API framework. Users engage seamlessly with the decision support system, as Fast API efficiently manages data processing, interfaces with machine learning algorithms, and delivers personalized recommendations to the ReactApp front. In essence, this interdisciplinary initiative epitomizes the fusion of nutritional science, statistical analysis, machine learning, and modern web technologies, providing a holistic and pioneering solution for selecting health-conscious beverage alternatives. One of the most noteworthy outcomes of our research lies in the compelling results we've achieved. Our predictive model has demonstrated exceptional performance: We achieved an impressive Accuracy Rate of 83.67%, signifying the high precision of our recommendations in identifying health-conscious beverage alternatives. Simultaneously, the F1-Score, which harmonizes precision and recall, stands at a commendable 82.21%, indicating a well-balanced and effective decision support system. Our model's ability to recall health-conscious choices is also strong, with a Recall rate of 83.67%. Furthermore, we assessed the model's agreement with observed data using Cohen's Kappa, which yielded a substantial score of 78.01%. This indicates not only the model's predictive power but also its ability to capture agreement beyond what might occur by chance alone. These results, seamlessly integrated into our research narrative, underscore the effectiveness and reliability of our approach. They constitute a rare and valuable contribution to the field, promising a novel pathway for informed and health-driven consumer choices, firmly supported by our rigorous methodology and advanced machine learning outcomes.

Is Trochlear Dysplasia Associated with a High Prevalence of Repaired Anterior Medial Portal Lesions in Anterior Cruciate Ligament-Deficient Knees?

Abstract Introduction Meniscal tears, especially in the medial meniscus, are common in anterior cruciate ligament (ACL) injuries. Repaired anterior medial portal (RAMP) lesions, specific meniscal injuries associated with ACL ruptures, affect the peripheral attachment of the posterior horn of the medial meniscus. While the relationship between femoral trochlear morphology and ACL tears is established, its association with RAMP lesions remains unexplored. This study investigates the association between femoral trochlear morphology and RAMP lesions in ACL-deficient knees using magnetic resonance imaging (MRI) and evaluates MRI's accuracy against arthroscopy. Methods This hospital-based, cross-sectional observational study was conducted in a tertiary care center over 18 months. It included 126 adult patients who underwent arthroscopic ACL reconstruction and had preoperative MRI. Variables included age, gender, body mass index, trauma nature, and various MRI parameters, including femoral trochlear morphology and presence of RAMP lesions. MRI findings were compared to arthroscopic findings to evaluate diagnostic accuracy. Results RAMP lesions were detected in 21 participants (16.7%) via MRI and 28 participants (22.2%) via arthroscopy. MRI showed 71.4% sensitivity and 99.0% specificity for detecting RAMP lesions. No significant association was found between femoral trochlear morphology and RAMP lesions. However, medial tibial slope (TS) and posterior medial tibial bone edema (PMTBE) were significantly associated with RAMP lesions. The agreement between MRI and arthroscopy was high (Cohen's kappa = 0.773, p &lt; 0.001). Conclusion This study found no significant association between femoral trochlear morphology and RAMP lesions in ACL-deficient knees. Instead, medial TS and PMTBE were significant predictors. MRI proved to have high specificity but moderate sensitivity compared to arthroscopy.

Using ChatGPT to Provide Patient-Specific Answers to Parental Questions in the PICU.

To determine if ChatGPT can incorporate patient-specific information to provide high-quality answers to parental questions in the PICU. We hypothesized that ChatGPT would generate high-quality, patient-specific responses. In this cross-sectional study, we generated assessments and plans for 3 PICU patients with respiratory failure, septic shock, and status epilepticus and paired them with 8 typical parental questions. We prompted ChatGPT with instructions, an assessment and plan, and 1 question. Six PICU physicians evaluated the responses for accuracy (1-6), completeness (yes/no), empathy (1-6), and understandability (Patient Education Materials Assessment Tool, PEMAT, 0% to 100%; Flesch-Kincaid grade level). We compared answer quality among scenarios and question types using the Kruskal-Wallis and Fischer's exact tests. We used percent agreement, Cohen's Kappa, and Gwet's agreement coefficient to estimate inter-rater reliability. All answers incorporated patient details, utilizing them for reasoning in 59% of sentences. Responses had high accuracy (median 5.0, [interquartile range (IQR), 4.0-6.0]), empathy (median 5.0, [IQR, 5.0-6.0]), completeness (97% of all questions), and understandability (PEMAT % median 100, [IQR, 87.5-100]; Flesch-Kincaid level 8.7). Only 4/144 reviewer scores were <4/6 in accuracy, and no response was deemed likely to cause harm. There was no difference in accuracy, completeness, empathy, or understandability among scenarios or question types. We found fair, substantial, and almost perfect agreement among reviewers for accuracy, empathy, and understandability, respectively. ChatGPT used patient-specific information to provide high-quality answers to parental questions in PICU clinical scenarios.

Equipping computational pathology systems with artifact processing pipelines: a showcase for computation and performance trade-offs

BackgroundHistopathology is a gold standard for cancer diagnosis. It involves extracting tissue specimens from suspicious areas to prepare a glass slide for a microscopic examination. However, histological tissue processing procedures result in the introduction of artifacts, which are ultimately transferred to the digitized version of glass slides, known as whole slide images (WSIs). Artifacts are diagnostically irrelevant areas and may result in wrong predictions from deep learning (DL) algorithms. Therefore, detecting and excluding artifacts in the computational pathology (CPATH) system is essential for reliable automated diagnosis.MethodsIn this paper, we propose a mixture of experts (MoE) scheme for detecting five notable artifacts, including damaged tissue, blur, folded tissue, air bubbles, and histologically irrelevant blood from WSIs. First, we train independent binary DL models as experts to capture particular artifact morphology. Then, we ensemble their predictions using a fusion mechanism. We apply probabilistic thresholding over the final probability distribution to improve the sensitivity of the MoE. We developed four DL pipelines to evaluate computational and performance trade-offs. These include two MoEs and two multiclass models of state-of-the-art deep convolutional neural networks (DCNNs) and vision transformers (ViTs). These DL pipelines are quantitatively and qualitatively evaluated on external and out-of-distribution (OoD) data to assess generalizability and robustness for artifact detection application.ResultsWe extensively evaluated the proposed MoE and multiclass models. DCNNs-based MoE and ViTs-based MoE schemes outperformed simpler multiclass models and were tested on datasets from different hospitals and cancer types, where MoE using (MobileNet) DCNNs yielded the best results. The proposed MoE yields 86.15 % F1 and 97.93% sensitivity scores on unseen data, retaining less computational cost for inference than MoE using ViTs. This best performance of MoEs comes with relatively higher computational trade-offs than multiclass models. Furthermore, we apply post-processing to create an artifact segmentation mask, a potential artifact-free RoI map, a quality report, and an artifact-refined WSI for further computational analysis. During the qualitative evaluation, field experts assessed the predictive performance of MoEs over OoD WSIs. They rated artifact detection and artifact-free area preservation, where the highest agreement translated to a Cohen Kappa of 0.82, indicating substantial agreement for the overall diagnostic usability of the DCNN-based MoE scheme.ConclusionsThe proposed artifact detection pipeline will not only ensure reliable CPATH predictions but may also provide quality control. In this work, the best-performing pipeline for artifact detection is MoE with DCNNs. Our detailed experiments show that there is always a trade-off between performance and computational complexity, and no straightforward DL solution equally suits all types of data and applications. The code and HistoArtifacts dataset can be found online at Github and Zenodo, respectively.