Clinical Diagnostic Tool Research Articles

P-2160. Where Does Karius Testing Have a Clinical Impact? A Retrospective Analysis of Next Generation Sequencing

Abstract Background Karius testing, a commercially available microbial cell-free DNA assay, has emerged as a new diagnostic tool for clinicians, particularly for management of culture negative infections and in immunocompromised patients. There remains uncertainty regarding the clinical impact of next-generation sequencing in patient management. Figure 1 Methods This was a retrospective multi-site observational cohort study over a five-year period from 2018 to 2023 at a community hospital system. Patients were included if they had a Karius test ordered with results available to treating clinicians during the time period. The primary outcome was a change in clinical management defined as either the addition or removal of antibiotics, based on results from Karius testing. Results 391 patients met inclusion criteria, with 283 (72%) in immunocompetent and 108 (28%) in immunocompromised patients. An infectious diseases (ID) consult was obtained in 355 (91%) cases. Fifty-two percent of Karius tests were positive, with an average of 1.1 pathogens reported. Concordance rate between positive cultures and Karius results was 12%. Pneumonia (24%), fever of unknown origin (FUO) (18%), and endocarditis (17%) were the most common indications for Karius orders (Fig 1). Testing changed management in 19% of all included individuals, and in 18% of immunocompromised patients. Karius testing was most impactful in neutropenic fever and endocarditis with 27% and 22% respectively changing management and least impactful when ordered for joint infection, where 12% resulted in a change in management (Fig 2). Conclusion The Karius test changed antibiotic treatment in 19% of patients. It was most clinically relevant in neutropenic fever and endocarditis management. Even with our sizeable sample population including both immunocompromised and immunocompetent patients across a large, multi-site community hospital system, findings remained consistent compared to other smaller studies. Further investigation would be beneficial to determine how best to incorporate this emerging tool in clinical practice. Disclosures All Authors: No reported disclosures

Hierarchical DNA Octahedral Nanoplatform for in Situ Biosensing and Clinical Monitoring of Acute Lymphoblastic Leukemia.

Developing nanoscale platforms with high integration, assembly efficiency, and structural stability for performing complex computations in specific cells remains a significant challenge. To address this, the Three-dimensional Hierarchical Octahedral Robotic (THOR) DNA nanoplatform is introduced, which integrates targeting, logic computation, and sensing modules within a single framework. This nanoplatform specifically binds to cancer cell surface proteins, releasing aptamer-linked fuel chains to initiate subsequent computational processes. Three logic gates efficiently compute any arbitrary binary combination of target proteins. The sensing module employs catalytic hairpin assembly for detecting specific miRNAs with high sensitivity. THOR demonstrates robust functionality both in vitro and in situ. As a proof-of-concept, this nanoplatform to distinguish acute lymphoblastic leukemia (ALL) patients from other leukemia subtypes and healthy participants, achieving 100% accuracy is applied. Additionally, this approach reliably monitored the therapeutic progress of ALL patients, showing strong concordance with bone marrow smear results. The THOR platform highlights the feasibility of constructing a reliable, hierarchical, and multifunctional analytical system based on a single DNA polyhedron. It offers a promising auxiliary tool for clinical diagnostics and therapeutic monitoring.

Development of a model for measuring sagittal plane parameters in 10–18-year old adolescents with idiopathic scoliosis based on RTMpose deep learning technology

PurposeThe study aimed to develop a deep learning model for rapid, automated measurement of full-spine X-rays in adolescents with Adolescent Idiopathic Scoliosis (AIS). A significant challenge in this field is the time-consuming nature of manual measurements and the inter-individual variability in these measurements. To address these challenges, we utilized RTMpose deep learning technology to automate the process.MethodsWe conducted a retrospective multicenter diagnostic study using 560 full-spine sagittal plane X-ray images from five hospitals in Inner Mongolia. The model was trained and validated using 500 images, with an additional 60 images for independent external validation. We evaluated the consistency of keypoint annotations among different physicians, the accuracy of model-predicted keypoints, and the accuracy of model measurement results compared to manual measurements.ResultsThe consistency percentages of keypoint annotations among different physicians and the model were 90–97% within the 4-mm range. The model's prediction accuracies for key points were 91–100% within the 4-mm range compared to the reference standards. The model's predictions for 15 anatomical parameters showed high consistency with experienced physicians, with intraclass correlation coefficients ranging from 0.892 to 0.991. The mean absolute error for SVA was 1.16 mm, and for other parameters, it ranged from 0.22° to 3.32°. A significant challenge we faced was the variability in data formats and specifications across different hospitals, which we addressed through data augmentation techniques. The model took an average of 9.27 s to automatically measure the 15 anatomical parameters per X-ray image.ConclusionThe deep learning model based on RTMpose can effectively enhance clinical efficiency by automatically measuring the sagittal plane parameters of the spine in X-rays of patients with AIS. The model's performance was found to be highly consistent with manual measurements by experienced physicians, offering a valuable tool for clinical diagnostics.

A simple and integrated fiber-optic real-time qPCR platform for remote and distributed detection of epidemic virus infection.

Quantitative polymerase chain reaction (qPCR) is a well-recognized technique for amplifying and quantifying nuclear acid, and its real-time monitoring capability, ultrahigh sensitivity, and accuracy make it a "golden-standard" tool in both molecular biology research and clinical diagnostics. However, current qPCR tests rely on bulky instrumentation and skilled laboratorians in centralized laboratories, which spatially and temporally separate the sample collection and test, leading to longer sample turnaround times (TATs) and limited working conditions. Herein, we propose an integrated optical fiber real-time polymerase chain reaction (iF-PCR) system that successfully allows convenient sample collection, rapid thermocycling, closed-loop thermal annealing, and real-time fluorescence detection in a tiny capillary reactor. By leveraging the easy-handling capillary-fiber structure and rapid photothermal actuation of the graphene-decorated fiber head, the whole TAT can be shortened, including sampling and 40 thermocycle amplifications, within 23min, in which an ultrasmall sample volume of ∼15μL is needed. Furthermore, the thermal amplification and fluorescence detection capability of the optical fiber system was cross-checked by a commercial qPCR instrument. The fiber-optic qPCR strategy can correctly distinguish between positive and negative samples of clinical respiratory syncytial virus (RSV) without the need for laboratory professional skills. Taking advantage of the distance signal transmission nature of optical fibers, the proposed strategy enables remote testing, which can eliminate the necessity of instrument deployment in crowded biosafety laboratories and facilitate potential distributed pathogen testing for coping with a new round of pandemics.

Role of Chromatography in Pharmaceutical Analysis: Trends and Future Perspectives

Abstract: Chromatography is a widely used analytical technique for separating and analyzing compounds present in complex mixtures. It operates on the principle of differential distribution of components between a stationary phase and a mobile phase, enabling the separation of substances based on their varying affinities for these phases. Over the years, chromatography has evolved into a versatile and indispensable tool in various fields, including chemistry, biochemistry, pharmacology, food science, environmental analysis, and clinical diagnostics. The primary types of chromatography include gas chromatography (GC), liquid chromatography (LC), ion-exchange chromatography (IEC), affinity chromatography, and size exclusion chromatography (SEC). Each type is suited to different applications, based on factors like the nature of the analytes, separation efficiency, and detection sensitivity. High-performance liquid chromatography (HPLC) and ultra-high-performance liquid chromatography (UHPLC) are particularly renowned for their resolution and speed, making them indispensable in pharmaceutical and clinical settings. This review provides an overview of the principles, types, applications, and recent advancements in chromatography. It highlights the integration of chromatography with other techniques, such as mass spectrometry (MS) and spectroscopy, to enhance its sensitivity and precision. Additionally, it discusses emerging trends, such as miniaturization and automation that are shaping the future of chromatographic analysis in research, quality control, and diagnostics.

Enhancing Alzheimer's Disease Diagnosis through Magnetic Resonance Imaging: An Analysis using VGG19 Architectures

Early detection of Alzheimer's disease (AD) is an area of much research since early diagnosis can offer patient better treatment and enhanced care. In this work we propose a deep learning approach to detect Alzheimer’s disease using the VGG-19 architecture, one of the state-of-the-art convolutional neural networks (CNN). In this work we utilized a dataset composed of a heterogeneous set of brain MRI images from healthy subjects and Alzheimer patients, they are part of the ADNI (Alzheimer's Disease Neuroimaging Initiative). The dataset was preprocessed with a few techniques such as image normalization, augmentation, and denoising to further increase the model's performance. These techniques also expanded the quality of the input data which, coupled with an impressive state-of-the art classification accuracy of 97 %, helped to achieve these results. The results showed deep learning can be effective for early detection of Alzheimer’s disease as a useful clinical diagnostic tool. Finally, this work demonstrates how CNNs such as VGG 19 are ready to be used in medical image analysis and renders a new benchmark on the accuracy of Alzheimer detection.

Intelligent imaging technology applications in multidisciplinary hospitals.

With the rapid development of artificial intelligence technology, its applications in medical imaging have become increasingly extensive. This review aimed to analyze the current development status and future direction of intelligent imaging technology by investigating its application in various medical departments. To achieve this, we conducted a comprehensive search of various data sources up to 2024, including PubMed, Web of Science, and Google Scholar, based on the principle of comprehensive search. A total of 332 articles were screened, and after applying the inclusion and exclusion criteria, 56 articles were selected for this study. According to the findings, intelligent imaging technology exhibits robust image recognition capabilities, making it applicable across diverse medical imaging modalities within hospital departments. This technology offers an efficient solution for the analysis of various medical images by extracting and accurately identifying complex features. Consequently, it significantly aids in the detection and diagnosis of clinical diseases. Its high accuracy, sensitivity, and specificity render it an indispensable tool in clinical diagnostics and related tasks, thereby enhancing the overall quality of healthcare services. The application of intelligent imaging technology in healthcare significantly enhances the efficiency of clinical diagnostics, resulting in more accurate and timely patient assessments. This advanced technology offers a faster and more precise diagnostic approach, ultimately improving patient care and outcomes. This review analyzed the socioeconomic changes brought about by intelligent imaging technology to provide a more comprehensive evaluation. Also, we systematically analyzed the current shortcomings of intelligent imaging technology and its future development directions, to enable future research.

C‐TUnet: A CNN‐Transformer Architecture‐Based Ultrasound Breast Image Classification Network

ABSTRACTUltrasound breast image classification plays a crucial role in the early detection of breast cancer, particularly in differentiating benign from malignant lesions. Traditional methods face limitations in feature extraction and global information capture, often resulting in lower accuracy for complex and noisy ultrasound images. This paper introduces a novel ultrasound breast image classification network, C‐TUnet, which combines a convolutional neural network (CNN) with a Transformer architecture. In this model, the CNN module initially extracts key features from ultrasound images, followed by the Transformer module, which captures global context information to enhance classification accuracy. Experimental results demonstrate that the proposed model achieves excellent classification performance on public datasets, showing clear advantages over traditional methods. Our analysis confirms the effectiveness of combining CNN and Transformer modules—a strategy that not only boosts the accuracy and robustness of ultrasound breast image classification but also offers a reliable tool for clinical diagnostics, holding substantial potential for real‐world application.

Electroanalytical Methods as a Tool for Determining the Antioxidant Capacity in Blood Samples

Aerobic metabolism is an essential process for energy production in cells, but it also generates reactive oxygen species (ROS), which, in excess, can cause cellular damage and contribute to the development of various diseases, such as cardiovascular diseases, cancer, and neurodegenerative disorders. The balance between ROS and antioxidants is crucial for health, as an imbalance can lead to oxidative stress, a significant risk factor for several clinical conditions. Electroanalytical techniques, such as cyclic voltammetry (CV), square wave voltammetry (SWV), and differential pulse voltammetry (DPV), have proven to be powerful tools for evaluating antioxidant capacity more efficiently and accurately than traditional spectrophotometric methods. These techniques offer notable advantages, such as high sensitivity, simplicity, and the ability to analyze complex biological samples rapidly. CV allows for the analysis of compounds with good sensitivity, DPV stands out for its high precision, while SWV offers excellent resolution and speed, making it ideal for clinical applications. The use of these methodologies has been expanded to analyze antioxidants in various biological matrices, enabling a more accurate assessment of oxidative stress and antioxidant status in specific clinical contexts. Additionally, electrochemical biosensors have become revolutionary tools in clinical diagnostics, allowing real-time monitoring of diseases related to oxidative stress. Innovations such as portable devices and integration with artificial intelligence promise to further enhance accessibility and treatment personalization, despite the challenges involved in device cost and standardization.

Solid phase extraction chromatography-based radiochemical isolation of cyclotron-produced 51Mn from enriched 54Fe targets.

We report DGA extraction chromatography isolation of 51Mn from isotopically enriched 54Fe. The method has been studied in semi-automated and automated realizations. The former achieves a decay corrected radiochemical yield of 78±1% (n=3) and a separation factor of (1.0±0.8) x 105 (n=3). With GE HealthCare's Solid Target Platform (STP) and FASTlab the latter, fully automated method achieves a decay corrected radiochemical yield of 87±1% (n=3) and a separation factor of (2.7±0.9) x 104 (n=3). Both setups efficiently isolate cyclotron-produced 51MnCl2 suitable for human administration as determined by developed Chemistry, Manufacturing, and Controls (CMC) acceptance criteria, and support exploration of 51Mn as a clinical diagnostic tool.

To Study the Thyroid Hormone Levels in Neonates of Rural and Urban Vidarbha Region

ABSTRACT Background: This study aimed to evaluate thyroid hormone levels in neonates from the rural and urban Vidarbha region. Over the past decade, neonatal hypothyroidism screening has been conducted using filter paper techniques to identify congenital hypothyroidism (CH). In some screening programs, only T4 is measured initially, with TSH assessed if the T4 levels fall below normal. Conversely, some programs exclusively measure TSH to screen for CH. Currently, initial TSH screening can identify subclinical hypothyroidism in cases where T4 levels are normal, while preliminary T4 testing can reveal congenital hypothyroidism when TSH is elevated. If either T4 or TSH is found to be abnormal during preliminary screening, the other parameter should be monitored. Early diagnosis and prompt treatment are crucial in preventing intellectual disabilities in newborns. Material and Method: The present cross-sectional research will be conducted in collaboration with the Departments of Biochemistry, Obstetrics and Gynecology, Community Medicine, and Pediatrics at Jawaharlal-Nehru Medical College Sawangi (Meghe) Wardha, along with the Datta-Meghe Institute of Medical Science (Deemed University) in Maharashtra, India, and Datta-Meghe Medical College, Shalinitai Meghe Hospital and Research Centre, Nagpur. Results: A total of 272 neonates were tested for Congenital Hypothyroidism (CH) from the study area of the Vidarba region in which 138 neonates were from the rural area and 134 neonates were from the urban area. A total of 144 (53%) males and 128 (47%) females were born in rural and urban areas. In the neonates, it was observed that the Mean ± SD for free T4 and TSH was 0.84 ± 0.15 and 3.83 ± 2.12, respectively. If the free T4 value was less than the lower limit, the TSH value CB ≥25 μIU/ml and plasma ≥10 μIU/ml was considered as screen positive in the present study. Conclusion: Newborn screening programs were established to assess neonates at birth by measuring (TSH) levels. In many screening programs, TSH is the primary marker for evaluating thyroid function and is the preferred test for clinicians diagnosing various thyroid disorders. However, TSH levels can vary significantly and exhibit a broad reference range within the general population, which presents challenges in interpreting the results during newborn screening and in using it later as a clinical diagnostic tool for thyroid function.

Portable SpectroChip-Based Immunoassay Platform for Rapid and Accurate Melamine Quantification in Urine Samples.

Growing concerns about the health risks of melamine adulteration in food products highlight the urgent need for reliable detection methods. However, the long-term effects of chronic low-level melamine exposure remain inadequately explored. This study introduces THE ONE InstantCare platform, a portable immunoassay analyzer integrating a SpectroChip-based spectral processing unit (SPU) with lateral flow immunoassay (LFIA) for sensitive and accurate quantification of melamine in human urine. This platform provides a cost-effective, rapid, and user-friendly point-of-care (POC) solution for melamine detection. Analytical evaluations across eight melamine concentrations (0-100 parts per billion, ppb) achieved a limit of detection (LOD) of 1.91 ppb. Validation with 24 human urine samples demonstrated strong concordance with liquid chromatography-mass spectrometry (LC-MS), yielding an intraclass correlation coefficient (ICC) of 0.9220, a Pearson correlation coefficient of 0.9389, and 95% agreement in Bland-Altman analysis. High reproducibility was observed, with an intraday coefficient of variation (CV) of 6.53% and acceptable interday CV values, while interference studies confirmed reliability in the presence of common biological substances. By delivering results in approximately 10 min, THE ONE InstantCare platform significantly reduces analysis time compared to LC-MS, which typically requires several hours. This novel platform enhances food safety surveillance and advances human health risk assessments, particularly for evaluating melamine-linked kidney damage. Its versatility and robust performance make it a promising tool for environmental monitoring and clinical diagnostics, enabling the detection of diverse biomarkers with high sensitivity and reproducibility.

Tyrosinase Oxidative Cross-Linking in the Cell-Like Crowded Microenvironment for Visible Inhibitor Screening.

Owing to many diseases and disorders being caused by dysfunctional or over/underexpressed enzymes, screening for inhibitors of pharmacologically relevant enzymes has emerged as a promising tool for drug discovery, clinical diagnostics, enzyme engineering, and other medical fields. However, despite the recent advances, most inhibitor screenings are still usually conducted in dilute media, at concentrations far from the media in which the enzymes are actually found, which may cause drugs to fail when translated to in vivo. Herein, we build a gel-like intracellular biological environment in vitro using a tyrosinase oxidative cross-linking hydrogel system that is closer to the real catalytic environment of enzymes. We report a straightforward and effective inhibitor evaluation strategy that can quickly compare the inhibitory strengths of inhibitors based on the principle that adding inhibitors causes color changes and mechanical changes in the system. Enabled by molecular docking, we further demonstrate the different performances of the inhibitors at different concentrations. By construction of the cell-like crowded environment in vitro, this system shows an appealing application prospect for new drug development.

Comprehensive profiling of T-cell exhaustion signatures and establishment of a prognostic model in lung adenocarcinoma through integrated RNA-sequencing analysis

Background T-cell exhaustion (TEX) in the tumor microenvironment causes immunotherapy resistance and poor prognosis. Objective We used bioinformatics to identify crucial TEX genes associated with the molecular classification and risk stratification of lung adenocarcinoma (LUAD). Methods Bulk RNA sequencing data of patients with LUAD were acquired from open sources. LUAD samples exhibited abnormal TEX gene expression, compared with normal samples. TEX gene-based prognostic signature was established and validated in both TCGA and GSE50081 datasets. Immune correlation and risk group-related functional analyses were also performed. Results Eight optimized TEX genes were identified using the LASSO algorithm: ERG, BTK, IKZF3, DCC, EML4, MET, LATS2, and LOX . Several crucial Kyoto encyclopedia of genes and genomes (KEGG) pathways were identified, such as T-cell receptor signaling, toll-like receptor signaling, leukocytes trans-endothelial migration, Fcγ R-mediated phagocytosis, and GnRH signaling. Eight TEX gene-based risk score models were established and validated. Patients with high-risk scores had worse prognosis (P < 0.001). A nomogram model comprising three independent clinical factors showed good predictive efficacy for survival rate in patients with LUAD. Correlation analysis revealed that the TEX signature significantly correlated with immune cell infiltration, tumor purity, stromal cells, estimate, and immunophenotype score. Conclusion TEX-derived risk score is a promising and effective prognostic factor that is closely correlated with the immune microenvironment and estimated score. TEX signature may be a useful clinical diagnostic tool for evaluating pre-immune efficacy in patients with LUAD.

Red Leg Dilemma: Development and Validation of Clinical Decision Tools for Non-Necrotizing Bacterial Dermohypodermitis, Necrotizing Fasciitis, and Eczematous Dermatitis.

Diagnosing red legs on first presentation is challenging. There exists a lack of robustly developed and validated diagnostic red leg tools in clinical practice. Physicians fear missing cases of infectious red legs and treat many patients unnecessarily with antibiotics. Develop and validate easy-to-use diagnostic tools applicable at bedside of patients to orient diagnosis of the commonest and most serious causes of infectious red legs (non-necrotizing bacterial dermohypodermitis (NNBDH), and necrotizing fasciitis (NF)) versus the commonest inflammatory cause (eczema). We collected data of patients presenting to our dermatology department from January first 2012 until May 17th 2017 with a diagnosis of red leg. Three models were developed using fast frugal trees. Validation was performed in a second cohort of patients. A total of 187 patients (mean age 56, SD = 21 years, 48.1% women) were included in the development phase and 62 patients (mean age 64, SD = 19, 52% women) in the validation phase. In the validation data set, sensitivity and specificity were respectively 67% and 91% for NNBDH, 83% and 66%, for NF and 88% and 93%, for eczema. Presentations of suspected lower-limb infections are commonly misdiagnosed, resulting in avoidable antibiotic prescription and hospitalization. We developed an easy-to-use clinical diagnostic tool applicable at the bedside of patients to help orient physicians in certain situations and avoid unnecessary initiation of antibiotics. Future work should focus on validating this tool in primary care to minimize misdiagnosis of red legs and overprescription of antibiotics.

Performance comparison of logistic regression and random forest model in bipolar disorder diagnosis

Abstract. Bipolar disorder (BD) is a significant psychiatric disease that has a large impact on patients living qualities. The diagnosis of BD is important for the early treatment and effective control that benefits the patients, their families, and society. With the development of machine learning algorithms, this paper makes use of the multinomial logistic regression (LR) and Random Forest (RF) models in predicting bipolar disorder (BD). The evaluation and comparison between the two models are conducted to present the possibility of using ML models to help with diagnosis of the BD. A dataset consisting of 120 individuals, including 28 BD Type I patients, 31 BD Type II patients, 31 depression patients, and 30 normal individuals was used in this study. The performance of the models was assessed using metrics such as accuracy, confusion matrix, cross-validation scores, classification report, and ROC-AUC curves. The multinomial LR model demonstrated superior performance with an accuracy of 83%, higher cross-validation scores, and better discriminative ability as indicated by ROC-AUC values. In contrast, the RF model achieved an accuracy of 79%, with lower precision and recall for certain classes. The findings suggest that the multinomial LR model is more effective in predicting bipolar disorder and its subtypes, making it a robust and reliable tool for clinical diagnostics.

Advanced Low-Cost Technology for Assessing Metal Accumulation in the Body of a Metropolitan Resident Based on a Neural Network Model.

This study is devoted to creating a neural network technology for assessing metal accumulation in the body of a metropolis resident with short-term and long-term intake from anthropogenic sources. Direct assessment of metal retention in the human body is virtually impossible due to the many internal mechanisms that ensure the kinetics of metals and the wide variety of organs, tissues, cellular structures, and secretions that ensure their functional redistribution, transport, and cumulation. We have developed an intelligent multi-neural network model capable of calculating the content of metals in the human body based on data on their environmental content. The model is two interconnected neural networks trained on actual measurement data. Since metals enter the body from the environment, the predictors of the model are metal content in drinking water and soil. In this case, water characterizes the short-term impact on the organism, and drinking water, combined with metal contents in soil, is a depository medium that accumulates metals from anthropogenic sources-the long-term impact. In addition, human physiological characteristics are taken into account in the calculations. Each period of exposure is taken into account by its neural network. Two variants of the model are proposed: open loop, where the calculation is performed by each neural network separately, and closed loop, where neural networks work together. The model built in this way was trained and tested on the data of real laboratory studies of 242 people living in different districts of Kazan. As a result, the accuracy of the neural network block for calculating long-term impact was 90% and higher, and the accuracy of the block for calculating short-term impact was 92% and higher. The closed double-loop model showed an accuracy of at least 96%. Conclusions: Our proposed method of assessing and quantifying metal accumulation in the body has high accuracy and reliability. It does not require expensive laboratory tests and allows quantifying the body's metal accumulation content based on readily available information. The calculation results can be used as a tool for clinical diagnostics and operational and planned management to reduce the levels of polymetallic contamination in urban areas.

ADVANCED PREDICTIVE MODELING FOR THE BIOFIRE FILMARRAY SYSTEM IN INFECTIOUS DISEASE DIAGNOSTICS

The Bio Fire Film Array System has emerged as a powerful tool in clinical diagnostics, offering rapid and multiplexed detection of infectious pathogens. This study presents the development of a comprehensive mathematical model aimed at predicting the accuracy and sensitivity of the BioFire FilmArray System under diverse conditions. Key factors such as sample type, pathogen load, specimen characteristics, and environmental variables were systematically analyzed to elucidate their impact on the system's performance. The model incorporates a range of parameters including sample volume, specimen complexity, microbial diversity, and system-specific variables to generate probabilistic estimates of diagnostic outcomes. By leveraging large datasets encompassing diverse clinical scenarios and pathogen profiles, the model achieves robustness and generalizability, thereby facilitating its applicability across different healthcare settings and populations. Validation of the model against independent datasets demonstrates its efficacy in accurately predicting the performance of the BioFire FilmArray System across a spectrum of conditions. Overall, this mathematical model represents a significant advancement in optimizing the utility of the BioFire FilmArray System for infectious disease diagnosis. Its integration into clinical practice holds promise for improving patient care, guiding sample processing protocols, and informing decision-making processes in infectious disease management.

Proteomic analysis of extracellular vesicles derived from canine mammary tumour cell lines identifies protein signatures specific for disease state

BackgroundCanine mammary tumours (CMT) are among the most common types of tumours in female dogs. Diagnosis currently requires invasive tissue biopsies and histological analysis. Tumour cells shed extracellular vesicles (EVs) containing RNAs and proteins with potential for liquid biopsy diagnostics. We aimed to identify CMT subtype-specific proteome profiles by comparing the proteomes of EVs isolated from epithelial cell lines derived from morphologically normal canine mammary tissue, adenomas, and carcinomas.MethodsWhole-cell protein lysates (WCLs) and EV-lysates were obtained from five canine mammary cell lines: MTH53A (non-neoplastic); ZMTH3 (adenoma); MTH52C (simple carcinoma); 1305, DT1406TB (complex carcinoma); and their proteins identified by LC-MS/MS analyses. Gene Ontology analysis was performed on differentially abundant proteins from each group to identify up- and down-regulated biological processes. To establish CMT subtype-specific proteomic profiles, weighted gene correlation network analysis (WGCNA) was carried out.ResultsWCL and EVs displayed distinct protein abundance signatures while still showing the same increase in adhesion, migration, and motility-related proteins in carcinoma-derived cell lines, and of RNA processing and RNA splicing factors in the adenoma cell line. WGCNA identified CMT stage-specific co-abundant EV proteins, allowing the identification of adenoma and carcinoma EV signatures not seen in WCLs.ConclusionsEVs from CMT cell lines exhibit distinct protein profiles reflecting malignancy state, allowing us to identify potential biomarkers for canine mammary carcinomas, such as biglycan. Our dataset could therefore potentially serve as a basis for the development of a less invasive clinical diagnostic tool for the characterisation of CMT.

Performance of Multimodal Artificial Intelligence Chatbots Evaluated on Clinical Oncology Cases

Multimodal artificial intelligence (AI) chatbots can process complex medical image and text-based information that may improve their accuracy as a clinical diagnostic and management tool compared with unimodal, text-only AI chatbots. However, the difference in medical accuracy of multimodal and text-only chatbots in addressing questions about clinical oncology cases remains to be tested. To evaluate the utility of prompt engineering (zero-shot chain-of-thought) and compare the competency of multimodal and unimodal AI chatbots to generate medically accurate responses to questions about clinical oncology cases. This cross-sectional study benchmarked the medical accuracy of multiple-choice and free-text responses generated by AI chatbots in response to 79 questions about clinical oncology cases with images. A unique set of 79 clinical oncology cases from JAMA Network Learning accessed on April 2, 2024, was posed to 10 AI chatbots. The primary outcome was medical accuracy evaluated by the number of correct responses by each AI chatbot. Multiple-choice responses were marked as correct based on the ground-truth, correct answer. Free-text responses were rated by a team of oncology specialists in duplicate and marked as correct based on consensus or resolved by a review of a third oncology specialist. This study evaluated 10 chatbots, including 3 multimodal and 7 unimodal chatbots. On the multiple-choice evaluation, the top-performing chatbot was chatbot 10 (57 of 79 [72.15%]), followed by the multimodal chatbot 2 (56 of 79 [70.89%]) and chatbot 5 (54 of 79 [68.35%]). On the free-text evaluation, the top-performing chatbots were chatbot 5, chatbot 7, and the multimodal chatbot 2 (30 of 79 [37.97%]), followed by chatbot 10 (29 of 79 [36.71%]) and chatbot 8 and the multimodal chatbot 3 (25 of 79 [31.65%]). The accuracy of multimodal chatbots decreased when tested on cases with multiple images compared with questions with single images. Nine out of 10 chatbots, including all 3 multimodal chatbots, demonstrated decreased accuracy of their free-text responses compared with multiple-choice responses to questions about cancer cases. In this cross-sectional study of chatbot accuracy tested on clinical oncology cases, multimodal chatbots were not consistently more accurate than unimodal chatbots. These results suggest that further research is required to optimize multimodal chatbots to make more use of information from images to improve oncology-specific medical accuracy and reliability.