Median Accuracy Research Articles

Precision pest control using purpose-built uncrewed aerial system (UAS) technology and a novel bait pod system

Controlling invasive species is critical due to their impact on disease transmission, endangerment of native species, and biodiversity loss. While crewed aircraft can effectively distribute bait over large areas to target pests, their use becomes impractical and costly in small, isolated regions with rugged terrain. Though feasible in smaller sites, ground control operations pose challenges such as high expenses, safety risks, and potential worker injuries in hazardous terrain. An innovative approach employing unmanned aerial systems (UASs) for precise bait deployment has been developed to address these issues. Our team engineered a purpose-built system designed specifically for deploying bait within innovative bait pods. Field trials in New Zealand validated its efficacy, with significant improvements observed in subsequent trials due to enhancements in bait pod design. The median deployment accuracy achieved was 1.91 meters from the target, with no statistically significant difference between open and forested areas. This advanced system enables precise bait placement, facilitating pest control in complex landscapes, challenging terrain, and dense vegetation. Its smart functionality and adaptability allow maximum accuracy and efficiency across various aircraft and autopilot systems. This innovative tool holds promise in managing invasive species, complementing existing strategies to expedite ecosystem restoration and safeguard biodiversity.

Cost-sensitive ordinal classification methods to predict SARS-CoV-2 pneumonia severity.

To study the suitability of costsensitive ordinal artificial intelligence-machine learning (AIML) strategies in the prognosis of SARS-CoV-2 pneumonia severity. Observational, retrospective, longitudinal, cohort study in 4 hospitals in Spain. Information regarding demographic and clinical status was supplemented by socioeconomic data and air pollution exposures. We proposed AI-ML algorithms for ordinal classification via ordinal decomposition and for cost-sensitive learning via resampling techniques. For performance-based model selection, we defined a custom score including per-class sensitivities and asymmetric misprognosis costs. 260 distinct AI-ML models were evaluated via 10 repetitions of 5×5 nested cross-validation with hyperparameter tuning. Model selection was followed by the calibration of predicted probabilities. Final overall performance was compared against five well-established clinical severity scores and against a 'standard' (non-cost sensitive, non-ordinal) AI-ML baseline. In our best model, we also evaluated its explainability with respect to each of the input variables. The study enrolled n = 1548 patients: 712 experienced low, 238 medium, and 598 high clinical severity. d = 131 variables were collected, becoming d ' = 148 features after categorical encoding. Model selection resulted in our best-performing AI-ML pipeline having: a) no imputation of missing data, b) no feature selection (i.e. using the full set of d ' features), c) 'Ordered Partitions' ordinal decomposition, d) cost-based reimbalance, and e) a Histogram-based Gradient Boosting classifier. This best model (calibrated) obtained a median accuracy of 68.1% [67.3%, 68.8%] (95% confidence interval), a balanced accuracy of 57.0% [55.6%, 57.9%], and an overall area under the curve (AUC) 0.802 [0.795, 0.808]. In our dataset, it outperformed all five clinical severity scores and the 'standard' AI-ML baseline. We conducted an exhaustive exploration of AI-ML methods designed for both ordinal and cost-sensitive classification, motivated by a real-world application domain (clinical severity prognosis) in which these topics arise naturally. Our model with the best classification performance exploited successfully the ordering information of ground truth classes, coping with imbalance and asymmetric costs. However, these ordinal and cost-sensitive aspects are seldom explored in the literature.

Single-channel EOG sleep staging on a heterogeneous cohort of subjects with sleep disorders

Objective. Sleep staging based on full polysomnography is the gold standard in the diagnosis of many sleep disorders. It is however costly, complex, and obtrusive due to the use of multiple electrodes. Automatic sleep staging based on single-channel electro-oculography (EOG) is a promising alternative, requiring fewer electrodes which could be self-applied below the hairline. EOG sleep staging algorithms are however yet to be validated in clinical populations with sleep disorders. Approach. We utilized the SOMNIA dataset, comprising 774 recordings from subjects with various sleep disorders, including insomnia, sleep-disordered breathing, hypersomnolence, circadian rhythm disorders, parasomnias, and movement disorders. The recordings were divided into train (574), validation (100), and test (100) groups. We trained a neural network that integrated transformers within a U-Net backbone. This design facilitated learning of arbitrary-distance temporal relationships within and between the EOG and hypnogram. Main results. For 5-class sleep staging, we achieved median accuracies of 85.0% and 85.2% and Cohen’s kappas of 0.781 and 0.796 for left and right EOG, respectively. The performance using the right EOG was significantly better than using the left EOG, possibly because in the recommended AASM setup, this electrode is located closer to the scalp. The proposed model is robust to the presence of a variety of sleep disorders, displaying no significant difference in performance for subjects with a certain sleep disorder compared to those without. Significance. The results show that accurate sleep staging using single-channel EOG can be done reliably for subjects with a variety of sleep disorders.

“Research on Music Emotion Analysis and Dance Creation Based on Neural Network”

The one-of-a-kind avenues for self-expression that dancing provides as a form of physical art. Movements of the body, both internal and external, and other forms of expression make up the bulk of a dancer's lexicon. Through passive observation, the viewer is able to catch up on the dancer's intended meaning through their body language. It's the one-of-a-kind way that each dancer finds to convey feeling and story through their body. In this piece, we intelligently oversee dance training and take on challenging nonlinear control problems by combining AI technology and the BP neural networks (BPNN) approach.In this study, the training dance teachers receive in dance language, talking music for dancing, and stage art was evaluated using the BPNN technique and the PCA-BPNN strategy. Median accuracy for the BPNN research model is 85.35 percent as age 80 approaches, whereas for the PCA-BPNN research model it is just 65.64 percent. This demonstrates that the BPNN grading model provides more precise results than the PCA-BPNN model. Using BPNN algorithm-based dancing performed within the context of artificial intelligence technology is one way to attain two goals: the spread and pleasure of beauty, and the harmonious combination of athletics and the arts. These two aims are not mutually exclusive.

Optimizing lightning location accuracy: A study of propagation velocity and time of arrival in long-range lightning location algorithms

The propagation velocity (v) and time of arrival (Ta) are the primary factors affecting long-range lightning location. In this paper, we have enhanced the Very Low Frequency Long-range Lightning Location Network (VLF-LLN) in China in terms of both pulse pairing methods and location algorithms. Then, the algorithm utilized in the VLF-LLN, and four other long-range lightning location algorithms were compared using lightning strike accidents, simulated lightning strike, and other lightning data as the reference values. Building on this foundation, the discussion centers on the effect of v and Ta on location accuracy, with the results showing that an appropriate Ta yields a more significant improvement in location accuracy compared to v. VLF-LLN uses the peak time of the ground wave as Ta avoids the significant lag caused by sky wave, which is present in other algorithms, and therefore has the best location accuracy with a median location accuracy of 1.81 km.

Deep Learning Based Fall Recognition and Forecasting for Reconfigurable Stair-Accessing Service Robots

This paper presents a comprehensive study on fall recognition and forecasting for reconfigurable stair-accessing robots by leveraging deep learning techniques. The proposed framework integrates machine learning algorithms and recurrent neural networks (RNNs), specifically Long Short-Term Memory (LSTM) and Bidirectional LSTM (BiLSTM), for fall detection of service robots on staircases. The reconfigurable stair-accessing robot sTetro serves as the platform, and the fall data required for training models are generated in a simulation environment. The two machine learning algorithms are compared and their effectiveness on the fall recognition task is reported. The results indicate that the BiLSTM model effectively classifies falls with a median categorical accuracy of 94.10% in simulation and 90.02% with limited experiments. Additionally, the BiLSTM model can be used for forecasting, which is practically valuable for making decisions well before the onset of a free fall. This study contributes insights into the design and implementation of fall detection systems for service robots used to navigate staircases through deep learning approaches. Our experimental and simulation data, along with the simulation steps, are available for reference and analysis via the shared link.

ChatGPT's Performance in Cardiac Arrest and Bradycardia Simulations Using the American Heart Association's Advanced Cardiovascular Life Support Guidelines: Exploratory Study.

ChatGPT is the most advanced large language model to date, with prior iterations having passed medical licensing examinations, providing clinical decision support, and improved diagnostics. Although limited, past studies of ChatGPT's performance found that artificial intelligence could pass the American Heart Association's advanced cardiovascular life support (ACLS) examinations with modifications. ChatGPT's accuracy has not been studied in more complex clinical scenarios. As heart disease and cardiac arrest remain leading causes of morbidity and mortality in the United States, finding technologies that help increase adherence to ACLS algorithms, which improves survival outcomes, is critical. This study aims to examine the accuracy of ChatGPT in following ACLS guidelines for bradycardia and cardiac arrest. We evaluated the accuracy of ChatGPT's responses to 2 simulations based on the 2020 American Heart Association ACLS guidelines with 3 primary outcomes of interest: the mean individual step accuracy, the accuracy score per simulation attempt, and the accuracy score for each algorithm. For each simulation step, ChatGPT was scored for correctness (1 point) or incorrectness (0 points). Each simulation was conducted 20 times. ChatGPT's median accuracy for each step was 85% (IQR 40%-100%) for cardiac arrest and 30% (IQR 13%-81%) for bradycardia. ChatGPT's median accuracy over 20 simulation attempts for cardiac arrest was 69% (IQR 67%-74%) and for bradycardia was 42% (IQR 33%-50%). We found that ChatGPT's outputs varied despite consistent input, the same actions were persistently missed, repetitive overemphasis hindered guidance, and erroneous medication information was presented. This study highlights the need for consistent and reliable guidance to prevent potential medical errors and optimize the application of ChatGPT to enhance its reliability and effectiveness in clinical practice.

An novel SDA-CNN few shot domain adaptation framework for silent speech recognition

In BCI (brain-computer interface) applications, it is difficult to obtain enough well-labeled EEG data because of the expensive annotation and time-consuming data capture procedure. Conventional classification techniques that repurpose EEG data across domains and subjects lead to significant decreases in silent speech recognition classification accuracy. This research provides a supervised domain adaptation using Convolutional Neural Network framework (SDA-CNN) to tackle this problem. The objective is to provide a solution for the distribution divergence issue in the categorization of speech recognition across domains. The suggested framework involves taking raw EEG data and deriving deep features from it and the proposed feature selection method also retrieves the statistical features from the corresponding channels. Moreover, it attempts to minimize the distribution divergence caused by variations in people and settings by aligning the correlation of both the source and destination EEG characteristic dissemination. In order to obtain minimal feature distribution divergence and discriminative classification performance, the last stage entails simultaneously optimizing the loss of classification and adaption loss. The usefulness of the suggested strategy in reducing distributed divergence among the source and target Electroencephalography (EEG) data is demonstrated by extensive experiments carried out on KaraOne datasets. The suggested method achieves an average accuracy for classification of 87.4% for single-subject classification and a noteworthy average class accuracy of 88.6% for cross-subject situations, which shows that it surpasses existing cutting-edge techniques in thinking tasks. Regarding the speaking task, the model’s median classification accuracy for single-subject categorization is 86.8%, while its average classification accuracy for cross-subject classification is 87.8%. These results underscore the innovative approach of SDA-CNN to mitigating distribution discrepancies while optimizing classification performance, offering a promising avenue to enhance accuracy and adaptability in brain-computer interface applications.

Predicting Male Infertility Using Artificial Neural Networks: A Review of the Literature.

Male infertility is a relevant public health problem, but there is no systematic review of the different machine learning (ML) models and their accuracy so far. The present review aims to comprehensively investigate the use of ML algorithms in predicting male infertility, thus reporting the accuracy of the used models in the prediction of male infertility as a primary outcome. Particular attention will be paid to the use of artificial neural networks (ANNs). A comprehensive literature search was conducted in PubMed, Scopus, and Science Direct between 15 July and 23 October 2023, conducted under the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines. We performed a quality assessment of the included studies using the recommended tools suggested for the type of study design adopted. We also made a screening of the Risk of Bias (RoB) associated with the included studies. Thus, 43 relevant publications were included in this review, for a total of 40 different ML models detected. The studies included reported a good quality, even if RoB was not always good for all the types of studies. The included studies reported a median accuracy of 88% in predicting male infertility using ML models. We found only seven studies using ANN models for male infertility prediction, reporting a median accuracy of 84%.

Surgical outcomes of intraocular lens iris suture fixation in eyes with residual capsule support

Purpose: To evaluate the safety and refractive outcomes of eyes after intraocular lens (IOL) iris suture fixation (ISF). Setting: Private practice, Los Angeles, California. Design: Nonrandomized and unmasked retrospective chart review. Methods: Eyes that underwent IOL exchange or repositioning with ISF with at least 270 degrees of capsular support were included. Eyes with less than 270 degrees of capsular support and eyes with iris damage were excluded. The primary outcome measures included incidence of cystoid macular edema (CME), IOL dislocation requiring refixation, and chronic inflammation. Secondary outcome measures included worsening intraocular pressure (IOP) control, retinal tear or detachment, worsening of corrected distance visual acuity (CDVA), and corneal decompensation. Refractive outcomes for 26 subgrouped eyes included mean and median spherical equivalent refraction accuracy (SERA), and percentage of eyes within 0.5 diopter (D) and 1 D of the refractive target. Results: The study included 53 eyes of 50 patients. CME: 2/53 (3.8%), IOL dislocation requiring refixation: 2/53 (3.8%), chronic inflammation: 1/53 (1.9%), worsening IOP control: 5/53 (9.4%), retinal tear or detachment: 2/53 (3.8%). No patient experienced worsening of CDVA from baseline or corneal decompensation. Mean SERA ± SD −0.35 ± 0.29 D, median SERA −0.37 D. Of the 26 eyes subgrouped for refractive analysis, 73% were within 0.5 D and 100% were within 1 D of the desired refractive outcome. Conclusions: ISF can offer stability for sulcus-fixated IOLs provided there is some residual capsule support. Although there are measurable complications, there is a relatively low side effect profile. The refractive error tended to be myopic, indicating the need for further refinement of IOL power predictive formulas.

MRPSO-Loc: Multi-targets UHF RFID localization system based on MRPSO algorithm

Existing RFID localization techniques mainly focus on the location of a single target, which do not fully utilize the multi-targets identification merit of UHF RFID systems. In this paper, a multi-targets localization algorithm called merged re-shaken particle swarm optimization (MRPSO) is proposed. The ranging information and the angle information, which are obtained by using multi-frequency phase information with the help of frequency hopping technique, are merged together during the iteration process of MRPSO. In addition, based on the traditional PSO algorithm, a low weight particle filtering method was proposed, and a re-shaken scheme was introduced to improve localization accuracy. By allocating particles to each target, MRPSO can obtain each target location simultaneously after PSO process, which avoids time consuming. The experimental results show that MRPSO can achieve single-station multi-target localization with a median accuracy of 9.64 cm.

A novel algorithm for ocean chlorophyll-a concentration using MODIS Aqua data

The ability to infer ocean chlorophyll-a concentrations (Chla) from spaceborne instruments is key to assessments of global ocean productivity and monitoring of water quality. Here, we present a novel parametric algorithm, OCG, trained on a set of global in situ high-performance liquid chromatography (HPLC) data that leverages Level-3 remote sensing reflectance (Rrs) products from the Moderate Resolution Imaging Spectroradiometer (MODIS) aboard the Aqua satellite. The OCG algorithm leverages more bands than existing algorithms and also provides pixel-wise uncertainty assessments that enable the calculation of the probability of exceeding specific Chla thresholds. This feature has significant implications for water quality management, particularly in monitoring harmful algal blooms. The OCG surpasses existing algorithms in bias and accuracy without overfitting, especially in coastal areas, where it outperforms the current standard product (CI OC3) by 20 % in median symmetric accuracy. Moreover, the OCG reduces the signed symmetric percentage bias (SSPB) in coastal regions from 41 % (CI OC3) to below 5 %. Globally, the OCG algorithm yields lower Chla in coastal regions, the Southern Ocean and the Mediterranean Sea, and higher values in the open ocean, particularly in ocean gyres and polar regions. For the Chesapeake Bay and the Baltic Sea, for example, daily OCG estimates for 2002 to 2021 are, on average, 2.9 g/L and 3.7 g/L lower than CI OC3 estimates, respectively. The presented approach also shows great potential for other existing and upcoming sensors, enabling widespread application in remote sensing.

The Role of Artificial Intelligence in Endocrine Management: Assessing ChatGPT's Responses to Prolactinoma Queries.

This research investigates the utility of Chat Generative Pre-trained Transformer (ChatGPT) in addressing patient inquiries related to hyperprolactinemia and prolactinoma. A set of 46 commonly asked questions from patients with prolactinoma were presented to ChatGPT and responses were evaluated for accuracy with a 6-point Likert scale (1: completely inaccurate to 6: completely accurate) and adequacy with a 5-point Likert scale (1: completely inadequate to 5: completely adequate). Two independent endocrinologists assessed the responses, based on international guidelines. Questions were categorized into groups including general information, diagnostic process, treatment process, follow-up, and pregnancy period. The median accuracy score was 6.0 (IQR, 5.4-6.0), and the adequacy score was 4.5 (IQR, 3.5-5.0). The lowest accuracy and adequacy score assigned by both evaluators was two. Significant agreement was observed between the evaluators, demonstrated by a weighted κ of 0.68 (p = 0.08) for accuracy and a κ of 0.66 (p = 0.04) for adequacy. The Kruskal-Wallis tests revealed statistically significant differences among the groups for accuracy (p = 0.005) and adequacy (p = 0.023). The pregnancy period group had the lowest accuracy score and both pregnancy period and follow-up groups had the lowest adequacy score. In conclusion, ChatGPT demonstrated commendable responses in addressing prolactinoma queries; however, certain limitations were observed, particularly in providing accurate information related to the pregnancy period, emphasizing the need for refining its capabilities in medical contexts.

Abstract 6560: Cell-free RNA analysis for non-invasive cancer detection and characterization

Abstract Blood-based liquid biopsies enable non-invasive characterization of cancers. Cell-free RNA (cfRNA) analysis could potentially complement circulating tumor DNA (ctDNA) analysis and allow broader molecular characterization of cancers but has not been extensively explored. Here, we describe RARE-Seq, a novel cfRNA sequencing method designed to maximize sensitivity by targeting rare, tissue-specific transcripts, and simultaneously detecting both tumor-derived expression signatures and somatic mutations in circulating tumor RNA (ctRNA). To interrogate samples for specific gene expression signatures, we developed and optimized a framework to measure tumor- and tissue-specific enrichment scores (ES) within cfRNA. We then empirically determined the limit of detection (LOD95) of RARE-Seq using admixtures of tumor cell line RNA and control cfRNA across a wide range of fractions. We found RARE-Seq to have an estimated LOD95 of 0.05% for cancer-derived RNA, which was 52-fold better than whole transcriptome cfRNA profiling. To explore its potential clinical utility, we profiled 192 samples from 170 subjects with various cancers and non-cancer controls using RARE-Seq. Non-small cell lung cancer (NSCLC) expression signatures were significantly detected in 40 out of 51 NSCLC samples (78% sensitivity at 90% specificity). Detection increased with stage, ranging from 40% in stage I to 87% in stage IV. In a subset of stage IV NSCLC patients with matching ctDNA data (n=16), ctRNA levels were significantly correlated with ctDNA allele frequencies (Pearson R=0.84, P=0.005). Intriguingly, 86% of samples with undetectable ctDNA contained detectable ctRNA, suggesting potential complementarity of ctRNA and ctDNA analysis for cancer detection. We also developed and benchmarked a direct ctRNA genotyping approach, allowing noninvasive detection of driver mutations in 42% of patients with advanced NSCLC. When applied to other tumors, RARE-Seq demonstrated 80-100% sensitivity at 90% specificity for detection of pancreatic adenocarcinomas (PAAD), prostate adenocarcinomas (PRAD), and hepatocellular carcinomas (LIHC). Furthermore, when evaluating tissue of origin (TOO) classification performance, RARE-Seq achieved 85% median accuracy for the top predicted tissue and 96% median accuracy for the top two tissues. These results highlight the potential value of ultrasensitive cfRNA analysis, providing proof-of-principle for detection of multiple tumor types using RARE-Seq. Citation Format: Monica C. Nesselbush, Bogdan A. Luca, Young-Jun Jeon, Isabel Jabara, Catherine B. Meador, Michael S. Binkley, Nova Xu, Angela B. Hui, Andrea Garofalo, William Y. Shi, Kevin J. Liu, Takeshi Sugio, Noah Kastelowitz, Emily G. Hamilton, Rene F. Bonilla, Yi Peng Wang, Alice Jiang, Brianna Lau, Jordan Eichholz, Mandeep Banwait, Joseph Schroers-Martin, Jan Boegeholz, Daniel A. King, Mohammad S. Esfahani, Tyler Raclin, Robert Tibshirani, Sandy Srinivas, Viswam S. Nair, James M. Isbell, Bob T. Li, Zofia Piotrowska, Lecia V. Sequist, Aaron N. Hata, Joel W. Neal, Heather A. Wakelee, Andrew J. Gentles, Ash A. Alizadeh, Maximilian Diehn. Cell-free RNA analysis for non-invasive cancer detection and characterization [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2024; Part 1 (Regular Abstracts); 2024 Apr 5-10; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2024;84(6_Suppl):Abstract nr 6560.

Predicting risk of endometrial failure: a biomarker signature that identifies a novel disruption independent of endometrial timing in patients undergoing hormonal replacement cycles

To propose a new gene expression signature that identifies endometrial disruptions independent of endometrial luteal phase timing and predicts if patients are at risk of endometrial failure. Multicentric, prospective study. Reproductive medicine research department in a public hospital affiliated with private fertility clinics and a reproductive genetics laboratory. Caucasian women (n = 281; 39.4 ± 4.8 years old with a body mass index of 22.9 ± 3.5 kg/m2) undergoing hormone replacement therapy between July 2018 and July 2021. Endometrial samples from 217 patients met RNA quality criteria for signature discovery and analysis. Endometrial biopsies collected in the mid-secretory phase. Endometrial luteal phase timing-corrected expression of 404 genes and reproductive outcomes of the first single embryo transfer (SET) after biopsy collection to identify prognostic biomarkers of endometrial failure. Removal of endometrial timing variation from gene expression data allowed patients to be stratified into poor (n = 137) or good (n = 49) endometrial prognosis groups on the basis of their clinical and transcriptomic profiles. Significant differences were found between endometrial prognosis groups in terms of reproductive rates: pregnancy (44.6% vs. 79.6%), live birth (25.6% vs. 77.6%), clinical miscarriage (22.2% vs. 2.6%), and biochemical miscarriage (20.4% vs. 0%). The relative risk of endometrial failure for patients predicted as a poor endometrial prognosis was 3.3 times higher than those with a good prognosis. The differences in gene expression between both profiles were proposed as a biomarker, coined the endometrial failure risk (EFR) signature. Poor prognosis profiles were characterized by 59 upregulated and 63 downregulated genes mainly involved in regulation (17.0%), metabolism (8.4%), immune response, and inflammation (7.8%). This EFR signature had a median accuracy of 0.92 (min = 0.88, max = 0.94), median sensitivity of 0.96 (min = 0.91, max = 0.98), and median specificity of 0.84 (min = 0.77, max = 0.88), positioning itself as a promising biomarker for endometrial evaluation. The EFR signature revealed a novel endometrial disruption, independent of endometrial luteal phase timing, present in 73.7% of patients. This EFR signature stratified patients into 2 significantly distinct and clinically relevant prognosis profiles providing opportunities for personalized therapy. Nevertheless, further validations are needed before implementing this gene signature as an artificial intelligence (AI)-based tool to reduce the risk of patients experiencing endometrial failure.

The performance of artificial intelligence models in generating responses to general orthodontic questions: ChatGPT vs Google Bard

This study aimed to evaluate and compare the performance of 2 artificial intelligence (AI) models, Chat Generative Pretrained Transformer-3.5 (ChatGPT-3.5; OpenAI, San Francisco, Calif) and Google Bidirectional Encoder Representations from Transformers (Google Bard; Bard Experiment, Google, Mountain View, Calif), in terms of response accuracy, completeness, generation time, and response length when answering general orthodontic questions. A team of orthodontic specialists developed a set of 100 questions in 10 orthodontic domains. One author submitted the questions to both ChatGPT and Google Bard. The AI-generated responses from both models were randomly assigned into 2 forms and sent to 5 blinded and independent assessors. The quality of AI-generated responses was evaluated using a newly developed tool for accuracy of information and completeness. In addition, response generation time and length were recorded. The accuracy and completeness of responses were high in both AI models. The median accuracy score was 9 (interquartile range [IQR]: 8-9) for ChatGPT and 8 (IQR: 8-9) for Google Bard (Median difference: 1; P<0.001). The median completeness score was similar in both models, with 8 (IQR: 8-9) for ChatGPT and 8 (IQR: 7-9) for Google Bard. The odds of accuracy and completeness were higher by 31% and 23% in ChatGPT than in Google Bard. Google Bard's response generation time was significantly shorter than that of ChatGPT by 10.4 second/question. However, both models were similar in terms of response length generation. Both ChatGPT and Google Bard generated responses were rated with a high level of accuracy and completeness to the posed general orthodontic questions. However, acquiring answers was generally faster using the Google Bard model.

Explainable 3D CNN based on baseline breast DCE-MRI to give an early prediction of pathological complete response to neoadjuvant chemotherapy

BackgroundSo far, baseline Dynamic Contrast-Enhanced Magnetic Resonance Imaging (DCE-MRI) has played a key role for the application of sophisticated artificial intelligence-based models using Convolutional Neural Networks (CNNs) to extract quantitative imaging information as earlier indicators of pathological Complete Response (pCR) achievement in breast cancer patients treated with neoadjuvant chemotherapy (NAC). However, these models did not exploit the DCE-MRI exams in their full geometry as 3D volume but analysed only few individual slices independently, thus neglecting the depth information. MethodThis study aimed to develop an explainable 3D CNN, which fulfilled the task of pCR prediction before the beginning of NAC, by leveraging the 3D information of post-contrast baseline breast DCE-MRI exams. Specifically, for each patient, the network took in input a 3D sequence containing the tumor region, which was previously automatically identified along the DCE-MRI exam. A visual explanation of the decision-making process of the network was also provided. ResultsTo the best of our knowledge, our proposal is competitive than other models in the field, which made use of imaging data alone, reaching a median AUC value of 81.8%, 95%CI [75.3%; 88.3%], a median accuracy value of 78.7%, 95%CI [74.8%; 82.5%], a median sensitivity value of 69.8%, 95%CI [59.6%; 79.9%] and a median specificity value of 83.3%, 95%CI [82.6%; 84.0%], respectively. The median and CIs were computed according to a 10-fold cross-validation scheme for 5 rounds. ConclusionFinally, this proposal holds high potential to support clinicians on non-invasively early pursuing or changing patient-centric NAC pathways.

Evaluation of ChatGPT for Pelvic Floor Surgery Counseling.

Large language models are artificial intelligence applications that can comprehend and produce human-like text and language. ChatGPT is one such model. Recent advances have increased interest in the utility of large language models in medicine. Urogynecology counseling is complex and time-consuming. Therefore, we evaluated ChatGPT as a potential adjunct for patient counseling. Our primary objective was to compare the accuracy and completeness of ChatGPT responses to information in standard patient counseling leaflets regarding common urogynecological procedures. Seven urogynecologists compared the accuracy and completeness of ChatGPT responses to standard patient leaflets using 5-point Likert scales with a score of 3 being "equally accurate" and "equally complete," and a score of 5 being "much more accurate" and much more complete, respectively. This was repeated 3 months later to evaluate the consistency of ChatGPT. Additional analysis of the understandability and actionability was completed by 2 authors using the Patient Education Materials Assessment Tool. Analysis was primarily descriptive. First and second ChatGPT queries were compared with the Wilcoxon signed rank test. The median (interquartile range) accuracy was 3 (2-3) and completeness 3 (2-4) for the first ChatGPT query and 3 (3-3) and 4 (3-4), respectively, for the second query. Accuracy and completeness were significantly higher in the second query (P < 0.01). Understandability and actionability of ChatGPT responses were lower than the standard leaflets. ChatGPT is similarly accurate and complete when compared with standard patient information leaflets for common urogynecological procedures. Large language models may be a helpful adjunct to direct patient-provider counseling. Further research to determine the efficacy and patient satisfaction of ChatGPT for patient counseling is needed.

Utility of artificial intelligence-based large language models in ophthalmic care.

With the introduction of ChatGPT, artificial intelligence (AI)-based large language models (LLMs) are rapidly becoming popular within the scientific community. They use natural language processing to generate human-like responses to queries. However, the application of LLMs and comparison of the abilities among different LLMs with their human counterparts in ophthalmic care remain under-reported. Hitherto, studies in eye care have demonstrated the utility of ChatGPT in generating patient information, clinical diagnosis and passing ophthalmology question-based examinations, among others. LLMs' performance (median accuracy, %) is influenced by factors such as the iteration, prompts utilised and the domain. Human expert (86%) demonstrated the highest proficiency in disease diagnosis, while ChatGPT-4 outperformed others in ophthalmology examinations (75.9%), symptom triaging (98%) and providing information and answering questions (84.6%). LLMs exhibited superior performance in general ophthalmology but reduced accuracy in ophthalmic subspecialties. Although AI-based LLMs like ChatGPT are deemed more efficient than their human counterparts, these AIs are constrained by their nonspecific and outdated training, no access to current knowledge, generation of plausible-sounding 'fake' responses or hallucinations, inability to process images, lack of critical literature analysis and ethical and copyright issues. A comprehensive evaluation of recently published studies is crucial to deepen understanding of LLMs and the potential of these AI-based LLMs. Ophthalmic care professionals should undertake a conservative approach when using AI, as human judgement remains essential for clinical decision-making and monitoring the accuracy of information. This review identified the ophthalmic applications and potential usages which need further exploration. With the advancement of LLMs, setting standards for benchmarking and promoting best practices is crucial. Potential clinical deployment requires the evaluation of these LLMs to move away from artificial settings, delve into clinical trials and determine their usefulness in the real world.

Histopathologic evaluation of gastric intestinal metaplasia in non-neoplastic biopsy specimens: Accuracy and interobserver reliability among general pathologists and pathology residents

ObjectivesThis study aimed to evaluate the accuracy and interobserver reliability of diagnosing and subtyping gastric intestinal metaplasia (IM) among general pathologists and pathology residents at a university hospital in Thailand, focusing on the challenges in the histopathologic evaluation of gastric IM for less experienced practitioners. MethodsThe study analyzed 44 non-neoplastic gastric biopsies, using a consensus diagnosis of gastrointestinal pathologists as the reference standard. Participants included 6 general pathologists and 9 pathology residents who assessed gastric IM and categorized its subtype (complete, incomplete, or mixed) on digital slides. After initial evaluations and receiving feedback, participants reviewed specific images of gastric IM, as agreed by experts. Following a one-month washout period, a reevaluation of the slides was conducted. ResultsDiagnostic accuracy, interobserver reliability, and time taken for diagnosis improved following training, with general pathologists showing higher accuracies than residents (median accuracy of gastric IM detection: 100 % vs. 97.7 %). Increased years of experience were associated with more IM detection accuracy (p-value<0.05). However, the overall median accuracy for diagnosing incomplete IM remained lower than for complete IM (86.4 % vs. 97.7 %). After training, diagnostic errors occurred in 6 out of 44 specimens (13.6 %), reported by over 40 % of participants. Errors involved omitting 5 slides with incomplete IM and 1 with complete IM, all showing a subtle presence of IM. ConclusionsThe study highlights the diagnostic challenges in identifying incomplete gastric IM, showing notable discrepancies in accuracy and interobserver agreement. It underscores the need for better diagnostic protocols and training to enhance detection and management outcomes.