Design of HL7 FHIR Profiles for Pathology Reports Integrated with Pathology Images.

This paper describes the development of Health Level Seven Fast Healthcare Interoperability Resource (FHIR) profiles for pathology reports integrated with whole slide images and clinical data to create a pathology research database. A report template was designed to collect structured reports, enabling pathologists to select structured terms based on a checklist, allowing for the standardization of terms used to describe tumor features. We gathered and analyzed 190 non-small-cell lung cancer pathology reports in free text format, which were then structured by mapping the itemized vocabulary to FHIR observation resources, using international standard terminologies, such as the International Classification of Diseases, LOINC, and SNOMED CT. The resulting FHIR profiles were published as an implementation guide, which includes 25 profiles for essential data elements, value sets, and structured definitions for integrating clinical data and pathology images associated with the pathology report. These profiles enable the exchange of structured data between systems and facilitate the integration of pathology data into electronic health records, which can improve the quality of care for patients with cancer.

Optimizing age-related hearing risk predictions: an advanced machine learning integration with HHIE-S.

OBJECTIVES: The elderly are disproportionately affected by age-related hearing loss (ARHL). Despite being a well-known tool for ARHL evaluation, the Hearing Handicap Inventory for the Elderly Screening version (HHIE-S) has only traditionally been used for direct screening using self-reported outcomes. This work uses a novel integration of machine learning approaches to improve the predicted accuracy of the HHIE-S tool for ARHL in older adults. METHODS: We employed a dataset that was gathered between 2016 and 2018 and included 1,526 senior citizens from several Taipei City Hospital branches. 80% of the data were used for training (n = 1220) and 20% were used for testing (n = 356). XGBoost, Gradient Boosting, and LightGBM were among the machine learning models that were only used and assessed on the training set. In order to prevent data leakage and overfitting, the Light Gradient Boosting Machine (LGBM) model-which had the greatest AUC of 0.83 (95% CI 0.81-0.85)-was then only used on the holdout testing data. RESULTS: On the testing set, the LGBM model showed a strong AUC of 0.82 (95% CI 0.79-0.86), far outperforming conventional techniques. Notably, several HHIE-S items and age were found to be significant characteristics. In contrast to traditional HHIE research, which concentrates on the psychological effects of hearing loss, this study combines cutting-edge machine learning techniques-specifically, the LGBM classifier-with the HHIE-S tool. The incorporation of SHAP values enhances the interpretability of the model's predictions and provides a more comprehensive comprehension of the significance of various aspects. CONCLUSIONS: Our methodology highlights the great potential that arises from combining machine learning with validated hearing evaluation instruments such as the HHIE-S. Healthcare practitioners can anticipate ARHL more accurately thanks to this integration, which makes it easier to intervene quickly and precisely.

Federated machine learning for predicting acute kidney injury in critically ill patients: a multicenter study in Taiwan.

Purpose: To address the contentious data sharing across hospitals, this study adopted a novel approach, federated learning (FL), to establish an aggregate model for acute kidney injury (AKI) prediction in critically ill patients in Taiwan. Methods: This study used data from the Critical Care Database of Taichung Veterans General Hospital (TCVGH) from 2015 to 2020 and electrical medical records of the intensive care units (ICUs) between 2018 and 2020 of four referral centers in different areas across Taiwan. AKI prediction models were trained and validated thereupon. An FL-based prediction model across hospitals was then established. Results: The study included 16,732 ICU admissions from the TCVGH and 38,424 ICU admissions from the other four hospitals. The complete model with 60 features and the parsimonious model with 21 features demonstrated comparable accuracies using extreme gradient boosting, neural network (NN), and random forest, with an area under the receiver-operating characteristic (AUROC) curve of approximately 0.90. The Shapley Additive Explanations plot demonstrated that the selected features were the key clinical components of AKI for critically ill patients. The AUROC curve of the established parsimonious model for external validation at the four hospitals ranged from 0.760 to 0.865. NN-based FL slightly improved the model performance at the four centers. Conclusion: A reliable prediction model for AKI in ICU patients was developed with a lead time of 24 h, and it performed better when the novel FL platform across hospitals was implemented. Supplementary Information: The online version contains supplementary material available at 10.1007/s13755-023-00248-5.

Using artificial intelligence algorithms to predict the overall survival of hemodialysis patients during the COVID-19 pandemic: A prospective cohort study.

BACKGROUND: Hemodialysis (HD) patients are a vulnerable population at high risk for severe complications from COVID-19. The impact of partial COVID-19 vaccination on the survival of HD patients remains uncertain. This prospective cohort study was designed to use artificial intelligence algorithms to predict the survival impact of partial COVID-19 vaccination in HD patients. METHODS: A cohort of 433 HD patients was used to develop machine-learning models based on a subset of clinical features assessed between July 1, 2021, and April 29, 2022. The patient cohort was randomly split into training (80%) and testing (20%) sets for model development and evaluation. Machine-learning models, including categorical boosting (CatBoost), light gradient boosting machines (LightGBM), RandomForest, and extreme gradient boosting models (XGBoost), were applied to evaluate their discriminative performance using the patient cohorts. RESULTS: Among these models, LightGBM achieved the highest F1 score of 0.95, followed by CatBoost, RandomForest, and XGBoost, with area under the receiver operating characteristic curve values of 0.94 on the testing dataset. The SHapley Additive explanation summary plot derived from the XGBoost model indicated that key features such as age, albumin, and vaccination details had a significant impact on survival. Furthermore, the fully vaccinated group exhibited higher levels of anti-spike (S) receptor-binding domain antibodies. CONCLUSION: This prospective cohort study involved using artificial intelligence algorithms to predict overall survival in HD patients during the COVID-19 pandemic. These predictive models assisted in identifying high-risk individuals and guiding vaccination strategies for HD patients, ultimately improving overall prognosis. Further research is warranted to validate and refine these predictive models in larger and more diverse populations of HD patients.

Survival benefit of a third dose of the COVID-19 vaccine among hemodialysis patients: A prospective cohort study.

BACKGROUND: Hemodialysis (HD) patients are particularly vulnerable to severe coronavirus disease 2019 (COVID-19) due to their immunocompromised state and comorbid conditions. Timely vaccination could be the most effective strategy to reduce morbidity and mortality. However, data on the survival benefit of the COVID-19 vaccine against severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infection and death among HD patients are limited, especially during the Omicron-dominant period. METHODS: In this prospective hospital-based cohort study, we identified HD patients from July 1, 2021, to April 29, 2022. The patients were divided into fully vaccinated and partially vaccinated groups. We compared the humoral response, risk of developing SARS-CoV-2 infection, and all-cause mortality between the two groups. RESULTS: Among the 440 HD patients included, 152 patients were fully vaccinated, and 288 patients were partially vaccinated. Patients in the fully vaccinated group exhibited higher anti-spike protein receptor-binding domain (S protein RBD) antibody levels and lower risks of all-cause mortality (adjusted hazard ratio, 0.35; 95% confidence interval, 0.17-0.73; p = 0.005) than the partially vaccinated group. However, the risk for SARS-CoV-2 infection did not significantly differ between the two groups. Irrespective of the number of vaccinations, the risk of all-cause mortality was lower in patients with anti-S protein RBD antibody levels in the higher tertile. CONCLUSION: A third dose of the COVID-19 vaccine was associated with a decreased risk of all-cause mortality among HD patients during the Omicron-dominant period. A higher post-vaccination anti-S protein RBD antibody level was also associated with a lower risk of mortality.

Fortifying Health Care Intellectual Property Transactions With Blockchain.

BACKGROUND: Intellectual property (IP) is a substantial competitive advantage in the health care industry. However, the COVID-19 pandemic highlighted the need for open innovation and collaboration for the greater good. Despite this, the industry faces challenges with innovation owing to organizational and departmental barriers. A secure platform is necessary to facilitate IP sharing without compromising the rights of IP owners. OBJECTIVE: This study proposes a blockchain-based framework to secure IP transactions in health care and bring social impact. METHODS: This study reviews existing researches, publications, practical cases, firm and organization websites, and conferences related to blockchain technology, blockchain in health care, blockchain in IP management, IP pledge research, and practice of IP management blockchain. The platform architecture has 7 components: pledgers, advanced research technology (ART), IP pledge platforms, IP databases, health care research, seeking ART, and transaction condition setting. These components work together seamlessly to support the sharing and pledging of ART and knowledge, while ensuring the platform's transparency, security, and trust. RESULTS: The open IP pledge framework can promote technology dissemination and use, reduce research and development costs, foster collaboration, and serve the public interest. Medical organizations' leadership and support and active participation from stakeholders are necessary for success. By leveraging blockchain technology, the platform ensures tamper-proof and transparent transactions and protects the rights of IP owners. In addition, the platform offers incentive mechanisms through pledge tokens that encourage stakeholders to share their ART and contribute to the platform. CONCLUSIONS: Overall, the proposed framework can facilitate technological innovation, tackle various challenges, and secure IP transactions. It provides a secure platform for stakeholders to share their IP without compromising their rights, promoting collaboration and progress in the health care industry. The implementation of the framework has the potential to revolutionize the industry's approach to innovation, allowing a more open and collaborative environment driven by the greater good.

Characteristics of breast cancers detected by screening mammography in Taiwan: a single institute's experience.

BACKGROUND/AIM: Breast cancer is the most common female malignancy in the world. Nearly ninety percent of screening-detected breast cancers were diagnosed with earlier stages of 0 to II in Taiwan. It's widely acknowledged that mammography screening of breast cancer can achieve the goal of early diagnosis and treatment in terms of preventive task while neglected interval cancers are associated with unfavorable tumor characteristics and worse outcomes. The purpose of this study was to explore the characteristics of screening-detected breast cancers in Taiwan. MATERIALS AND METHODS: Both screening and diagnostic mammography examinations with accompanied BI-RADS (breast imaging-reporting and data system) classification were extracted from the health information system and linked to cancer registry in Taiwan. Enrolled population included those attending their first mammography between 2012 and 2016, excluding subjects with previous breast cancer, or with missing or incomplete data. We compared treatment outcomes between breast cancers with either initial screening or diagnostic mammography (control group), and investigated the compositions of breast cancers detected by screening mammography through direct chart reviews. RESULTS: A total of 84,246 screening and 61,230 diagnostic mammography sessions were performed from 2010 to 2020. More positive results (BI-RADS 0, 3, 4 and 5) were observed for those attending the first diagnostic than the first screening mammography (43.58% versus 16.12%, p < 0.001). Earlier stages (0 and I) distribution (92% versus 81%, p < 0.0001), better survivorship (overall survival: 96.91% versus 92.17%, p = 0.007) and a lower HER2 (human epidermal growth factor receptor II) positive status and lower tumor grade were noted in breast cancers with initial screening rather than diagnostic mammography. Among 26,103 mammography screening invitees between 2012 and 2016, 325 breast cancers were ascertained from cancer registry. Of these, 234 had one, 72 had two and 19 had three episodes of mammography before cancer diagnosis. Extensive chart reviews revealed that women with and without breast symptoms constituted 29.9 and 70.1% of the 325 screening-detected breast cancers, with the latter further divided into false negative results (interval cancer), diagnosed at the first mammography, diagnostic at the secondary or subsequent mammography and those with a delayed biopsy or confirmatory imaging constituted (5.2, 47, 10.5 and 7.4%). CONCLUSION: Screening-detected breast cancers were a mixture of women with and without symptoms, those with a false negative result, true negative results with cancer detected at subsequent mammography and non-adherers. Despite this, efficacy of mammography screening was ascertained in Taiwan from this study. To further enhance earlier detection and decrease false negativity, the impact of repeated mammography, and additional sonography for symptomatic women, compliance following a positive screening mammography should not be overemphasized.

Prediction of the risk of developing end-stage renal diseases in newly diagnosed type 2 diabetes mellitus using artificial intelligence algorithms.

OBJECTIVES: Type 2 diabetes mellitus (T2DM) imposes a great burden on healthcare systems, and these patients experience higher long-term risks for developing end-stage renal disease (ESRD). Managing diabetic nephropathy becomes more challenging when kidney function starts declining. Therefore, developing predictive models for the risk of developing ESRD in newly diagnosed T2DM patients may be helpful in clinical settings. METHODS: We established machine learning models constructed from a subset of clinical features collected from 53,477 newly diagnosed T2DM patients from January 2008 to December 2018 and then selected the best model. The cohort was divided, with 70% and 30% of patients randomly assigned to the training and testing sets, respectively. RESULTS: The discriminative ability of our machine learning models, including logistic regression, extra tree classifier, random forest, gradient boosting decision tree (GBDT), extreme gradient boosting (XGBoost), and light gradient boosting machine were evaluated across the cohort. XGBoost yielded the highest area under the receiver operating characteristic curve (AUC) of 0.953, followed by extra tree and GBDT, with AUC values of 0.952 and 0.938 on the testing dataset. The SHapley Additive explanation summary plot in the XGBoost model illustrated that the top five important features included baseline serum creatinine, mean serum creatine within 1 year before the diagnosis of T2DM, high-sensitivity C-reactive protein, spot urine protein-to-creatinine ratio and female gender. CONCLUSIONS: Because our machine learning prediction models were based on routinely collected clinical features, they can be used as risk assessment tools for developing ESRD. By identifying high-risk patients, intervention strategies may be provided at an early stage.

Is There an Association between Concurrent Epstein-Barr Virus Infection and Sudden Hearing Loss?-A Case-Control Study in an East Asian Population.

Viral infection serves as the crucial etiology for the development of sudden sensorineural hearing loss (SSNHL). We aimed to investigate whether there is an association between concurrent Epstein-Barr virus (EBV) infection and SSNHL in an East Asian population. Patients who were older than 18 years of age and met the criteria of sudden hearing loss without an identifiable etiology were enrolled from July 2021 until June 2022, followed by the serological testing of IgA antibody responses against EBV-specific early antigen (EA) and viral capsid antigen (VCA) with an indirect hemagglutination assay (IHA) and real-time quantitative polymerase chain reaction (qPCR) of EBV DNA in serum before the treatment was initiated. After the treatment for SSNHL, post-treatment audiometry was performed to record the treatment response and degree of recovery. Among the 29 patients included during enrollment, 3 (10.3%) had a positive qPCR result for EBV. In addition, a trend of poor recovery of hearing thresholds was noted for those patients with a higher viral PCR titer. This is the first study to use real-time PCR to detect possible concurrent EBV infection in SSNHL. Our study demonstrated that approximately one-tenth of the enrolled SSNHL patients had evidence of concurrent EBV infection, as reflected by the positive qPCR test results, and a negative trend between hearing gain and the viral DNA PCR level was found within the affected cohort after steroid therapy. These findings indicate a possible role for EBV infection in East Asian patients with SSNHL. Further larger-scale research is needed to better understand the potential role and underlying mechanism of viral infection in the etiology of SSNHL.

Is N -acetylcysteine effective in treating patients with coronavirus disease 2019? A meta-analysis.

BACKGROUND: Coronavirus disease 2019 (COVID-19) is a global pandemic caused by severe acute respiratory syndrome coronavirus type 2 (SARS-CoV-2). It has brought tremendous challenges to public health and medical systems around the world. The current strategy for drug repurposing has accumulated some evidence on the use of N -acetylcysteine (NAC) in treating patients with COVID-19. However, the evidence remains debated. METHODS: We performed the systematic review and meta-analysis that complies with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines. Five databases and reference lists were searched from inception to May 14, 2022. Studies evaluating the efficacy of NAC in treating patients with COVID-19 were regarded as eligible. The review was registered prospectively on PROSPERO (CRD42022332791). RESULTS: Of 778 records identified from the preliminary search, four studies were enrolled in the final qualitative review and quantitative meta-analysis. A total of 355 patients were allocated into the NAC group and the control group. The evaluated outcomes included intubation rate, improvement, duration of intensive unit stay and hospital stay and mortality. The pooled results showed nonsignificant differences in intubation rate (OR, 0.55; 95% CI, 0.16-1.89; p = 0.34; I2 = 75%), improvement of oxygenation ([MD], 80.84; 95% CI, -38.16 to 199.84; p = 0.18; I2 = 98%), ICU stay (MD, -0.74; 95% CI, -3.19 to 1.71; p = 0.55; I2 = 95%), hospital stay (MD, -1.05; 95% CI, -3.02 to 0.92; p = 0.30; I2 = 90%), and mortality (OR, 0.58; 95% CI, 0.23-1.45; p = 0.24; I2 = 54%). Subsequent trial sequential analysis (TSA) showed conclusive nonsignificant results for mortality, while the TSA for the other outcomes suggested that a larger sample size is essential. CONCLUSIONS: The current evidence reveals NAC is not beneficial for treating patients with COVID- 19 with regard to respiratory outcome, mortality, duration of ICU stay and hospital stay.

Evaluating the hearing screening effectiveness of active noise cancellation technology among young adults: A pilot study.

BACKGROUND: The population of young adults who are hearing impaired increases yearly, and a device that enables convenient hearing screening could help monitor their hearing. However, background noise is a critical issue that limits the capabilities of such a device. Therefore, this study evaluated the effectiveness of commercial active noise cancellation (ANC) headphones for hearing screening applications in the presence of background noise. In particular, six confounders were used for a comprehensive evaluation. METHODS: We enrolled 12 young adults (a total of 23 ears with normal hearing) to participate in this study. A cross-sectional self-controlled study was conducted to explore the effectiveness of hearing screening in the presence of background noise, with a total of 240 test conditions (=3 ANC models × 2 ANC function statuses × 2 noise types × 5 noise levels × 4 frequencies) for each test ear. Subsequently, a linear regression model was used to prove the effectiveness of ANC headphones for hearing screening applications in the presence of background noise with six confounders. RESULTS: The experimental results showed that, on average, the ANC function of headphones can improve the effectiveness of hearing screening tasks in the presence of background noise. Specifically, the statistical analysis showed that the ANC function enabled a significant 10% improvement ( p < 0.001) compared with no ANC function. CONCLUSION: This study confirmed the effectiveness of ANC headphones for young adult hearing screening applications in the presence of background noise. Furthermore, the statistical results confirmed that as confounding variables, noise type, noise level, hearing screening frequency, ANC headphone model, and sex all affect the effectiveness of the ANC function. These findings suggest that ANC is a potential means of helping users obtain high-accuracy hearing screening results in the presence of background noise. Moreover, we present possible directions of development for ANC headphones in future studies.

Training a Deep Contextualized Language Model for International Classification of Diseases, 10th Revision Classification via Federated Learning: Model Development and Validation Study.

BACKGROUND: The automatic coding of clinical text documents by using the International Classification of Diseases, 10th Revision (ICD-10) can be performed for statistical analyses and reimbursements. With the development of natural language processing models, new transformer architectures with attention mechanisms have outperformed previous models. Although multicenter training may increase a model's performance and external validity, the privacy of clinical documents should be protected. We used federated learning to train a model with multicenter data, without sharing data per se. OBJECTIVE: This study aims to train a classification model via federated learning for ICD-10 multilabel classification. METHODS: Text data from discharge notes in electronic medical records were collected from the following three medical centers: Far Eastern Memorial Hospital, National Taiwan University Hospital, and Taipei Veterans General Hospital. After comparing the performance of different variants of bidirectional encoder representations from transformers (BERT), PubMedBERT was chosen for the word embeddings. With regard to preprocessing, the nonalphanumeric characters were retained because the model's performance decreased after the removal of these characters. To explain the outputs of our model, we added a label attention mechanism to the model architecture. The model was trained with data from each of the three hospitals separately and via federated learning. The models trained via federated learning and the models trained with local data were compared on a testing set that was composed of data from the three hospitals. The micro F1 score was used to evaluate model performance across all 3 centers. RESULTS: The F1 scores of PubMedBERT, RoBERTa (Robustly Optimized BERT Pretraining Approach), ClinicalBERT, and BioBERT (BERT for Biomedical Text Mining) were 0.735, 0.692, 0.711, and 0.721, respectively. The F1 score of the model that retained nonalphanumeric characters was 0.8120, whereas the F1 score after removing these characters was 0.7875-a decrease of 0.0245 (3.11%). The F1 scores on the testing set were 0.6142, 0.4472, 0.5353, and 0.2522 for the federated learning, Far Eastern Memorial Hospital, National Taiwan University Hospital, and Taipei Veterans General Hospital models, respectively. The explainable predictions were displayed with highlighted input words via the label attention architecture. CONCLUSIONS: Federated learning was used to train the ICD-10 classification model on multicenter clinical text while protecting data privacy. The model's performance was better than that of models that were trained locally.

Objective Signal Analysis for Investigating Feasibility of Active Noise Cancellation in Hearing Screening.

With the development of active noise cancellation (ANC) technology, ANC has been used to mitigate the effects of environmental noise on audiometric results. However, objective evaluation methods supporting the accuracy of audiometry for ANC exposure to different levels of noise have not been reported. Accordingly, the audio characteristics of three different ANC headphone models were quantified under different noise conditions and the feasibility of ANC in noisy environments was investigated. Steady (pink noise) and non-steady noise (cafeteria babble noise) were used to simulate noisy environments. We compared the integrity of pure-tone signals obtained from three different ANC headphone models after processing under different noise scenarios and analyzed the degree of ANC signal correlation based on the Pearson correlation coefficient compared to pure-tone signals in quiet. The objective signal correlation results were compared with audiometric screening results to confirm the correspondence. Results revealed that ANC helped mitigate the effects of environmental noise on the measured signal and the combined ANC headset model retained the highest signal integrity. The degree of signal correlation was used as a confidence indicator for the accuracy of hearing screening in noise results. It was found that the ANC technique can be further improved for more complex noisy environments.

Combination Therapy and Single-Modality Treatment for Acute Low-Tone Hearing Loss: A Meta-Analysis with Trial Sequential Analysis.

Acute low-tone hearing loss (ALHL) is a common clinical disease and was first proposed by Abe in 1981 as sensorineural hearing loss confined to low frequencies. The best strategy for initiating medication is still unclear, as the superiority of steroids and diuretics is still debated, and combination therapy might yield additional benefits. However, no study regarding combination therapy has been published. The objective of this study was to evaluate the efficacy of steroid therapy versus combination therapy of diuretics with steroids by conducting a systematic review with a meta-analysis and trial sequential analysis (TSA). Studies enrolling patients with a diagnosis of acute low-tone hearing loss were considered eligible. After searching the PubMed, Cochrane Library, Embase, Scopus and Web of Science databases from inception to 31 December 2021, five studies including 433 patients were enrolled. Overall, the comparison between combination therapy with steroids and diuretics and single-modality treatment with steroids (OR, 1.15; 95% CI, 0.51 to 2.59; p = 0.74; I2 = 34%) and the comparison between combination therapy and treatment with diuretics alone (OR, 1.73; 95% CI, 0.93 to 3.23; p = 0.09; I2 = 5%) showed that combination therapy did not confer significant benefits when compared to single-modality treatments. A trial sequential analysis (TSA) showed conclusive nonsignificant results of the comparison between the combination of steroids and diuretics and a single-modality treatment. In conclusion, we reported that the combination of steroids and diuretics did not yield significant benefits when compared to single-modality treatment with steroids or diuretics. We suggest that treatment should be initiated with steroids or diuretics alone to avoid potential adverse effects.

Smartphone-based artificial intelligence using a transfer learning algorithm for the detection and diagnosis of middle ear diseases: A retrospective deep learning study.

Background: Middle ear diseases such as otitis media and middle ear effusion, for which diagnoses are often delayed or misdiagnosed, are among the most common issues faced by clinicians providing primary care for children and adolescents. Artificial intelligence (AI) has the potential to assist clinicians in the detection and diagnosis of middle ear diseases through imaging. Methods: Otoendoscopic images obtained by otolaryngologists from Taipei Veterans General Hospital in Taiwan between Jany 1, 2011 to Dec 31, 2019 were collected retrospectively and de-identified. The images were entered into convolutional neural network (CNN) training models after data pre-processing, augmentation and splitting. To differentiate sophisticated middle ear diseases, nine CNN-based models were constructed to recognize middle ear diseases. The best-performing models were chosen and ensembled in a small CNN for mobile device use. The pretrained model was converted into the smartphone-based program, and the utility was evaluated in terms of detecting and classifying ten middle ear diseases based on otoendoscopic images. A class activation map (CAM) was also used to identify key features for CNN classification. The performance of each classifier was determined by its accuracy, precision, recall, and F1-score. Findings: A total of 2820 clinical eardrum images were collected for model training. The programme achieved a high detection accuracy for binary outcomes (pass/refer) of otoendoscopic images and ten different disease categories, with an accuracy reaching 98.0% after model optimisation. Furthermore, the application presented a smooth recognition process and a user-friendly interface and demonstrated excellent performance, with an accuracy of 97.6%. A fifty-question questionnaire related to middle ear diseases was designed for practitioners with different levels of clinical experience. The AI-empowered mobile algorithm's detection accuracy was generally superior to that of general physicians, resident doctors, and otolaryngology specialists (36.0%, 80.0% and 90.0%, respectively). Our results show that the proposed method provides sufficient treatment recommendations that are comparable to those of specialists. Interpretation: We developed a deep learning model that can detect and classify middle ear diseases. The use of smartphone-based point-of-care diagnostic devices with AI-empowered automated classification can provide real-world smart medical solutions for the diagnosis of middle ear diseases and telemedicine. Funding: This study was supported by grants from the Ministry of Science and Technology (MOST110-2622-8-075-001, MOST110-2320-B-075-004-MY3, MOST-110-2634-F-A49 -005, MOST110-2745-B-075A-001A and MOST110-2221-E-075-005), Veterans General Hospitals and University System of Taiwan Joint Research Program (VGHUST111-G6-11-2, VGHUST111c-140), and Taipei Veterans General Hospital (V111E-002-3).

A pressure ulcers assessment system for diagnosis and decision making using convolutional neural networks.

BACKGROUND/PURPOSE: Pressure ulcers are a common problem in hospital care and long-term care. Pressure ulcers are caused by prolonged compression of soft tissues, which can cause local tissue damage and even lead to serious infections. This study uses a deep learning algorithm to construct a system that diagnoses pressure ulcers and assists in making treatment decisions, thus providing additional reference for first-line caregivers. METHODS: We performed a retrospective research of medical records to find photos of patients with pressure ulcers at National Taiwan University Hospital from 2016 to 2020. We used photos from 2016 to 2019 for training and after removing the photos which were vague, underexposed, or overexposed, 327 photos were obtained. The photos were then labeled as "erythema" or "non-erythema" for the first classification task and "extensive necrosis", "moderate necrosis" or "limited necrosis" for the second, by consensus of three recruited physicians. An Inception-ResNet-v2 model, a kind of Convolutional Neural Network (CNN), was applied for training these two classification tasks to construct an assessment system. Finally, we tested the model with the photos of pressure ulcers taken from 2019 to 2020 to verify its accuracy. RESULTS: For the task of classification of erythema and non-erythema wounds, our CNN model achieved an accuracy of about 98.5%. For the task of classification of necrotic tissue, our model achieved accuracy of about 97%. CONCLUSION: Our CNN model, which was based on Inception-ResNet-v2, achieved high accuracy when classifying different types of pressure ulcers, making it applicable in clinical circumstances.

Comparison of personal sound amplification products and conventional hearing aids for patients with hearing loss: A systematic review with meta-analysis.

Background: Hearing loss is a common morbidity that requires a hearing device to improve quality of life and prevent sequelae, such as dementia, depression falls, and cardiovascular disease. However, conventional hearing aids have some limitations, including poor accessibility and unaffordability. Consequently, personal sound amplification products (PSAPs) are considered a potential first-line alternative remedy for patients with hearing loss. The main objective of this study was to compare the efficacy of PSAPs and conventional hearing aids regarding hearing benefits in patients with hearing loss. Methods: This systematic review and meta-analysis followed the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines. Five databases and reference lists were searched from inception to January 12, 2022. Studies including randomised, controlled trials; nonrandomised, controlled trials; or observational studies comparing PSAPs and hearing aids with regard to hearing gain performance (e.g., speech intelligence) were considered eligible. The review was registered prospectively on PROSPERO (CRD42021267187). Findings: Of 599 records identified in the preliminary search, five studies were included in the review and meta-analysis. A total of 124 patients were divided into the PSAP group and the conventional hearing aid group. Five studies including seven groups compared differences for speech intelligence in the signal-noise ratio (SNR) on the hearing in noise test (HINT) between PSAPs and conventional hearing aids. The pooled results showed nonsignificant differences in speech intelligence (SMD, 0.14; 95% CI, -0.19 to 0.47; P = .41; I 2=65%), sound quality (SMD, -0.37; 95% CI, -0.87 to 0.13; P = .15; I 2=77%) and listening effort (SMD 0.02; 95% CI, -0.24 to 0.29; P = .86; I 2=32%). Nonsignificant results were also observed in subsequent analyses after excluding patients with moderately severe hearing loss. Complete sensitivity analyses with all of the possible combinations suggested nonsignificant results in most of the comparisons between PSAPs and conventional hearing aids. Interpretation: PSAPs are potentially beneficial as conventional hearing aids are in patients with hearing loss. The different features among PSAPs should be considered for patients indicated for hearing devices. Funding: This work was supported by grants from Ministry of Science and Technology (MOST-10-2622-8-075-001) and Veterans General Hospitals and University System of Taiwan Joint Research Program (VGHUST111-G6-11-2 and VGHUST111c-140).

Smartphone-Enabled versus Conventional Otoscopy in Detecting Middle Ear Disease: A Meta-Analysis.

Traditional otoscopy has some limitations, including poor visualization and inadequate time for evaluation in suboptimal environments. Smartphone-enabled otoscopy may improve examination quality and serve as a potential diagnostic tool for middle ear diseases using a telemedicine approach. The main objectives are to compare the correctness of smartphone-enabled otoscopy and traditional otoscopy and to evaluate the diagnostic confidence of the examiner via meta-analysis. From inception through 20 January 2022, the Cochrane Library, PubMed, EMBASE, Web of Science, and Scopus databases were searched. Studies comparing smartphone-enabled otoscopy with traditional otoscopy regarding the outcome of interest were eligible. The relative risk (RR) for the rate of correctness in diagnosing ear conditions and the standardized mean difference (SMD) in diagnostic confidence were extracted. Sensitivity analysis and trial sequential analyses (TSAs) were conducted to further examine the pooled results. Study quality was evaluated by using the revised Cochrane risk of bias tool 2. Consequently, a total of 1840 examinees were divided into the smartphone-enabled otoscopy group and the traditional otoscopy group. Overall, the pooled result showed that smartphone-enabled otoscopy was associated with higher correctness than traditional otoscopy (RR, 1.26; 95% CI, 1.06 to 1.51; p = 0.01; I2 = 70.0%). Consistently significant associations were also observed in the analysis after excluding the simulation study (RR, 1.10; 95% CI, 1.00 to 1.21; p = 0.04; I2 = 0%) and normal ear conditions (RR, 1.18; 95% CI, 1.01 to 1.40; p = 0.04; I2 = 65.0%). For the confidence of examiners using both otoscopy methods, the pooled result was nonsignificant between the smartphone-enabled otoscopy and traditional otoscopy groups (SMD, 0.08; 95% CI, -0.24 to 0.40; p = 0.61; I2 = 16.3%). In conclusion, smartphone-enabled otoscopy was associated with a higher rate of correctness in the detection of middle ear diseases, and in patients with otologic complaints, the use of smartphone-enabled otoscopy may be considered. More large-scale studies should be performed to consolidate the results.

Sepsis and the Risks of Long-Term Renal Adverse Outcomes in Patients With Chronic Kidney Disease.

Background: Sepsis is known to cause renal function fluctuations during hospitalization, but whether these patients discharged from sepsis were still at greater risks of long-term renal adverse outcomes remains unknown. Methods: From 2011 to 2018, we included 1,12,628 patients with chronic kidney disease (CKD) aged ≥ 20 years. The patients with CKD were further divided into 11,661 sepsis group and 1,00,967 non-sepsis group. The following outcome of interest was included: all-cause mortality, readmission for acute kidney injury, estimated glomerular filtration rate decline ≥50% or doubling of serum creatinine, and end-stage renal disease. Results: After propensity score matching, the sepsis group was at higher risks of all-cause mortality [hazard ratio (HR) 1.39, 95% CI, 1.31-1.47], readmission for acute kidney injury (HR 1.67, 95% CI 1.58-1.76), eGFR decline ≥ 50% or doubling of serum creatinine (HR 3.34, 95% CI 2.78-4.01), and end-stage renal disease (HR 1.43, 95% CI 1.34-1.53) than non-sepsis group. Conclusions: Our study found that patients with CKD discharged from hospitalization for sepsis have higher risks of subsequent renal adverse events.

Artificial Intelligence for Risk Prediction of End-Stage Renal Disease in Sepsis Survivors with Chronic Kidney Disease.

Sepsis may lead to kidney function decline in patients with chronic kidney disease (CKD), and the deleterious effect may persist in patients who survive sepsis. We used a machine learning approach to predict the risk of end-stage renal disease (ESRD) in sepsis survivors. A total of 11,661 sepsis survivors were identified from a single-center database of 112,628 CKD patients between 2010 and 2018. During a median follow-up of 3.5 years, a total of 1366 (11.7%) sepsis survivors developed ESRD after hospital discharge. We adopted the random forest, extra trees, extreme gradient boosting, light gradient boosting machine (LGBM), and gradient boosting decision tree (GBDT) algorithms to predict the risk of ESRD development among these patients. GBDT yielded the highest area under the receiver operating characteristic curve of 0.879, followed by LGBM (0.868), and extra trees (0.865). The GBDT model revealed the strong effect of estimated glomerular filtration rates <25 mL/min/1.73 m2 at discharge in predicting ESRD development. In addition, hemoglobin and proteinuria were also essential predictors. Based on a large-scale dataset, we established a machine learning model computing the risk for ESRD occurrence among sepsis survivors with CKD. External validation is required to evaluate the generalizability of this model.