Estimating the Male to Female Ratio in Autistic Spectrum Disorder, Defining and Calculating the Biases and Finding the True Number of Girls.

The ratio of males to females with ASD is generally quoted as 4:1 though it is believed that there are biases preventing females being diagnosed and that the true ratio is lower. These biases have not been clearly identified or quantified. Starting with a clinical dataset of 1711 children <18 years old four different methods were employed in an inductive study to identify and quantify the biases and calculate the proportion of females missed. A mathematical model was constructed to compare the findings with current published data. The true male to female ratio appears to be 3:4. Eighty per cent of females remain undiagnosed at age 18 which has serious consequences for the mental health of young women.

The Relationship Between Impedance, Programming and Word Recognition in a Large Clinical Dataset of Cochlear Implant Recipients

Cochlear implant programming typically involves measuring electrode impedance, selecting a speech processing strategy and fitting the dynamic range of electrical stimulation. This study retrospectively analyzed a clinical dataset of adult cochlear implant recipients to understand how these variables relate to speech recognition. Data from 425 implanted post-lingually deafened ears with Advanced Bionics devices were analyzed. A linear mixed-effects model was used to infer how impedance, programming and patient factors were associated with monosyllabic word recognition scores measured in quiet. Additional analyses were conducted on subsets of data to examine the role of speech processing strategy on scores, and the time taken for the scores of unilaterally implanted patients to plateau. Variation in basal impedance was negatively associated with word score, suggesting importance in evaluating the profile of impedance. While there were small, negative bivariate correlations between programming level metrics and word scores, these relationships were not clearly supported by the model that accounted for other factors. Age at implantation was negatively associated with word score, and duration of implant experience was positively associated with word score, which could help to inform candidature and guide expectations. Electrode array type was also associated with word score. Word scores measured with traditional continuous interleaved sampling and current steering speech processing strategies were similar. The word scores of unilaterally implanted patients largely plateaued within 6-months of activation. However, there was individual variation which was not related to initially measured impedance and programming levels.

Commercial AI solutions in detecting COVID-19 pneumonia in chest CT: not yet ready for clinical implementation?

Objectives In response to the COVID-19 pandemic, many researchers have developed artificial intelligence (AI) tools to differentiate COVID-19 pneumonia from other conditions in chest CT. However, in many cases, performance has not been clinically validated. The aim of this study was to evaluate the performance of commercial AI solutions in differentiating COVID-19 pneumonia from other lung conditions. Methods Four commercial AI solutions were evaluated on a dual-center clinical dataset consisting of 500 CT studies; COVID-19 pneumonia was microbiologically proven in 50 of these. Sensitivity, specificity, positive and negative predictive values, and AUC were calculated. In a subgroup analysis, the performance of the AI solutions in differentiating COVID-19 pneumonia from other conditions was evaluated in CT studies with ground-glass opacities (GGOs). Results Sensitivity and specificity ranges were 62–96% and 31–80%, respectively. Negative and positive predictive values ranged between 82–99% and 19–25%, respectively. AUC was in the range 0.54–0.79. In CT studies with GGO, sensitivity remained unchanged. However, specificity was lower, and ranged between 15 and 53%. AUC for studies with GGO was in the range 0.54–0.69. Conclusions This study highlights the variable specificity and low positive predictive value of AI solutions in diagnosing COVID-19 pneumonia in chest CT. However, one solution yielded acceptable values for sensitivity. Thus, with further improvement, commercial AI solutions currently under development have the potential to be integrated as alert tools in clinical routine workflow. Randomized trials are needed to assess the true benefits and also potential harms of the use of AI in image analysis. Key Points • Commercial AI solutions achieved a sensitivity and specificity ranging from 62 to 96% and from 31 to 80%, respectively, in identifying patients suspicious for COVID-19 in a clinical dataset. • Sensitivity remained within the same range, while specificity was even lower in subgroup analysis of CT studies with ground-glass opacities, and interrater agreement between the commercial AI solutions was minimal to nonexistent. • Thus, commercial AI solutions have the potential to be integrated as alert tools for the detection of patients with lung changes suspicious for COVID-19 pneumonia in a clinical routine workflow, if further improvement is made.

Multiple Imputation by Chained Equations–K-Nearest Neighbors and Deep Neural Network Architecture for Kidney Disease Prediction

Chronic kidney disease (CKD) is a health concern that affects people all over the world. Kidney dysfunction or impaired kidney functions are the causes of CKD. The machine learning-based prediction models are used to determine the risk level of CKD and assist healthcare practitioners in delaying and preventing the disease’s progression. The researchers proposed many prediction models for determining the CKD risk level. Although these models performed well, their precision is limited since they do not handle missing values in the clinical dataset adequately. The missing values of a clinical dataset can degrade the training outcomes that leads to false predictions. Thus, imputing missing values increases the prediction model performance. This proposed work developed a novel imputation technique by combining Multiple Imputation by Chained Equations and [Formula: see text]-Nearest Neighbors (MICE–KNN) for imputing the missing values. The experimental results show that MICE–KNN accurately predicts the missing values, and the Deep Neural Network (DNN) improves the prediction performance of the CKD model. Various metrics like mean absolute error, accuracy, specificity, Matthews correlation coefficient, the area under the curve, [Formula: see text]-score, sensitivity, and precision have been used to evaluate the proposed CKD model performance. The performance analysis exhibits that MICE–KNN with deep learning outperforms other classifiers. According to our experimental study, the MICE–KNN imputation algorithm with DNN is more appropriate for predicting the kidney disease.

“A novel paradigm for optimal mass feature peak picking in large scale LC-MS datasets using the ‘isopair’: isoLock, autoCredential and anovAlign”

Untargeted metabolomics enables direct quantification of metabolites without apriori knowledge of their identity. Liquid chromatography mass spectrometry (LC-MS), a popular method to implement untargeted metabolomics, identifies metabolites via combined mass/charge (m/z) and retention time as mass features. Improvements in the sensitivity of mass spectrometers has increased the complexity of data produced, leading to computational obstacles. One outstanding challenge is calling metabolite mass feature peaks rapidly and accurately in large LC-MS datasets (dozens to thousands of samples) in the presence of measurement and other noise. While existing algorithms are useful, they have limitations that become pronounced at scale and lead to false positive metabolite predictions as well as signal dropouts. To overcome some of these shortcomings, biochemists have developed hybrid computational and carbon labeling techniques, such as credentialing. Credentialing can validate metabolite signals, but is laborious and its applicability is limited. We have developed a suite of three computational tools to overcome the challenges of unreliable algorithms and inefficient validation protocols: isolock, autoCredential and anovAlign. Isolock uses isopairs, or metabolite-istopologue pairs, to calculate and correct for mass drift noise across LC-MS runs. autoCredential leverages statistical features of LC-MS data to amplify naturally present 13C isotopologues and validate metabolites through isopairs. This obviates the need to artificially introduce carbon labeling. anovAlign, an anova-derived algorithm, is used to align retention time windows across samples to accurately delineate retention time windows for mass features. Using a large published clinical dataset as well as a plant dataset with biological replicates across time, genotype and treatment, we demonstrate that this suite of tools is more sensitive and reproducible than both an open source metabolomics pipelines, XCMS, and the commercial software progenesis QI. This software suite opens a new era for enhanced accuracy and increased throughput for untargeted metabolomics.

Generalizability and Quality Control of Deep Learning-Based 2D Echocardiography Segmentation Models in a Large Clinical Dataset

Use of machine learning for automated annotation of heart structures from echocardiographic videos is an active research area, but understanding of comparative, generalizable performance among models is lacking. This study aimed to 1) assess the generalizability of five state-of-the-art machine learning-based echocardiography segmentation models within a large clinical dataset, and 2) test the hypothesis that a quality control (QC) method based on segmentation uncertainty can further improve segmentation results. Five models were applied to 47,431 echocardiography studies that were independent from any training samples. Chamber volume and mass from model segmentations were compared to clinically-reported values. The median absolute errors (MAE) in left ventricular (LV) volumes and ejection fraction exhibited by all five models were comparable to reported inter-observer errors (IOE). MAE for left atrial volume and LV mass were similarly favorable to respective IOE for models trained for those tasks. A single model consistently exhibited the lowest MAE in all five clinically-reported measures. We leveraged the 10-fold cross-validation training scheme of this best-performing model to quantify segmentation uncertainty for potential application as QC. We observed that filtering segmentations with high uncertainty improved segmentation results, leading to decreased volume/mass estimation errors. The addition of contour-convexity filters further improved QC efficiency. In conclusion, five previously published echocardiography segmentation models generalized to a large, independent clinical dataset—segmenting one or multiple cardiac structures with overall accuracy comparable to manual analyses—with variable performance. Convexity-reinforced uncertainty QC efficiently improved segmentation performance and may further facilitate the translation of such models.

554 The panorama of tumor intrinsic immune regulators exhibited by genome-wide CRPISR immune screen integrated with comprehensive clinical dataset analysis

BackgroundDespite approval of immunotherapy for wide ranges of cancers, the majority of patients fail to respond to immunotherapy or relapse following initial response which is partially attributed to immunosuppression co-opted by tumor cells. However, it is challenging to utilize conventional methods to systematically evaluate the potential of tumor intrinsic factors to act as immune regulators in cancer patients.MethodsIn this study, an unbiased integrative strategy were designed to leverage the complementary strength of in vitro functional genomic screens and multi-omics clinical dataset to assess roles of individual tumor-intrinsic factors in regulating T cell tumor infiltration and T cell-mediated tumor killing, the two most important rate-limiting steps of cancer immunotherapy. Initially, a genome-wide CRISPR-Cas9 screening system using paired murine tumors and tumor-reactive T cells was employed to globally screen tumor intrinsic factors modulating the tumor sensitivity to T cell-mediated killing. Then, findings from the screening were integrated with the bioinformatics analysis of clinical datasets to further evaluate roles of each tumor intrinsic factor in governing antitumor immunity.ResultsThe integrative analysis successfully identified several novel tumor intrinsic factors as effectors of immune resistance, but also demonstrated distinct roles of these factors in controlling immune cell trafficking and tumor sensitivity to T cell-mediated killing. Among these factors, candidates controlling both rate-limiting steps of T cell tumor infiltration and T cell-mediated tumor killing were termed as ”Dual immune resistance regulators” and the remaining factors whose expression levels were not associated with tumor immune infiltration were termed as ”Cytotoxicity resistance regulators”. By selecting PRMT1 and RIPK1 as the representatives of these two groups respectively, we confirmed that genetically depletion of PRMT1 and RIPK1 sensitized tumors to T-cell mediated killing via two independent experimental approaches. Furthermore, inhibiting Prmt1 or Ripk1 sensitizes tumors to cancer immunotherapy, such as anti-PD-1 and anti-OX40 treatments (Tumor size (mm2) on day 21 after tumor inoculation: for anti-PD-1 treatment, Ctrl 84.05±23.10, PRMT1 KO 7.30±7.81, RIPK1 KO 2.03±4.96; for anti-OX40 treatment, Ctrl 81.04±7.72, PRMT1 KO 55.80±15.74, RIPK1 KO 38.78±14.06) and extended the survival of tumor-bearing mice. Moreover, by using a RIPK1-specific inhibitor, GSK2982772, we demonstrated that targeting cytotoxicity resistance regulators could enhance the antitumor activity of T cell-based cancer immunotherapy, despite limited impact on T cell tumor infiltration.ConclusionsCollectively, our data not only demonstrate distinct immunoregulatory roles and therapeutic potentials of PRMT1 and RIPK1 in T cell-mediated antitumor activity, but also provide a rich resource of novel targets for rational immuno-oncology combinations.

Automated diagnosis of atrial fibrillation in 24-hour Holter recording based on deep learning:a study with randomized and real-world data validation

SummaryBackgroundWith the increasing demand for atrial fibrillation (AF) screening, clinicians spend a significant amount of time in identifying the AF signals from massive electrocardiogram (ECG) data in long-term dynamic ECG monitoring. In this study, we aim to reduce clinicians’ workload and promote AF screening by using artificial intelligence (AI) to automatically detect AF episodes and identify AF patients in 24 h Holter recording.MethodsWe used a total of 22 979 Holter recordings (24 h) from 22 757 adult patients and established accurate annotations for AF by cardiologists. First, a randomized clinical cohort of 3 000 recordings (1 500 AF and 1 500 non-AF) from 3000 patients recorded between April 2012 and May 2020 was collected and randomly divided into training, validation and test sets (10:1:4). Then, a deep-learning-based AI model was developed to automatically detect AF episode using RR intervals and was tested with the test set. Based on AF episode detection results, AF patients were automatically identified by using a criterion of at least one AF episode of 6 min or longer. Finally, the clinical effectiveness of the model was verified with an independent real-world test set including 19 979 recordings (1 006 AF and 18 973 non-AF) from 19 757 consecutive patients recorded between June 2020 and January 2021.FindingsOur model achieved high performance for AF episode detection in both test sets (sensitivity: 0.992 and 0.972; specificity: 0.997 and 0.997, respectively). It also achieved high performance for AF patient identification in both test sets (sensitivity:0.993 and 0.994; specificity: 0.990 and 0.973, respectively). Moreover, it obtained superior and consistent performance in an external public database.InterpretationOur AI model can automatically identify AF in long-term ECG recording with high accuracy. This cost-effective strategy may promote AF screening by improving diagnostic effectiveness and reducing clinical workload.Research in contextEvidence before this studyWe searched Google Scholar and PubMed for research articles on artificial intelligence-based diagnosis of atrial fibrillation (AF) published in English between Jan 1, 2016 and Aug 1, 2021, using the search terms “deep learning” OR “deep neural network” OR “machine learning” OR “artificial intelligence” AND “atrial fibrillation”. We found that most of the previous deep learning models in AF detection were trained and validated on benchmark datasets (such as the PhysioNet database, the Massachusetts Institute of Technology Beth Israel Hospital AF database or Long-Term AF database), in which there were less than 100 patients or the recordings contained only short ECG segments (30-60s). Our search did not identify any articles that explored deep neural networks for AF detection in large real-world dataset of 24 h Holter recording, nor did we find articles that can automatically identify patients with AF in 24 h Holter recording.Added value of this studyFirst, long-term Holter monitoring is the main method of AF screening, however, most previous studies of automatic AF detection mainly tested on short ECG recordings. This work focused on 24 h Holter recording data and achieved high accuracy in detecting AF episodes. Second, AF episodes detection did not automatically transform to AF patient identification in 24 h Holter recording, since at present, there is no well-recognized criterion for automatically identifying AF patient. Therefore, we established a criterion to identify AF patients by use of at least one AF episode of 6 min or longer, as this condition led to significantly increased risk of thromboembolism. Using this criterion, our method identified AF patients with high accuracy. Finally, and more importantly, our model was trained on a randomized clinical dataset and tested on an independent real-world clinical dataset to show great potential in clinical application. We did not exclude rare or special cases in the real-world dataset so as not to inflate our AF detection performance. To the best of our knowledge, this is the first study to automatically identifies both AF episodes and AF patients in 24 h Holter recording of large real-world clinical dataset.Implications of all the available evidenceOur deep learning model automatically identified AF patient with high accuracy in 24 h Holter recording and was verified in real-world data, therefore, it can be embedded into the Holter analysis system and deployed at the clinical level to assist the decision making of Holter analysis system and clinicians. This approach can help improve the efficiency of AF screening and reduce the cost for AF diagnosis. In addition, our RR-interval-based model achieved comparable or better performance than the raw-ECG-based method, and can be widely applied to medical devices that can collect heartbeat information, including not only the multi-lead and single-lead Holter devices, but also other wearable devices that can reliably measure the heartbeat signals.