Differential prognostic value of tumor and plasma T790M mutations in EGFR TKI-treated advanced NSCLC.

Background: Substitution of methionine for threonine at codon 790 (T790M) of epidermal growth factor receptor (EGFR) represents the major mechanism of resistance to EGFR tyrosine kinase inhibitors (TKIs) in EGFR-mutant non-small-cell lung cancer. We determined the prognostic impact and association of secondary T790M mutations with the outcomes of osimertinib and chemotherapy. Methods: Patients (n = 460) progressing from first-line EGFR-TKI treatment were assessed. Tissue and/or liquid biopsies were used to determine T790M status; post-progression overall survival (OS) was analyzed. Results: Overall, 143 (31.1%) patients were T790M positive, 95 (20.7%) were T790M negative, and 222 (48.2%) had unknown T790M status. T790M status [T790M positive versus T790M negative: hazard ratio (HR) 0.48 (95% confidence interval (CI), 0.32-0.70); p < 0.001, T790M unknown versus T790M negative: HR 1.97 (95% CI, 1.47-2.64); p < 0.001] was significantly associated with post-progression OS. T790M positivity rates were similar for tissue (90/168, 53.6%) and liquid (53/90, 58.9%) biopsies (Fisher's exact test, p = 0.433). Tumor T790M-positive patients had significantly longer post-progression OS than tumor T790M-negative patients (34.1 versus 17.1 months; log-rank test, p = 8 × 10-5). Post-progression OS was similar between plasma T790M-positive and -negative patients (17.4 versus not reached; log-rank test, p = 0.600). In tumor T790M-positive patients, post-progression OS was similar after osimertinib and chemotherapy [34.1 versus 29.1 months; log-rank test, p = 0.900; HR 1.06 (95% CI, 0.44-2.57); p = 0.897]. Conclusion: T790M positivity predicts better post-progression OS than T790M negativity; tumor T790M positivity has a stronger prognostic impact than plasma T790M positivity. Osimertinib and chemotherapy provide similar OS benefits in patients with T790M-positive tumors.

Different prognostic meaning of tumor resistant gene detected from tumor or blood in patients with EGFR-mutant lung cancer The study demonstrates that patients with EGFR-mutant lung cancer who develop resistance due to a secondary T790M mutation, defined by tumor or blood T790M positivity, achieve better survival than patients without secondary T790M mutation; this association was mainly contributed by tumour T790M positivity. Oismertinib and chemotherapy led to similar survival in tumour T790M-positive patients. However, compared to osimertinib, chemotherapy was associated with longer survival in blood T790M-positive patients.

Bayesian at heart: Towards autonomic outflow estimation via generative state-space modelling of heart rate dynamics.

Recent research is revealing how cognitive processes are supported by a complex interplay between the brain and the rest of the body, which can be investigated by the analysis of physiological features such as breathing rhythms, heart rate, and skin conductance. Heart rate dynamics are of particular interest as they provide a way to track the sympathetic and parasympathetic outflow from the autonomic nervous system, which is known to play a key role in modulating attention, memory, decision-making, and emotional processing. However, extracting useful information from heartbeats about the autonomic outflow is still challenging due to the noisy estimates that result from standard signal-processing methods. To advance this state of affairs, we propose a novel approach in how to conceptualise and model heart rate: instead of being a mere summary of the observed inter-beat intervals, we introduce a modelling framework that views heart rate as a hidden stochastic process that drives the observed heartbeats. Moreover, by leveraging the rich literature of state-space modelling and Bayesian inference, our proposed framework delivers a description of heart rate dynamics that is not a point estimate but a posterior distribution of a generative model. We illustrate the capabilities of our method by showing that it recapitulates linear properties of conventional heart rate estimators, while exhibiting a better discriminative power for metrics of dynamical complexity compared across different physiological states.

Brain metastasis, EGFR mutation subtype and generation of EGFR-TKI jointly influence the treatment outcome of patient with EGFR-mutant NSCLC.

Non-small cell lung cancer (NSCLC) with epidermal growth factor receptor (EGFR) mutation is brain metastasis (BM)-prone. We determined the impact of this hallmark, along with EGFR subtype and generation of tyrosine kinase inhibitor (TKI) treatment, on patients' outcome. 553 metastatic EGFR-mutant NSCLC patients received front-line EGFR-TKI treatment. Progression-free survival (PFS), overall survival (OS) and secondary T790M rate were analysed. BM was observed in 211 (38.2%) patients. BM (HR 1.20 [95% CI 0.99-1.48]; p = 0.053), ECOG PS 0-1 (HR 0.71 [95% CI 0.54-0.93]; p = 0.014) and afatinib treatment (HR 0.81 [95% CI 0.66-0.99]; p = 0.045) were associated with PFS. Afatinib-treated patients without BM demonstrated a significantly longer PFS (16.3 months) compared to afatinib-treated patients with BM (13.7 months) and to gefitinib/erlotinib-treated patients with (11.1 months) or without BM (14.2 months; p < 0.001). CNS-only progression trended higher in afatinib-treated patients. ECOG PS 0-1 (HR 0.41 [95% CI 0.31-0.56]; p < 0.001) and EGFR L858R mutation (HR 1.46 [95% CI 1.13-1.88]; p = 0.003), but not BM, were the predictors for OS. BM (OR 2.02 [95% CI 1.02-4.08]; p = 0.040), afatinib treatment (OR 0.26 [95% CI 0.12-0.50]; p < 0.001) and EGFR L858R mutation (OR 0.55 [95% CI 0.28-1.05]; p = 0.070) were associated with secondary T790M rate. In BM patients, gefitinib/erlotinib-treated ones with 19 deletion mutation and afatinib-treated ones with L858R mutation had the highest and the lowest T790M rate (94.4% vs. 27.3%, p < 0.001), respectively. BM and generation of EGFR-TKI jointly impact PFS and secondary T790M rate in patients with EGFR-mutant NSCLC, whereas OS was mainly associated with EGFR subtype.

Real-time updating of dynamic social networks for COVID-19 vaccination strategies.

Vaccination strategy is crucial in fighting the COVID-19 pandemic. Since the supply is still limited in many countries, contact network-based interventions can be most powerful to set an efficient strategy by identifying high-risk individuals or communities. However, due to the high dimension, only partial and noisy network information can be available in practice, especially for dynamic systems where contact networks are highly time-variant. Furthermore, the numerous mutations of SARS-CoV-2 have a significant impact on the infectious probability, requiring real-time network updating algorithms. In this study, we propose a sequential network updating approach based on data assimilation techniques to combine different sources of temporal information. We then prioritise the individuals with high-degree or high-centrality, obtained from assimilated networks, for vaccination. The assimilation-based approach is compared with the standard method (based on partially observed networks) and a random selection strategy in terms of vaccination effectiveness in a SIR model. The numerical comparison is first carried out using real-world face-to-face dynamic networks collected in a high school, followed by sequential multi-layer networks generated relying on the Barabasi-Albert model emulating large-scale social networks with several communities.

Analyzing drop coalescence in microfluidic devices with a deep learning generative model.

Predicting drop coalescence based on process parameters is crucial for experimental design in chemical engineering. However, predictive models can suffer from the lack of training data and more importantly, the label imbalance problem. In this study, we propose the use of deep learning generative models to tackle this bottleneck by training the predictive models using generated synthetic data. A novel generative model, named double space conditional variational autoencoder (DSCVAE) is developed for labelled tabular data. By introducing label constraints in both the latent and the original space, DSCVAE is capable of generating consistent and realistic samples compared to the standard conditional variational autoencoder (CVAE). Two predictive models, namely random forest and gradient boosting classifiers, are enhanced on synthetic data and their performances are evaluated based on real experimental data. Numerical results show that a considerable improvement in prediction accuracy can be achieved by using synthetic data and the proposed DSCVAE clearly outperforms the standard CVAE. This research clearly provides more insights into handling imbalanced data for classification problems, especially in chemical engineering.

Federated Active Learning for Multicenter Collaborative Disease Diagnosis.

Current computer-aided diagnosis system with deep learning method plays an important role in the field of medical imaging. The collaborative diagnosis of diseases by multiple medical institutions has become a popular trend. However, large scale annotations put heavy burdens on medical experts. Furthermore, the centralized learning system has defects in privacy protection and model generalization. To meet these challenges, we propose two federated active learning methods for multicenter collaborative diagnosis of diseases: the Labeling Efficient Federated Active Learning (LEFAL) and the Training Efficient Federated Active Learning (TEFAL). The proposed LEFAL applies a task-agnostic hybrid sampling strategy considering data uncertainty and diversity simultaneously to improve data efficiency. The proposed TEFAL evaluates the client informativeness with a discriminator to improve client efficiency. On the Hyper-Kvasir dataset for gastrointestinal disease diagnosis, with only 65% of labeled data, the LEFAL achieves 95% performance on the segmentation task with whole labeled data. Moreover, on the CC-CCII dataset for COVID-19 diagnosis, with only 50 iterations, the accuracy and F1-score of TEFAL are 0.90 and 0.95, respectively on the classification task. Extensive experimental results demonstrate that the proposed federated active learning methods outperform state-of-the-art methods on segmentation and classification tasks for multicenter collaborative disease diagnosis.

Th2 high and mast cell gene signatures are associated with corticosteroid sensitivity in COPD.

RATIONALE: Severe asthma and chronic obstructive pulmonary disease (COPD) share common pathophysiological traits such as relative corticosteroid insensitivity. We recently published three transcriptome-associated clusters (TACs) using hierarchical analysis of the sputum transcriptome in asthmatics from the Unbiased Biomarkers for the Prediction of Respiratory Disease Outcomes (U-BIOPRED) cohort comprising one Th2-high inflammatory signature (TAC1) and two Th2-low signatures (TAC2 and TAC3). OBJECTIVE: We examined whether gene expression signatures obtained in asthma can be used to identify the subgroup of patients with COPD with steroid sensitivity. METHODS: Using gene set variation analysis, we examined the distribution and enrichment scores (ES) of the 3 TACs in the transcriptome of bronchial biopsies from 46 patients who participated in the Groningen Leiden Universities Corticosteroids in Obstructive Lung Disease COPD study that received 30 months of treatment with inhaled corticosteroids (ICS) with and without an added long-acting ß-agonist (LABA). The identified signatures were then associated with longitudinal clinical variables after treatment. Differential gene expression and cellular convolution were used to define key regulated genes and cell types. MEASUREMENTS AND MAIN RESULTS: Bronchial biopsies in patients with COPD at baseline showed a wide range of expression of the 3 TAC signatures. After ICS±LABA treatment, the ES of TAC1 was significantly reduced at 30 months, but those of TAC2 and TAC3 were unaffected. A corticosteroid-sensitive TAC1 signature was developed from the TAC1 ICS-responsive genes. This signature consisted of mast cell-specific genes identified by single-cell RNA-sequencing and positively correlated with bronchial biopsy mast cell numbers following ICS±LABA. Baseline levels of gene transcription correlated with the change in RV/TLC %predicted following 30-month ICS±LABA. CONCLUSION: Sputum-derived transcriptomic signatures from an asthma cohort can be recapitulated in bronchial biopsies of patients with COPD and identified a signature of airway mast cells as a predictor of corticosteroid responsiveness.

Macrophage migration inhibitory factor promotes glucocorticoid resistance of neutrophilic inflammation in a murine model of severe asthma.

BACKGROUND: Severe neutrophilic asthma is resistant to treatment with glucocorticoids. The immunomodulatory protein macrophage migration inhibitory factor (MIF) promotes neutrophil recruitment to the lung and antagonises responses to glucocorticoids. We hypothesised that MIF promotes glucocorticoid resistance of neutrophilic inflammation in severe asthma. METHODS: We examined whether sputum MIF protein correlated with clinical and molecular characteristics of severe neutrophilic asthma in the Unbiased Biomarkers for the Prediction of Respiratory Disease Outcomes (U-BIOPRED) cohort. We also investigated whether MIF regulates neutrophilic inflammation and glucocorticoid responsiveness in a murine model of severe asthma in vivo. RESULTS: MIF protein levels positively correlated with the number of exacerbations in the previous year, sputum neutrophils and oral corticosteroid use across all U-BIOPRED subjects. Further analysis of MIF protein expression according to U-BIOPRED-defined transcriptomic-associated clusters (TACs) revealed increased MIF protein and a corresponding decrease in annexin-A1 protein in TAC2, which is most closely associated with airway neutrophilia and NLRP3 inflammasome activation. In a murine model of severe asthma, treatment with the MIF antagonist ISO-1 significantly inhibited neutrophilic inflammation and increased glucocorticoid responsiveness. Coimmunoprecipitation studies using lung tissue lysates demonstrated that MIF directly interacts with and cleaves annexin-A1, potentially reducing its biological activity. CONCLUSION: Our data suggest that MIF promotes glucocorticoid-resistance of neutrophilic inflammation by reducing the biological activity of annexin-A1, a potent glucocorticoid-regulated protein that inhibits neutrophil accumulation at sites of inflammation. This represents a previously unrecognised role for MIF in the regulation of inflammation and points to MIF as a potential therapeutic target for the management of severe neutrophilic asthma.

Protocol for Brain Magnetic Resonance Imaging and Extraction of Imaging-Derived Phenotypes from the China Phenobank Project.

Imaging-derived phenotypes (IDPs) have been increasingly used in population-based cohort studies in recent years. As widely reported, magnetic resonance imaging (MRI) is an important imaging modality for assessing the anatomical structure and function of the brain with high resolution and excellent soft-tissue contrast. The purpose of this article was to describe the imaging protocol of the brain MRI in the China Phenobank Project (CHPP). Each participant underwent a 30-min brain MRI scan as part of a 2-h whole-body imaging protocol in CHPP. The brain imaging sequences included T1-magnetization that prepared rapid gradient echo, T2 fluid-attenuated inversion-recovery, magnetic resonance angiography, diffusion MRI, and resting-state functional MRI. The detailed descriptions of image acquisition, interpretation, and post-processing were provided in this article. The measured IDPs included volumes of brain subregions, cerebral vessel geometrical parameters, microstructural tracts, and function connectivity metrics.

Clinical utility of automatic phenotype annotation in unstructured clinical notes: intensive care unit use.

OBJECTIVE: Clinical notes contain information that has not been documented elsewhere, including responses to treatment and clinical findings, which are crucial for predicting key outcomes in patients in acute care. In this study, we propose the automatic annotation of phenotypes from clinical notes as a method to capture essential information to predict outcomes in the intensive care unit (ICU). This information is complementary to typically used vital signs and laboratory test results. METHODS: In this study, we developed a novel phenotype annotation model to extract the phenotypical features of patients, which were then used as input features of predictive models to predict ICU patient outcomes. We demonstrated and validated this approach by conducting experiments on three ICU prediction tasks, including in-hospital mortality, physiological decompensation and length of stay (LOS) for over 24 000 patients using the Medical Information Mart for Intensive Care (MIMIC-III) dataset. RESULTS: The predictive models incorporating phenotypical information achieved 0.845 (area under the curve-receiver operating characteristic (AUC-ROC)) for in-hospital mortality, 0.839 (AUC-ROC) for physiological decompensation and 0.430 (kappa) for LOS, all of which consistently outperformed the baseline models using only vital signs and laboratory test results. Moreover, we conducted a thorough interpretability study showing that phenotypes provide valuable insights at both the patient and cohort levels. CONCLUSION: The proposed approach demonstrates that phenotypical information complements traditionally used vital signs and laboratory test results and significantly improves the accuracy of outcome prediction in the ICU.

Phenotyping in clinical text with unsupervised numerical reasoning for patient stratification.

Phenotypic information of patients, as expressed in clinical text, is important in many clinical applications such as identifying patients at risk of hard-to-diagnose conditions. Extracting and inferring some phenotypes from clinical text requires numerical reasoning, for example, a temperature of 102°F suggests the phenotype Fever. However, while current state-of-the-art phenotyping models using natural language processing (NLP) are in general very efficient in extracting phenotypes, they struggle to extract phenotypes that require numerical reasoning. In this article, we propose a novel unsupervised method that leverages external clinical knowledge and contextualized word embeddings by ClinicalBERT for numerical reasoning in different phenotypic contexts. Experiments show that the proposed method achieves significant improvement against unsupervised baseline methods with absolute increase in generalized Recall and F1 scores of up to 79% and 71%, respectively. Also, the proposed method outperforms supervised baseline methods with absolute increase in generalized Recall and F1 scores of up to 70% and 44%, respectively. In addition, we validate the methodology on clinical use cases where the detected phenotypes significantly contribute to patient stratification systems for a set of diseases, namely, HIV and myocardial infarction (heart attack). Moreover, we find that these phenotypes from clinical text can be used to impute the missing values in structured data, which enrich and improve data quality.

Scale-Consistent Fusion: From Heterogeneous Local Sampling to Global Immersive Rendering.

Image-based geometric modeling and novel view synthesis based on sparse large-baseline samplings are challenging but important tasks for emerging multimedia applications such as virtual reality and immersive telepresence. Existing methods fail to produce satisfactory results due to the limitation on inferring reliable depth information over such challenging reference conditions. With the popularization of commercial light field (LF) cameras, capturing LF images (LFIs) is as convenient as taking regular photos, and geometry information can be reliably inferred. This inspires us to use a sparse set of LF captures to render high-quality novel views globally. However, the fusion of LF captures from multiple angles is challenging due to the scale inconsistency caused by various capture settings. To overcome this challenge, we propose a novel scale-consistent volume rescaling algorithm that robustly aligns the disparity probability volumes (DPV) among different captures for scale-consistent global geometry fusion. Based on the fused DPV projected to the target camera frustum, novel learning-based modules (i.e., the attention-guided multi-scale residual fusion module, and the disparity field-guided deep re-regularization module), which comprehensively regularize noisy observations from heterogeneous captures for high-quality rendering of novel LFIs, have been proposed. Both quantitative and qualitative experiments over the Stanford Lytro Multi-view LF dataset show that the proposed method outperforms state-of-the-art methods significantly under different experiment settings for disparity inference and LF synthesis.

Ensemble latent assimilation with deep learning surrogate model: application to drop interaction in a microfluidics device.

A major challenge in the field of microfluidics is to predict and control drop interactions. This work develops an image-based data-driven model to forecast drop dynamics based on experiments performed on a microfluidics device. Reduced-order modelling techniques are applied to compress the recorded images into low-dimensional spaces and alleviate the computational cost. Recurrent neural networks are then employed to build a surrogate model of drop interactions by learning the dynamics of compressed variables in the reduced-order space. The surrogate model is integrated with real-time observations using data assimilation. In this paper we developed an ensemble-based latent assimilation algorithm scheme which shows an improvement in terms of accuracy with respect to the previous approaches. This work demonstrates the possibility to create a reliable data-driven model enabling a high fidelity prediction of drop interactions in microfluidics device. The performance of the developed system is evaluated against experimental data (i.e., recorded videos), which are excluded from the training of the surrogate model. The developed scheme is general and can be applied to other dynamical systems.

Clinical and transcriptomic features of persistent exacerbation-prone severe asthma in U-BIOPRED cohort.

BACKGROUND: Exacerbation-prone asthma is a feature of severe disease. However, the basis for its persistency remains unclear. OBJECTIVES: To determine the clinical and transcriptomic features of frequent exacerbators (FEs) and persistent FEs (PFEs) in the U-BIOPRED cohort. METHODS: We compared features of FE (≥2 exacerbations in past year) to infrequent exacerbators (IE, <2 exacerbations) and of PFE with repeat ≥2 exacerbations during the following year to persistent IE (PIE). Transcriptomic data in blood, bronchial and nasal epithelial brushings, bronchial biopsies and sputum cells were analysed by gene set variation analysis for 103 gene signatures. RESULTS: Of 317 patients, 62.4% had FE, of whom 63.6% had PFE, while 37.6% had IE, of whom 61.3% had PIE. Using multivariate analysis, FE was associated with short-acting beta-agonist use, sinusitis and daily oral corticosteroid use, while PFE was associated with eczema, short-acting beta-agonist use and asthma control index. CEA cell adhesion molecule 5 (CEACAM5) was the only differentially expressed transcript in bronchial biopsies between PE and IE. There were no differentially expressed genes in the other four compartments. There were higher expression scores for type 2, T-helper type-17 and type 1 pathway signatures together with those associated with viral infections in bronchial biopsies from FE compared to IE, while there were higher expression scores of type 2, type 1 and steroid insensitivity pathway signatures in bronchial biopsies of PFE compared to PIE. CONCLUSION: The FE group and its PFE subgroup are associated with poor asthma control while expressing higher type 1 and type 2 activation pathways compared to IE and PIE, respectively.

The different overall survival between single-agent EGFR-TKI treatment and with bevacizumab in non-small cell lung cancer patients with brain metastasis.

Comparison of epidermal growth factor receptor tyrosine kinase inhibitor (EGFR-TKI) monotherapy or with bevacizumab in real-world non-small cell lung cancer (NSCLC) patients was lacking. 310 patients of advanced NSCLC with common EGFR mutation receiving first-generation EGFR-TKI monotherapy or with bevacizumab were included and propensity-score matched. Progression-free survival (PFS), overall survival (OS) and secondary T790M mutation were analysed. Patients receiving EGFR-TKI and bevacizumab were significantly younger, had better performance status and with high incidence of brain metastasis (55.8%). In the propensity-score matched cohort, PFS (13.5 vs. 13.7 months; log-rank p = 0.700) was similar between the two groups. The OS (61.3 vs. 34.2 months; log-rank p = 0.010) and risk reduction of death (HR 0.42 [95% CI 0.20-0.85]; p = 0.017) were significantly improved in EGFR-TKI plus bevacizumab group. Analysis of treatment by brain metastasis status demonstrated EGFR-TKI plus bevacizumab in patients with brain metastasis was associated with significant OS benefit compared to other groups (log-rank p = 0.030) and these patients had lower early-CNS and early-systemic progressions. The secondary T790M did not significantly differ between EGFR-TKI plus bevacizumab and EGFR-TKI monotherapy groups (66.7% vs. 75.0%, p = 0.460). Forty-one (31.1%) and 31 (23.5%) patients received subsequent osimertinib and chemotherapy, respectively. The post-progression OS of osimertinib and chemotherapy were 22.1 and 44.9 months in EGFR-TKI plus bevacizumab group and were 10.0 and 14.1 months in EGFR-TKI monotherpay group, respectively. First-generation EGFR-TKI with bevacizumab improved treatment efficacy in real-world patients of NSCLC with EGFR mutation. Patients with brain metastasis received additional OS benefit from this treatment.

Bayesian data assimilation for estimating instantaneous reproduction numbers during epidemics: Applications to COVID-19.

Estimating the changes of epidemiological parameters, such as instantaneous reproduction number, Rt, is important for understanding the transmission dynamics of infectious diseases. Current estimates of time-varying epidemiological parameters often face problems such as lagging observations, averaging inference, and improper quantification of uncertainties. To address these problems, we propose a Bayesian data assimilation framework for time-varying parameter estimation. Specifically, this framework is applied to estimate the instantaneous reproduction number Rt during emerging epidemics, resulting in the state-of-the-art 'DARt' system. With DARt, time misalignment caused by lagging observations is tackled by incorporating observation delays into the joint inference of infections and Rt; the drawback of averaging is overcome by instantaneously updating upon new observations and developing a model selection mechanism that captures abrupt changes; the uncertainty is quantified and reduced by employing Bayesian smoothing. We validate the performance of DARt and demonstrate its power in describing the transmission dynamics of COVID-19. The proposed approach provides a promising solution for making accurate and timely estimation for transmission dynamics based on reported data.

Suggestive annotation of brain MR images with gradient-guided sampling.

Machine learning has been widely adopted for medical image analysis in recent years given its promising performance in image segmentation and classification tasks. The success of machine learning, in particular supervised learning, depends on the availability of manually annotated datasets. For medical imaging applications, such annotated datasets are not easy to acquire, it takes a substantial amount of time and resource to curate an annotated medical image set. In this paper, we propose an efficient annotation framework for brain MR images that can suggest informative sample images for human experts to annotate. We evaluate the framework on two different brain image analysis tasks, namely brain tumour segmentation and whole brain segmentation. Experiments show that for brain tumour segmentation task on the BraTS 2019 dataset, training a segmentation model with only 7% suggestively annotated image samples can achieve a performance comparable to that of training on the full dataset. For whole brain segmentation on the MALC dataset, training with 42% suggestively annotated image samples can achieve a comparable performance to training on the full dataset. The proposed framework demonstrates a promising way to save manual annotation cost and improve data efficiency in medical imaging applications.

Afatinib Treatment Alone or with Bevacizumab in a Real-World Cohort of Non-Small Cell Lung Cancer Patients with Epidermal Growth Factor Receptor Mutation.

BACKGROUND: Treatment outcome between afatinib alone or with bevacizumab in non-small cell lung cancer (NSCLC) patient with epidermal growth factor receptor (EGFR) mutation remains insufficiently reported. METHODS: A total of 405 advanced NSCLC patients with sensitizing-EGFR mutation receiving first-line single-agent afatinib or with bevacizumab were grouped and propensity score-matched. Progression-free survival (PFS), overall survival (OS) and secondary T790M mutation were analyzed. RESULTS: In the original cohort, 367 (90.6%) patients received afatinib treatment alone and 38 (9.4%) patients received afatinib plus bevacizumab. Patients who received bevacizumab combination were significantly younger (54.6 ± 10.9 vs. 63.9 ± 11.5; p < 0.001) compared to the afatinib alone group. After propensity score matching, the afatinib alone and afatinib plus bevacizumab groups contained 118 and 34 patients, respectively. A non-significantly higher objective response was noted in the afatinib plus bevacizumab group (82.4% vs. 67.8%; p = 0.133). In the propensity score-matched cohort, a bevacizumab add-on offered no increased PFS (16.1 vs. 15.0 months; p = 0.500), risk reduction of progression (HR 0.85 [95% CI, 0.52-1.40]; p = 0.528), OS benefit (32.1 vs. 42.0 months; p = 0.700), nor risk reduction of death (HR 0.85 [95% CI, 0.42-1.74] p = 0.660) compared to the single-agent afatinib. The secondary T790M rate in afatinib plus bevacizumab and afatinib alone groups was similar (56.3% vs. 49.4%, p = 0.794). Multivariate analysis demonstrated that EGFR L858R (OR 0.51 [95% CI, 0.26-0.97]; p = 0.044), EGFR uncommon mutation (OR 0.14 [95% CI, 0.02-0.64]; p = 0.021), and PFS longer than 12 months (OR 2.71 [95% CI, 1.39-5.41]; p = 0.004) were independent predictors of secondary T790M positivity. CONCLUSION: Bevacizumab treatment showed moderate efficacy in real-world, afatinib-treated NSCLC patients with EGFR-sensitizing mutation.

Digital twins based on bidirectional LSTM and GAN for modelling the COVID-19 pandemic.

The outbreak of the coronavirus disease 2019 (COVID-19) has now spread throughout the globe infecting over 150 million people and causing the death of over 3.2 million people. Thus, there is an urgent need to study the dynamics of epidemiological models to gain a better understanding of how such diseases spread. While epidemiological models can be computationally expensive, recent advances in machine learning techniques have given rise to neural networks with the ability to learn and predict complex dynamics at reduced computational costs. Here we introduce two digital twins of a SEIRS model applied to an idealised town. The SEIRS model has been modified to take account of spatial variation and, where possible, the model parameters are based on official virus spreading data from the UK. We compare predictions from one digital twin based on a data-corrected Bidirectional Long Short-Term Memory network with predictions from another digital twin based on a predictive Generative Adversarial Network. The predictions given by these two frameworks are accurate when compared to the original SEIRS model data. Additionally, these frameworks are data-agnostic and could be applied to towns, idealised or real, in the UK or in other countries. Also, more compartments could be included in the SEIRS model, in order to study more realistic epidemiological behaviour.