Clinical characteristics of airway impairment assessed by impulse oscillometry in patients with chronic obstructive pulmonary disease: findings from the ECOPD study in China.

BACKGROUND: The role of airway impairment assessed by impulse oscillometry (IOS) in patients with chronic obstructive pulmonary disease (COPD) remains unclear. Therefore, this study aimed to analyze the proportion and clinical characteristics of airway impairment assessed by IOS across COPD severities, and explore whether airway impairment is a subtype of COPD. METHODS: This study was based on cross-sectional data from the ECOPD cohort in Guangdong, China. Subjects were consecutively recruited from July 2019 to August 2021. They filled out questionnaires and underwent lung function tests, IOS and computed tomography (CT). COPD was defined as post-bronchodilator forced expiratory volume in one second/forced vital capacity < lower limit of normal (LLN). Meanwhile, airway impairment was defined as IOS parameters > upper limit of normal or < LLN. On the one hand, Poisson regression was employed to analyze the associations between acute exacerbations of COPD (AECOPD) in the previous year and airway impairment. On the other hand, logistic regression was used to assess differences in CT imaging between patients with IOS parameters' abnormalities and patients with normal IOS parameters. RESULTS: 768 COPD subjects were finally enrolled in the study. The proportion of airway impairment assessed by R5, R20, R5-R20, X5, AX, and Fres was 59.8%, 29.7%, 62.5%, 52.9%, 60.9% and 67.3%, respectively. Airway impairment assessed by IOS parameters (R5, R5-R20, X5, AX, and Fres) in patients with COPD was present across all severities of COPD, particularly in GOLD 3-4 patients. Compared with patients with normal IOS parameters, patients with IOS parameters' abnormalities had more respiratory symptoms, more severe airway obstruction and imaging structural abnormalities. Patients with IOS parameters' abnormalities assessed by R5 [risk ratio (RR): 1.58, 95% confidential interval (CI): 1.13-2.19, P = 0.007], R5-R20 [RR: 1.73, 95%CI: 1.22-2.45, P = 0.002], X5 [RR: 2.11, 95%CI: 1.51-2.95, P < 0.001], AX [RR: 2.20, 95%CI: 1.53-3.16, P < 0.001], and Fres [RR: 2.13, 95%CI: 1.44-3.15, P < 0.001] had a higher risk of AECOPD in the previous year than patients with normal IOS parameters. CONCLUSIONS: Airway impairment assessed by IOS may be a subtype of COPD. Future studies are warranted to identify the underlying mechanisms and longitudinal progression of airway impairment.

The association between small airway dysfunction and aging: a cross-sectional analysis from the ECOPD cohort.

BACKGROUND: Aging has been evidenced to bring about some structural and functional lung changes, especially in COPD. However, whether aging affects SAD, a possible precursor of COPD, has not been well characterized. OBJECTIVE: We aimed to comprehensively assess the relationship between aging and SAD from computed tomography, impulse oscillometry, and spirometry perspectives in Chinese. METHODS: We included 1859 participants from ECOPD, and used a linear-by-linear association test for evaluating the prevalence of SAD across various age subgroups, and multivariate regression models for determining the impact of age on the risk and severity of SAD. We then repeated the analyses in these subjects stratified by airflow limitation. RESULTS: The prevalence of SAD increases over aging regardless of definitional methods. After adjustment for other confounding factors, per 10-yrs increase in age was significantly associated with the risk of CT-defined SAD (OR 2.57, 95% CI 2.13 to 3.10) and the increase in the severity of air trapping (ß 2.09, 95% CI - 0.06 to 4.25 for LAA-856), airway reactance (ß - 0.02, 95% CI - 0.04 to - 0.01 for X5; ß 0.30, 95% CI 0.13 to 0.47 for AX; ß 1.75, 95% CI 0.85 to 2.66 for Fres), as well as the decrease in expiratory flow rates (ß - 3.95, 95% CI - 6.19 to - 1.71 for MMEF%predicted; ß - 5.42, 95% CI - 7.88 to - 2.95 for FEF50%predicted) for SAD. All these associations were generally maintained in SAD defined by IOS or spirometry. After stratification of airflow limitation, we further found that the effect of age on LAA-856 was the most significant among almost all subgroups. CONCLUSIONS: Aging is significantly associated with the prevalence, increased risk, as well as worse severity of SAD. CT may be a more optimal measure to assess aging-related SAD. The molecular mechanisms for the role of aging in SAD need to be explored in the future. Trial registration Chinese Clinical Trial Registry ChiCTR1900024643. Registered on 19 July 2019.

Novel Convolutional Neural Network with Variational Information Bottleneck for P300 Detection.

In the area of brain-computer interfaces (BCI), the detection of P300 is a very important technique and has a lot of applications. Although this problem has been studied for decades, it is still a tough problem in electroencephalography (EEG) signal processing owing to its high dimension features and low signal-to-noise ratio (SNR). Recently, neural networks, like conventional neural networks (CNN), has shown excellent performance on many applications. However, standard convolutional neural networks suffer from performance degradation on dealing with noisy data or data with too many redundant information. In this paper, we proposed a novel convolutional neural network with variational information bottleneck for P300 detection. Wiht the CNN architecture and information bottleneck, the proposed network termed P300-VIB-Net could remove the redundant information in data effectively. The experimental results on BCI competition data sets show that P300-VIB-Net achieves cutting-edge character recognition performance. Furthermore, the proposed model is capable of restricting the flow of irrelevant information adaptively in the network from perspective of information theory. The experimental results show that P300-VIB-Net is a promising tool for P300 detection.

Characteristics of inflammatory phenotypes in patients with chronic obstructive pulmonary disease: a cross-sectional study.

BACKGROUND: The relationship between airway inflammation in chronic obstructive pulmonary disease (COPD) and clinical characteristics remains unclear. This study aimed to investigate the airway inflammatory phenotypes in COPD and their association with clinical characteristics. METHODS: 895 patients with COPD were recruited from Guangdong Province, China in this study. Each patient underwent questionnaire interviews, spirometry testing, CT scans and induced sputum examination. Classification of airway inflammation phenotypes was based on sputum inflammatory cell counts. Covariance analysis was applied to assess associations with airway inflammation phenotypes. RESULTS: In this study, we found that neutrophilic phenotype (NP, 58.0%) was the most common airway inflammation phenotype in patients with COPD, followed by mixed granulocytic phenotype (MGP, 32.6%), eosinophilic phenotype (EP, 5.4%) and paucigranulocytic phenotype (PP, 4.0%). Compared with NP patients, those with MGP exhibited more frequent chronic respiratory symptoms, and a higher proportion of individuals classified under Global Initiative for Chronic Obstructive Lung Disease stages 3 and 4. After adjusting for confounding factors, MGP patients had lower lung function, and more severe emphysema and air trapping. On the contrary, patients with PP had the best pulmonary function and less emphysema and air trapping. CONCLUSIONS: NP was the most common airway inflammation phenotype in patients with COPD. Patients with MGP had more respiratory symptoms, greater loss of lung function, and more severe emphysema and gas trapping compared with those with NP. Meanwhile, PP may be a phenotype of mild damage to lung structure in patients with COPD.

Association Between Serum Total Bilirubin Level and Lung Function Decline in Patients with COPD: Results from a Pooled Study.

Background: Serum total bilirubin has been reported to have antioxidant properties against chronic respiratory diseases. The objective of our study is to evaluate the association of total bilirubin (TB) with annual lung function decline in COPD patients with different GOLD stages. Methods: This study used pooled data from two observational and prospective cohorts of 612 COPD patients whose TB levels were measured at baseline. The associations between TB and postbronchodilator FEV1, FEV1pred, FVC, FVCpred, FEV1/FVC, and the rate of their decline were all determined using linear regression models in the total population and strata of GOLD stages. Results: Serum TB was positively related to FEV1 and FVC in the total group (ß 0.02, 95% CI 0.001~0.02, P = 0.025 and ß 0.02, 95% CI 0.002~0.03, P = 0.022, respectively). Additionally, TB was inversely associated with the annual decline in FEV1 and FEV1pred (ß 4.91, 95% CI 1.68~8.14, P = 0.025 and ß 0.21, 95% CI 0.06~0.36, P = 0.022, respectively) when adjusted for multivariables. After stratification, the significant associations merely persisted in COPD patients with GOLD 2 and GOLD 3-4. Conclusion: Increased TB level was related to less annual decline in FEV1 as well as FEV1pred in moderate-to-severe COPD but not mild COPD, which indicated the different status of TB in different COPD severity and the possible role as potential biomarker merely in moderate-to-severe COPD. Future researches to determine whether TB could be served as biomarker for COPD and the mechanisms should be focused on some target patients with a certain disease severity.

Association Between Serum Total Bilirubin and COPD: Results from a Cross-Sectional Study and a Bidirectional Mendelian Randomization Analysis.

Background: The potential protective role of serum total bilirubin (TB) for chronic obstructive pulmonary disease (COPD) is controversial. We aimed to investigate whether serum TB could prevent airflow limitation (reduce the risk of COPD) and whether these associations were causal or reversely causal. Methods: We conducted a multi-center and cross-sectional study including 3069 participants. Logistic regression model (LRM) with restricted cubic spline (RCS) and priori defined quintile categories were used to assess the associations of TB with COPD. Besides, ordinary least squares (OLS) regression model with RCS curves were applied to assess the dose-response relationship between serum TB and airflow limitation (FEV1/FVC). To verify the causal direction between TB and COPD, a bidirectional Mendelian randomization analysis was carried out with GWAS data from European ancestry. Results: In the cross-sectional study, the relationship between levels of TB and COPD risk was U shaped (P=0.001), and the low and high concentrations of TB apparently increasing the risk of COPD (OR 1.40, 95% CI 1.07 to 1.82 for less than 9 µmol/L; OR 1.36, 95% CI 1.06 to 1.76 for 9.01-1 0.88 µmol/L; OR 1.50, 95% CI 1.16 to 1.95 for more than 13 µmol/L). There was a significant non-linear relationship between TB and FEV1/FVC (non-linear p=0.004). Furthermore, results of bidirectional Mendelian randomization analysis (OR 1.000; 95% CI 0.983 to 1.017 for MR and OR 0.998; 95% CI 0.976 to 1.020 for reversal MR) did not support the causal effects between serum TB and FEV1/FVC after controlling the effect of potential confounders and revised causality. Conclusion: Our study reveals that there was non-linear does-response pattern between serum TB and COPD. However, there was little evidence for the linear causal associations of serum TB with airflow limitation. The relationship of TB with COPD needs further study and careful interpretation.

Traffic-related air pollution is a risk factor in the development of chronic obstructive pulmonary disease.

Background: Outdoor traffic-related air pollution has negative effects on respiratory health. In this study, we aimed to explore the effect of outdoor traffic-related air pollution on chronic obstructive pulmonary disease (COPD) in Guangzhou. Methods: We enrolled 1,460 residents aged 40 years or older between 21 January 2014 and 31 January 2018. We administered questionnaires and spirometry tests. The distance of participants' residences or locations of outdoor activities from busy roads (as indicators of outdoor traffic-related air pollution), indoor air pollution, and smoking history were queried in the questionnaires. Results: Of the 1,460 residents with valid survey and test results, 292 were diagnosed with COPD, with a detection rate of 20%. Participants who lived and did their outdoor activities near busy roads had a higher detection rate of COPD. Among residents living at distances of <50 meters, 50-199 meters, and more than 200 meters from busy roads, the detection rates were 20.6, 21.2, and 14.8%, respectively; the rates for outdoor activities at these distances were 23.8, 24.5, and 13.7%, respectively (p < 0.05). After adjusting for sex, age, smoking status, family history, and smoking index, the distance of outdoor activities from busy roads was an independent risk factor for COPD. Participants whose outdoor activities were conducted <50 meters and 50-199 meters of main roads had odds ratios of 1.54 (95% confidence interval 1.01-2.36) and 1.84 (95% interval 1.23-2.76) for the risk of COPD in comparison with a distance of more than 200 meters from busy roads. Conclusions: Residents of Guangzhou whose outdoor activities were close to busy roads had a high risk of COPD. Traffic-related air pollution presents a risk to human health and a risk of COPD.

Predictors of High Sputum Eosinophils in Chronic Obstructive Pulmonary Disease.

Background: Eosinophils are involved in the development of chronic obstructive pulmonary disease (COPD) and inhaled corticosteroid responsiveness. We evaluated clinical predictors of high sputum eosinophil levels in a COPD cohort in China. Methods: We conducted an observational, prospective, population-based, cross-sectional study. Participants were tested for COPD and underwent spirometry, computed tomography scans, and a blood test. Participants also produced induced sputum and responded to an information-gathering questionnaire. High sputum eosinophils were defined as ≥3.0%. Multivariate logistic regression was used to identify predictors of high sputum eosinophil levels. Results: We recruited 895 patients with complete and quality control data. The median percentage of sputum eosinophil abundance was 2.00% (interquartile range: 0.75-5.00) and the prevalence of COPD with high sputum eosinophils was 38.0%. Covariance analysis indicated that the high sputum eosinophil group had lower lung function, more severe emphysema, and air trapping. Multivariate logistic regression indicated that high blood eosinophil levels, severe respiratory symptoms, being a former smoker, and a family history of respiratory diseases were associated with high sputum eosinophil levels. Conclusion: High blood eosinophil levels, severe respiratory symptoms, being a former smoker, and a family history of respiratory diseases may be predictors of high sputum eosinophil levels in Chinese COPD patients. High sputum eosinophils were associated with lower lung function, more emphysema, and gas trapping.

Massive-training support vector regression and Gaussian process for false-positive reduction in computer-aided detection of polyps in CT colonography.

PURPOSE: A massive-training artificial neural network (MTANN) has been developed for the reduction of false positives (FPs) in computer-aided detection (CADe) of polyps in CT colonography (CTC). A major limitation of the MTANN is the long training time. To address this issue, the authors investigated the feasibility of two state-of-the-art regression models, namely, support vector regression (SVR) and Gaussian process regression (GPR) models, in the massive-training framework and developed massive-training SVR (MTSVR) and massive-training GPR (MTGPR) for the reduction of FPs in CADe of polyps. METHODS: The authors applied SVR and GPR as volume-processing techniques in the distinction of polyps from FP detections in a CTC CADe scheme. Unlike artificial neural networks (ANNs), both SVR and GPR are memory-based methods that store a part of or the entire training data for testing. Therefore, their training is generally fast and they are able to improve the efficiency of the massive-training methodology. Rooted in a maximum margin property, SVR offers excellent generalization ability and robustness to outliers. On the other hand, GPR approaches nonlinear regression from a Bayesian perspective, which produces both the optimal estimated function and the covariance associated with the estimation. Therefore, both SVR and GPR, as the state-of-the-art nonlinear regression models, are able to offer a performance comparable or potentially superior to that of ANN, with highly efficient training. Both MTSVR and MTGPR were trained directly with voxel values from CTC images. A 3D scoring method based on a 3D Gaussian weighting function was applied to the outputs of MTSVR and MTGPR for distinction between polyps and nonpolyps. To test the performance of the proposed models, the authors compared them to the original MTANN in the distinction between actual polyps and various types of FPs in terms of training time reduction and FP reduction performance. The authors' CTC database consisted of 240 CTC data sets obtained from 120 patients in the supine and prone positions. The training set consisted of 27 patients, 10 of which had 10 polyps. The authors selected 10 nonpolyps (i.e., FP sources) from the training set. These ten polyps and ten nonpolyps were used for training the proposed models. The testing set consisted of 93 patients, including 19 polyps in 7 patients and 86 negative patients with 474 FPs produced by an original CADe scheme. RESULTS: With the MTSVR, the training time was reduced by a factor of 190, while a FP reduction performance [by-polyp sensitivity of 94.7% (18/19) with 2.5 (230/93) FPs/patient] comparable to that of the original MTANN [the same sensitivity with 2.6 (244/93) FPs/patient] was achieved. The classification performance in terms of the area under the receiver-operating-characteristic curve value of the MTGPR (0.82) was statistically significantly higher than that of the original MTANN (0.77), with a two-sided p-value of 0.03. The MTGPR yielded a 94.7% (18/19) by-polyp sensitivity at a FP rate of 2.5 (235/93) per patient and reduced the training time by a factor of 1.3. CONCLUSIONS: Both MTSVR and MTGPR improve the efficiency of the training in the massive-training framework while maintaining a comparable performance.

Computer-aided measurement of liver volumes in CT by means of geodesic active contour segmentation coupled with level-set algorithms.

PURPOSE: Computerized liver extraction from hepatic CT images is challenging because the liver often abuts other organs of a similar density. The purpose of this study was to develop a computer-aided measurement of liver volumes in hepatic CT. METHODS: The authors developed a computerized liver extraction scheme based on geodesic active contour segmentation coupled with level-set contour evolution. First, an anisotropic diffusion filter was applied to portal-venous-phase CT images for noise reduction while preserving the liver structure, followed by a scale-specific gradient magnitude filter to enhance the liver boundaries. Then, a nonlinear grayscale converter enhanced the contrast of the liver parenchyma. By using the liver-parenchyma-enhanced image as a speed function, a fast-marching level-set algorithm generated an initial contour that roughly estimated the liver shape. A geodesic active contour segmentation algorithm coupled with level-set contour evolution refined the initial contour to define the liver boundaries more precisely. The liver volume was then calculated using these refined boundaries. Hepatic CT scans of 15 prospective liver donors were obtained under a liver transplant protocol with a multidetector CT system. The liver volumes extracted by the computerized scheme were compared to those traced manually by a radiologist, used as "gold standard." RESULTS: The mean liver volume obtained with our scheme was 1504 cc, whereas the mean gold standard manual volume was 1457 cc, resulting in a mean absolute difference of 105 cc (7.2%). The computer-estimated liver volumetrics agreed excellently with the gold-standard manual volumetrics (intraclass correlation coefficient was 0.95) with no statistically significant difference (F = 0.77; p(F < or = f) = 0.32). The average accuracy, sensitivity, specificity, and percent volume error were 98.4%, 91.1%, 99.1%, and 7.2%, respectively. Computerized CT liver volumetry would require substantially less completion time (compared to an average of 39 min per case by manual segmentation). CONCLUSIONS: The computerized liver extraction scheme provides an efficient and accurate way of measuring liver volumes in CT.

A new nonlinear similarity measure for multichannel signals.

We propose a novel similarity measure, called the correntropy coefficient, sensitive to higher order moments of the signal statistics based on a similarity function called the cross-correntopy. Cross-correntropy nonlinearly maps the original time series into a high-dimensional reproducing kernel Hilbert space (RKHS). The correntropy coefficient computes the cosine of the angle between the transformed vectors. Preliminary experiments with simulated data and multichannel electroencephalogram (EEG) signals during behaviour studies elucidate the performance of the new measure versus the well-established correlation coefficient.

A new classifier based on information theoretic learning with unlabeled data.

Supervised learning is conventionally performed with pairwise input-output labeled data. After the training procedure, the adaptive system's weights are fixed while the testing procedure with unlabeled data is performed. Recently, in an attempt to improve classification performance unlabeled data has been exploited in the machine learning community. In this paper, we present an information theoretic learning (ITL) approach based on density divergence minimization to obtain an extended training algorithm using unlabeled data during the testing. The method uses a boosting-like algorithm with an ITL based cost function. Preliminary simulations suggest that the method has the potential to improve the performance of classifiers in the application phase.

Max-AUC feature selection in computer-aided detection of polyps in CT colonography.

We propose a feature selection method based on a sequential forward floating selection (SFFS) procedure to improve the performance of a classifier in computerized detection of polyps in CT colonography (CTC). The feature selection method is coupled with a nonlinear support vector machine (SVM) classifier. Unlike the conventional linear method based on Wilks' lambda, the proposed method selected the most relevant features that would maximize the area under the receiver operating characteristic curve (AUC), which directly maximizes classification performance, evaluated based on AUC value, in the computer-aided detection (CADe) scheme. We presented two variants of the proposed method with different stopping criteria used in the SFFS procedure. The first variant searched all feature combinations allowed in the SFFS procedure and selected the subsets that maximize the AUC values. The second variant performed a statistical test at each step during the SFFS procedure, and it was terminated if the increase in the AUC value was not statistically significant. The advantage of the second variant is its lower computational cost. To test the performance of the proposed method, we compared it against the popular stepwise feature selection method based on Wilks' lambda for a colonic-polyp database (25 polyps and 2624 nonpolyps). We extracted 75 morphologic, gray-level-based, and texture features from the segmented lesion candidate regions. The two variants of the proposed feature selection method chose 29 and 7 features, respectively. Two SVM classifiers trained with these selected features yielded a 96% by-polyp sensitivity at false-positive (FP) rates of 4.1 and 6.5 per patient, respectively. Experiments showed a significant improvement in the performance of the classifier with the proposed feature selection method over that with the popular stepwise feature selection based on Wilks' lambda that yielded 18.0 FPs per patient at the same sensitivity level.

[Cited status and its correlated factors of articles published in Chinese Journal of Pediatrics from 2001 to 2010].

OBJECTIVE: The present study was designed to explore the cited status and its correlated factors of articles published in Chinese Journal of Pediatrics. METHOD: Articles published in Chinese Journal of Pediatrics from 2001 to 2010 were searched using Wanfang Medical Online database, and the relationship between cited number and column and funding status were analyzed. RESULTS: Totally 3209 articles were published by Chinese Journal of Pediatrics from 2001 to 2010. Two thousand and seventy-three articles (64.60%) were cited. The total cited number was 18 546 (mean rate: 5.78 per paper). Standard/protocol/guideline was the most often cited column (mean rate: 62.92 per paper). Featured articles including editorials (mean rate: 7.12 per paper), special articles (mean rate: 6.50 per paper) and original articles (mean rate: 7.90 per paper) had higher cited rate. Cited rate of original papers in featured articles was higher than other original articles (mean rate: 6.03 per paper). Those with academic perspectives, such as Opinion/Debate/Discussion, were well cited (mean rate: 7.09 per paper). All of original articles in Rapid Pathway were well cited (mean rate: 16.20 per paper). Mean cited number of grant-supported articles was 4.81 per paper, lower than that of all kinds of papers (mean rate: 5.78 per paper) and non-grant-supported articles (mean rate: 6.06 per paper). However, it was slightly higher than that of articles except Standard/protocol/guideline (mean rate: 4.36 per paper). CONCLUSION: Cited status varies among columns. The invited papers should be increased in number to raise commentary papers and original articles with high quality.Grant-supported or non-grant-supported papers should be reviewed based on the same standard after submission.

Massive-training artificial neural network coupled with Laplacian-eigenfunction-based dimensionality reduction for computer-aided detection of polyps in CT colonography.

A major challenge in the current computer-aided detection (CAD) of polyps in CT colonography (CTC) is to reduce the number of false-positive (FP) detections while maintaining a high sensitivity level. A pattern-recognition technique based on the use of an artificial neural network (ANN) as a filter, which is called a massive-training ANN (MTANN), has been developed recently for this purpose. The MTANN is trained with a massive number of subvolumes extracted from input volumes together with the teaching volumes containing the distribution for the "likelihood of being a polyp;" hence the term "massive training." Because of the large number of subvolumes and the high dimensionality of voxels in each input subvolume, the training of an MTANN is time-consuming. In order to solve this time issue and make an MTANN work more efficiently, we propose here a dimension reduction method for an MTANN by using Laplacian eigenfunctions (LAPs), denoted as LAP-MTANN. Instead of input voxels, the LAP-MTANN uses the dependence structures of input voxels to compute the selected LAPs of the input voxels from each input subvolume and thus reduces the dimensions of the input vector to the MTANN. Our database consisted of 246 CTC datasets obtained from 123 patients, each of whom was scanned in both supine and prone positions. Seventeen patients had 29 polyps, 15 of which were 5-9 mm and 14 were 10-25 mm in size. We divided our database into a training set and a test set. The training set included 10 polyps in 10 patients and 20 negative patients. The test set had 93 patients including 19 polyps in seven patients and 86 negative patients. To investigate the basic properties of a LAP-MTANN, we trained the LAP-MTANN with actual polyps and a single source of FPs, which were rectal tubes. We applied the trained LAP-MTANN to simulated polyps and rectal tubes. The results showed that the performance of LAP-MTANNs with 20 LAPs was advantageous over that of the original MTANN with 171 inputs. To test the feasibility of the LAP-MTANN, we compared the LAP-MTANN with the original MTANN in the distinction between actual polyps and various types of FPs. The original MTANN yielded a 95% (18/19) by-polyp sensitivity at an FP rate of 3.6 (338/93) per patient, whereas the LAP-MTANN achieved a comparable performance, i.e., an FP rate of 3.9 (367/93) per patient at the same sensitivity level. With the use of the dimension reduction architecture, the time required for training was reduced from 38 h to 4 h. The classification performance in terms of the area under the receiver-operating-characteristic curve of the LAP-MTANN (0.84) was slightly higher than that of the original MTANN (0.82) with no statistically significant difference (p-value =0.48).