Interpretable Aspect-Aware Capsule Network for Peer Review Based Citation Count Prediction

Citation count prediction is an important task for estimating the future impact of research papers. Most of the existing works utilize the information extracted from the paper itself. In this article, we focus on how to utilize another kind of useful data signal (i.e., peer review text) to improve both the performance and interpretability of the prediction models. Specially, we propose a novel aspect-aware capsule network for citation count prediction based on review text. It contains two major capsule layers, namely the feature capsule layer and the aspect capsule layer, with two different routing approaches, respectively. Feature capsules encode the local semantics from review sentences as the input of aspect capsule layer, whereas aspect capsules aim to capture high-level semantic features that will be served as final representations for prediction. Besides the predictive capacity, we also enhance the model interpretability with two strategies. First, we use the topic distribution of the review text to guide the learning of aspect capsules so that each aspect capsule can represent a specific aspect in the review. Then, we use the learned aspect capsules to generate readable text for explaining the predicted citation count. Extensive experiments on two real-world datasets have demonstrated the effectiveness of the proposed model in both performance and interpretability.

Case Report: Favorable Response and Manageable Toxicity to the Combination of Camrelizumab, Oxaliplatin, and Oral S-1 in a Patient With Advanced Epstein–Barr Virus-Associated Gastric Cancer

Some pertinent studies have demonstrated that Epstein–Barr virus-associated gastric cancer (EBVaGC) patients showed a favorable clinical outcome to immunotherapy and Epstein–Barr virus (EBV)-positive status might be a potential biomarker for immunotherapy in gastric cancer (GC). However, knowledge of given exposure to EBVaGC to the first-line immunotherapy is largely inadequate. Moreover, whether camrelizumab can be as effective as other PD-1 inhibitors in the treatment of advanced EBVaGC has not been reported. We report a case of advanced EBVaGC patient with a positive expression of PD-L1, enriched PD-L1+CD68+macrophages, and high TMB who had a long-term partial response and manageable toxicity to the combined approach of camrelizumab (a novel PD-1 inhibitor) and oxaliplatin plus oral S-1 (SOX). As the first-line treatment of advanced EBVaGC patients, camrelizumab combined with SOX regimen may provide a novel combined approach with favorable response and manageable safety. Combination of multiple biomarkers could have a higher effective predictive capacity to immunotherapy. Integrated treatment (chemo-immunotherapy and radiotherapy) might be the optimal strategy for patients with oligometastasis. It deserves prospective research to further validate the efficacy.

A Combined-Radiomics Approach of CT Images to Predict Response to Anti-PD-1 Immunotherapy in NSCLC: A Retrospective Multicenter Study

ObjectiveBased on non-contrast-enhanced (NCE)/contrast-enhanced (CE) computed tomography (CT) images, we try to identify a combined-radiomics model and evaluate its predictive capacity regarding response to anti-PD1 immunotherapy of patients with non-small-cell lung cancer (NSCLC).Methods131 patients with NSCLC undergoing anti-PD1 immunotherapy were retrospectively enrolled from 7 institutions. Using largest lesion (LL) and target lesions (TL) approaches, we performed a radiomics analysis based on pretreatment NCE-CT (NCE-radiomics) and CE-CT images (CE-radiomics), respectively. Meanwhile, a combined-radiomics model based on NCE-CT and CE-CT images was constructed. Finally, we developed their corresponding nomograms incorporating clinical factors. ROC was used to evaluate models’ predictive performance in the training and testing set, and a DeLong test was employed to compare the differences between different models.ResultsFor TL approach, both NCE-radiomics and CE-radiomics performed poorly in predicting response to immunotherapy. For LL approach, NCE-radiomics nomograms and CE-radiomics nomograms incorporating with clinical factor of distant metastasis all showed satisfactory results, reflected by the AUCs in the training (AUC=0.84, 95% CI: 0.75-0.92; AUC=0.77, 95% CI: 0.67-0.87) and test sets (AUC=0.78, 95% CI: 0.64-0.92, AUC=0.73, 95% CI: 0.57-0.88), respectively. Compared with the NCE-radiomics nomograms, the combined-radiomics nomogram showed incremental predictive capacity in the training set (AUC=0.85, 95% CI: 0.77-0.92) and test set (AUC=0.81, 95% CI: 0.67-0.94), respectively, but no statistical difference (P=0.86, P=0.79).ConclusionCompared with radiomics based on single NCE or CE-CT images, the combined-radiomics model has potential advantages to identify patients with NSCLC most likely to benefit from immunotherapy, and may effectively improve more precise and individualized decision support.

Predictive Model of Quality of Life in Patients with Parkinson’s Disease

Parkinson’s disease is a chronic, progressive, and disabling neurodegenerative disease which evolves until the end of life and triggers different mood and organic alterations that influence health-related quality of life. The objective of our study was to identify the factors that negatively impact the quality of life of patients with Parkinson’s disease and construct a predictive model of health-related quality of life in these patients. Methods: An analytical, prospective observational study was carried out, including Parkinson’s patients at different stages in the Albacete Health Area. The sample consisted of 155 patients (T0) who were followed up at one (T1) and two years (T2). The instruments used were a purpose-designed data collection questionnaire and the “Parkinson’s Disease Questionnaire” (PDQ-39), with a global index where a higher score indicates a worse quality of life. A multivariate analysis was performed by multiple linear regression at T0. Next, the model’s predictive capacity was evaluated at T1 and T2 using the area under the ROC curve (AUROC). Results: Predictive factors were: sex, living in a residence, using a cane, using a wheelchair, having a Parkinson’s stage of HY > 2, having Alzheimer’s disease or a major neurocognitive disorder, having more than five non-motor symptoms, polypharmacy, and disability greater than 66%. This model showed good predictive capacity at one year and two years of follow-up, with an AUROC of 0.89 (95% CI: 0.83–0.94) and 0.83 (95% CI: 0.76–0.89), respectively. Conclusions: A predictive model constructed with nine variables showed a good discriminative capacity to predict the quality of life of patients with Parkinson’s disease at one and two years of follow-up.

Clinical Presentation of the SARS-CoV-2 Virus Infection and Predictive Validity of the PCR Test in Primary Health Care Worker Patients of the Spanish National Health System

Background: Despite the impact that the SARS-CoV-2 virus infection has presented in Spain, data on the diagnostic capacity of the symptoms associated with this infection are limited, especially among patients with mild symptoms and who are detected in the primary care field (PC). The objective of the present study was to know the associated symptoms and their predictive criterial validity in SARS-CoV-2 infection among professionals working in PC. Methods: A cross-sectional, multicenter study was carried out in the Spanish National Health System, through an epidemiological survey directed to patients who underwent the PCR test for SARS-CoV-2 in the PC setting. Results: A total of 1612 patients participated, of which 86.6% were PC healthcare professionals, and of these, 67.4% family doctors. Hyposmia, with a sensitivity of 42.69% (95% CI: 37.30–48.08) and a specificity of 95.91% (95% CI: 94.78–97.03), and ageusia with a sensitivity of 39.47% (34.15–44.80) and a specificity of 95.20% (93.98–96.41) were the symptoms with the highest criteria validity indexes. Conclusions: This study identifies the specific symptoms of loss of smell or taste as the most frequently associated with SARS-CoV-2 infection, essential in the detection of COVID-19 given its high frequency and predictive capacity.

Kennard-Stone method outperforms the Random Sampling in the selection of calibration samples in SNPs and NIR data

ABSTRACT: Splitting the whole dataset into training and testing subsets is a crucial part of optimizing models. This study evaluated the influence of the choice of the training subset in the construction of predictive models, as well as on their validation. For this purpose we assessed the Kennard-Stone (KS) and the Random Sampling (RS) methods in near-infrared spectroscopy data (NIR) and marker data SNPs (Single Nucleotide Polymorphisms). It is worth noting that in SNPs data, there is no knowledge of reports in the literature regarding the use of the KS method. For the construction and validation of the models, the partial least squares (PLS) estimation method and the Bayesian Lasso (BLASSO) proved to be more efficient for NIR data and for marker data SNPs, respectively. The evaluation of the predictive capacity of the models obtained after the data partition occurred through the correlation between the predicted and the observed values, and the corresponding square root of the mean squared error of prediction. For both datasets, results indicated that the results from KS and RS methods differ statistically from each other by the F test (P-value < 0.01). The KS method showed to be more efficient than RS in practically all repetitions. Also, KS method has the advantage of being easy and fast to be applied and also to select the same samples, which provides excellent benefits in the following analyses.

Evaluation of Four Lymph Node Classifications for the Prediction of Survival in Hilar Cholangiocarcinoma

Background An important prognostic indicator of hilar cholangiocarcinoma (HCCA) in patients after surgery is metastasis of lymph nodes (LN). However, there are many types of LN staging systems to the issue of a better determination of the prognosis of patients through the lymphatic staging system which needs research. Based on the above, we tried to re-evaluate the staging system of HCCA LNs. We compared the American Joint Committee on Cancer (AJCC), number of metastatic LNs (MLN), ratio of LN (LNR), and log odds of MLNs (LODDS) in individuals undergoing curative resection to determine the best LN staging system. Methods In the current study, we retrospectively analyzed 229 patients undergoing curative resection. We evaluated the impact of the stage of AJCC pN, LNR, LODDS, and MLN on OS (overall survival) and RFS (recurrence-free survival). According to the curve of receiver operating characteristic (ROC), we compared the predictive capacity of different staging systems of LN for survival and recurrence. Results Multivariate analysis results revealed that LODDS > − 0.45 (95% CI = 1.115–2.709, P = 0.015; 95% CI = 1.187–2.780, P = 0.006) are independent risk factors affecting OS and RFS, respectively. Compared with LN status, AJCC pN stage, MLN, and LNR, the variable having the highest area under the ROC curve (AUC) was LODDS when predicting 1-year, 3-year, and 5-year OS and RFS. Conclusion This study shows that metastasis of LNs is a key indicator for predicting patient death and recurrence. Among them, LODDS is the best LN staging system for the prognostic evaluation of HCCA patients after surgery. Clinicians can incorporate LODDS into HCCA patient lymphatic staging system for a more accurate prognosis of HCCA patients post-surgery.

The Dynamic Effects of Anchoring on Portfolio Return Volatility; Are there any economic issues in the developing world?

This study aims to see the anchoring effect on portfolio return volatility in the case of KSE-30. Business anomalies such as overreaction and under-reaction are affected by a variety of psychological causes. The use of anchors or baseline values known as the anchoring effect causes market under-reaction and overreaction. This research used nearness to 52-week high and nearness to historical high as proxies for under and over-reaction, respectively, to analyze the psychological causes for under and over-reaction. On the KSE-30, the findings revealed that proximity to the 52-week peak positively predicts future returns, whereas proximity to the historical high negatively predicts future returns. KSE-30 was used for rigorous testing. Similarly, the three macroeconomic variables used as control variables are the exchange rate, inflation rate, and interest rate to provide a more robust model of strong prediction capacity. The findings revealed that proximity to the 52-week maximum and proximity to the historical high and other macroeconomic factors had a forecast capacity of around 62 percent. Similarly, focused on volatility clusters, the GARCH (1, 1) model was used to measure the association between potential and past returns. The results show that there is a first order autoregressive function in the GARCH (1, 1) model. The findings also show that their predictive capacity decreases when the study's individual variables are moved from every day to annual Periods.

SCORE2 Assessment in the Calculation of Cardiovascular Risk in Patients with Rheumatoid Arthritis

Patients with rheumatoid arthritis (RA) are at increased risk for cardiovascular disease (CVD). Risk chart algorithms, such as the Systematic Coronary Risk Assessment (SCORE), often underestimate the risk of CVD in patients with RA. In this sense, the use of noninvasive tools, such as the carotid ultrasound, has made it possible to identify RA patients at high risk of CVD who had subclinical atherosclerosis disease and who had been included in the low or moderate CVD risk categories when the SCORE risk tables were applied. The 2003 SCORE calculator was recently updated to a new prediction model: SCORE2. This new algorithm improves the identification of individuals from the general population at high risk of developing CVD in Europe. Our objective was to compare the predictive capacity between the original SCORE and the new SCORE2 to identify RA patients with subclinical atherosclerosis and, consequently, high risk of CVD. 1168 non-diabetic patients with RA and age > 40 years were recruited. Subclinical atherosclerosis was searched for by carotid ultrasound. The presence of carotid plaque and the carotid intima media wall thickness (cIMT) were evaluated. SCORE and SCORE2 were also calculated. The relationships of SCORE and SCORE2 to each other and to the presence of subclinical carotid atherosclerosis were studied. The correlation between SCORE and SCORE2 was found to be high in patients with RA (Spearman’s Rho = 0.961, p < 0.001). Both SCORE (Spearman’s Rho = 0.524) and SCORE2 (Spearman’s Rho = 0.521) were similarly correlated with cIMT (p = 0.92). Likewise, both calculators showed significant and comparable discriminations for the presence of carotid plaque: SCORE AUC 0.781 (95%CI 0.755–0.807) and SCORE2 AUC 0.774 (95%CI 0.748–0.801). Using SCORE, 80% and 20% of the patients were in the low or moderate and high or very high CVD risk categories, respectively. However, when the same categories were evaluated using SCORE2, the percentages were different (58% and 42%, respectively). Consequently, the number of RA patients included in the high or very high CVD risk categories was significantly higher with SCORE2 compared to the original SCORE. (p < 0.001). In conclusion, although predictive capacity for the presence of carotid plaque is equivalent between SCORE and SCORE2, SCORE2 identifies a significantly higher proportion of patients with RA who are at high or very high risk of CVD.