The partial area under the summary ROC curve

Association between neutrophil to high-density lipoprotein cholesterol ratio and disease severity in patients with acute biliary pancreatitis

BACKGROUND

The neutrophil to high-density lipoprotein cholesterol ratio (NHR) is independently associated with the severity of various diseases. However, its association with acute biliary pancreatitis (ABP) remains unknown.

METHODS

This study included 1335 eligible patients diagnosed with ABP from April 2016 to December 2022. Patients were divided into low- and high-NHR level groups using an optimal cut-off value determined utilizing Youden's index. Multivariate logistic regression analysis was used to investigate the correlation between NHR and ABP severity. Multivariate analysis-based limited restricted cubic spline (RCS) method was used to evaluate the nonlinear relationship between NHR and the risk of developing moderate or severe ABP.

RESULTS

In this study, multivariate logistic regression analysis indicated an independent association between NHR and ABP severity (p < .001). The RCS analysis showed a linear correlation between NHR and the risk of developing moderate or severe ABP (P for non-linearity > 0.05), and increased NHR was found to be independently associated with a more severe form of the disease.

CONCLUSIONS

Our study suggests that NHR is a simple and practical independent indicator of disease severity, serving as a potential novel predictor for patients with ABP.

Relationship between platelet-to-lymphocyte ratio and early rebleeding after endoscopic variceal ligation: a bicenter retrospective study

BACKGROUND

Endoscopic variceal ligation (EVL) is the primary treatment for esophageal variceal bleeding in patients with liver cirrhosis (LC). Postoperative rebleeding is a complication of EVL, contributing to over 20% of bleeding-related deaths. This study aims to examine the association between platelet-to-lymphocyte ratio (PLR) and rebleeding within 6 weeks after EVL in patients with LC.

METHODS

The study included 145 eligible patients who underwent their first EVL procedure at Yijishan Hospital of Wannan Medical College between January 2016 and August 2022 (YJS cohort). An external validation cohort comprising 338 eligible patients from NO.2 People's Hospital of Fuyang City (FY cohort) between July 2018 and August 2022 was also utilized.

RESULTS

In the YJS cohort, Multivariate logistic analysis indicated that high PLR is independently associated with early rebleeding after EVL. The restricted cubic spline analysis demonstrated that the risk of rebleeding increases with rising PLR, stabilizing at PLR values greater than 150. Similar findings were validated in the FY cohort.

CONCLUSIONS

Our results have the potential to aid in the identification of high-risk patients for early rebleeding after EVL, thereby enabling improved clinical management and outcomes for these individuals.

Usefulness of driver's eye movement measurement to detect potential risks under combined conditions of taking second-generation antihistamines and calling tasks

BACKGROUND

Concerns persist regarding the potential reduction in driving performance due to taking second-generation antihistamines or performing hands-free calling. Previous studies have indicated a potential risk to driving performance under an emergency event when these two factors are combined, whereas a non-emergency event was operated effectively. Currently, there is a lack of a discriminative index capable of detecting the potential risks of driving performance impairment. This study aims to investigate the relationship between driving performance and eye movements under combined conditions of taking second-generation antihistamines and a calling task, and to assess the usefulness of eye movement measurements as a discriminative index for detecting potential risks of driving performance impairment.

METHODS

Participants engaged in a simulated driving task, which included a calling task, both under taking or not taking second-generation antihistamines. Driving performance and eye movements were monitored during both emergency and non-emergency events, assessing their correlation between driving performance and eye movements. The study further evaluated the usefulness of eye movement as a discriminative index for potential driving impairment risk through receiver operating characteristic (ROC) analysis.

RESULTS

In the case of a non-emergency event, no correlation was observed between driving performance and eye movement under the combined conditions. Conversely, a correlation was observed during an emergency event. The ROC analysis, conducted to assess the discriminative index capability of eye movements in detecting the potential risk of driving performance impairment, demonstrated a high discriminative power, with an area under the curve of 0.833.

CONCLUSIONS

The findings of this study show the correlation between driving performance and eye movements under the concurrent influence of second-generation antihistamines and a calling task, suggesting the usefulness of eye movement measurement as a discriminant index for detecting potential risks of driving performance impairment.

The conundrum of differentiating Cushing's syndrome from non-neoplastic hypercortisolism: a systematic review and meta-analysis

CONTEXT

Once hypercortisolemia is confirmed, differential diagnosis between Cushing's syndrome (CS) due to neoplastic endogenous hypercortisolism and non-neoplastic hypercortisolism (NNH, pseudo-Cushing's syndrome) is crucial. Due to worldwide corticotropin-releasing hormone (CRH) unavailability, accuracy of alternative tests to dexamethasone (Dex)-CRH, is clearly needed.

OBJECTIVE

Assess the diagnostic accuracy of Dex-CRH test, desmopressin stimulation test, midnight serum cortisol (MSC), and late-night salivary cortisol (LNSC) levels to distinguish CS from NNH.

METHODS

Articles through March 2022 were identified from Scopus, Web of Science, MEDLINE, EMBASE, and PubMed. All steps through the systematic review were performed independently and in duplicate and strictly adhered to the updated PRISMA-DTA checklist.

DATA SYNTHESIS

A total of 24 articles (1900 patients) were included. Dex-CRH had a pooled sensitivity and specificity of 91% (95%CI 87-94%; I2 0%) and 82% (73-88%; I2 50%), desmopressin test 86% (81-90%; I2 28%) and 90% (84-94%; I2 15%), MSC 91% (85-94%; I2 66%) and 81% (70-89%; I2 71%), and LNSC 80% (67-89%; I2 57%) and 90% (84-93%; I2 21%), respectively. Summary receiver operating characteristics areas under the curve were Dex-CRH 0.949, desmopressin test 0.936, MSC 0.942, and LNSC 0.950 without visual or statistical significance. The overall risk of studies bias was moderate.

CONCLUSION

Dex-CRH, the desmopressin stimulation test, and MSC have similar diagnostic accuracy, with Dex-CRH and MSC having slightly higher sensitivity, and the desmopressin test being more specific. LNSC was the least accurate, probably due to high heterogeneity, intrinsic variability, different assays, and lack of consistent reported cutoffs. When facing this challenging differential diagnosis, the results presented here should increase clinicians' confidence when deciding which test to perform.

The Potential MicroRNA Diagnostic Biomarkers in Oral Squamous Cell Carcinoma of the Tongue

Oral squamous cell carcinoma (OSCC) of the tongue is a common type of head and neck malignancy with a poor prognosis, underscoring the urgency for early detection. MicroRNAs (miRNAs) have remarkable stability and are easily measurable. Thus, miRNAs may be a promising biomarker candidate among biomarkers in cancer diagnosis. Biomarkers have the potential to facilitate personalized medicine approaches by guiding treatment decisions and optimizing therapy regimens for individual patients. Utilizing data from The Cancer Genome Atlas, we identified 13 differentially expressed upregulated miRNAs in OSCC of the tongue. Differentially expressed miRNAs were analyzed by enrichment analysis to reveal underlying biological processes, pathways, or functions. Furthermore, we identified miRNAs associated with the progression of OSCC of the tongue, utilizing receiver operating characteristic analysis to evaluate their potential as diagnostic biomarkers. A total of 13 upregulated miRNAs were identified as differentially expressed in OSCC of the tongue. Five of these miRNAs had high diagnostic power. In particular, miR-196b has the potential to serve as one of the most effective diagnostic biomarkers. Then, functional enrichment analysis for the target gene of miR-196b was performed, and a protein-protein interaction network was constructed. This study assessed an effective approach for identifying miRNAs as early diagnostic markers for OSCC of the tongue.

Effectiveness of metagenomic next-generation sequencing in the diagnosis of infectious diseases: A systematic review and meta-analysis

OBJECTIVES

Early diagnosis of infectious diseases remains a challenge. This study assessed the diagnostic value of mNGS in infections and explored the effect of various factors on the accuracy of mNGS.

METHODS

An electronic article search of PubMed, Cochrane Library, and Embase was performed. A total of 85 papers were eligible for inclusion and analysis. Stata 12.0 was used for statistical calculation to evaluate the efficacy of mNGS for the diagnosis of infectious diseases.

RESULTS

The AUC of 85 studies was 0.88 (95%CI, 0.85-0.90). The AUC of the clinical comprehensive diagnosis and conventional test groups was 0.92 (95%CI, 0.89-0.94) and 0.82 (95%CI, 0.78-0.85), respectively. The results of subgroup analysis indicated that the PLR and NLR were 12.67 (95%CI, 6.01-26.70) and 0.05 (95%CI, 0.03-0.10), respectively, in arthrosis infections. The PLR was 24.41 (95%CI, 5.70-104.58) in central system infections and the NLR of immunocompromised patients was 0.08 (95%CI, 0.01-0.62).

CONCLUSION

mNGS demonstrated satisfactory diagnostic performance for infections, especially for bone and joint infections and central system infections. Moreover, mNGS also has a high value in the exclusion of infection in immunocompromised patients.

Strategies for improving the performance of prediction models for response to immune checkpoint blockade therapy in cancer

Immune checkpoint blockade (ICB) therapy holds promise for bringing long-lasting clinical gains for the treatment of cancer. However, studies show that only a fraction of patients respond to the treatment. In this regard, it is valuable to develop gene expression signatures based on RNA sequencing (RNAseq) data and machine learning methods to predict a patient's response to the ICB therapy, which contributes to more personalized treatment strategy and better management of cancer patients. However, due to the limited sample size of ICB trials with RNAseq data available and the vast number of candidate gene expression features, it is challenging to develop well-performed gene expression signatures. In this study, we used several published melanoma datasets and investigated approaches that can improve the construction of gene expression-based prediction models. We found that merging datasets from multiple studies and incorporating prior biological knowledge yielded prediction models with higher predictive accuracies. Our finding suggests that these two strategies are of high value to identify ICB response biomarkers in future studies.

miR-221 and Parkinson's disease: A biomarker with therapeutic potential

Parkinson's disease (PD) is a neurodegenerative disorder characterized by the loss of dopaminergic neurons in the substantia nigra, leading to various motor and non-motor symptoms. Several cellular and molecular mechanisms such as alpha-synuclein (α-syn) accumulation, mitochondrial dysfunction, oxidative stress and neuroinflammation are involved in the pathogenesis of this disease. MicroRNAs (miRNAs) play important roles in post-transcriptional gene regulation. They are typically about 21-25 nucleotides in length and are involved in the regulation of gene expression by binding to the messenger RNA (mRNA) molecules. miRNAs like miR-221 play important roles in various biological processes, including development, cell proliferation, differentiation and apoptosis. miR-221 promotes neuronal survival against oxidative stress and neurite outgrowth and neuronal differentiation. Additionally, the role of miR-221 in PD has been investigated in several studies. According to the results of these studies, (1) miR-221 protects PC12 cells against oxidative stress induced by 6-hydroxydopamine; (2) miR-221 prevents Bax/caspase-3 signalling activation by stopping Bim; (3) miR-221 has moderate predictive power for PD; (4) miR-221 directly targets PTEN, and PTEN over-expression eliminates the protective action of miR-221 on p-AKT expression in PC12 cells; and (5) miRNA-221 controls cell viability and apoptosis by manipulating the Akt signalling pathway in PD. This review study suggested that miR-221 has the potential to be used as a clinical biomarker for PD diagnosis and stage assignment.

Measurement Properties and Optimal Cutoff Point of the WHO-5 Among Chinese Healthcare Students

Purpose

The World Health Organization-Five Well-Being Index (WHO-5) is widely used to assess subjective well-being. Nevertheless, measurement invariance and optimal cutoff point of the WHO-5 have not been examined in Chinese samples. We aimed to assess measurement properties of the Chinese version of the WHO-5 (WHO-5-C) among healthcare students.

Patients and Methods

A two-wave longitudinal assessment was conducted among 343 Chinese healthcare students from September to November 2022. Measurement properties of the WHO-5-C were assessed through structural validity using confirmatory factor analysis (CFA), measurement invariance using multigroup CFA (MGCFA) and longitudinal CFA (LCFA), convergent validity using correlation analysis with the Self-Rated Health Questionnaire (SRHQ) and Patient Health Questionnaire-4 (PHQ-4), reliability using internal consistency and test-retest reliability, and optimal cutoff point using receiver operating characteristic (ROC) analysis.

Results

The WHO-5-C demonstrated satisfactory structural validity with comparative fit index (CFI) of 0.968 at baseline and 0.980 at follow-up, and adequate measurement invariance in different sociodemographic variables at baseline (gender, age, major, home location, being only child, monthly household income, part-time job, physical exercise, hobby, frequency of visiting home, and stress coping strategy) (CFI changes [ΔCFI] = -0.009-0.003) and over a week (ΔCFI = -0.006-0.000). The WHO-5-C also had good internal consistency (Cronbach's α = 0.907-0.934; McDonald's ω = 0.908-0.935) and test-retest reliability (intraclass correlation coefficient [ICC] = 0.803). Convergent validity was supported by moderate correlations of the WHO-5-C with the SRHQ and PHQ-4. The optimal cutoff point of the WHO-5-C was found to be 50, with an area under the ROC curve of 0.882 at baseline data, with sensitivity of 0.803 and specificity of 0.762 at follow-up.

Conclusion

The WHO-5-C demonstrated adequate measurement properties, especially concerning cross-sectional and longitudinal measurement invariance, with a recommended optimal cutoff point of ≥ 50 for assessing adequate level of psychological well-being in healthcare students.

A meta-analysis of the value of MRI-based VBQ scores for evaluating osteoporosis

Objective

Osteoporosis is the most common skeletal disease in humans. Early onset of osteoporosis is usually asymptomatic, so early diagnosis is critical. The purpose of this study was to analyze the value of MRI-based VBQ scores for evaluating osteoporosis.

Methods

We searched PubMed, Embase, the Cochrane Library databases, Web of Science, and some Chinese electronic databases for published articles and the ClinicalTrials.gov site for completed but unpublished studies on evaluating the value of MRI-based VBQ scores for evaluating osteoporosis. We calculated the summarized sensitivity, specificity, the ROC curve (AUC) values and their 95% confidence intervals (CIs) using MetaDiSc 1.4 software and STATA.

Results

Our study included 8 studies involving 999 patients of which 660 patients were diagnosed with osteopenia/osteoporosis, and 339 patients were identified as having normal BMD. The pooled sensitivity was 0.809 (95% CI, 0.777-0.838, I 2  = 78.8%), the pooled specificity was 0.640 (95% CI, 0.587-0.691, I 2  = 85.9%), and the pooled AUC was 0.8375.

Conclusion

MRI-based VBQ scores provided high sensitivity and moderate specificity in detecting osteoporosis. Opportunistic use of VBQ scores could be considered, e.g. before lumbar spine surgery.

Prospero registration number

CRD42022377024.

Noninvasive Staging of Lymph Node Status in Breast Cancer Using Machine Learning: External Validation and Further Model Development

BACKGROUND

Most patients diagnosed with breast cancer present with a node-negative disease. Sentinel lymph node biopsy (SLNB) is routinely used for axillary staging, leaving patients with healthy axillary lymph nodes without therapeutic effects but at risk of morbidities from the intervention. Numerous studies have developed nodal status prediction models for noninvasive axillary staging using postoperative data or imaging features that are not part of the diagnostic workup. Lymphovascular invasion (LVI) is a top-ranked predictor of nodal metastasis; however, its preoperative assessment is challenging.

OBJECTIVE

This paper aimed to externally validate a multilayer perceptron (MLP) model for noninvasive lymph node staging (NILS) in a large population-based cohort (n=18,633) and develop a new MLP in the same cohort. Data were extracted from the Swedish National Quality Register for Breast Cancer (NKBC, 2014-2017), comprising only routinely and preoperatively available documented clinicopathological variables. A secondary aim was to develop and validate an LVI MLP for imputation of missing LVI status to increase the preoperative feasibility of the original NILS model.

METHODS

Three nonoverlapping cohorts were used for model development and validation. A total of 4 MLPs for nodal status and 1 LVI MLP were developed using 11 to 12 routinely available predictors. Three nodal status models were used to account for the different availabilities of LVI status in the cohorts and external validation in NKBC. The fourth nodal status model was developed for 80% (14,906/18,663) of NKBC cases and validated in the remaining 20% (3727/18,663). Three alternatives for imputation of LVI status were compared. The discriminatory capacity was evaluated using the validation area under the receiver operating characteristics curve (AUC) in 3 of the nodal status models. The clinical feasibility of the models was evaluated using calibration and decision curve analyses.

RESULTS

External validation of the original NILS model was performed in NKBC (AUC 0.699, 95% CI 0.690-0.708) with good calibration and the potential of sparing 16% of patients with node-negative disease from SLNB. The LVI model was externally validated (AUC 0.747, 95% CI 0.694-0.799) with good calibration but did not improve the discriminatory performance of the nodal status models. A new nodal status model was developed in NKBC without information on LVI (AUC 0.709, 95% CI: 0.688-0.729), with excellent calibration in the holdout internal validation cohort, resulting in the potential omission of 24% of patients from unnecessary SLNBs.

CONCLUSIONS

The NILS model was externally validated in NKBC, where the imputation of LVI status did not improve the model's discriminatory performance. A new nodal status model demonstrated the feasibility of using register data comprising only the variables available in the preoperative setting for NILS using machine learning. Future steps include ongoing preoperative validation of the NILS model and extending the model with, for example, mammography images.

Diagnostic performance of contrast-enhanced ultrasound in diagnosing hepatic artery occlusion after liver transplantation: A systematic review and meta-analysis

INTRODUCTION

Hepatic artery occlusion (HAO) is a significant complication post-liver transplantation. Doppler ultrasound (DUS) has been widely used as an initial screening test for detecting HAO; however, its performance is often not sufficient. Although other diagnostic tests such as computed tomography angiography (CTA), magnetic resonance angiography (MRA), and angiogram are more accurate, they are invasive and have several limitations. Contrast-enhanced ultrasound (CEUS) is an emerging tool for detecting HAO; however, the results from previous studies were limited due to a small number of patients. Therefore, we aimed to evaluate its performance by performing a meta-analysis.

METHOD

We performed a systemic review and meta-analysis of studies evaluating the performance of CEUS for the detection of HAO in an adult population. A literature search of EMBASE, Scopus, CINAHL, and Medline was conducted through March 2022. Pooled sensitivity, specificity, log diagnostic odd ratio (LDOR), and area under summary receiver operating curve (AUC) were calculated. Publication bias was assessed by Deeks' funnel plot.

RESULT

Eight studies were included, with 434 CEUS performed. Using a combination of CTA, MRA, angiography, clinical follow-up, and surgery as the gold standard, the sensitivity, specificity, and LDOR of CEUS for detection of HAO were .969 (.938, .996), .991 (.981, 1.001), and 5.732 (4.539, 6.926), respectively. AUC was .959. The heterogeneity between studies appeared universally low, and no significant publication bias was found (p = .44).

CONCLUSION

CEUS appeared to have an excellent performance for the detection of HAO and could be considered as an alternative when DUS is non-diagnostic or when CTA, MRA, and angiogram are not feasible.

Strategies for Improving the Performance of Prediction Models for Response to Immune Checkpoint Blockade Therapy in Cancer

Immune checkpoint blockade (ICB) therapy holds promise for bringing long-lasting clinical gains for the treatment of cancer. However, studies show that only a fraction of patients respond to the treatment. In this regard, it is valuable to develop gene expression signatures based on RNA sequencing (RNAseq) data and machine learning methods to predict patients' response to the ICB therapy, which contributes to more personalized treatment strategy and better management of cancer patients. However, due to the limited sample size of ICB trials with RNAseq data available and the vast number of candidate gene expression features, it is challenging to develop well-performed gene expression signatures. In this study, we used several published melanoma datasets and investigated approaches that can improve the construction of gene expression-based prediction models. We found that merging datasets from multiple studies and incorporating prior biological knowledge yielded prediction models with higher predictive accuracies. Our finding suggests that these two strategies are of high value to identify ICB response biomarkers in future studies.

Current and future potential distribution of the invasive scale Ceroplastes rusci (L., 1758) (Hemiptera: Coccidae) under climate niche

BACKGROUND

The fig wax scale, Ceroplastes rusci is an invasive pest that feeds on more than 94 genera from 52 families that is spread across 60 countries, causing negative impacts to agriculture and forestry. Understanding the potential distribution of invasive species under climate change is crucial for the management and monitoring purposes. Thus, we predicted the potential distribution areas of C. rusci using Maximum Entropy (MaxEnt) based on the occurrence data and environmental variables under current and future climatic scenarios.

RESULTS

Our results showed that the temperature annual range (Bio 7) and mean temperature of the warmest quarter (Bio 10) attributed to a higher contribution to the current model of the distribution of C. rusci. The potential distribution maps illustrated the main concentrated areas of C. rusci which included South America, Africa, Asia, and Oceania. In addition, potential range expansions or reductions were predicted under different future climate change scenarios, which showed that the total suitable areas of the fig wax scale presented an increasing trend until 2100.

CONCLUSION

Our study provides significant data to understand the potential distribution of C. rusci around the world. It also serves as an early warning for the highly suitable habitat areas that even offers a platform to the currently non-infested regions or countries who are yet to develop monitoring strategies in response to the possible C. rusci outbreak. © 2022 Society of Chemical Industry.

Melatonin modulates rectal and body surface temperatures and their circadian rhythmicity in donkeys (Equus asinus) subjected to packing during the hot-dry season

The study aimed to evaluate the effects of melatonin administration on rectal and body surface temperature (RT and BST, respectively) responses and their circadian rhythmicity in donkeys subjected to packing (load carrying) during the hot-dry season. Twenty donkeys were divided into two equal groups randomly: Groups 1 (packing + melatonin) and 2 (packing - melatonin), subjected to packing and both covered 20 km. The RT, BST, and thermal environmental parameters were measured before and after packing. The procedure was carried out three times within the week, one day apart. This was followed 16-h after the last (third) packing procedure by 27-h recording period of all the parameters at 3-h intervals. The RT (37.77 ± 0.1℃) recorded in packing + melatonin donkeys was lower (P < 0.05) than in (packing - melatonin) (38.29 ± 0.1℃) post-packing, while the BSTs in packing + melatonin donkeys were lower than in (packing - melatonin) donkeys, especially the neck (33.07 ± 0.6℃ vs 35.4 ± 0.7℃, respectively) and coronary band (30.58 ± 0.5℃ vs 33.38 ± 0.7℃, respectively) temperatures. In both groups 16-h post-packing, overall mean RT and BST values were not different, except the coronary band temperature (26.61 ± 0.8℃) that was lower (P < 0.05) in packing + melatonin donkeys than (packing - melatonin) donkeys (28.78 ± 1.4℃). Melatonin-exerted biphasic effects on circadian rhythms of RT and BSTs by reducing body temperatures during the photophase and increasing the values during the scotophase in pack donkeys. In conclusion, melatonin may enhance packing output in donkeys during the hot-dry season.

Prognostic nutritional index: A potential biomarker for predicting the prognosis of decompensated liver cirrhosis

Background

Prognostic nutritional index (PNI) is an independent predictor of the prognosis of various diseases. However, the prognosis value of PNI in patients with decompensated liver cirrhosis (DLC) remains unknown. The study aimed to investigate the prognostic significance of PNI in patients with DLC.

Methods

A total of 214 eligible patients were enrolled in the study's development cohort between January 2018 and March 2021. The clinical primary study endpoints were mortality at 3 and 6 months. Receiver operating characteristic (ROC) curve analysis was used to assess the PNI's prediction accuracy, and Youden's index was utilized to determine the PNI's optimal cut-off value. Moreover, based on the optimal cut-off value, patients were categorized into high and low PNI groups. Multivariate logistic regression analysis was used to determine independent risk factors for mortality, while the relationship between PNI and the risk of death was identified and demonstrated using restricted cubic splines (RCS). A validation cohort of 139 patients was to verify the predictive power of the PNI.

Results

In the development cohort, the mortality rate at 3 and 6 months were 10.3% (22) and 14.0% (30), respectively. The PNI had comparable predictive power with the MELD score at all follow-up endpoints. Decreased PNI was an independent predictor of adverse prognosis at all follow-up endpoints. The RCS revealed a linear correlation between PNI and the risk of death. We confirmed that lower PNI was an independent predictor of poor prognosis in the validation cohort.

Conclusion

The findings showed that lower PNI is an independent factor of poor outcomes and might be utilized as a potentially promising prognostic predictor in patients with DLC.

Defining the extent of gene function using ROC curvature

MOTIVATION

Interactions between proteins help us understand how genes are functionally related and how they contribute to phenotypes. Experiments provide imperfect 'ground truth' information about a small subset of potential interactions in a specific biological context, which can then be extended to the whole genome across different contexts, such as conditions, tissues or species, through machine learning methods. However, evaluating the performance of these methods remains a critical challenge. Here, we propose to evaluate the generalizability of gene characterizations through the shape of performance curves.

RESULTS

We identify Functional Equivalence Classes (FECs), subsets of annotated and unannotated genes that jointly drive performance, by assessing the presence of straight lines in ROC curves built from gene-centric prediction tasks, such as function or interaction predictions. FECs are widespread across data types and methods, they can be used to evaluate the extent and context-specificity of functional annotations in a data-driven manner. For example, FECs suggest that B cell markers can be decomposed into shared primary markers (10-50 genes), and tissue-specific secondary markers (100-500 genes). In addition, FECs suggest the existence of functional modules that span a wide range of the genome, with marker sets spanning at most 5% of the genome and data-driven extensions of Gene Ontology sets spanning up to 40% of the genome. Simple to assess visually and statistically, the identification of FECs in performance curves paves the way for novel functional characterization and increased robustness in the definition of functional gene sets.

AVAILABILITY AND IMPLEMENTATION

Code for analyses and figures is available at https://github.com/yexilein/pyroc.

SUPPLEMENTARY INFORMATION

Supplementary data are available at Bioinformatics online.

A potential novel inflammation biomarker for predicting the prognosis of decompensated liver cirrhosis

OBJECTIVE

This study aimed to explore the prognostic value of the lymphocyte (LYM)-to-white blood cell (WBC) ratio (LWR) in patients with decompensated liver cirrhosis (DLC).

METHODS

This study was conducted by recruiting 214 patients with DLC with different aetiologies (development cohort). Receiver operating characteristic (ROC) curve analyses were used to assess the predictive accuracy of the LWR, and Youden's index was used to determine the optimal cut-off values of the LWR based on the ROC curve. Next, patients were divided into high- and low-LWR groups according to the cut-off values. Multivariate logistic analyses were performed to determine the independent predictors for the 1-, 3- and 6-month mortality. Restricted cubic spline (RCS) was used to determine and visualize the association between LWR and the risk of death. We verified the predictive ability of LWR in the validation cohort of 139 patients.

RESULTS

In the development cohort, there were 16 (7.5%), 22 (10.3%) and 30 patients (14.0%) who died at 1, 3 and 6 months, respectively. The LWR was significantly lower in non-survivors than in survivors and was an independent predictor of poor outcomes. The ROC analyses with the Delong test showed that the LWR had comparable predictive power with the Model for End-Stage Liver Disease (MELD) score, neutrophil-to-LYM ratio (NLR) and Chronic Liver Failure consortium score for acute decompensated (CLIF-C ADs). RCS showed a non-linear relationship between the LWR and the risk of death at 1 and 3 months, whereas a linear relationship was observed between the LWR and the risk of death at 6 months. We verified that the decreased LWR was an independent predictor of adverse outcomes at 3-, and 6-month follow-up endpoints in the validation cohort.

CONCLUSIONS

Our findings indicate that a lower LWR is an independent factor for unfavourable outcomes and may serve as a potential novel prognostic predictor in patients with DLC.KEY MESSAGESThis study is the first report on the prognostic value of the lymphocyte (LYM)-to-white blood cell (WBC) ratio (LWR) in patients with decompensated liver cirrhosis (DLC).Decreased LWR is an independent factor for adverse outcomes in patients with DLC.

Discovery and Analytical Validation of a Vocal Biomarker to Monitor Anosmia and Ageusia in Patients With COVID-19: Cross-sectional Study

Background

The COVID-19 disease has multiple symptoms, with anosmia and ageusia being the most prevalent, varying from 75% to 95% and from 50% to 80% of infected patients, respectively. An automatic assessment tool for these symptoms will help monitor the disease in a fast and noninvasive manner.

Objective

We hypothesized that people with COVID-19 experiencing anosmia and ageusia had different voice features than those without such symptoms. Our objective was to develop an artificial intelligence pipeline to identify and internally validate a vocal biomarker of these symptoms for remotely monitoring them.

Methods

This study used population-based data. Participants were assessed daily through a web-based questionnaire and asked to register 2 different types of voice recordings. They were adults (aged >18 years) who were confirmed by a polymerase chain reaction test to be positive for COVID-19 in Luxembourg and met the inclusion criteria. Statistical methods such as recursive feature elimination for dimensionality reduction, multiple statistical learning methods, and hypothesis tests were used throughout this study. The TRIPOD (Transparent Reporting of a multivariable prediction model for Individual Prognosis Or Diagnosis) Prediction Model Development checklist was used to structure the research.

Results

This study included 259 participants. Younger (aged <35 years) and female participants showed higher rates of ageusia and anosmia. Participants were aged 41 (SD 13) years on average, and the data set was balanced for sex (female: 134/259, 51.7%; male: 125/259, 48.3%). The analyzed symptom was present in 94 (36.3%) out of 259 participants and in 450 (27.5%) out of 1636 audio recordings. In all, 2 machine learning models were built, one for Android and one for iOS devices, and both had high accuracy—88% for Android and 85% for iOS. The final biomarker was then calculated using these models and internally validated.

Conclusions

This study demonstrates that people with COVID-19 who have anosmia and ageusia have different voice features from those without these symptoms. Upon further validation, these vocal biomarkers could be nested in digital devices to improve symptom assessment in clinical practice and enhance the telemonitoring of COVID-19–related symptoms.

Trial Registration

Clinicaltrials.gov NCT04380987; https://clinicaltrials.gov/ct2/show/NCT04380987

Protocol for a systematic review and meta-analysis of the diagnostic accuracy of artificial intelligence for grading of ophthalmology imaging modalities

BACKGROUND

With the rise of artificial intelligence (AI) in ophthalmology, the need to define its diagnostic accuracy is increasingly important. The review aims to elucidate the diagnostic accuracy of AI algorithms in screening for all ophthalmic conditions in patient care settings that involve digital imaging modalities, using the reference standard of human graders.

METHODS

This is a systematic review and meta-analysis. A literature search will be conducted on Ovid MEDLINE, Ovid EMBASE, and Wiley Cochrane CENTRAL from January 1, 2000, to December 20, 2021. Studies will be selected via screening the titles and abstracts, followed by full-text screening. Articles that compare the results of AI-graded ophthalmic images with results from human graders as a reference standard will be included; articles that do not will be excluded. The systematic review software DistillerSR will be used to automate part of the screening process as an adjunct to human reviewers. After the full-text screening, data will be extracted from each study via the categories of study characteristics, patient information, AI methods, intervention, and outcomes. Risk of bias will be scored using Quality Assessment of Diagnostic Accuracy Studies (QUADAS-2) by two trained independent reviewers. Disagreements at any step will be addressed by a third adjudicator. The study results will include summary receiver operating characteristic (sROC) curve plots as well as pooled sensitivity and specificity of artificial intelligence for detection of any ophthalmic conditions based on imaging modalities compared to the reference standard. Statistics will be calculated in the R statistical software.

DISCUSSION

This study will provide novel insights into the diagnostic accuracy of AI in new domains of ophthalmology that have not been previously studied. The protocol also outlines the use of an AI-based software to assist in article screening, which may serve as a reference for improving the efficiency and accuracy of future large systematic reviews.

TRIAL REGISTRATION

PROSPERO, CRD42021274441.

A comprehensive method for improvement of water quality index (WQI) models for coastal water quality assessment

Here, we present an improved water quality index (WQI) model for assessment of coastal water quality using Cork Harbour, Ireland, as the case study. The model involves the usual four WQI components - selection of water quality indicators for inclusion, sub-indexing of indicator values, sub-index weighting and sub-index aggregation - with improvements to make the approach more objective and data-driven and less susceptible to eclipsing and ambiguity errors. The model uses the machine learning algorithm, XGBoost, to rank and select water quality indicators for inclusion based on relative importance to overall water quality status. Of the ten indicators for which data were available, transparency, dissolved inorganic nitrogen, ammoniacal nitrogen, BOD₅, chlorophyll, temperature and orthophosphate were selected for summer, while total organic nitrogen, dissolved inorganic nitrogen, pH, transparency and dissolved oxygen were selected for winter. Linear interpolation functions developed using national recommended guideline values for coastal water quality are used for sub-indexing of water quality indicators and the XGBoost rankings are used in combination with the rank order centroid weighting method to determine sub-index weight values. Eight sub-index aggregation functions were tested - five from existing WQI models and three proposed by the authors. The computed indices were compared with those obtained using a multiple linear regression (MLR) approach and R² and RMSE used as indicators of aggregation function performance. The weighted quadratic mean function (R² = 0.91, RMSE = 4.4 for summer; R² = 0.97, RMSE = 3.1 for winter) and the unweighted arithmetic mean function (R² = 0.92, RMSE = 3.2 for summer; R² = 0.97, RMSE = 3.2 for winter) proposed by the authors were identified as the best functions and showed reduced eclipsing and ambiguity problems compared to the others.

Evaluating the Prognostic Implication of the Collins Histology Scoring System in a Pediatric Eastern Ontario Population With Eosinophilic Esophagitis

INTRODUCTION

Collins et al developed a histology scoring system (EoE HSS) to assess multiple pathologic features. The aim of this study is to identify if the EoE HSS can better detect endoscopic and symptom improvement vs the Peak Eosinophilic Count (PEC).

METHODS

A retrospective chart review was performed for patients during 2014-2016. All patients ≤18 years old with a diagnosis of EoE and whose records included initial and follow-up upper gastrointestinal endoscopies were included. Severity and extent of endoscopic features were scored using 8 parameters, from normal to maximum change for each location of the esophageal biopsy.

RESULTS

Forty patients with EoE were included in the study, of which 35 (87.5%) patients demonstrated symptom and 25 (62.5%) endoscopic improvement at the time of follow-up. In the proximal esophagus, the EoE HSS outperformed the change in eosinophil count of the Children's Hospital of Eastern Ontario (CHEO) practice in predicting endoscopic improvement by 16.8% when examining the change in grade and 17.1% when examining the change in stage scores.

CONCLUSIONS

At our institution, adoption of the EoE HSS in assessing biopsies of EoE patients might be warranted, compared to the traditional practice. However, a bigger sample size may give a more robust difference in all locations.

Towards Conserving Crop Wild Relatives along the Texas–Mexico Border: The Case of Manihot walkerae

Walker’s Manihot, Manihot walkerae, is an endangered species endemic to south Texas and northeastern Mexico and is a Crop Wild Relative (CWR) of the international and economically important crop cassava (M. esculenta). Manihot walkerae is globally endangered (IUCN’s Redlist, Texas list, USA); however, it is not recognized on the Mexican list of endangered species (NOM-059-SEMARNAT). We assessed the status of M. walkerae in Mexico and re-evaluated its global status. According to our analysis, M. walkerae should be considered an endangered species based on the IUCN’s assessment method and a threatened species in Mexico based on the Mexican criteria. Our findings encourage the establishment of sound conservation plans for M. walkerae along the Texas–Mexico border.

A new measure to quantify sedentary behavior using accelerometer data: Application to the Hispanic Community Health Study/Study of Latinos

Availability of accelerometer data has made it possible to objectively and continuously monitor sedentary behavior. Various summaries of the extensive accelerometer data have been used to understand the relationship between sedentary behavior and health. However, the widely used summary measures on sedentary bouts, average bout length or its derivatives, fail to reveal patterns of accumulated sedentary behavior over time. Studies have suggested that prolonged uninterrupted sedentary behavior can be an important metric that is related to health states. Yet existing measures to capture the prolonged sedentary patterns either rely on parametric assumptions on the underlying distribution of sedentary bout length or have to categorize sedentary bout length into somewhat arbitrary categories. Gini index was also used; however, it only measures the variability in bout lengths but not the actual length. To overcome these limitations, we proposed a non-parametric weighted survival function to characterize uninterrupted sedentary behavior over time in a continuous fashion and used the area under the survival curve as a new summary measure to quantify sedentary behavior. We showed that this measure is a weighted average of bout length and contains the information on both the mean and variability of bout lengths. We demonstrated in the simulation studies that the proposed measure could better identify prolonged uninterrupted sedentary behavior and predict health outcomes. We applied this new measure and existing sedentary measures to data from the Hispanic Community Health Study/Study of Latinos to examine the association between sedentary behavior and overweight/obesity.

Reporting on the Value of Artificial Intelligence in Predicting the Optimal Embryo for Transfer: A Systematic Review including Data Synthesis

Artificial intelligence (AI) has been gaining support in the field of in vitro fertilization (IVF). Despite the promising existing data, AI cannot yet claim gold-standard status, which serves as the rationale for this study. This systematic review and data synthesis aims to evaluate and report on the predictive capabilities of AI-based prediction models regarding IVF outcome. The study has been registered in PROSPERO (CRD42021242097). Following a systematic search of the literature in Pubmed/Medline, Embase, and Cochrane Central Library, 18 studies were identified as eligible for inclusion. Regarding live-birth, the Area Under the Curve (AUC) of the Summary Receiver Operating Characteristics (SROC) was 0.905, while the partial AUC (pAUC) was 0.755. The Observed: Expected ratio was 1.12 (95%CI: 0.26–2.37; 95%PI: 0.02–6.54). Regarding clinical pregnancy with fetal heartbeat, the AUC of the SROC was 0.722, while the pAUC was 0.774. The O:E ratio was 0.77 (95%CI: 0.54–1.05; 95%PI: 0.21–1.62). According to this data synthesis, the majority of the AI-based prediction models are successful in accurately predicting the IVF outcome regarding live birth, clinical pregnancy, clinical pregnancy with fetal heartbeat, and ploidy status. This review attempted to compare between AI and human prediction capabilities, and although studies do not allow for a meta-analysis, this systematic review indicates that the AI-based prediction models perform rather similarly to the embryologists’ evaluations. While AI models appear marginally more effective, they still have some way to go before they can claim to significantly surpass the clinical embryologists’ predictive competence.

Gene-environment interaction identification via penalized robust divergence

In high-throughput cancer studies, gene-environment interactions associated with outcomes have important implications. Some commonly adopted identification methods do not respect the "main effect, interaction" hierarchical structure. In addition, they can be challenged by data contamination and/or long-tailed distributions, which are not uncommon. In this article, robust methods based on γ $\gamma$ -divergence and density power divergence are proposed to accommodate contaminated data/long-tailed distributions. A hierarchical sparse group penalty is adopted for regularized estimation and selection and can identify important gene-environment interactions and respect the "main effect, interaction" hierarchical structure. The proposed methods are implemented using an effective group coordinate descent algorithm. Simulation shows that when contamination occurs, the proposed methods can significantly outperform the existing alternatives with more accurate identification. The proposed approach is applied to the analysis of The Cancer Genome Atlas (TCGA) triple-negative breast cancer data and Gene Environment Association Studies (GENEVA) Type 2 Diabetes data.

Performance of bone tracer for diagnosis and differentiation of transthyretin cardiac amyloidosis: a systematic review and meta-analysis

PURPOSE

Bone tracers have been validated for many years in detecting transthyretin cardiac amyloidosis (TTR-CA). However, several new studies suggest conflicting results. Our study aimed to systematically evaluate the accuracy of bone radiotracers for diagnosis and differentiation of TTR-CA via a systematic review and meta-analysis.

METHODS

We retrieved articles assessing the performance of bone tracer in diagnosing and differentiating TTR-CA from PubMed, the Cochrane Library, ScienceDirect, and DOAJ databases, dating up to 10 July 2020. The meta-analysis was conducted through Stata 16 software, and the risk of bias for the included studies was assessed by the QUADAS-2 tool. Moreover, we made a comprehensive review.

RESULTS

Fourteen articles were included in the systematic review, and 9 in the meta-analysis. The pooled sensitivity was 0.97 (95% confidence interval [95% CI] 0.85-0.99) with heterogeneity (I2=73.5, 95% CI 55.6-91.2), and the specificity was 0.92 (95% CI 0.82-0.96) with heterogeneity (I2=42.0, 95% CI 0.0-86.9). The pooled positive and negative likelihood ratios were 11.49 (95% CI 5.07-26.0) and 0.03 (95% CI 0.01-0.18), respectively. The diagnostic odds ratio was 341 (95% CI 53-2194), and the area under the receiver operating characteristic curve was 0.96 (95% CI 0.94-0.97).

CONCLUSION

The findings evidence that the bone radiotracer is a valuable noninvasive approach that provides high accuracy for diagnosing TTR-CA and plays a modest role in differentiating TTR-CA from immunoglobulin amyloid light-chain cardiac amyloidosis. 99mTc-HMDP may be more accurate than 99mTc-PYP, 99mTc-DPD, and 18F-NaF in the TTR-CA detecting process, and 18F-NaF is a promising bone tracer to diagnose and differentiate TTR-CA.

Radiomics, machine learning, and artificial intelligence-what the neuroradiologist needs to know

PURPOSE

Artificial intelligence (AI) is playing an ever-increasing role in Neuroradiology.

METHODS

When designing AI-based research in neuroradiology and appreciating the literature, it is important to understand the fundamental principles of AI. Training, validation, and test datasets must be defined and set apart as priorities. External validation and testing datasets are preferable, when feasible. The specific type of learning process (supervised vs. unsupervised) and the machine learning model also require definition. Deep learning (DL) is an AI-based approach that is modelled on the structure of neurons of the brain; convolutional neural networks (CNN) are a commonly used example in neuroradiology.

RESULTS

Radiomics is a frequently used approach in which a multitude of imaging features are extracted from a region of interest and subsequently reduced and selected to convey diagnostic or prognostic information. Deep radiomics uses CNNs to directly extract features and obviate the need for predefined features.

CONCLUSION

Common limitations and pitfalls in AI-based research in neuroradiology are limited sample sizes ("small-n-large-p problem"), selection bias, as well as overfitting and underfitting.

Deep learning with digital holographic microscopy discriminates apoptosis and necroptosis

Regulated cell death modalities such as apoptosis and necroptosis play an important role in regulating different cellular processes. Currently, regulated cell death is identified using the golden standard techniques such as fluorescence microscopy and flow cytometry. However, they require fluorescent labels, which are potentially phototoxic. Therefore, there is a need for the development of new label-free methods. In this work, we apply Digital Holographic Microscopy (DHM) coupled with a deep learning algorithm to distinguish between alive, apoptotic and necroptotic cells in murine cancer cells. This method is solely based on label-free quantitative phase images, where the phase delay of light by cells is quantified and is used to calculate their topography. We show that a combination of label-free DHM in a high-throughput set-up (~10,000 cells per condition) can discriminate between apoptosis, necroptosis and alive cells in the L929sAhFas cell line with a precision of over 85%. To the best of our knowledge, this is the first time deep learning in the form of convolutional neural networks is applied to distinguish-with a high accuracy-apoptosis and necroptosis and alive cancer cells from each other in a label-free manner. It is expected that the approach described here will have a profound impact on research in regulated cell death, biomedicine and the field of (cancer) cell biology in general.

Performance of echocardiography for detection of portopulmonary hypertension among liver transplant candidates: Meta-analysis

BACKGROUND

Evaluation of pulmonary arterial pressure is crucial among cirrhotic patients, considering that moderate portopulmonary hypertension (POPH) is a contraindication for liver transplantation. Although right heart catheterization (RHC) is the most accurate method to diagnose POPH, it is invasive.

OBJECTIVE

The aim of the study is to evaluate the performance of echocardiography in detecting POPH in liver transplant candidates.

METHODS

A Literature search was performed, and pooled sensitivity, specificity, positive likelihood ratio (LR), negative LR, and area under the summary receiver operating curve (AUC) were calculated. Subgroup analyses were performed based on different cutoff values for echocardiography and diagnostic criteria of RHC.

RESULTS

Sensitivity, specificity, positive LR, negative LR, and AUC of echocardiography for detection of POPH were 0.86 (0.74, 0.94), 0.87 (0.84, 0.90), 7.17 (3.59, 14.31), 0.22 (0.13, 0.38), and 0.807 while they were 0.82 (0.74, 0.89), 0.81 (0.78, 0.84), 117.75 (16.03, 865.08), 0.28 (0.16, 0.50), and 0.876for detection of moderate POPH, respectively. Performance of echocardiography was not significantly different in the subgroup analyses of stringency of POPH criteria and pulmonary arterial systolic pressure (ePASP) cutoffs.

CONCLUSIONS

Our meta-analysis supports utilization of echocardiography for screening of POPH. However, RHC remains essential in highly suspicious cases. Echocardiographic data other than ePASP should be evaluated in future studies.

A threshold-free summary index for quantifying the capacity of covariates to yield efficient treatment rules

When data on treatment assignment, outcomes, and covariates from a randomized trial are available, a question of interest is to what extent covariates can be used to optimize treatment decisions. Statistical hypothesis testing of covariate-by-treatment interaction is ill-suited for this purpose. The application of decision theory results in treatment rules that compare the expected benefit of treatment given the patient's covariates against a treatment threshold. However, determining treatment threshold is often context-specific, and any given threshold might seem arbitrary when the overall capacity towards predicting treatment benefit is of concern. We propose the Concentration of Benefit index (C_b ), a threshold-free metric that quantifies the combined performance of covariates towards finding individuals who will benefit the most from treatment. The construct of the proposed index is comparing expected treatment outcomes with and without knowledge of covariates when one of a two randomly selected patients is to be treated. We show that the resulting index can also be expressed in terms of the integrated efficiency of individualized treatment decision over the entire range of treatment thresholds. We propose parametric and semiparametric estimators, the latter being suitable for out-of-sample validation and correction for optimism. We used data from a clinical trial to demonstrate the calculations in a step-by-step fashion. The proposed index has intuitive and theoretically sound interpretation and can be estimated with relative ease for a wide class of regression models. Beyond the conceptual developments, various aspects of estimation and inference for such a metric need to be pursued in future research. R code that implements the method for a variety of regression models is provided at (https://github.com/msadatsafavi/txBenefit).

Memory complaints inventory profiles: Differentiating neurocognitive impairment, depression, and non-credible performance

The Memory Complaints Inventory (MCI) is a symptom validity measure designed to assess exaggerated memory complaints. The aim of current study was to develop memory complaint profiles on the MCI to distinguish between various neurocognitive disorders, depression, and non-credible performance. This study utilized MCI scores (N = 244) from a neuropsychology clinic to determine the presence of, and difference between, subjective memory complaints between a depression group, non-credible group, and subgroups of cognitive impairment (Alzheimer's Dementia, Vascular Dementia, and Mild Cognitive Impairment). Significant differences were found on MCI endorsement between cognitive impairment, depression, and non-credible groups. This pattern indicated fewer memory complaints for cognitive impairment groups when compared to depression and non-credible groups; the non-credible group had the highest MCI scores overall. ROC analyses revealed recommended clinical cutoff values with high specificity for distinguishing between the non-credible group and other groups. The findings provided further evidence for the MCI as a symptom validity measure, given its ability to differentiate between a non-credible group and clinical groups. Replication of the study's findings would result in reliable genuine subjective memory complaint profiles to provide additional diagnostic and prognostic specificity in neuropsychological practice.

Predicting instructed simulation and dissimulation when screening for depressive symptoms

The intentional distortion of test results presents a fundamental problem to self-report-based psychiatric assessment, such as screening for depressive symptoms. The first objective of the study was to clarify whether depressed patients like healthy controls possess both the cognitive ability and motivation to deliberately influence results of commonly used screening measures. The second objective was the construction of a method derived directly from within the test takers' responses to systematically detect faking behavior. Supervised machine learning algorithms posit the potential to empirically learn the implicit interconnections between responses, which shape detectable faking patterns. In a standardized design, faking bad and faking good were experimentally induced in a matched sample of 150 depressed and 150 healthy subjects. Participants completed commonly used questionnaires to detect depressive and associated symptoms. Group differences throughout experimental conditions were evaluated using linear mixed-models. Machine learning algorithms were trained on the test results and compared regarding their capacity to systematically predict distortions in response behavior in two scenarios: (1) differentiation of authentic patient responses from simulated responses of healthy participants; (2) differentiation of authentic patient responses from dissimulated patient responses. Statistically significant convergence of the test scores in both faking conditions suggests that both depressive patients and healthy controls have the cognitive ability as well as the motivational compliance to alter their test results. Evaluation of the algorithmic capability to detect faking behavior yielded ideal predictive accuracies of up to 89%. Implications of the findings, as well as future research objectives are discussed. Trial Registration The study was pre-registered at the German registry for clinical trials (Deutsches Register klinischer Studien, DRKS; DRKS00007708).

sDeepFM: Multi-Scale Stacking Feature Interactions for Click-Through Rate Prediction

For estimating the click-through rate of advertisements, there are some problems in that the features cannot be automatically constructed, or the features built are relatively simple, or the high-order combination features are difficult to learn under sparse data. To solve these problems, we propose a novel structure multi-scale stacking pooling (MSSP) to construct multi-scale features based on different receptive fields. The structure stacks multi-scale features bi-directionally from the angles of depth and width by constructing multiple observers with different angles and different fields of view, ensuring the diversity of extracted features. Furthermore, by learning the parameters through factorization, the structure can ensure high-order features being effectively learned in sparse data. We further combine the MSSP with the classical deep neural network (DNN) to form a unified model named sDeepFM. Experimental results on two real-world datasets show that the sDeepFM outperforms state-of-the-art models with respect to area under the curve (AUC) and log loss.

Deep learning workflow in radiology: a primer

Interest for deep learning in radiology has increased tremendously in the past decade due to the high achievable performance for various computer vision tasks such as detection, segmentation, classification, monitoring, and prediction. This article provides step-by-step practical guidance for conducting a project that involves deep learning in radiology, from defining specifications, to deployment and scaling. Specifically, the objectives of this article are to provide an overview of clinical use cases of deep learning, describe the composition of multi-disciplinary team, and summarize current approaches to patient, data, model, and hardware selection. Key ideas will be illustrated by examples from a prototypical project on imaging of colorectal liver metastasis. This article illustrates the workflow for liver lesion detection, segmentation, classification, monitoring, and prediction of tumor recurrence and patient survival. Challenges are discussed, including ethical considerations, cohorting, data collection, anonymization, and availability of expert annotations. The practical guidance may be adapted to any project that requires automated medical image analysis.

A Comparative Study of Kernel Logistic Regression, Radial Basis Function Classifier, Multinomial Naïve Bayes, and Logistic Model Tree for Flash Flood Susceptibility Mapping

Risk of flash floods is currently an important problem in many parts of Vietnam. In this study, we used four machine-learning methods, namely Kernel Logistic Regression (KLR), Radial Basis Function Classifier (RBFC), Multinomial Naïve Bayes (NBM), and Logistic Model Tree (LMT) to generate flash flood susceptibility maps at the minor part of Nghe An province of the Center region (Vietnam) where recurrent flood problems are being experienced. Performance of these four methods was evaluated to select the best method for flash flood susceptibility mapping. In the model studies, ten flash flood conditioning factors, namely soil, slope, curvature, river density, flow direction, distance from rivers, elevation, aspect, land use, and geology, were chosen based on topography and geo-environmental conditions of the site. For the validation of models, the area under Receiver Operating Characteristic (ROC), Area Under Curve (AUC), and various statistical indices were used. The results indicated that performance of all the models is good for generating flash flood susceptibility maps (AUC = 0.983–0.988). However, performance of LMT model is the best among the four methods (LMT: AUC = 0.988; KLR: AUC = 0.985; RBFC: AUC = 0.984; and NBM: AUC = 0.983). The present study would be useful for the construction of accurate flash flood susceptibility maps with the objectives of identifying flood-susceptible areas/zones for proper flash flood risk management.

A new concordant partial AUC and partial c statistic for imbalanced data in the evaluation of machine learning algorithms

BACKGROUND

In classification and diagnostic testing, the receiver-operator characteristic (ROC) plot and the area under the ROC curve (AUC) describe how an adjustable threshold causes changes in two types of error: false positives and false negatives. Only part of the ROC curve and AUC are informative however when they are used with imbalanced data. Hence, alternatives to the AUC have been proposed, such as the partial AUC and the area under the precision-recall curve. However, these alternatives cannot be as fully interpreted as the AUC, in part because they ignore some information about actual negatives.

METHODS

We derive and propose a new concordant partial AUC and a new partial c statistic for ROC data-as foundational measures and methods to help understand and explain parts of the ROC plot and AUC. Our partial measures are continuous and discrete versions of the same measure, are derived from the AUC and c statistic respectively, are validated as equal to each other, and validated as equal in summation to whole measures where expected. Our partial measures are tested for validity on a classic ROC example from Fawcett, a variation thereof, and two real-life benchmark data sets in breast cancer: the Wisconsin and Ljubljana data sets. Interpretation of an example is then provided.

RESULTS

Results show the expected equalities between our new partial measures and the existing whole measures. The example interpretation illustrates the need for our newly derived partial measures.

CONCLUSIONS

The concordant partial area under the ROC curve was proposed and unlike previous partial measure alternatives, it maintains the characteristics of the AUC. The first partial c statistic for ROC plots was also proposed as an unbiased interpretation for part of an ROC curve. The expected equalities among and between our newly derived partial measures and their existing full measure counterparts are confirmed. These measures may be used with any data set but this paper focuses on imbalanced data with low prevalence.

FUTURE WORK

Future work with our proposed measures may: demonstrate their value for imbalanced data with high prevalence, compare them to other measures not based on areas; and combine them with other ROC measures and techniques.

Identification of gene-environment interactions with marginal penalization

Gene-environment (G-E) interaction analysis has been extensively conducted for complex diseases. In marginal analysis, the common practice is to conduct likelihood-based (and other "standard") estimation with each marginal model, and then select significant G-E interactions and main effects based on p values and multiple comparisons adjustment. One limitation of this approach is that the identification results often do not respect the "main effects, interactions" hierarchy, which has been stressed in recent G-E interaction analyses. There is some recent effort tackling this problem, however, with very complex formulations. Another limitation of the common practice is that it may not perform well when regularization is needed, for example, because of "non-normal" distributions. In this article, we propose a marginal penalization approach which adopts a novel penalty to directly tackle the aforementioned problems. The proposed approach has a framework more coherent with that of the recently developed joint analysis methods and an intuitive formulation, and can be effectively realized. In simulation, it outperforms the popular significance-based analysis and simple penalization-based alternatives. Promising findings are made in the analysis of a single-nucleotide polymorphism and a gene expression data.

Predictive Factors and Clinical Prediction Score for Serious Intracranial Causes in Acute Nontraumatic Headache at an Emergency Department

Purpose

The objectives of this study were to investigate the predictive factors and develop a clinical prediction score to identify serious intracranial causes in acute nontraumatic headache (NTH).

Methods

A retrospective chart review study was conducted from 2013 to 2018 in acute NTH patients who visited the emergency department. The patients were divided into serious intracranial headache and nonserious intracranial headache groups. The two groups were compared in regard to the baseline characteristics, clinical presentation, physical examination, investigation, and diagnosis. The significant factors to predict a serious intracranial cause were examined using a multivariate logistic regression model. The coefficients from the multivariate logistic regression were used to plot the receiver operating characteristic curve to develop a clinical prediction score.

Results

From 2,372 patients, 454 met the inclusion criteria. Of the 454 patients with acute NTH, 88 (19.4%) patients were serious intracranial cause. The seven significant factors that predicted serious intracranial cause were abrupt onset (odds ratio (OR) 7.96, 95% confidence interval (CI) 2.77‒22.91), awakening pain (OR 3.14, 95% CI 4.15-6.82), duration of headache >1 week (OR 10.59, 95% CI 2.9-38.7), fever (OR 6.01, 95% CI 2.07-17.46), worst headache ever (OR 12.95, 95% CI 5.69-29.45), alteration of consciousness (OR 13.55, 95% CI 2.07‒88.88), and localizing neurological deficit (OR 5.28, 95% CI 1.6‒17.46). A score ≥3 out of 10 points of the clinical prediction score was likely to identify a serious intracranial cause of acute NTH with a sensitivity and specificity of 87.50% (95% CI 78.73-93.59%) and 87.70% (95% CI 83.90-90.89%), respectively. The area under the curve was 0.933.

Conclusion

Abrupt onset, awakening pain, duration of headache >1 week, fever, worst headache ever, alteration of consciousness, and localizing neurological deficit were the significant predictive factors for serious intracranial cause of acute NTH.

Red cell distribution width to platelet ratio for liver fibrosis: a systematic review and meta-analysis of diagnostic accuracy

Introduction: Red cell distribution width to platelet ratio (RPR) may be a useful marker for the evaluation of liver fibrosis in chronic liver disease (CLD). We sought to investigate its value in fibrosis-related outcomes in a meta-analysis of diagnostic accuracy. Areas covered: We searched MEDLINE (1966-2019), Clinicaltrials.gov (2008-2019), Cochrane Central Register of Controlled Trials (CENTRAL) (1999-2019), Google Scholar (2004-2019) and WHO (International Clinical Trials Register Platform) databases using a structured algorithm. The articles were assessed by Quality Assessment of Diagnostic Accuracy Studies tool (QUADAS-2). In over 1,800 patients for each outcome, pooled sensitivity and specificity for a) significant fibrosis, b) advanced fibrosis and c) cirrhosis were: a) 0.635 and 0.769 with an AUC of 0.747, b) 0.607 and 0.783 with an AUC of 0.773, c) 0.739 and 0.768 with an AUC of 0.818 respectively. Similar results were found for chronic hepatitis B in all outcomes. Subgroup analysis indicated a high specificity for advanced fibrosis detection in primary biliary cirrhosis. Sensitivity analysis did not alter the results. Expert opinion: RPR is a good predictor of fibrosis, especially as severity of chronic liver disease progresses. Future research should elucidate its value in specific etiologies of chronic liver disease.

Performance of ultrasound for detection of transjugular intrahepatic portosystemic shunt dysfunction: a meta-analysis

PURPOSE

Although ultrasound has been widely used to evaluate transjugular intrahepatic portosystemic shunts (TIPS) patency, several studies have reported conflicting data regarding its performance. Therefore, we aimed to evaluate performance of ultrasound for detection of TIPS dysfunction by performing a meta-analysis.

METHODS

Literature search was performed for studies evaluating ultrasound for TIPS dysfunction, stenosis, and occlusion using PubMed, EMBASE, Scopus, and Cochrane Library through February 2019. Pooled sensitivity, specificity, log diagnostic odds ratio (LDOR), and area under curve (AUC) of summary receiver-operating characteristic were calculated. Subgroup analyses were performed according to ultrasonographic criteria and type of stent.

RESULTS

In total, 21 studies were evaluated. Pooled sensitivity, specificity, and LDOR of ultrasound for detection of TIPS dysfunction were 0.82 (0.67, 0.93), 0.58 (0.46, 0.70), and 1.77 (1.20, 2.35). Pooled sensitivity, specificity, and LDOR for TIPS stenosis were 0.80 (0.69, 0.90), 0.80 (0.69, 0.91), and 2.83 (1.88, 3.78). Pooled sensitivity, specificity, and LDOR for TIPS occlusion were 0.96 (0.92, 0.99), 1 (0.99, 1.00), and 6.28 (4.96, 7.60). AUCs of ultrasound for TIPS dysfunction, stenosis, and occlusion were 0.77, 0.86, and 0.95, respectively.

CONCLUSIONS

Although ultrasound had excellent performance for TIPS occlusion and acceptable performance for TIP stenosis, most studies utilized bare metal stent, and therefore, application to current practice is limited. Ultrasound for TIPS dysfunction in the setting of covered metal stent appeared to have acceptable sensitivity of 0.82, but limited specificity of 0.58 and low LDOR of 1.77. A new noninvasive tool is needed for detection of TIPS dysfunction in the era of covered metal stent.

Flash flood susceptibility modeling using an optimized fuzzy rule based feature selection technique and tree based ensemble methods

The main objective of the present study was to provide a novel methodological approach for flash flood susceptibility modeling based on a feature selection method (FSM) and tree based ensemble methods. The FSM, used a fuzzy rule based algorithm FURIA, as attribute evaluator, whereas GA were used as the search method, in order to obtain optimal set of variables used in flood susceptibility modeling assessments. The novel FURIA-GA was combined with LogitBoost, Bagging and AdaBoost ensemble algorithms. The performance of the developed methodology was evaluated at the Bao Yen district and the Bac Ha district of Lao Cai Province in the Northeast region of Vietnam. For the case study, 654 floods and twelve geo-environmental variables were used. The predictive performance of each model was estimated through the calculation of the classification accuracy, the sensitivity, the specificity, the success and predictive rate curve and the area under the curves (AUC). The FURIA-GA FSM compared to a conventional rule based method gave more accurate predictive results. Also, the FURIA-GA based models, presented higher learning and predictive ability compared to the ensemble models that had not undergone a FSM. Based on the predictive classification accuracy, FURIA-GA-Bagging (93.37%) outperformed FURIA-GA-LogitBoost (92.35%) and FURIA-GA-AdaBoost (89.03%). FURIA-GA-Bagging showed also the highest sensitivity (96.94%) and specificity (89.80%). On the other hand, the FURIA-GA-LogitBoost showed the lowest percentage in very high susceptible zone and the highest relative flash-flood density, whereas the FURIA-GA-AdaBoost achieved the highest prediction AUC value (0.9740), based on the prediction rate curve, followed by FURIA-GA-Bagging (0.9566), and FURIA-GA-LogitBoost (0.8955). It can be concluded that the usage of different statistical metrics, provides different outcomes concerning the best prediction model, which mainly could be attributed to sites specific settings. The proposed models could be considered as a novel alternative investigation tools appropriate for flash flood susceptibility mapping.

Metazoan parasite infracommunities of the dusky flounder (Syacium papillosum) as bioindicators of environmental conditions in the continental shelf of the Yucatan Peninsula, Mexico

Background

We assessed metrics of the metazoan parasite infracommunities of the dusky flounder (Syacium papillosum) as indicators of aquatic environmental health of the Yucatan Shelf (YS) prior to oil extraction. We sampled the dusky flounder and its parasites along the YS, mostly during the 2015 north wind season (November–April). Our aims were: (i) to determine whether the parasite infracommunity metrics of S. papillosum exhibit significant differences among YS subregions; (ii) to determine whether the probability of the occurrence of its parasite species and individuals were affected by environmental variables, nutrients, heavy metals and hydrocarbons at the seascape level; and (iii) to determine whether there were statistical differences between the parasite infracommunity metrics of S. papillosum from YS and those of Syacium gunteri from the Campeche Sound. Multivariate statistical analyses and generalised additive models (GAMs) were used to examine the potential statistical associations between the contaminants, environmental variables and parasite community metrics, and the maximum entropy algorithm (MaxEnt) was used to characterise the habitat’s suitability for the parasite’s probability of occurrence.

Results

We recovered 48 metazoan parasite species from 127 S. papillosum, with larval cestodes and digeneans being the most numerically-dominant. Multivariate analyses showed significant differences in parasite infracommunity metrics among Western YS, Mid YS and Caribbean subregions, with the latter being the richest in species but not in individuals. The GAM and MaxEnt results indicated a negative effect of top predators (e.g. sharks and rays) removal on parasite metrics. The parasite infracommunities of S. papillosum were twice as rich in the number of species and individuals as those reported for S. gunteri from the Campeche Sound.

Conclusions

The significant differences among subregions in parasite metrics were apparently due to the interruption of the Yucatan current during the north wind season. The fishing of top predators in combination with an influx of nutrients and hydrocarbons in low concentrations coincides with an increase in larval cestodes and digeneans in S. papillosum. The dusky flounder inhabits a region (YS) with a larger number of metazoan parasite species compared with those available for S. gunteri in the Campeche Sound, suggesting better environmental conditions for transmission in the YS.

Electronic supplementary material

The online version of this article (10.1186/s13071-019-3524-6) contains supplementary material, which is available to authorized users.

Glaucoma Diagnosis with Machine Learning Based on Optical Coherence Tomography and Color Fundus Images

This study aimed to develop a machine learning-based algorithm for glaucoma diagnosis in patients with open-angle glaucoma, based on three-dimensional optical coherence tomography (OCT) data and color fundus images. In this study, 208 glaucomatous and 149 healthy eyes were enrolled, and color fundus images and volumetric OCT data from the optic disc and macular area of these eyes were captured with a spectral-domain OCT (3D OCT-2000, Topcon). Thickness and deviation maps were created with a segmentation algorithm. Transfer learning of convolutional neural network (CNN) was used with the following types of input images: (1) fundus image of optic disc in grayscale format, (2) disc retinal nerve fiber layer (RNFL) thickness map, (3) macular ganglion cell complex (GCC) thickness map, (4) disc RNFL deviation map, and (5) macular GCC deviation map. Data augmentation and dropout were performed to train the CNN. For combining the results from each CNN model, a random forest (RF) was trained to classify the disc fundus images of healthy and glaucomatous eyes using feature vector representation of each input image, removing the second fully connected layer. The area under receiver operating characteristic curve (AUC) of a 10-fold cross validation (CV) was used to evaluate the models. The 10-fold CV AUCs of the CNNs were 0.940 for color fundus images, 0.942 for RNFL thickness maps, 0.944 for macular GCC thickness maps, 0.949 for disc RNFL deviation maps, and 0.952 for macular GCC deviation maps. The RF combining the five separate CNN models improved the 10-fold CV AUC to 0.963. Therefore, the machine learning system described here can accurately differentiate between healthy and glaucomatous subjects based on their extracted images from OCT data and color fundus images. This system should help to improve the diagnostic accuracy in glaucoma.