Hierarchical Summary Receiver Operating Characteristic Research Articles

Accuracy of screening tests for gestational diabetes mellitus in Southeast Asia

To investigate the accuracy of screening tests for gestational diabetes mellitus (GDM) in Southeast Asian pregnant women. We searched PubMed (MEDLINE), Web of Science, Cochrane Library, ClinicalTrials.gov, Google Scholar, and Google for relevant articles published in English up to November 2018 using search terms related to GDM, screening tests for GDM and diagnostic performance. The studies were independently screened and selected by both authors. The methodological quality of the included studies was independently assessed by quality assessment of diagnostic accuracy studies 2. A hierarchical summary receiver operating characteristic (HSROC) model was created to estimate the HSROC curve. The summary sensitivity, specificity, positive likelihood ratio, negative likelihood ratio, and diagnostic odds ratio were calculated in a meta-analysis using bivariate random-effects model. A total of 19 studies were included in which the 100 g oral glucose tolerance test (OGTT) and 75 g OGTT were the two common reference standards for diagnosis of GDM. Most points of diagnostic performance in the HSROC 50 g GCT curve compared with the 100 g OGTT reference standard were clustered in the upper left-hand quadrant. The pooled sensitivity and specificity of the 50 g GCT were 79% (95% confidence interval [CI] 64%-89%) and 74% (95% CI 59%-85%), respectively. For the 75 g OGTT reference standard, the non-fasting 2-hour plasma glucose showed quite similar sensitivity the 50 g GCT compared with the 100 g OGTT reference standard. The pooled sensitivities and specificities of the fasting plasma glucose and hemoglobin A1c were 81% (95% CI 76%-86%) and 70% (95% CI 67%-72%), and 80% (95% CI 66%-90%) and 69% (95% CI 58%-78%), respectively. Our findings indicate that the 50 g GCT using the threshold of 140 mg/dL is a good screening test for identifying GDM at 24 to 28 weeks' gestational age for both high-risk and universal screening strategies in Southeast Asian countries. The non-fasting 2-hour PG, fasting plasma glucose or hemoglobin A1c are alternative choices for screening.

Clinical tools used in young infants born very preterm to predict motor and cognitive delay (not cerebral palsy): a systematic review.

This systematic review evaluates the accuracy of clinical tools used at a corrected age of 6months or younger to predict motor and cognitive delay (not cerebral palsy) at 24months' corrected age, in infants born very preterm. Six databases were searched. Quality was evaluated using the Quality Assessment of Diagnostic Accuracy Studies tool. Predictive analysis included calculation of sensitivity and specificity, inspection of summary receiver operating characteristics curves, and bivariate meta-analysis. Six assessments were identified in 10 studies of 992 infants. Overall prevalence of motor delay was 13.8% and cognitive delay was 11.7%. Methodological quality was variable for patient selection, reference standard, flow, and timing. All studies had a low risk of bias for the index test. General Movement Assessment (GMA) predicted motor and cognitive outcomes with good accuracy for mild, moderate, and severe delays (fidgety age: pooled diagnostic odds ratio=12.3 [5.9-29.8]; hierarchical summary receiver operating characteristics curve=0.733). The Hammersmith Infant Neurological Examination (HINE) demonstrated excellent predictive accuracy for severe motor delay (3mo and 6mo; sensitivity 93% [68-100%], specificity 100% [96-100%]) but showed limited ability to predict milder delays. In the population of infants born very preterm, few assessment tools used at 6 months or younger corrected age have proven predictive accuracy for cognitive and motor delay at 24 months' corrected age. Only the GMA and HINE demonstrated useful predictive validity. General movements have predictive validity for both motor and cognitive dysfunction in infants born very preterm. The Hammersmith Infant Neurological Examination showed the highest predictive accuracy for severe motor delay. The General Movement Assessment was the best tool to predict mild-to-moderate motor and cognitive delays.

Diagnostic Test Accuracy of the Red Reflex Test for Ocular Pathology in Infants

Red reflex testing is a simple and inexpensive method implemented in many countries as an important part of infant screening for ocular pathologies. To review the literature on the diagnostic accuracy of the red reflex test in infant screening for ocular pathologies and to perform meta-analyses to provide summary estimates. The following literature databases were searched for English-language, peer-reviewed literature, published until April 19, 2020: Cochrane Central, PubMed/MEDLINE, Embase, Web of Science Core Collection, BIOSIS Previews, Current Contents Connect, Data Citation Index, Derwent Innovations Index, KCI-Korean Journal Database, Russian Science Citation Index, SciELO Citation Index, CINAHL (Cumulative Index to Nursing and Allied Health Literature), and ClinicalTrials.gov. Eligibility criteria were defined according to population (studies of consecutively screened infants), exposure (red reflex or Brückner test as the index test), comparator (any ophthalmological examination), and study type (any study with diagnostic test accuracy data). The Preferred Reporting Items for Systematic Reviews and Meta-Analysis of Diagnostic Test Accuracy Studies (PRISMA-DTA) guidelines were followed. Data were extracted independently by 2 authors. For summary estimates of diagnostic test accuracy, the hierarchical summary receiver operating characteristics curve was used. Prevalence of ocular pathologies was introduced for a prevalence meta-analysis, which was then used in calculations of diagnostic accuracy of the red reflex test when applied in infant screening. True-positive, false-positive, true-negative, and false-negative findings; sensitivity; specificity; and positive and negative predictive values. In this meta-analysis, 8713 unique infants from 5 unique studies were eligible for qualitative and quantitative review. All studies used the red reflex test without pupillary dilation and were compared with a reference test performed with pupillary dilation. For any ocular pathology, an estimated sensitivity of 7.5% (95% CI, 7.4%-7.5%) and specificity of 97.5% (95% CI, 97.5%-97.5%) was found. Focusing on ocular pathologies that required a medical or surgical intervention, sensitivity improved to 17.5% (95% CI, 0.8%-84.8%) and specificity remained high at 97.6% (95% CI, 87.7%-99.6%). These findings suggest that an abnormal red reflex finding most likely reflects an underlying ocular pathology. However, a normal red reflex finding during screening does not exclude ocular disease.

Gonadotropin-Releasing Hormone Analogue Stimulation Test Versus Venous Sampling in Postmenopausal Hyperandrogenism.

Postmenopausal hyperandrogenism can be due to excessive androgen secretion from adrenal or ovarian virilizing tumors or nonneoplastic conditions. The etiology of postmenopausal hyperandrogenism can be difficult to discern because of limited accuracy of current diagnostic tests. This systematic review compares the diagnostic accuracy of the gonadotropin-releasing hormone (GnRH) analogue stimulation test against selective ovarian and adrenal vein sampling of androgens in distinguishing neoplastic from nonneoplastic causes of postmenopausal hyperandrogenism. Diagnostic test accuracy studies on these index tests in postmenopausal women were selected based on preestablished criteria. The true positive, false positive, false negative, and true negative values were extracted and meta-analysis was conducted using the hierarchical summary receiver operator characteristics curve method. The summary sensitivity of the GnRH analogue stimulation test is 10% (95% confidence interval [CI], 1.1%-46.7%) and that for selective venous sampling is 100% (95% CI, 0%-100%). Both tests have 100% specificity. There is limited evidence for the use of either test in identifying virilizing tumors in postmenopausal hyperandrogenism.

Diagnostic accuracy of non-contact infrared thermometers and thermal scanners: a systematic review and meta-analysis.

Infrared thermal screening, via the use of handheld non-contact infrared thermometers (NCITs) and thermal scanners, has been widely implemented all over the world. We performed a systematic review and meta-analysis to investigate its diagnostic accuracy for the detection of fever. We searched PubMed, Embase, the Cochrane Library, medRxiv, bioRxiv, ClinicalTrials.gov, COVID-19 Open Research Dataset, COVID-19 research database, Epistemonikos, EPPI-Centre, World Health Organization International Clinical Trials Registry Platform, Scopus and Web of Science databases for studies where a non-contact infrared device was used to detect fever against a reference standard of conventional thermometers. Forest plots and Hierarchical Summary Receiver Operating Characteristics curves were used to describe the pooled summary estimates of sensitivity, specificity and diagnostic odds ratio. From a total of 1063 results, 30 studies were included in the qualitative synthesis, of which 19 were included in the meta-analysis. The pooled sensitivity and specificity were 0.808 (95%CI 0.656-0.903) and 0.920 (95%CI 0.769-0.975), respectively, for the NCITs (using forehead as the site of measurement), and 0.818 (95%CI 0.758-0.866) and 0.923 (95%CI 0.823-0.969), respectively, for thermal scanners. The sensitivity of NCITs increased on use of rectal temperature as the reference. The sensitivity of thermal scanners decreased in a disease outbreak/pandemic setting. Changes approaching statistical significance were also observed on the exclusion of neonates from the analysis. Thermal screening had a low positive predictive value, especially at the initial stage of an outbreak, while the negative predictive value (NPV) continued to be high even at later stages. Thermal screening has reasonable diagnostic accuracy in the detection of fever, although it may vary with changes in subject characteristics, setting, index test, and the reference standard used. Thermal screening has a good NPV even during a pandemic. The policymakers must take into consideration the factors surrounding the screening strategy while forming ad-hoc guidelines.

Nutritional screening tools for adult cancer patients: A hierarchical Bayesian latent-class meta-analysis

Cancer treatment requires attentiveness to its broader effect on the body. Cancer's effect on appetite, strength, and body composition is contained in the summary term malnutrition. The tools used to detect malnutrition are a critical part of effective cancer care. In clinical care, selection of any specific tool is random. The relative validity of these tools have not been systematically compared. Using hierarchical Bayesian latent-class meta-analysis methods, this report compares three tools used for adult cancer patients - the Mini Nutritional Assessment (MNA), the Nutritional Risk Screening 2002 (NRS-2002) and the Patient Generated Subjective Global Assessment (PG-SGA). Drawing from English and Chinese language databases, a broad pool of eligible studies were identified for further selection and assessment. Using the hierarchical summary receiver operating characteristic (HSROC) model, pooled sensitivity, specificity, and other measurements the accuracy of the three tools were compared. A total of 37 eligible studies involving the MNA, NRS-2002 and PG-SGA were included in this meta-analysis. The pooled sensitivity was 0.910 (95% CI: 0.763 to 0.970) for MNA, 0.747 (95% CI: 0.680 to 0.804) for NRS-2002, and 0.964 (95% CI: 0.913 to 0.986) for PG-SGA. The pooled specificity was 0.720 (95% CI: 0.623 to 0.800) for MNA, 0.854 (95% CI: 0.808 to 0.891) for NRS-2002, 0.905 (95% CI: 0.807 to 0.956) for PG-SGA, respectively. The back-calculated likelihood ratio (LR) showed that MNA had a low negative likelihood ratio (LR-), NRS-2002 corresponded to a high positive likelihood ratio (LR+) and PG-SGA represented the best LR+ and LR-. While there is no standard approach to assessment of malnutrition, the PG-SGA has the best diagnostic performance with cancer patients. Further work is needed to refine the utility of these tools in larger clinical samples.

VP46.05: Risk factors, histopathology and diagnostic accuracy of posterior placenta accreta spectrum disorders: systematic review and meta‐analysis

To elucidate the risk factor, histopathological correlations and diagnostic accuracy of prenatal imaging in detecting posterior PAS. MEDLINE, Embase and CINAHL were searched. The outcomes explored were: risk factor for posterior PAS, histopathological correlation, and diagnostic accuracy of ultrasound and MRI in detecting these anomalies. Random-effect meta-analyses of proportions and summary estimates of sensitivity, specificity, positive and negative likelihood ratios (LR+ and LR–) and diagnostic odds ratio (DOR) using the hierarchical summary receiver–operating characteristics (HSROC) model were used to analyse the data. 20 studies were included. Placenta previa was present in 92.8% pregnancies complicated by posterior PAS, while 76.1% of women had a prior uterine surgery, mainly a CS or curettage. When considering the histopathological analysis of women affected by posterior PAS, 77.5% had placenta accreta, 19.5% placenta increta and 9.3 placenta percreta. 56.4% of posterior PAS disorders were detected prenatally on ultrasound, while 46.7% were diagnosed only at birth. When exploring the distribution of the classic ultrasound signs of PAS, placental lacunae were present in 39.0%, loss of the clear zone 41.15 and bladder wall interruption in 16.6% of women, while none of the included cases showed hypervascularisation at the bladder wall interface. When assessing the role of MRI in detecting posterior PAS, 73.5% of cases were detected at prenatal MRI, while 26.5% were discovered only at the time of CS. Placenta previa and prior uterine surgery represent the most commonly reported risk factors for posterior PAS. Ultrasound had a very low diagnostic accuracy in detecting these disorders prenatally, while MRI seems to perform better.

Diagnostic Test Accuracy of the 4AT for Delirium Detection: A Systematic Review and Meta-Analysis.

Under-recognition of delirium is an international problem. For the early detection of delirium, a feasible and valid screening tool for healthcare professionals is needed. This study aimed to present a scientific reason for using the 4 ‘A’s Test (4AT) through a systematic review and meta-analysis of studies on the diagnostic test accuracy. We systematically searched articles in the EMBASE, MEDLINE, CINAHL, and PsycINFO databases and selected relevant articles on the basis of the predefined inclusion criteria. The quality of the included articles was evaluated using the Quality Assessment of the Diagnostic Accuracy Studies-2 tool. We estimated the pooled values of diagnostic test accuracy by employing the bivariate model and the hierarchical summary receiver operating characteristic (HSROC) model in data synthesis. A total of 3729 patients of 13 studies were included in the analysis. The pooled estimates of sensitivity and specificity of the 4AT were 81.5% (95% confidence interval: 70.7%, 89.0%) and 87.5% (79.5%, 92.7%), respectively. Given the 4AT’s evidence of accuracy and practicality, we suggest healthcare professionals to utilize this tool for routine screening of delirium. However, for detecting delirium in the dementia population, further work is required to evaluate the 4AT with other cut-off points or scoring methods in order for it to be more sensitive and specific.

Screening Accuracy of SARC-F for Sarcopenia in the Elderly: A Diagnostic Meta-Analysis.

Sarcopenia is an age-related disease, which is characterized by a decline in muscle mass and function. It is one of the most important health issues in the elderly and often leads to a high rate and variety of adverse outcomes. To evaluate the screening accuracy of SARC-F for sarcopenia in the elderly. We conducted a meta-analysis using articles available in 6 databases including PubMed (Medline), Web of Science, Embase, Cochrane Controlled Register of Trials (CENTRAL), China Knowledge Resource Integrated Database (CNKI), and Wanfang databases from inception to May 2020. Adults aged 60 years and older. Sarcopenia was defined by EWGSOP2, EWGSOP, AWGS, FNIH and IWGS. Two authors independently extracted data based on predefined criteria. Where data were available we calculated pooled summary estimates of sensitivity, specificity, positive likelihood ratio (PLR), negative likelihood ratio (NLR), diagnostic odds ratio (DOR) and their 95% confidence interval (CI) based on different criteria using the hierarchical logistic regression modeling including bivariate modeling and hierarchical summary receiver operating characteristic (HSROC) modeling. We included 20 studies, with the prevalence of sarcopenia ranging from 6.42% to 21.56%. The number of the literatures using EWGSOP, EWGSOP2, AWGS, IWGS and FNIH as diagnostic criteria was 13, 4, 13, 8, 7, respectively. Bivariate analysis yielded a pooled sensitivity of 32% (95%CI:19%-47%), 77% (95%CI: 49%-92%), 27% (95%CI: 16%-42%), 39% (95%CI: 27%-52%), 35% (95%CI: 23%-49%) and a pooled specificity of 86% (95%CI:77%-92%), 63% (95%CI: 43%-79%), 91% (95%CI: 85%-95%), 86% (95%CI: 76%-92%), 89% (95%CI: 81%-93%), respectively. The area under the HSROC curve were 0.68 (95%CI: 0.64-0.72), 0.75 (95%CI: 0.71-0.78), 0.73 (95%CI: 0.69-0.77), 0.67 (95%CI: 0.62-0.71), 0.70 (95%CI: 0.65-0.73), respectively. The screening accuracy of SARC-F was various based on different diagnostic criteria. There were some limitations for SARC-F, however, considering the higher practicability and specificity for screening sarcopenia in practice, SARC-F was still an effective screening tool for sarcopenia in the elderly. And the screening accuracy of SARC-F needs further exploration when EWGSOP2 is applied as diagnostic criteria and geriatric inpatients are the target participants.

The diagnostic performance of CT and MRI for detecting extranodal extension in patients with head and neck squamous cell carcinoma: a systematic review and diagnostic meta-analysis.

To review the diagnostic performance of CT and MRI for detecting extranodal extension (ENE) in head and neck squamous cell carcinoma (HNSCC) patients. MEDLINE and EMBASE databases were searched up to October 7, 2019. Studies that evaluated the diagnostic performance of CT and/or MRI for detecting ENE in HNSCC patients were included. A 2 × 2 table was reconstructed for each study. Pooled sensitivity and specificity were calculated using the bivariate model and hierarchical summary receiver operating characteristic (HSROC) model. Subgroup analyses were performed according to HPV status and radiological features. Pooled correlation coefficient for interobserver agreement was calculated. Twenty-two studies including 2478 patients were included. The pooled sensitivity and specificity for detecting ENE were 73% (95% CI, 62-82%) and 83% (95% CI, 75-89%), respectively, for CT, and 60% (95% CI, 49-70%) and 96% (95% CI, 85-99%), respectively, for MRI. There was substantial heterogeneity for both CT and MRI. A threshold effect was present for MRI. On subgroup analysis, the pooled specificity of CT was significantly lower in patients with HPV+ OPSCC than in patients with HPV‑ oral cavity cancer or all HNSCC (74% vs. 87%; p = 0.01). Central node necrosis showed significantly higher pooled sensitivity (81% vs. 51%; p = 0.02), while infiltration of adjacent planes showed significantly higher pooled specificity (94% vs. 65%; p = 0.03). The pooled correlation coefficient was 0.72 (95% CI, 0.60-0.81). Both CT and MRI show reasonable diagnostic performance for detecting ENE in HNSCC patients and interobserver agreement was substantial. • Pooled sensitivity and specificity were 73% and 83% for CT and 60% and 96% for MRI without significant difference. • Pooled specificity was lower for HPV+ OPSCC than for HPV‑ oral cavity cancer or all HNSCC (74% vs. 87%, p = 0.01), likely due to central node necrosis. • Central node necrosis showed higher sensitivity (81% vs. 51%; p = 0.02), while infiltration of adjacent planes showed higher specificity (94% vs. 65%; p = 0.03).

Echocardiography versus computed tomography and cardiac magnetic resonance for the detection of left heart thrombosis: a systematic review and meta-analysis

Accurate and reproducible diagnostic techniques are essential to detect left-sided cardiac thrombi [either in the left ventricle (LV) or in the left atrial appendage (LAA)] and to guide the onset and duration of antithrombotic treatment while minimizing the risk for thromboembolic and hemorrhagic events. We conducted a systematic review and meta-analysis aiming to compare the diagnostic performance of transthoracic echocardiography (TTE) vs. cardiac magnetic resonance (CMR) for the detection of LV thrombi, and transesophageal echocardiography (TEE) vs. computed tomography (CT) for the detection of LAA thrombi. Six studies were included in the first meta-analysis (TTE vs. CMR for LV thrombosis). Pooled sensitivity and specificity values were 62% [95% confidence interval (CI), 37-81%] and 97% (95% CI, 94-99%). The shape of the hierarchical summary receiver operating characteristic (HSROC) curve and the area under the curve (AUC) of 0.96 suggested a high accuracy. Ten studies were included in the second meta-analysis (CT versus TEE for LAA thrombosis). The pooled values of sensitivity and specificity were 97% (95% CI, 77-100%) and 94% (95% CI, 87-98%). The pooled diagnostic odds ratio (DOR) was 500 (95% CI, 52-4810), and the pooled likelihood ratios (LR + and LR-) were 17% (95% CI, 7-40%) and 3% (95% CI, 0-28%). The shape of the HSROC curve and 0.99 AUC suggested a high accuracy of CT vs. TEE. TTE is a fair alternative to DE-CMR for the identification of LV thrombi, while CT has a good accuracy compared to TEE for the detection of LAA thrombosis. CRD42020185842.

Prostate-specific membrane antigen positron emission tomography (PSMA-PET) for local staging of prostate cancer: a systematic review and meta-analysis

PurposeProstate-specific membrane antigen positron emission tomography (PSMA-PET) has shown promise for detecting nodal and distant prostate cancer (PCa) metastases. However, its performance for local tumor staging is not as well established. The purpose of this study was to review the diagnostic performance of PSMA-PET for determining seminal vesical invasion (SVI) and extraprostatic extension (EPE).MethodsPubmed and Embase databases were searched until January 12, 2020. Studies assessing accuracy of PSMA-PET in determining SVI and EPE were included. Study quality was evaluated with the revised Quality Assessment of Diagnostic Accuracy Studies-2 tool. Pooled sensitivity and specificity were calculated using hierarchical summary receiver operating characteristics modeling. Heterogeneity was explored using meta-regression analyses for anatomical imaging component (MRI vs CT) and by testing for a threshold effect.ResultsTwelve studies (615 patients) were included. Pooled sensitivity and specificity were 0.68 (95% CI 0.53-0.81) and 0.94 (95% CI 0.90-0.96) for SVI and 0.72 (95% CI 0.56-0.84) and 0.87 (95% CI 0.72-0.94) for EPE. Meta-regression analyses showed that for SVI, PET/MRI demonstrated greater sensitivity than PET/CT (0.87 [95% CI 0.75-0.98] vs 0.60 [95% CI 0.47-0.74]; p = 0.02 for joint model) while specificity was comparable (0.91 [95% CI 0.84-0.97] vs. 0.96 [95% CI 0.93-0.99]) but not for EPE (p = 0.08). A threshold effect was present for studies assessing EPE (correlation coefficient = 0.563 [95% CI, −0.234-0.908] between sensitivity and false-positive rate).ConclusionPSMA-PET has moderate sensitivity and excellent specificity for assessing local tumor extent in patients with PCa. PET/MRI showed potential for greater sensitivity than PET/CT in assessing SVI.

Early laboratory biomarkers for severity in acute pancreatitis; A systematic review and meta-analysis

Background/ObjectivesAcute pancreatitis is complicated by local and systemic complications in 20–30% of the patients. Accurate prediction of severity may be important for clinical decision making. Our aim is to identify and compare the accuracy of laboratory biomarkers that predict severity and complications in adult patients. MethodsMedline, EMBASE, Web of Science and Cochrane Library (1993 to August 2020) were searched for studies with an unselected population of patients with acute pancreatitis, that contains accuracy data for ≥1 laboratory biomarker(s) and/or APACHE-II score for the prediction of a patient outcomes of interest during the first 48 h of admission. The primary outcome is moderate severe or severe acute pancreatitis (MSAP/SAP). Secondary outcomes are severe acute pancreatitis, pancreatic necrosis and organ failure. Risk of bias was assed using QUADAS-2. Biomarkers extracted from ≥3 unique sources, were analyzed using hierarchical summary receiver operating characteristic (HSROC) and bivariate model analysis. ResultsIn total, 181 studies were included in the qualitative analysis reporting on 29 biomarkers. For the primary outcome at admission, summary sensitivities and specificities were, respectively, 87% (95% CI 69–95%) and 88% (95% CI 80–93%) for IL-6 at a threshold of >50 pg/ml, 72% (95% CI 64–79%) and 76% (95% CI 67–84%) for an APACHE-II score of ≥8, and 53% (95% CI 35–71%) and 82% (95% CI 74–88%) for CRP >150 mg/l. HSROC curve analysis confirmed these results. ConclusionThis study indicates superiority of IL-6 for the early prediction of MSAP/SAP and may be used for to guide clinical decision making.

Artificial Intelligence in Drug-Resistant Tuberculosis Diagnosis. Systematic Review & Meta Analysis

Abstract Background Artificial Intelligence (AI) models possess potential screening or diagnosing Drug-Resistant Tuberculosis. This meta-analysis aimed to elaborate on the performance of AI models in predicting drug-resistant TB compared to the gold standard. Methods A systematic search of full-paper articles from Pubmed/MEDLINE and SCOPUS published between January 2005 and April 2020 was conducted. Studies involving whole-genome sequencing or human subject data that used drug-susceptibility test results as a reference standard were retrieved. Hierarchical summary receiver-operating characteristic (HSROC) and bivariate model were performed to calculate pooled sensitivity and specificity. Heterogeneity and publication bias were also assessed. Results 25 Models form four genome-based studies, four radiology-based studies, and two clinical and demographic-based studies were included. The radiology-based models, mainly built with the Convolutional Neural Network (CNN), possessed pooled sensitivity of 58% (95%CI 50%-65%), and a pooled specificity of 75% (95% CI 47%-91%) for detecting MDR-TB. Genome-based models outperformed other models despite developed with simpler classifiers (CART/GBT/LR). The genome-based studies reached pooled sensitivity of 92% (95%CI 90-93%), pooled specificity of 98 (95% CI 96%-98%) in predicting isoniazid resistance. For predicting rifampicin resistance, the pooled sensitivity and specificity were 93% (95%CI 90%-94%) and 98% (95% CI 96%-98%) respectively. The Artificial Neural Network (ANN) demonstrated superiority rather than the Classification and Regression Tree (CART), and logistic regression (LR) on the clinical and demographic model in predicting drug-resistant TB. Conclusions Simple classifier performs better in genomic data whereas the CNN model works best in high-dimensional data, such as radiology images. The ANN model indicates a preferred model for data obtained from clinical and demographic parameters. (PROSPERO number CRD42020167439) Key messages Artificial Intelligence model through machine learning and deep learning possess a good diagnostic performance for drug-resistant tuberculosis. Artificial Neural Network model outperforms other technique. The model built from genomic data shows the best performance, followed by model from patient-based data and radiology/imaging data.

Diagnostic accuracy and safety of inhalation challenge tests for bird fancier's lung-Systematic review and meta-analysis.

To assess the diagnostic accuracy of inhalation challenge tests for bird fancier's lung and related adverse reactions. We performed a systematic review and meta-analysis for the diagnostic test accuracy of inhalation challenge tests and a systematic review for adverse events of the tests. We evaluated the risk of bias and applicability of the included articles for diagnostic test accuracy with the modified Quality Assessment of Diagnostic Accuracy Studies-2 tool. We used hierarchical summary receiver operating characteristic (HSROC) curve analysis to evaluate the sensitivity and specificity of challenge tests and assessed subjective adverse reactions and steroid treatment use. Sensitivity was calculated by fixing specificity at 99% from the HSROC curve. We searched for articles evaluating the diagnostic accuracy of inhalation challenge tests or describing adverse reactions in Medline, Embase, the Cochrane Library, the International Clinical Trials Registry Platform, and Web of Science. We included randomized control trials, prospective or retrospective cohort studies, or case-control studies assessing the diagnostic test accuracy of inhalation challenge tests for bird fancier's lung. For adverse effects review, we included the same articles, and case series or case reports reporting adverse reactions of inhalation challenge tests. In our review of 12 articles, the diagnostic accuracy of inhalation challenge tests was substantially high. Point estimate of sensitivity when calculated with a fixed specificity of 99% was 99%. Among 873 patients, 6 needed steroid treatment for adverse reactions; however, no death occurred due to acute exacerbation following a challenge test. Inhalation challenges for bird fancier's lung could be accurate and safe diagnostic procedures and may be considered both as rule-in and rule-out tests in tertiary care centres. However, caution is required regarding overestimation of diagnostic yield due to risk of bias. Systematic review registration: University hospital Medical Information Network Clinical Trials Registry (UMIN000038799).

Performance of deep neural network-based artificial intelligence method in diabetic retinopathy screening: a systematic review and meta-analysis of diagnostic test accuracy.

Automatic diabetic retinopathy screening system based on neural networks has been used to detect diabetic retinopathy (DR). However, there is no quantitative synthesis of performance of these methods. We aimed to estimate the sensitivity and specificity of neural networks in DR grading. Medline, Embase, IEEE Xplore, and Cochrane Library were searched up to 23 July 2019. Studies that evaluated performance of neural networks in detection of moderate or worse DR or diabetic macular edema using retinal fundus images with ophthalmologists' judgment as reference standard were included. Two reviewers extracted data independently. Risk of bias of eligible studies was assessed using QUDAS-2 tool. Twenty-four studies involving 235 235 subjects were included. Quantitative random-effects meta-analysis using the Rutter and Gatsonis hierarchical summary receiver operating characteristics (HSROC) model revealed a pooled sensitivity of 91.9% (95% CI: 89.6% to 94.3%) and specificity of 91.3% (95% CI: 89.0% to 93.5%). Subgroup analyses and meta-regression did not provide any statistically significant findings for the heterogeneous diagnostic accuracy in studies with different image resolutions, sample sizes of training sets, architecture of convolutional neural networks, or diagnostic criteria. State-of-the-art neural networks could effectively detect clinical significant DR. To further improve diagnostic accuracy of neural networks, researchers might need to develop new algorithms rather than simply enlarge sample sizes of training sets or optimize image quality.

Comparative accuracy of non-invasive imaging versus right heart catheterization for the diagnosis of pulmonary hypertension: A systematic review and meta-analysis.

BackgroundRight heart catheterization (RHC) is the gold-standard in the diagnosis of pulmonary hypertension (PH) but at the cost of procedure-related complications. We sought to determine the comparative accuracy of RHC versus non-invasive imaging techniques such as computed tomography (CT), magnetic resonance imaging (MRI), and transthoracic echocardiography (TTE).MethodsPulmonary hypertension was defined as a mean pulmonary artery pressure (mPAP) of>20 mmHg. Multiple databases were queried for relevant articles. Raw data were pooled using a bivariate model to calculate the measures of diagnostic accuracy and to estimate Hierarchical Summary Receiver Operating Characteristic (HSROC) on Stata 13.ResultsA total of 51 studies with a total patient population of 3947 were selected. The pooled sensitivity and specificity of MRI for diagnosing PH was 0.92(95% confidence interval (CI) 0.88–0.96) and 0.86 (95% CI, 0.77–0.95), respectively. The net sensitivities for CT scan and TTE were 0.79 (95% CI 0.72–0.89) and 0.85 (95% CI 0.83–0.91), respectively. The overall specificity was 0.82 (0.76–0.92) for the CT scan and 0.71 (95% CI 0.61–0.84) for TTE. The diagnostic odds ratio (DOR) for MRI was 124 (95% CI 36–433) compared to 30 (95% CI 11–78) and 24 (95% 11–38) for CT scan and TTE, respectively. Chi-squared (x2) test showed moderate heterogeneity on the test for equality of sensitivities and specificities.ConclusionsMRI has the highest sensitivity and specificity compared to CT and TTE. MRI can potentially serve as a surrogate technique to RHC for the diagnosis of PH.

Clinicopathological significance of Ki67 expression in colorectal cancer: A protocol of systematic review and meta-analysis.

Background:This study will investigate the diagnostic accuracy of Ki67 expression in colorectal cancer (CC).Methods:A comprehensive search in electronic bibliographic databases (MEDLINE, EMBASE, Cochrane Library, Web of Science, Chinese Biomedical Literature Database, and China National Knowledge Infrastructure) will be performed from inception to the February 29, 2020 with no restrictions to the language and publication status. Two authors will examine the collected studies, extract essential data, and appraise study quality separately. If possible, we will estimate receiver operating characteristic (ROC), sensitivity and specificity by utilizing bivariate random effects and hierarchical summary ROC models.Results:This study will summarize present evidence to explore the diagnostic accuracy of Ki67 expression in CC.Conclusion:The findings of this study will clarify the diagnostic accuracy of Ki67 expression in CC.Systematic review registration:INPLASY202030009.

Use of Vesical Imaging-Reporting and Data System (VI-RADS) for detecting the muscle invasion of bladder cancer: a diagnostic meta-analysis.

To comprehensively assess the diagnostic performance of Vesical Imaging-Reporting and Data System (VI-RADS) score for detecting the muscle invasion of bladder cancer. PubMed, Web of Science, and Embase were searched up to November 20, 2019. QUADAS-2 tool assessed the quality of included studies. The diagnostic estimates including sensitivity, specificity, positive likelihood ratio, negative likelihood ratio, and the area under the curve (AUC) of hierarchical summary receiver operating characteristic (HSROC) were calculated. Further subgroup analysis, meta-regression and sensitivity analysis were conducted. Six studies with 1064 patients were finally included. The pooled sensitivity, specificity, and AUC value were 0.90 (95% CI 0.86-0.94), 0.86 (95% CI 0.71-0.94), and 0.93 (95% CI 0.91-0.95) for VI-RADS 3 as the cutoff value. The corresponding estimates were 0.77 (95% CI 0.65-0.86), 0.97 (95% CI 0.88-0.99), and 0.92 (95% CI 0.89-0.94) for VI-RADS 4 as the cutoff value. Meta-regression analysis revealed that study design (p value 0.01) and surgical pattern of reference standard (p value 0.02) were source of the heterogeneity of pooled sensitivity. No publication bias was observed. The VI-RADS score can provide a good predictive ability for detecting the muscle invasiveness of primary bladder cancer with VI-RADS 3 or VI-RADS 4 as the cutoff value. • VI-RADS score has high sensitivity and specificity for predicting muscle invasion. • The diagnostic efficiencies of VI-RADS 3 and VI-RADS 4 as the cutoff value are similar. • VI-RADS score could be used for detecting muscle invasion of bladder cancer in clinical practice.

Predictive value of cystatin C and neutrophil gelatinase-associated lipocalin in contrast-induced nephropathy: Ameta-analysis.

There are still limited studies comprehensively examining the diagnostic performance of neutrophil gelatinase-associated lipocalin (NGAL) and cystatin C in contrast-induced nephropathy (CIN). The study aimed to investigate and compare the predictive value of NGAL and cystatin C in the early diagnosis of CIN. We searched the PubMed, EMBASE and Cochrane Library databases until November 10, 2019. The methodological quality of the included studies was assessed by the Quality Assessment of Diagnostic Accuracy Studies 2 (QUADAS-2) tool. Bivariate modeling and hierarchical summary receiver operating characteristic (HSROC) modeling were performed to summarize and compare the diagnostic performance of blood/urine NGAL and serum cystatin C in CIN. Subgroup and meta-regression analyses were performed according to the study and patient characteristics. Thirty-seven studies from thirty-one original studies were included (blood NGAL, 1840 patients in 9 studies; urine NGAL, 1701 patients in 10 studies; serum cystatin C, 5509 patients in 18 studies). Overall, serum cystatin C performed better than serum/urine NGAL (pooled DOR: 43 (95%CI: 12-152); AUROC: 0.93; λ: 3.79); serum and urine NGAL had a similar diagnostic performance (pooled DOR: 25 (95%CI: 6-108)/22(95%CI: 8-64); AUROC: 0.90/0.89; λ: 3.20/3.08). Meta-regression analysis indicated that the sources of heterogeneity might be CIN definition, assays, and nationalities. Both NGAL and cystatin C can serve as early diagnostic indicators of CIN, while cystatin C may perform better than NGAL.